ARM® Cortex®-A72 MPCore Processor

Revision: r0p3

Technical Reference Manual

Copyright © 2014-2016 ARM Limited or its affiliates. All rights reserved.

ARM 100095_0003_06_en
 ARM® Cortex®-A72 MPCore Processor

ARM® Cortex®-A72 MPCore Processor

Technical Reference Manual
Copyright © 2014-2016 ARM Limited or its affiliates. All rights reserved.
Release Information

Document History

Issue Date Confidentiality Change

0000-01 23 October 2014 Confidential First release for r0p0.
0001-02 20 February 2015 Non-Confidential First release for r0p1.
0002-03 05 June 2015 Non-Confidential First release for r0p2.
0002-04 02 February 2016 Non-Confidential Second release for r0p2.
0003-05 22 April 2016 Non-Confidential First release for r0p3.
0003-06 01 December 2016 Non-Confidential Second release for r0p3.

Non-Confidential Proprietary Notice

This document is protected by copyright and other related rights and the practice or implementation of the information contained in
this document may be protected by one or more patents or pending patent applications. No part of this document may be
reproduced in any form by any means without the express prior written permission of ARM. No license, express or implied, by
estoppel or otherwise to any intellectual property rights is granted by this document unless specifically stated.

Your access to the information in this document is conditional upon your acceptance that you will not use or permit others to use
the information for the purposes of determining whether implementations infringe any third party patents.

THIS DOCUMENT IS PROVIDED “AS IS”. ARM PROVIDES NO REPRESENTATIONS AND NO WARRANTIES,
EXPRESS, IMPLIED OR STATUTORY, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
MERCHANTABILITY, SATISFACTORY QUALITY, NON-INFRINGEMENT OR FITNESS FOR A PARTICULAR PURPOSE
WITH RESPECT TO THE DOCUMENT. For the avoidance of doubt, ARM makes no representation with respect to, and has
undertaken no analysis to identify or understand the scope and content of, third party patents, copyrights, trade secrets, or other
rights.

This document may include technical inaccuracies or typographical errors.

TO THE EXTENT NOT PROHIBITED BY LAW, IN NO EVENT WILL ARM BE LIABLE FOR ANY DAMAGES,
INCLUDING WITHOUT LIMITATION ANY DIRECT, INDIRECT, SPECIAL, INCIDENTAL, PUNITIVE, OR
CONSEQUENTIAL DAMAGES, HOWEVER CAUSED AND REGARDLESS OF THE THEORY OF LIABILITY, ARISING
OUT OF ANY USE OF THIS DOCUMENT, EVEN IF ARM HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
DAMAGES.

This document consists solely of commercial items. You shall be responsible for ensuring that any use, duplication or disclosure of
this document complies fully with any relevant export laws and regulations to assure that this document or any portion thereof is
not exported, directly or indirectly, in violation of such export laws. Use of the word “partner” in reference to ARM’s customers is
not intended to create or refer to any partnership relationship with any other company. ARM may make changes to this document at
any time and without notice.

If any of the provisions contained in these terms conflict with any of the provisions of any signed written agreement covering this
document with ARM, then the signed written agreement prevails over and supersedes the conflicting provisions of these terms.
This document may be translated into other languages for convenience, and you agree that if there is any conflict between the
English version of this document and any translation, the terms of the English version of the Agreement shall prevail.

Words and logos marked with ® or ™ are registered trademarks or trademarks of ARM Limited or its affiliates in the EU and/or
elsewhere. All rights reserved. Other brands and names mentioned in this document may be the trademarks of their respective
owners. Please follow ARM’s trademark usage guidelines at http://www.arm.com/about/trademark-usage-guidelines.php

Copyright © 2014-2016, ARM Limited or its affiliates. All rights reserved.

ARM Limited. Company 02557590 registered in England.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2
reserved.
Non-Confidential
 ARM® Cortex®-A72 MPCore Processor

110 Fulbourn Road, Cambridge, England CB1 9NJ.

LES-PRE-20349
Confidentiality Status
This document is Non-Confidential. The right to use, copy and disclose this document may be subject to license restrictions in
accordance with the terms of the agreement entered into by ARM and the party that ARM delivered this document to.

Unrestricted Access is an ARM internal classification.

Product Status
The information in this document is Final, that is for a developed product.
Web Address
http://www.arm.com

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 3
reserved.
Non-Confidential
Contents
ARM® Cortex®-A72 MPCore Processor Technical
Reference Manual

Preface
About this book ...................................................... ...................................................... 9
Feedback .................................................................................................................... 12

Chapter 1 Introduction
1.1 About the Cortex-A72 processor ...................................... ...................................... 1-14
1.2 Compliance .............................................................................................................. 1-15
1.3 Features ......................................................... ......................................................... 1-17
1.4 Interfaces ........................................................ ........................................................ 1-18
1.5 Implementation options ............................................. ............................................. 1-19
1.6 Test features ............................................................................................................ 1-20
1.7 Product documentation and design flow .................................................................. 1-21
1.8 Product revisions .................................................. .................................................. 1-23

Chapter 2 Functional Description

2.1 About the Cortex-A72 processor functions .............................................................. 2-25
2.2 Interfaces ........................................................ ........................................................ 2-29
2.3 Clocking and resets ................................................ ................................................ 2-32
2.4 Power management ................................................ ................................................ 2-42

Chapter 3 Programmers Model

3.1 About the programmers model ................................................................................ 3-60

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4
reserved.
Non-Confidential
 3.2 ARMv8-A architecture concepts .............................................................................. 3-61
3.3 ThumbEE instruction set ............................................ ............................................ 3-69
3.4 Jazelle implementation ............................................................................................ 3-70
3.5 Memory model .................................................... .................................................... 3-72

Chapter 4 System Control

4.1 About system control ............................................... ............................................... 4-74
4.2 AArch64 register summary ...................................................................................... 4-75
4.3 AArch64 register descriptions .................................................................................. 4-88
4.4 AArch32 register summary .................................................................................... 4-214
4.5 AArch32 register descriptions ................................................................................ 4-239

Chapter 5 Memory Management Unit

5.1 About the MMU ...................................................................................................... 5-276
5.2 TLB organization ................................................. ................................................. 5-277
5.3 TLB match process ................................................................................................ 5-278
5.4 Memory access sequence .......................................... .......................................... 5-279
5.5 MMU enabling and disabling ........................................ ........................................ 5-281
5.6 Intermediate table walk caches ...................................... ...................................... 5-282
5.7 External aborts ................................................... ................................................... 5-283

Chapter 6 Level 1 Memory System

6.1 About the L1 memory system ................................................................................ 6-285
6.2 Cache organization ................................................................................................ 6-286
6.3 L1 instruction memory system ....................................... ....................................... 6-287
6.4 L1 data memory system ........................................................................................ 6-289
6.5 Program flow prediction ............................................ ............................................ 6-295
6.6 L1 RAM memories ................................................ ................................................ 6-297

Chapter 7 Level 2 Memory System

7.1 About the L2 memory system ................................................................................ 7-299
7.2 Cache organization ................................................................................................ 7-300
7.3 L2 RAM memories ................................................ ................................................ 7-306
7.4 L2 cache prefetcher ............................................... ............................................... 7-307
7.5 Cache coherency ................................................. ................................................. 7-308
7.6 Asynchronous errors .............................................. .............................................. 7-309
7.7 External coherent interfaces .................................................................................. 7-310
7.8 ACP ........................................................... ........................................................... 7-317

Chapter 8 Generic Interrupt Controller CPU Interface

8.1 About the GIC ........................................................................................................ 8-319
8.2 GIC functional description .......................................... .......................................... 8-320
8.3 GIC programmers model ........................................... ........................................... 8-325

Chapter 9 Generic Timer

9.1 About the Generic Timer ........................................... ........................................... 9-339
9.2 Generic Timer functional description .................................. .................................. 9-340
9.3 Generic Timer register summary ..................................... ..................................... 9-341

Chapter 10 Debug
10.1 About debug ........................................................................................................ 10-344

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 5
reserved.
Non-Confidential
 10.2 Debug register interfaces .......................................... .......................................... 10-346
10.3 AArch64 debug register summary ................................... ................................... 10-348
10.4 AArch64 debug register descriptions ................................. ................................. 10-350
10.5 AArch32 debug register summary ................................... ................................... 10-356
10.6 AArch32 debug register descriptions ................................. ................................. 10-358
10.7 Memory-mapped register summary .................................. .................................. 10-362
10.8 Memory-mapped register descriptions ................................................................ 10-366
10.9 Debug events ................................................... ................................................... 10-380
10.10 External debug interface ...................................................................................... 10-381
10.11 ROM table ..................................................... ..................................................... 10-384

Chapter 11 Performance Monitor Unit

11.1 About the PMU .................................................. .................................................. 11-396
11.2 PMU functional description .................................................................................. 11-397
11.3 AArch64 PMU register summary .................................... .................................... 11-399
11.4 AArch64 PMU register descriptions .................................. .................................. 11-401
11.5 AArch32 PMU register summary .................................... .................................... 11-406
11.6 Memory-mapped register summary .................................. .................................. 11-408
11.7 Memory-mapped register descriptions ................................ ................................ 11-411
11.8 Events .................................................................................................................. 11-428
11.9 Interrupts .............................................................................................................. 11-432
11.10 Exporting PMU events ............................................ ............................................ 11-433

Chapter 12 Cross Trigger

12.1 About the cross trigger ............................................ ............................................ 12-435
12.2 Trigger inputs and outputs ......................................... ......................................... 12-436
12.3 CTI ........................................................... ........................................................... 12-437
12.4 CTM .......................................................... .......................................................... 12-438
12.5 Cross trigger register summary ..................................... ..................................... 12-439
12.6 Cross trigger register descriptions ................................... ................................... 12-442

Chapter 13 Embedded Trace Macrocell

13.1 About ETM ..................................................... ..................................................... 13-458
13.2 ETM trace generation options and resources ...................................................... 13-459
13.3 ETM functional description .................................................................................. 13-461
13.4 Reset ......................................................... ......................................................... 13-462
13.5 ETM register interfaces ........................................... ........................................... 13-463
13.6 Register summary ............................................... ............................................... 13-464
13.7 Register descriptions ............................................. ............................................. 13-468
13.8 Interaction with debug and the Performance Monitor Unit ................. ................. 13-510

Chapter 14 Advanced SIMD and Floating-point

14.1 About Advanced SIMD and Floating-point ............................. ............................. 14-512
14.2 Programmers model for Advanced SIMD and Floating-point .............................. 14-513
14.3 AArch64 register summary .................................................................................. 14-514
14.4 AArch64 register descriptions .............................................................................. 14-515
14.5 AArch32 register summary .................................................................................. 14-524
14.6 AArch32 register descriptions .............................................................................. 14-525

Appendix A Signal Descriptions

A.1 About the signal descriptions ................................... ................................... Appx-A-530

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 6
reserved.
Non-Confidential
 A.2 Clock signals ................................................................................................ Appx-A-531
A.3 Reset signals ............................................... ............................................... Appx-A-532
A.4 Configuration signals ......................................... ......................................... Appx-A-533
A.5 GIC CPU interface signals ..................................... ..................................... Appx-A-535
A.6 Generic Timer signals .................................................................................. Appx-A-537
A.7 Power control signals ......................................... ......................................... Appx-A-538
A.8 ACE and CHI interface signals .................................................................... Appx-A-540
A.9 CHI interface signals .................................................................................... Appx-A-543
A.10 ACE interface signals ......................................... ......................................... Appx-A-548
A.11 ACP interface signals ......................................... ......................................... Appx-A-553
A.12 Debug interface signals ....................................... ....................................... Appx-A-556
A.13 ETM interface ............................................... ............................................... Appx-A-559
A.14 Cross trigger channel interface .................................................................... Appx-A-561
A.15 PMU signals ................................................ ................................................ Appx-A-562
A.16 DFT and MBIST signals ....................................... ....................................... Appx-A-563

Appendix B AArch32 Unpredictable Behaviors

B.1 Unpredictable behaviors .............................................................................. Appx-B-565
B.2 Debug UNPREDICTABLE behaviors ................................ ................................ Appx-B-567

Appendix C Revisions
C.1 Revisions .................................................. .................................................. Appx-C-575

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7
reserved.
Non-Confidential
Preface

This preface introduces the ARM® Cortex®-A72 MPCore Processor Technical Reference Manual.
It contains the following:
• About this book on page 9.
• Feedback on page 12.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8
reserved.
Non-Confidential
 Preface
About this book

About this book

This document describes the ARM® Cortex®-A72 processor.

Product revision status

The rmpn identifier indicates the revision status of the product described in this book, for example, r1p2,
where:
rm Identifies the major revision of the product, for example, r1.
pn Identifies the minor revision or modification status of the product, for example, p2.

Intended audience
This document is written for system designers, system integrators, and programmers who are designing
or programming a System-on-Chip (SoC) that uses the Cortex-A72 processor.

Using this book

This book is organized into the following chapters:
Chapter 1 Introduction
This chapter introduces the Cortex-A72 processor and its features.
Chapter 2 Functional Description
This chapter describes the functionality of the Cortex-A72 processor.
Chapter 3 Programmers Model
This chapter describes the processor registers and provides information for programming the
processor.
Chapter 4 System Control
This chapter describes the System registers, their structure, operation, and how to use them.
Chapter 5 Memory Management Unit
This section describes the Memory Management Unit (MMU).
Chapter 6 Level 1 Memory System
This section describes the Level 1 (L1) memory system.
Chapter 7 Level 2 Memory System
This chapter describes the Level 2 (L2) memory system.
Chapter 8 Generic Interrupt Controller CPU Interface
This section describes the Cortex-A72 processor implementation of the GIC CPU interface.
Chapter 9 Generic Timer
This chapter describes the Cortex-A72 processor implementation of the ARM Generic Timer.
Chapter 10 Debug
This section describes the Cortex-A72 processor debug registers and shows examples of how to
use them.
Chapter 11 Performance Monitor Unit
This section describes the Performance Monitor Unit (PMU) and the registers that it uses.
Chapter 12 Cross Trigger
This chapter describes the cross trigger interfaces for the Cortex-A72 processor.
Chapter 13 Embedded Trace Macrocell
This section describes the Embedded Trace Macrocell (ETM) for the Cortex-A72 processor.
Chapter 14 Advanced SIMD and Floating-point
This chapter describes the Advanced SIMD and Floating-point features and registers in the
Cortex-A72 processor.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 9
reserved.
Non-Confidential
 Preface
About this book

Appendix A Signal Descriptions

This section describes the Cortex-A72 processor signals.
Appendix B AArch32 Unpredictable Behaviors
This appendix describes specific Cortex-A72 processor UNPREDICTABLE behaviors that are of
particular interest.
Appendix C Revisions
This appendix describes the technical changes between released issues of this book.

Glossary
The ARM Glossary is a list of terms used in ARM documentation, together with definitions for those
terms. The ARM Glossary does not contain terms that are industry standard unless the ARM meaning
differs from the generally accepted meaning.
See the ARM Glossary for more information.

Typographic conventions
italic
Introduces special terminology, denotes cross-references, and citations.
bold
Highlights interface elements, such as menu names. Denotes signal names. Also used for terms
in descriptive lists, where appropriate.
monospace
Denotes text that you can enter at the keyboard, such as commands, file and program names,
and source code.
monospace
Denotes a permitted abbreviation for a command or option. You can enter the underlined text
instead of the full command or option name.
monospace italic
Denotes arguments to monospace text where the argument is to be replaced by a specific value.
monospace bold
Denotes language keywords when used outside example code.
<and>
Encloses replaceable terms for assembler syntax where they appear in code or code fragments.
For example:
MRC p15, 0, <Rd>, <CRn>, <CRm>, <Opcode_2>

SMALL CAPITALS
Used in body text for a few terms that have specific technical meanings, that are defined in the
ARM glossary. For example, IMPLEMENTATION DEFINED, IMPLEMENTATION SPECIFIC, UNKNOWN, and
UNPREDICTABLE.

Timing diagrams
The following figure explains the components used in timing diagrams. Variations, when they occur,
have clear labels. You must not assume any timing information that is not explicit in the diagrams.
Shaded bus and signal areas are undefined, so the bus or signal can assume any value within the shaded
area at that time. The actual level is unimportant and does not affect normal operation.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10
reserved.
Non-Confidential
 Preface
About this book

Clock

HIGH to LOW

Transient

HIGH/LOW to HIGH

Bus stable

Bus to high impedance

Bus change

High impedance to stable bus

Figure 1 Key to timing diagram conventions

Signals
The signal conventions are:
Signal level
The level of an asserted signal depends on whether the signal is active-HIGH or active-LOW.
Asserted means:
• HIGH for active-HIGH signals.
• LOW for active-LOW signals.

Lowercase n
At the start or end of a signal name denotes an active-LOW signal.

Additional reading
This book contains information that is specific to this product. See the following documents for other
relevant information.
ARM publications
• ARM® AMBA® APB Protocol Specification (ARM IHI 0024).
• ARM® AMBA® 3 ATB Protocol Specification (ARM IHI 0032).
• ARM® AMBA® AXI and ACE Protocol Specification (ARM IHI 0022).
• ARM® AMBA® AXI4-Stream Protocol Specification (ARM IHI 0051).
• ARM® Architecture Reference Manual ARMv8, for ARMv8-A architecture
profile (ARM DDI 0487).
• ARM® CoreSight™ SoC-400 Technical Reference Manual (ARM DII 0480).
• ARM® Debug Interface Architecture Specification ADIv5.0 to ADIv5.2 (ARM IHI 0031).
• ARM® Embedded Trace Macrocell Architecture Specification, ETMv4 (ARM IHI 0064).
• ARM® Generic Interrupt Controller Architecture Specification GICv3 (ARM IHI 0048).
The following confidential books are only available to licensees:
• ARM® CoreSight™ Architecture Specification (ARM IHI 0029).
• ARM® AMBA® 5 CHI Protocol Specification (ARM IHI 0050).
• ARM® Cortex-A72 MPCore Processor Configuration and Sign-off Guide (ARM 100098).
• ARM® Cortex-A72 MPCore Processor Integration Manual (ARM 100096).
• ARM® Cortex-A72 MPCore Processor Cryptography Extension Technical Reference
Manual (ARM 100097).
Other publications
• ANSI/IEEE, IEEE Standard for Binary Floating-Point Arithmetic, Std 754-1985.
• ANSI/IEEE, IEEE Standard for Floating-Point Arithmetic, Std 754-2008.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11
reserved.
Non-Confidential
 Preface
Feedback

Feedback

Feedback on this product

If you have any comments or suggestions about this product, contact your supplier and give:
• The product name.
• The product revision or version.
• An explanation with as much information as you can provide. Include symptoms and diagnostic
procedures if appropriate.

Feedback on content
If you have comments on content then send an e-mail to [email protected]. Give:
• The title ARM® Cortex®-A72 MPCore Processor Technical Reference Manual.
• The number ARM 100095_0003_06_en.
• If applicable, the page number(s) to which your comments refer.
• A concise explanation of your comments.
ARM also welcomes general suggestions for additions and improvements.
Note
ARM tests the PDF only in Adobe Acrobat and Acrobat Reader, and cannot guarantee the quality of the
represented document when used with any other PDF reader.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12
reserved.
Non-Confidential
Chapter 1
Introduction

This chapter introduces the Cortex-A72 processor and its features.

It contains the following sections:

• 1.1 About the Cortex-A72 processor on page 1-14.
• 1.2 Compliance on page 1-15.
• 1.3 Features on page 1-17.
• 1.4 Interfaces on page 1-18.
• 1.5 Implementation options on page 1-19.
• 1.6 Test features on page 1-20.
• 1.7 Product documentation and design flow on page 1-21.
• 1.8 Product revisions on page 1-23.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 1-13
reserved.
Non-Confidential
 1 Introduction
1.1 About the Cortex-A72 processor

1.1 About the Cortex-A72 processor

The Cortex-A72 processor is a high-performance, low-power processor that implements the ARMv8-A
architecture. It has one to four cores in a single processor device with L1 and L2 cache subsystems.
The following figure shows an example block diagram of a Cortex-A72 processor configuration with
four cores.
Cortex-A72 processor

APB ATB TimerEvents Interrupts AXI4 Stream Protocol

Debug and Generic

Trace GIC CPU interface
CTI Timer

Miscellaneous

Core 0 Core 1 Core 2 Core 3

L1 L1 L1 L1 L1 L1 L1 L1
TLBs TLBs TLBs TLBs
ICache DCache ICache DCache ICache DCache ICache DCache

Snoop
Slave Master Control L2 Cache
Unit

ACP ACE or CHI Level 2 memory system

Figure 1-1 Example Cortex-A72 processor configuration

See 2.1.1 Components of the processor on page 2-26 for a description of the Cortex-A72 processor
functional components.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 1-14
reserved.
Non-Confidential
 1 Introduction
1.2 Compliance

1.2 Compliance
The Cortex-A72 processor complies with, or implements, the specifications described in this section.
This TRM complements architecture reference manuals, architecture specifications, protocol
specifications, and relevant external standards. It does not duplicate information from these sources.
This section contains the following subsections:
• 1.2.1 ARM architecture on page 1-15.
• 1.2.2 Advanced Microcontroller Bus Architecture (AMBA) on page 1-15.
• 1.2.3 CHI architecture on page 1-15.
• 1.2.4 Generic Interrupt Controller architecture on page 1-15.
• 1.2.5 Generic Timer architecture on page 1-16.
• 1.2.6 Debug architecture on page 1-16.
• 1.2.7 Embedded Trace Macrocell architecture on page 1-16.

1.2.1 ARM architecture

The Cortex-A72 processor implements the ARMv8-A architecture. This includes:
• Support for both AArch32 and AArch64 Execution states.
• Support for all Exception levels, EL0, EL1, EL2, and EL3, in each Execution state.
• The A32 instruction set, previously called the ARM instruction set.
• The T32 instruction set, previously called the Thumb instruction set.
• The A64 instruction set.
The Cortex-A72 processor supports the following features:
• Advanced Single Instruction Multiple Data (SIMD) operations.
• Floating-point operations.
• Optional Cryptography Extension.
The Cortex-A72 processor does not support the T32EE (ThumbEE) instruction set.
See the ARM® Architecture Reference Manual ARMv8, for ARMv8-A architecture profile for more
information.

1.2.2 Advanced Microcontroller Bus Architecture (AMBA)

The Cortex-A72 processor complies with the:
• AMBA 4 AXI Coherency Extensions (ACE) protocol if the processor memory interface implements
an ACE bus interface. See the ARM® AMBA® AXI and ACE Protocol Specification.
• AMBA 4 Advanced eXtensible Interface (AXI) protocol for the Accelerator Coherency Port (ACP)
slave interface. See the ARM® AMBA® AXI and ACE Protocol Specification.
• AMBA 3 Advanced Peripheral Bus (APB) protocol. See the ARM® AMBA® APB Protocol
Specification.
• AMBA 3 Advanced Trace Bus (ATB) protocol. See the ARM® AMBA® 3 ATB Protocol Specification.

1.2.3 CHI architecture

The processor complies with the CHI architecture if the memory interface implements a CHI bus
interface.

1.2.4 Generic Interrupt Controller architecture

The Cortex-A72 processor implements the ARM Generic Interrupt Controller (GIC) v3 architecture.
The processor includes only the GIC CPU interface. See the ARM® Generic Interrupt Controller
Architecture Specification, GICv3.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 1-15
reserved.
Non-Confidential
 1 Introduction
1.2 Compliance

1.2.5 Generic Timer architecture

The processor implements the ARM Generic Timer architecture. See the ARM® Architecture Reference
Manual ARMv8 for more information.

1.2.6 Debug architecture

The processor implements version 8 of the ARM Debug architecture that complies with the CoreSight
architecture.
See the following for more information:
• ARM® CoreSight™ Architecture Specification.
• ARM® Architecture Reference Manual ARMv8, for ARMv8-A architecture profile.

1.2.7 Embedded Trace Macrocell architecture

The processor implements the ETMv4 architecture.
See the ARM® Embedded Trace Macrocell Architecture Specification ETMv4.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 1-16
reserved.
Non-Confidential
 1 Introduction
1.3 Features

1.3 Features
The Cortex-A72 processor includes the following features:
• Full implementation of the ARMv8-A architecture profile. See 1.2 Compliance on page 1-15.
• Superscalar, variable-length, out-of-order pipeline.
• Dynamic branch prediction with Branch Target Buffer (BTB) and Global History Buffer (GHB)
RAMs, a return stack, and an indirect predictor.
• 48-entry fully-associative L1 instruction Translation Lookaside Buffer (TLB) with native support for
4KB, 64KB, and 1MB page sizes.
• 32-entry fully-associative L1 data TLB with native support for 4KB, 64KB, and 1MB page sizes.
• 4-way set-associative unified 1024-entry Level 2 (L2) TLB in each processor.
• Fixed 48K L1 instruction cache and 32K L1 data cache.
• Shared L2 cache of 512KB, 1MB, 2MB or 4MB configurable size.
• Optional Error Correction Code (ECC) protection for L2 cache, and optional ECC protection for L1
data cache and parity protection for L1 instruction cache.
• AMBA 4 AXI Coherency Extensions (ACE) or CHI master interface.
• Optional Accelerator Coherency Port (ACP) implemented as an AXI4 slave interface.
• Embedded Trace Macrocell (ETM) based on the ETMv4 architecture.
• Performance Monitor Unit (PMU) support based on the PMUv3 architecture.
• Cross Trigger Interface (CTI) for multiprocessor debugging.
• Optional Cryptography engine.
• Optional Generic Interrupt Controller (GIC) CPU interface.
• Support for power management with multiple power domains.

Note
The optional Cryptography engine is not included in the base product of the Cortex-A72 processor. ARM
requires licensees to have contractual rights to obtain the Cortex-A72 processor Cryptography engine.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 1-17
reserved.
Non-Confidential
 1 Introduction
1.4 Interfaces

1.4 Interfaces
The Cortex-A72 processor has the following external interfaces:
• Memory interface that implements either an ACE or CHI interface.
• Optional ACP that implements an AXI slave interface.
• Optional GIC CPU interface that implements an AXI4-Stream interface
• Debug interface that implements an APB slave interface.
• Trace interface that implements an ATB interface.
• PMU interface.
• Generic Timer interface.
• Cross trigger interface.
• Power management interface.
• Design For Test (DFT).
• Memory Built-In Self Test (MBIST).
See 2.2 Interfaces on page 2-29 for more information.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 1-18
reserved.
Non-Confidential
 1 Introduction
1.5 Implementation options

1.5 Implementation options

The following table lists the options that the implementer can choose when implementing the Cortex-
A72 processor in an SoC.

Table 1-1 Cortex-A72 processor implementation options

Feature Range of options

Number of cores 1-4

Cryptography engine Included or Not

L2 cache size 512KB, 1MB, 2MB, or 4MB

L2 Tag RAM register slice 0 or 1

L2 Data RAM register slice 0, 1, or 2

L2 arbitration register slice 0 or 1

L2 FEQ size 20, 24, or 28 entries

Regional gated clock a Included or Not

ECC or parity support Supported in L1 and L2, L2 only, or none.

Bus interface ACE or CHI

ACP Included or Not

GIC CPU interface Included or Not

Note
• All the cores share an integrated L2 cache and optional GIC CPU interface. Each core has the same
configuration for the Cryptography engine and L1 ECC or parity.
• The optional Cryptography engine is not included in the base product of the Cortex-A72 processor.
ARM requires licensees to have contractual rights to obtain the Cortex-A72 processor Cryptography
engine.
• The L2 Tag RAM register slice option adds register slices to the L2 Tag RAMs. The L2 Data RAM
register slice option adds register slices to the L2 Data RAMs. The following table lists valid
combinations of the L2 Tag RAM and L2 Data RAM register slice options.

Table 1-2 Valid combinations of L2 Tag and Data RAM register slice

L2 Tag RAM register slice L2 Data RAM register slice

0 0

0 1

1 1

0 2

1 2
• If the L2 arbitration register slice is included then it adds an additional pipeline stage in the
processor-L2 arbitration logic interface.
• The Cortex-A72 processor must be configured with a CHI interface to connect to a CHI interconnect.

a See Regional clock gating on page 2-50 for more information.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 1-19
reserved.
Non-Confidential
 1 Introduction
1.6 Test features

1.6 Test features

The Cortex-A72 processor provides test signals that enable the use of both Automatic Test Pattern
Generation (ATPG) and MBIST to test the processor and its memory arrays.
See Appendix A Signal Descriptions on page Appx-A-529 for more information.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 1-20
reserved.
Non-Confidential
 1 Introduction
1.7 Product documentation and design flow

1.7 Product documentation and design flow

This section describes the Cortex-A72 processor books and how they relate to the design flow in
documentation and design flow.
See Additional reading on page 11 for more information about the books described in this section. For
information on the relevant architectural standards and protocols, see 1.2 Compliance on page 1-15.
This section contains the following subsections:
• 1.7.1 Documentation on page 1-21.
• 1.7.2 Design flow on page 1-22.

1.7.1 Documentation
The Cortex-A72 processor documentation is as follows:
Technical Reference Manual
The Technical Reference Manual (TRM) describes the functionality and the effects of functional
options on the behavior of the processor. It is required at all stages of the design flow. The
choices made in the design flow can mean that some behavior described in the TRM is not
relevant. If you are programming the multiprocessor, additional information must be obtained
from:
• The implementer to determine the build configuration of the implementation.
• The integrator to determine the pin configuration of the device that you are using.

Note
• The out-of-order design of the Cortex-A72 processor pipeline makes it impossible to provide
accurate timing information for complex instructions. The timing of an instruction can be
affected by factors such as:
— Other concurrent instructions.
— Memory system activity.
— Events outside the instruction flow.
• Timing information has been provided in the past for some ARM processors to assist in the
hand tuning of performance critical code sequences or in the development of an instruction
scheduler within a compiler. This timing information is not required for producing optimized
instruction sequences on the Cortex-A72 processor. The out-of-order pipeline of the
processor can schedule and execute the instructions in an optimal fashion without any
instruction reordering required.

Configuration and Sign-off Guide

The Configuration and Sign-off Guide (CSG) describes:
• The available build configuration options and related issues in selecting them.
• How to configure the Register Transfer Level (RTL) source files with the build configuration
options.
• How to integrate RAM arrays.
• How to run test vectors.
• The processes to sign off the configured design.
The ARM product deliverables include reference scripts and information about using them to
implement your design. Reference methodology flows supplied by ARM are example reference
implementations. For EDA tool support, contact your EDA vendor.
The CSG is a confidential book that is only available to licensees.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 1-21
reserved.
Non-Confidential
 1 Introduction
1.7 Product documentation and design flow

Integration Manual
The Integration Manual (IM) describes how to integrate the processor into an SoC. It describes
the signals that the integrator must tie off to configure the macrocell for the required integration.
Some of the implementation options might affect which integration options are available.
The IM is a confidential book that is only available to licensees.

1.7.2 Design flow

The Cortex-A72 processor is delivered as synthesizable RTL. Before the processor can be used in a
product, it must go through the following process:
Implementation
The implementer configures and synthesizes the RTL to produce a hard macrocell. This might
include integrating the cache RAMs into the design.
Integration
The integrator connects the configured design into a SoC. This includes connecting it to a
memory system and peripherals.
Programming
This is the last process. The system programmer develops the:
• Software to configure the Cortex-A72 processor.
• Software to initialize the Cortex-A72 processor.
• Application software and the SoC tests.
Each process:
• Can be performed by a different party.
• Can include implementation and integration choices that affect the behavior and features of the
processor.
The operation of the final device depends on:
Build configuration
The implementer chooses the options that affect how the RTL source files are preprocessed.
These options usually include or exclude logic that can affect one or more of the area, maximum
frequency, and features of the resulting macrocell.
Configuration inputs
The integrator configures some features of the processor by tying inputs to specific values.
These configurations affect the start-up behavior before any software configuration is made.
They can also limit the options available to the software.
Software configuration
The programmer configures the processor by programming particular values into registers. This
affects the behavior of the processor.

Note
This manual refers to IMPLEMENTATION DEFINED features that apply to build configuration options.
Reference to a feature that is included means that the appropriate build and signal configuration options
have been selected. Reference to an enabled feature means that the feature has also been configured by
software.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 1-22
reserved.
Non-Confidential
 1 Introduction
1.8 Product revisions

1.8 Product revisions

This section describes the differences in functionality between product revisions.
r0p0 First release.
r0p1 Adds support for 4MB L2 RAM.
r0p2 Errata fixes.
r0p3 Errata fixes.
RTL change to improve performance of VA-based data prefetcher on some workloads.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 1-23
reserved.
Non-Confidential
Chapter 2
Functional Description

This chapter describes the functionality of the Cortex-A72 processor.

It contains the following sections:

• 2.1 About the Cortex-A72 processor functions on page 2-25.
• 2.2 Interfaces on page 2-29.
• 2.3 Clocking and resets on page 2-32.
• 2.4 Power management on page 2-42.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-24
reserved.
Non-Confidential
 2 Functional Description
2.1 About the Cortex-A72 processor functions

2.1 About the Cortex-A72 processor functions

The following figure shows a top-level functional diagram of the Cortex-A72 processor.
Trace Message
Debug output TimerEvents Interrupts interrupts

Cortex-A72 processor
APB ATB
AXI4-Stream

Debug Generic GIC

Trace
and CTI Timer CPU Interface†

Miscellaneous

Debug ETM PMU

Rename Integer execute

Virtual to Dispatch Adv SIMD and FP

Decode
physical stages Writeback
register Cryptography†
pool
Load Data Data
store TLB cache
Branch
prediction Instruction
fetch Register
System registers
Ind Pred file

Return Instruction Instruction

L2-TLB
stack cache TLB

Processor arbitration (1st level)

Core 0
Core 1†
Core 2†
Core 3†

L2 arbitration (2nd level)

Fill / Evict L2 data RAM Snoop TAG Auto prefetch

Slave Snoop Master
buffers (512KB to 4MB) RAM engine

Level 2 memory system

AXI ACE or CHI

ACP† Memory †
Depending on the implementation, this feature might not be available
interface

Figure 2-1 Block diagram

This section contains the following subsections:
• 2.1.1 Components of the processor on page 2-26.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-25
reserved.
Non-Confidential
 2 Functional Description
2.1 About the Cortex-A72 processor functions

2.1.1 Components of the processor

The main components of the processor are:
• Instruction fetch.
• Instruction decode.
• Instruction dispatch.
• Integer execute.
• Load/Store unit.
• L2 memory system.
• Advanced SIMD and Floating-point unit.
• GIC CPU interface.
• Generic Timer.
• Debug and trace.

Instruction fetch
The instruction fetch unit fetches instructions from L1 instruction cache and delivers up to three
instructions per cycle to the instruction decode unit. It supports dynamic and static branch prediction.
The instruction fetch unit includes:
• L1 instruction cache that is a 48KB 3-way set-associative cache with a 64-byte cache line and
optional dual-bit parity protection per 32 bits in the Data RAM and 36 bits in the Tag RAM.
• 48-entry fully-associative L1 instruction Translation Lookaside Buffer (TLB) with native support for
4KB, 64KB, and 1MB page sizes.
• 2-level dynamic predictor with Branch Target Buffer (BTB) for fast target generation.
• Static branch predictor.
• Indirect predictor.
• Return stack.

Instruction decode
The instruction decode unit decodes the following instruction sets:
• A32.
• T32.
• A64.
The instruction decode unit supports the A32, T32, and A64 Advanced SIMD and Floating-point
instruction sets. The instruction decode unit also performs register renaming to facilitate out-of-order
execution by removing Write-After-Write (WAW) and Write-After-Read (WAR) hazards.

Instruction dispatch
The instruction dispatch unit controls when the decoded instructions are dispatched to the execution
pipelines and when the returned results are retired. It includes:
• The ARM core general-purpose registers.
• The Advanced SIMD and Floating-point register set.
• The AArch32 CP15 and AArch64 System registers.

Integer execute
The integer execute unit includes:
• Two Arithmetic Logical Unit (ALU) pipelines.
• Integer multiply-accumulate and ALU pipelines.
• Iterative integer divide hardware.
• Branch and instruction condition codes resolution logic.
• Result forwarding and comparator logic.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-26
reserved.
Non-Confidential
 2 Functional Description
2.1 About the Cortex-A72 processor functions

Load/Store unit
The Load/Store (LS) execution unit executes load and store instructions and encompasses the L1 data
side memory system. It also services memory coherency requests from the L2 memory system.
The load/store unit includes:
• L1 data cache that is a 32KB 2-way set-associative cache with a 64-byte cache line and optional
Error Correction Code (ECC) protection per 32 bits.
• 32-entry fully-associative L1 data TLB with native support for 4KB, 64KB, and 1MB page sizes.
• Automatic hardware prefetcher that generates prefetches targeting the L1D cache and the L2 cache.

L2 memory system
The L2 memory system services L1 instruction and data cache misses from each processor. It manages
requests on the AMBA 4 AXI Coherency Extensions (ACE) or CHI master interface and the optional
Accelerator Coherency Port (ACP) slave interface.
The L2 memory system includes:
• L2 cache that is:
— 512KB, 1MB, 2MB, or 4MB configurable size.
— 16-way set-associative cache with optional data ECC protection per 64 bits.
• Duplicate copy of L1 data cache Tag RAMs from each processor for handling snoop requests.
• 4-way set-associative of 1024-entry L2 TLB in each processor.
• Automatic hardware prefetcher with programmable instruction fetch distance.

Related information
Chapter 7 Level 2 Memory System on page 7-298.

Advanced SIMD and Floating-point unit

The Advanced SIMD and Floating-point unit provides support for the ARMv8 Advanced SIMD and
Floating-point execution. In addition, the Advanced SIMD and Floating-point unit provides support for
the optional Cryptography engine.

Note
The optional Cryptography engine is not included in the base product of the Cortex-A72 processor. ARM
requires licensees to have contractual rights to obtain the Cortex-A72 processor Cryptography engine.

Related information
Chapter 14 Advanced SIMD and Floating-point on page 14-511.

GIC CPU interface

The Generic Interrupt Controller (GIC) CPU interface delivers interrupts to the processor.

Related information
Chapter 8 Generic Interrupt Controller CPU Interface on page 8-318.

Generic Timer
The Generic Timer provides the ability to schedule events and trigger interrupts.

Related information
Chapter 9 Generic Timer on page 9-338.

Debug and trace

The debug and trace unit includes:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-27
reserved.
Non-Confidential
 2 Functional Description
2.1 About the Cortex-A72 processor functions

• Support for ARMv8 Debug architecture with an AMBA Advanced Peripheral Bus (APB) slave
interface for access to the debug registers.
• Performance Monitor Unit (PMU) based on the PMUv3 architecture.
• Embedded Trace Macrocell (ETM) based on the ETMv4 architecture and an AMBA Advanced Trace
Bus (ATB) interface for each processor.
• Cross trigger interfaces for core debugging.

Related information
Chapter 10 Debug on page 10-343.
Chapter 11 Performance Monitor Unit on page 11-395.
Chapter 12 Cross Trigger on page 12-434.
Chapter 13 Embedded Trace Macrocell on page 13-457.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-28
reserved.
Non-Confidential
 2 Functional Description
2.2 Interfaces

2.2 Interfaces
This section describes the external interfaces within the Cortex-A72 processor.
This section contains the following subsections:
• 2.2.1 Memory interface on page 2-29.
• 2.2.2 Optional Accelerator Coherency Port on page 2-29.
• 2.2.3 Optional GIC CPU interface on page 2-29.
• 2.2.4 Debug interface on page 2-29.
• 2.2.5 Trace interface on page 2-30.
• 2.2.6 PMU interface on page 2-30.
• 2.2.7 Generic Timer interface on page 2-30.
• 2.2.8 Cross trigger interface on page 2-30.
• 2.2.9 Power management interface on page 2-30.
• 2.2.10 DFT on page 2-30.
• 2.2.11 MBIST on page 2-30.

2.2.1 Memory interface

The processor has a memory interface that implements either an AMBA 4 ACE or CHI bus interface:
• ACE is an extension to the Advanced eXtensible Interface (AXI) protocol and provides the following
enhancements:
— Support for hardware cache coherency.
— Barrier transactions that guarantee transaction ordering.
— Distributed virtual memory messaging, enabling management of a virtual memory system.
See the ARM® AMBA® AXI and ACE Protocol Specification for more information.
• CHI is a protocol that provides an architecture for connecting multiple nodes using a scalable
interconnect. The nodes on the interconnect might be cores, core clusters, I/O bridges, memory
controllers, or graphics processors. See the ARM® AMBA® 5 CHI Protocol Specification.

2.2.2 Optional Accelerator Coherency Port

The processor implements an optional Accelerator Coherency Port (ACP). This is an AMBA 4 AXI
slave interface.
The ACP slave interface supports memory coherent accesses to the Cortex-A72 processor memory
system, but cannot receive coherent requests, barriers, or distributed virtual memory messages.

Related information
10.10 External debug interface on page 10-381.
AMBA AXI and ACE Protocol Specification.

2.2.3 Optional GIC CPU interface

The processor implements an optional GIC CPU interface with an AMBA AXI-4 Stream interface.
See Chapter 8 Generic Interrupt Controller CPU Interface on page 8-318 for more information.

2.2.4 Debug interface

The Cortex-A72 processor implements an AMBA 3 APB slave interface that enables access to the debug
registers.

Related information
10.10 External debug interface on page 10-381.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-29
reserved.
Non-Confidential
 2 Functional Description
2.2 Interfaces

2.2.5 Trace interface

The processor implements dedicated AMBA 3 ATB interfaces for each core that outputs trace
information for debugging. The ATB interface is compatible with the CoreSight architecture.

Related information
13.3 ETM functional description on page 13-461.

2.2.6 PMU interface

The processor implements Performance Monitors and provides an interrupt output and an event interface
for each core.

Related information
Chapter 11 Performance Monitor Unit on page 11-395.

2.2.7 Generic Timer interface

The processor has a global timer input and each core has four timer interrupt outputs.

Related information
Chapter 9 Generic Timer on page 9-338.

2.2.8 Cross trigger interface

The processor implements a single cross trigger channel interface. This external interface is connected to
the CoreSight Cross Trigger Interface (CTI) corresponding to each core through a simplified Cross
Trigger Matrix (CTM).

Related information
Chapter 12 Cross Trigger on page 12-434.

2.2.9 Power management interface

The processor provides Q-channel interfaces that enable an external power controller to control the
retention state of each core and the L2 RAMs.

Related information
Processor dynamic retention on page 2-46.
L2 RAMs dynamic retention on page 2-48.

2.2.10 DFT
The processor implements a Design For Test (DFT) interface that enables an industry-standard
Automatic Test Pattern Generation (ATPG) tool to test logic outside of the embedded memories.

Related information
A.16.1 DFT signals on page Appx-A-563.
ARM Cortex-A72 MPCore Processor Integration Manual.

2.2.11 MBIST
The Memory Built-In Self Test (MBIST) interface provides support for manufacturing testing of the
memories embedded in the Cortex-A72 processor. MBIST is the industry-standard method of testing
embedded memories. MBIST works by performing sequences of reads and writes to the memory based
on test algorithms.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-30
reserved.
Non-Confidential
 2 Functional Description
2.2 Interfaces

Related information
A.16 DFT and MBIST signals on page Appx-A-563.
ARM Cortex-A72 MPCore Processor Integration Manual.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-31
reserved.
Non-Confidential
 2 Functional Description
2.3 Clocking and resets

2.3 Clocking and resets

Clocks and resets used within the processor are described in this section.
This section contains the following subsections:
• 2.3.1 Clocks on page 2-32.
• 2.3.2 Resets on page 2-36.

2.3.1 Clocks
The processor has the following clock inputs:
CLK
This is the main clock of the Cortex-A72 processor. All cores, the shared L2 memory system
logic, the GIC, and the Generic Timer are clocked with a distributed version of CLK.
PCLKDBG
This is the APB clock that controls the Debug APB, CTI, and CTM logic in the PCLKDBG
domain. PCLKDBG is asynchronous to CLK.
The processor has the following clock enable inputs:
ACLKENM
The AXI master interface is a synchronous AXI interface that can operate at any integer
multiple that is equal to or slower than the processor clock, CLK, using the ACLKENM signal.
For example, you can set the CLK to ACLKM frequency ratio to 1:1, 2:1, or 3:1, where
ACLKM is the AXI master clock. ACLKENM asserts one CLK cycle prior to the rising edge
of ACLKM. The CLK to ACLKM frequency ratio can be changed dynamically using
ACLKENM.
The following figure shows a timing example of ACLKENM that changes the CLK to
ACLKM frequency ratio from 3:1 to 1:1.
CLK

ACLKENM

ACLKENM is HIGH, one CLK cycle

before the rising edge of ACLKM
ACLKM
CLK:ACLKM = 3:1
CLK:ACLKM = 1:1

Figure 2-2 ACLKENM with CLK:ACLKM ratio changing from 3:1 to 1:1

Note
• The previous figure shows the timing relationship between the AXI master clock, ACLKM
and ACLKENM, where ACLKENM asserts one CLK cycle before the rising edge of
ACLKM. It is important that the relationship between ACLKM and ACLKENM is
maintained.
• The input signal ACLKENM exists in the processor if it is configured to include the ACE
interface.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-32
reserved.
Non-Confidential
 2 Functional Description
2.3 Clocking and resets

SCLKEN
The CHI interface is a synchronous interface that can operate at any integer multiple that is
equal to or slower than the processor clock, CLK, using the SCLKEN signal. For example, you
can set the CLK to SCLK frequency ratio to 1:1, 2:1, or 3:1, where SCLK is the CHI clock.
SCLKEN asserts one CLK cycle prior to the rising edge of SCLK. The CLK to SCLK
frequency ratio can be changed dynamically using SCLKEN.
The following figure shows a timing example of SCLKEN that changes the CLK to SCLK
frequency ratio from 3:1 to 1:1.
CLK

SCLKEN

SCLKEN is HIGH, one CLK cycle

before the rising edge of SCLK
SCLK
CLK:SCLK = 3:1
CLK:SCLK = 1:1

Figure 2-3 SCLKEN with CLK:SCLK ratio changing from 3:1 to 1:1

Note
• The previous figure shows the timing relationship between the CHI clock, SCLK and
SCLKEN, where SCLKEN asserts one CLK cycle before the rising edge of SCLK. It is
important that the relationship between SCLK and SCLKEN is maintained.
• The input signal SCLKEN exists in the processor if it is configured to include the CHI
interface.

ACLKENS
ACP is a synchronous AXI slave interface that can operate at any integer multiple that is equal
to or slower than the processor clock, CLK, using the ACLKENS signal. For example, the
CLK to ACLKS frequency ratio can be 1:1, 2:1, or 3:1, where ACLKS is the AXI slave clock.
ACLKENS asserts one CLK cycle before the rising edge of ACLKS. The CLK to ACLKS
frequency ratio can be changed dynamically using ACLKENS.
The following figure shows a timing example of ACLKENS that changes the CLK to ACLKS
frequency ratio from 3:1 to 1:1.
CLK

ACLKENS

ACLKENS is HIGH, one CLK cycle

before the rising edge of ACLKS
ACLKS
CLK:ACLKS = 3:1
CLK:ACLKS = 1:1

Figure 2-4 ACLKENS with CLK:ACLKS ratio changing from 3:1 to 1:1

Note
The previous figure shows the timing relationship between the ACP clock, ACLKS and
ACLKENS, where ACLKENS asserts one CLK cycle before the rising edge of ACLKS. It is
important that the relationship between ACLKS and ACLKENS is maintained.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-33
reserved.
Non-Confidential
 2 Functional Description
2.3 Clocking and resets

PCLKENDBG
The Debug APB interface is an asynchronous interface that can operate at any integer multiple
that is equal to or slower than the APB clock, PCLKDBG, using the PCLKENDBG signal. For
example, the PCLKDBG to internal PCLKDBG frequency ratio can be 1:1, 2:1, or 3:1.
PCLKENDBG asserts one PCLKDBG cycle before the rising edge of the internal
PCLKDBG. The PCLKDBG to internal PCLKDBG frequency ratio can be changed
dynamically using PCLKENDBG.
The following figure shows a timing example of PCLKENDBG that changes the PCLKDBG
to internal PCLKDBG frequency ratio from 2:1 to 1:1.

PCLKDBG

PCLKENDBG
PCLKENDBG is HIGH, one
PCLKDBG cycle before the rising edge
of internal PCLKDBG
internal PCLKDBG

PCLKDBG:internal PCLKDBG = 2:1 PCLKDBG:internal PCLKDBG = 1:1

Figure 2-5 PCLKENDBG with PCLKDBG:internal PCLKDBG ratio changing from 2:1 to 1:1
ATCLKEN
The ATB interface is a synchronous interface that can operate at any integer multiple that is
slower than the processor clock, CLK, using the ATCLKEN signal. For example, the CLK to
ATCLK frequency ratio can be 2:1, 3:1, or 4:1, where ATCLK is the ATB bus clock.
ATCLKEN asserts three CLK cycles before the rising edge of ATCLK. Three CLK cycles are
required to allow propagation delay from the ATCLKEN input to the processor. The CLK to
ATCLK frequency ratio can be changed dynamically using ATCLKEN.
The following figure shows a timing example of ATCLKEN where the CLK to ATCLK
frequency ratio is 2:1.
3 CLK cycles
CLK

ATCLKEN
ATCLKEN asserts three CLK cycles
before the rising edge of ATCLK
ATCLK

CLK:ATCLK = 2:1

Figure 2-6 ATCLKEN with CLK:ATCLK ratio at 2:1

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-34
reserved.
Non-Confidential
 2 Functional Description
2.3 Clocking and resets

CNTCLKEN
The CNTVALUEB is a synchronous 64-bit binary encoded counter value that can operate at
any integer multiple that is equal to or slower than the processor clock, CLK, using the
CNTCLKEN signal. For example, you can set the CLK to CNTCLK frequency ratio to 1:1,
2:1, or 3:1, where CNTCLK is the system counter clock. CNTCLKEN asserts one CLK cycle
prior to the rising edge of CNTCLK.
The following figure shows a timing example of CNTCLKEN where the CLK to CNTCLK
frequency ratio is 2:1.

CLK

CNTCLKEN
CNTCLKEN is HIGH, one CLK cycle
before the rising edge of CNTCLK
CNTCLK

CLK:CNTCLK = 2:1

Figure 2-7 CNTCLKEN with CLK:CNTCLK ratio at 2:1

CLKEN
This is the clock enable for all internal clocks in the processor that are derived from CLK. The
CLKEN signal must be asserted at least one cycle before applying CLK to the processor.
When all the cores and L2 are in WFI low-power state, you can place the processor in a low-
power state using the CLKEN input. Setting CLKEN LOW disables all of the internal clocks,
excluding the asynchronous Debug APB PCLKDBG domain.

Related information
L2 Wait for Interrupt on page 2-44.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-35
reserved.
Non-Confidential
 2 Functional Description
2.3 Clocking and resets

2.3.2 Resets
The Cortex-A72 processor has the following reset inputs:
nCPUPORESET[N:0]
Initializes the entire core logic, including Debug, ETM, breakpoint and watchpoint logic in the
processor CLK domain. Each core has one nCPUPORESET reset input.
nCORERESET[N:0]
Initializes the entire core but excludes the Debug, ETM, breakpoint and watchpoint logic. Each
core has one nCORERESET reset input.
nPRESETDBG
Initializes the shared Debug APB, CTI, and CTM logic in the PCLKDBG domain.
nL2RESET
Initializes the shared L2 memory system, GIC, and Timer logic.
nMBISTRESET
Performs an MBIST mode reset.
All resets are active-LOW inputs. The reset signals enable you to reset different areas of the processor
independently. The following table shows the areas of the processor controlled by the various reset
signals.

Table 2-1 Areas that the reset signals control

Reset signal Coreb (CLK) Debug and Debug APB, CTI, and L2 memory system, Individual processor
ETM (CLK) CTM (PCLKDBG) shared GIC and GIC and Timer logic
Timer logic

nCPUPORESET Reset Reset - - Reset

nCORERESET Reset - - - Reset

nPRESETDBG - - Reset - -

nL2RESET - - - Reset Reset

The following table shows the valid reset combinations the processor supports. The core which is being
reset is indicated by [n].

Table 2-2 Valid reset combinations

Reset combination Signals Value Description

Full powerup reset for the nCPUPORESET[N:0] all = 0 All logic is held in reset.
processor
nCORERESET[N:0] all = 0
nPRESETDBG 0
nL2RESET 0
nMBISTRESET 1

Individual core powerup reset nCPUPORESET[N:0] [n] = 0 Individual core in the CLK domain and Debug in the PCLKDBG
with Debug (PCLKDBG) domain are held in reset, so that the core and Debug PCLKDBG
nCORERESET[N:0] [n] = 0c
reset domain can be powered up.
nPRESETDBG 0
nL2RESET 1
nMBISTRESET 1

b Core logic, excluding Debug, ETM, breakpoint, and watchpoint logic.

c For powerup reset or core reset, nCPUPORESET must be asserted. nCORERESET can be asserted, but is not required.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-36
reserved.
Non-Confidential
 2 Functional Description
2.3 Clocking and resets

Table 2-2 Valid reset combinations (continued)

Reset combination Signals Value Description

All core and L2 reset with nCPUPORESET[N:0] all = 0 All cores and L2 are held in reset, so they can be powered up. This
Debug (PCLKDBG) active enables external debug over powerdown for all cores.
nCORERESET[N:0] all = 0c
nPRESETDBG 1
nL2RESET 0
nMBISTRESET 1

Individual core powerup reset nCPUPORESET[N:0] [n] = 0 Individual core is held in reset, so that the core can be powered up.
with Debug (PCLKDBG) This enables external debug over powerdown for the processor
nCORERESET[N:0] [n] = 0c
active that is held in reset.
nPRESETDBG 1
nL2RESET 1
nMBISTRESET 1

All cores Warm reset nCPUPORESET[N:0] all = 1 All logic, excluding Debug and ETM (CLK and PCLKDBG) and
L2, is held in reset. All breakpoints and watchpoints are retained.
nCORERESET[N:0] all = 0
nPRESETDBG 1
nL2RESET 1
nMBISTRESET 1

All cores Warm reset and L2 nCPUPORESET[N:0] all = 1 All logic, excluding Debug and ETM (CLK and PCLKDBG), is
reset held in reset. All breakpoints and watchpoints are retained.
nCORERESET[N:0] all = 0
nPRESETDBG 1
nL2RESET 0
nMBISTRESET 1

Individual core Warm reset nCPUPORESET[N:0] [n] = 1 Individual core logic, excluding the ETM and Debug in the CLK
domain, is held in reset. Breakpoints and watchpoints for that core
nCORERESET[N:0] [n] = 0
are retained.
nPRESETDBG 1
nL2RESET 1
nMBISTRESET 1

Debug (PCLKDBG) reset nCPUPORESET[N:0] all = 1 Debug in the PCLKDBG domain is held in reset, so that the
Debug PCLKDBG domain can be powered up.
nCORERESET[N:0] all = 1
nPRESETDBG 0
nL2RESET 1
nMBISTRESET 1

Run mode nCPUPORESET[N:0] 1 No logic is held in reset.

nCORERESET[N:0] 1
nPRESETDBG 1
nL2RESET 1
nMBISTRESET 1

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-37
reserved.
Non-Confidential
 2 Functional Description
2.3 Clocking and resets

Note
• nL2RESET resets the shared L2 memory system logic, GIC, and Generic Timer that is common to
all cores. This reset must not assert while any individual processor is active.
• nPRESETDBG resets the shared Debug, PCLKDBG, that is common to all cores. This reset must
not assert while any individual core is actively being debugged in normal operating mode or during
external debug over powerdown.

There are specific requirements that you must meet to reset each reset area listed in Table 2-1 Areas that
the reset signals control on page 2-36. Not adhering to these requirements can lead to a reset area that is
not functional.
The reset sequences in the following sections are the only reset sequences that ARM recommends. Any
deviation from these sequences might cause an improper reset of that reset domain. The supported reset
sequences are:
• Powerup reset on page 2-38.
• Warm reset on page 2-39.
• Debug PCLKDBG reset on page 2-39.
• WARMRSTREQ and DBGRSTREQ on page 2-40.
• Memory arrays reset on page 2-40.

Powerup reset

Powerup reset is also known as Cold reset. This section describes the sequence for:
• A full powerup reset.
• An individual core powerup reset.
The full powerup reset initializes all logic in the processor. You must apply powerup reset to the
processor when power is first applied to the SoC. Logic in all clock domains are placed in a benign state
following the deassertion of the reset sequence.
The following figure shows the full powerup reset sequence for the Cortex-A72 processor.

CLK

nCPUPORESET[N:0] 16 CLK cycles minimum

nCORERESET[N:0]
16 CLK cycles minimum
optional
nL2RESET 16 CLK cycles minimum

nPRESETDBG 16 PCLKDBG cycles minimum

Figure 2-8 Powerup reset timing

On full powerup reset for the processor, perform the following reset sequence:
1. Apply nCPUPORESET, nL2RESET, and nPRESETDBG. The remaining core reset,
nCORERESET can assert, but is not required.
2. nCPUPORESET and nL2RESET must assert for at least 16 CLK cycles. nPRESETDBG must
assert for at least 16 PCLKDBG cycles. Holding the resets for this duration ensures that the resets
propagate to all locations within the processor.
3. nL2RESET must deassert in the same cycle as the core resets, or before any of the core resets
deassert.
Individual core powerup reset initializes all logic in a single core. You must apply the powerup reset
when the individual core is being powered up, so that power to the core can be safely applied. You must
apply the correct sequence before applying a powerup reset to that core.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-38
reserved.
Non-Confidential
 2 Functional Description
2.3 Clocking and resets

For individual core powerup reset:

• nCPUPORESET for that core must assert for at least 16 CLK cycles.
• nL2RESET must not assert while any individual core is active.
• nPRESETDBG must not assert while any individual core is actively being debugged in normal
operating mode or during external debug over powerdown.

Note
If core dynamic retention using the CPU Q-channel interface is used, the core must be in quiescent state
with STANDBYWFI asserted and CPUQREQn, CPUQACCEPTn, and CPUQACCEPT must be
LOW before nCPUPORESET is applied.

Warm reset

The Warm reset initializes all logic in the individual core apart from the Debug and ETM logic in the
CLK domain. All breakpoints and watchpoints are retained during a Warm reset sequence.
The following figure shows the Warm reset sequence for the Cortex-A72 processor.

CLK

nCPUPORESET[N:0]

nL2RESET

nPRESETDBG
16 CLK cycles minimum
nCORERESET[N:0]

Figure 2-9 Warm reset timing

Individual core Warm reset initializes all logic in a single core apart from its Debug, ETM, breakpoint,
and watchpoint logic. Breakpoints and watchpoints for that core are retained. You must apply the correct
sequence before applying Warm reset to that core.
For individual processor Warm reset:
• You must apply steps 1 on page 2-53 to 12 on page 2-53 in the core powerdown sequence, see
Individual core powerdown on page 2-52, and wait until STANDBYWFI is asserted, indicating that
the core is idle, before asserting nCORERESET for that core.
• nCORERESET for that core must assert for at least 16 CLK cycles.
• nL2RESET must not assert while any individual core is active.
• nPRESETDBG must not assert while any individual core is actively being debugged in normal
operating mode.

Note
If core dynamic retention using the CPU Q-channel interface is used, the core must be in quiescent state
with STANDBYWFI asserted and CPUQREQn, CPUQACCEPTn, and CPUQACCEPT must be
LOW before nCORERESET is applied.

Debug PCLKDBG reset

Use nPRESETDBG to reset the Debug APB, CTI, and CTM logic in the PCLKDBG domain. This
reset holds the Debug PCLKDBG unit in a reset state so that the power to the unit can be safely applied
during powerup.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-39
reserved.
Non-Confidential
 2 Functional Description
2.3 Clocking and resets

To safely reset the Debug PCLKDBG unit, nPRESETDBG must assert for a minimum of 16
PCLKDBG cycles.
The following figure shows the Debug PCLKDBG reset sequence.

PCLKDBG

nCPUPORESET[N:0]

nCORERESET[N:0]

nL2RESET
16 PCLKDBG cycles minimum
nPRESETDBG

Figure 2-10 Debug PCLKDBG reset timing

WARMRSTREQ and DBGRSTREQ

The ARMv8-A architecture provides a mechanism to configure whether a processor uses AArch32 or
AArch64 at EL3 as a result of a Warm reset. When the Reset Request bit in the RMR or RMR_EL3
register is set to 1, the processor asserts the WARMRSTREQ signal and the SoC reset controller can use
this request to trigger a Warm reset of the core and change the register width state. The AA64 bit in the
RMR or RMR_EL3 register selects the register width at the next Warm reset, at the highest Exception
level, EL3.
See the ARM® Architecture Reference Manual ARMv8, for ARMv8-A architecture profile for information
about the recommended code sequence to use, to request a Warm reset.
You must apply steps 1 on page 2-53 to 12 on page 2-53 in the core powerdown sequence, and wait
until STANDBYWFI asserts indicating the processor is idle, before asserting nCORERESET for that
core. nCORERESET must satisfy the timing requirements described in the Warm reset section.
The Core Warm Reset Request (CWRR) bit in the External Debug Power/Reset Control Register,
EDPRCR, controls the DBGRSTREQ signal. An external debugger can use this bit to request a Warm
reset of the processor, if it does not have access to the core Warm reset signal. See the ARM® Architecture
Reference Manual ARMv8, for ARMv8-A architecture profile for more information about the EDPRCR.

Related information
4.3.61 Reset Management Register, EL3 on page 4-175.
Individual core powerdown.
Warm reset on page 2-39.

Memory arrays reset

During a core reset, the following memory arrays in the core are invalidated:
• Branch Prediction arrays such as BTB, GHB, and Indirect Predictor.
• L1 instruction and data TLBs.
• L1 instruction and data caches.
• L2 unified TLB.
In addition to these core memory arrays, during a powerup reset, the following shareable memory arrays
are invalidated:
• L2 duplicate Snoop Tag RAM.
• L2 unified cache RAM, if L2RSTDISABLE is tied LOW.
The L1 instruction and data cache resets can take up to 128 CLK cycles after the deasserting edge of the
reset signals, with each array being reset in parallel. Depending on the size of the L2 cache, the L2 cache

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-40
reserved.
Non-Confidential
 2 Functional Description
2.3 Clocking and resets

reset can take 640 CLK cycles for a 512KB L2 cache or 5120 CLK cycles for a 4MB L2 cache. The L2
cache reset occurs in the background, in parallel with the L1 cache resets. The core can begin execution
in Non-cacheable state, but any attempt to perform Cacheable transactions stalls the core until the
appropriate cache reset is complete.
The branch prediction arrays require 512 CLK cycles to reset after the deasserting edge of reset. The
core begins execution with branch prediction disabled, any resolved branches do not update the branch
predictor until the reset sequence completes.
The processor input signal, L2RSTDISABLE, controls the L2 cache hardware reset process. The usage
models for the L2RSTDISABLE signal are as follows:
• When the processor powers up for the first time, L2RSTDISABLE must be held LOW to invalidate
the L2 cache using the L2 cache hardware reset mechanism.
• For systems that do not retain the L2 cache RAM contents while the L2 memory system is powered
down, L2RSTDISABLE must be held LOW to invalidate the L2 cache using the L2 cache hardware
reset mechanism.
• For systems that retain the L2 cache RAM contents while the L2 memory system is powered down,
L2RSTDISABLE must be held HIGH to disable the L2 cache hardware reset mechanism.
The L2RSTDISABLE signal is sampled during nL2RESET assertion and must be held a minimum of
32 CLK cycles after the deasserting edge of nL2RESET.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-41
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

2.4 Power management

The Cortex-A72 processor provides mechanisms and support to control both dynamic and static power
dissipation.
This section contains the following subsections:
• 2.4.1 Dynamic power management on page 2-42.
• 2.4.2 Power domains on page 2-51.
• 2.4.3 Power modes on page 2-52.
• 2.4.4 Using SMPEN as a power mode indicator on page 2-57.

2.4.1 Dynamic power management

This section describes the following dynamic power management features in the processor:
• Core Wait for Interrupt on page 2-42.
• Core Wait for Event on page 2-43.
• Event communication using WFE and SEV instructions on page 2-43.
• CLREXMON request and acknowledge signaling on page 2-43.
• L2 Wait for Interrupt on page 2-44.
• L2 hardware cache flush on page 2-45.
• Processor dynamic retention on page 2-46.
• L2 RAMs dynamic retention on page 2-48.
• Advanced SIMD and FP clock gating on page 2-49.
• L2 control and tag banks clock gating on page 2-49.
• Regional clock gating on page 2-50.

Core Wait for Interrupt

Wait for Interrupt (WFI) is a feature of the ARMv8-A architecture that puts the core in a low-power state
by disabling the clocks in the core while keeping the core powered up. This reduces the power drawn to
the static leakage when the core is in WFI low-power state.
A core enters into WFI low-power state by executing the WFI instruction.
When executing the WFI instruction, the core waits for all instructions in the core to retire before entering
the idle or low-power state. The WFI instruction ensures that all explicit memory accesses occurred
before the WFI instruction in program order, have retired. For example, the WFI instruction ensures that
the following instructions receive the required data or responses from the L2 memory system:
• Load instructions.
• Cache and TLB maintenance operations.
• Store-Exclusive instructions.
In addition, the WFI instruction ensures that store instructions update the cache or are issued to the L2
memory system.
While the core is in WFI low-power state, the clocks in the core are temporarily enabled without causing
the core to exit WFI low-power state, when any of the following events are detected:
• An L2 snoop request that must be serviced by the core L1 data cache.
• A cache, TLB, or BTB maintenance operation that must be serviced by the core L1 instruction cache,
data cache, instruction TLB, data TLB, or BTB.
• An APB access to the debug or trace registers residing in the core power domain.
The core exits from WFI low-power state when it detects a reset or a WFI wake-up event occurs. See the
ARM® Architecture Reference Manual ARMv8, for ARMv8-A architecture profile for information about
the various WFI wake-up events.
On entry into WFI low-power state, STANDBYWFI for that core is asserted. STANDBYWFI continues
to assert even if the clocks in the core are temporarily enabled because of an L2 snoop request, cache,
TLB, and BTB maintenance operation or an APB access.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-42
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

Core Wait for Event

Wait for Event (WFE) is a feature of the ARMv8-A architecture that uses a locking mechanism based on
events to put the core in a low-power state by disabling the clocks in the core while keeping the core
powered up. This reduces the power drawn to the static leakage current, when the core is in WFE low-
power state.
A core enters into WFE low-power state by executing the WFE instruction. When executing the WFE
instruction, the core waits for all instructions in the core to complete before entering the idle or low-
power state. The WFE instruction ensures that all explicit memory accesses occurred before the WFE
instruction in program order, have completed.
While the core is in WFE low-power state, the clocks in the core are temporarily enabled without causing
the core to exit WFE low-power state, when any of the following events are detected:
• An L2 snoop request that must be serviced by the core L1 data cache.
• A cache, TLB, or BTB maintenance operation that must be serviced by the core L1 instruction cache,
data cache, instruction TLB, data TLB, or BTB.
• An APB access to the debug or trace registers residing in the core power domain.
The cores exits from WFE low-power state when:
• It detects a reset.
• The EVENTI input signal asserts.
• The CLREXMONREQ input signal asserts.
• A WFE wake-up event occurs. See the ARM® Architecture Reference Manual ARMv8, for ARMv8-A
architecture profile for information about the various WFE wake-up events.
On entry into WFE low-power state, STANDBYWFE for that core is asserted. STANDBYWFE
continues to assert even if the clocks in the core are temporarily enabled because of an L2 snoop request,
cache, TLB, and BTB maintenance operation or an APB access.

Event communication using WFE and SEV instructions

The EVENTI signal enables an external agent to participate in the WFE and SEV event communication.
When this signal is asserted, it sends an event message to all the cores in the processor. This is similar to
executing an SEV instruction on one core in the processor. This enables the external agent to signal to the
core that it has released a semaphore and that the core can leave the WFE low-power state. The EVENTI
input signal must remain HIGH for at least one CLK cycle to be visible by the cores.
The external agent can determine that at least one of the cores in the processor has executed an SEV
instruction by checking the EVENTO signal. When any of the cores in the processor executes an SEV
instruction, an event is signaled to all the cores in the processor, and the EVENTO signal is asserted.
This signal is asserted HIGH for three CLK cycles when any of the cores executes an SEV instruction.

CLREXMON request and acknowledge signaling

The CLREXMONREQ signal has a corresponding CLREXMONACK response signal. This forms a
standard 2-wire, 4-phase handshake that can be used to signal across the voltage and frequency boundary
between the core and system.
When the CLREXMONREQ input is asserted, it signals the clearing of an external global exclusive
monitor and acts as WFE wake-up event to all the cores in the processor.
The following figure shows the CLREXMON request and acknowledge handshake. When the request
signal is asserted, it continues to assert until an acknowledge is received. When the request is deasserted,
the acknowledge can then deassert.
Note
If a global exclusive monitor does not exist in your system, tie the CLREXMONREQ input LOW.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-43
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

CLREXMONREQ

CLREXMONACK

Figure 2-11 CLREXMON request and acknowledge handshake

L2 Wait for Interrupt

When all the cores are in WFI low-power state, the shared L2 memory system logic that is common to all
the cores can also enter a WFI low-power state.
Entry into L2 WFI low-power state can only occur if specific requirements are met and the following
sequence applied:
1. All cores are in the WFI low-power state, so all the core STANDBYWFI outputs are asserted.
2. When ACP is present and all outstanding ACP requests are complete, the SoC asserts the AINACTS
input to idle the ACP slave interface. When AINACTS has been asserted, the SoC must not assert
ARVALIDS, AWVALIDS, or WVALIDS.
3. If the processor implements:
An ACE interface
When all outstanding snoop requests are complete, the SoC asserts the ACINACTM input
signal to idle the AXI master snoop interface. This prevents the L2 memory system from
accepting any new requests from the AXI master snoop interface. When ACINACTM has
been asserted, the SoC must not assert ACVALIDM.
A CHI interface
When all outstanding snoop requests are complete, the SoC asserts the SINACT input signal
indicating that the processor is removed from the coherency domain and does not receive
any more snoops. This triggers the L2 to deactivate the TX and RX links. When the TX and
RX links are in their respective stop states, the L2 memory system does not accept any new
requests from the CHI interface.
4. When the L2 memory system completes the outstanding transactions for ACE and CHI interfaces, it
can then enter the L2 WFI low-power state. On entry into L2 WFI low-power state,
STANDBYWFIL2 is asserted. Assertion of STANDBYWFIL2 guarantees that the L2 is idle and
does not accept any new transactions.
5. The SoC can then choose to deassert the CLKEN input to the processor to stop all remaining internal
clocks within the core that are derived from CLK. All clocks in the shared L2 memory system logic,
GIC, and Timer, are disabled.
If CLKEN is deasserted, the SoC must assert the CLKEN input on a WFI wake-up event to enable the
L2 memory system and potentially the core. There are two classes of wake-up events:
• An event that requires only the L2 memory system to be enabled.
• An event that requires both the L2 memory and the core to be enabled.
The following wake-up events cause both the L2 memory system and the core to exit WFI low-power
state:
• A physical IRQ or FIQ interrupt.
• A debug event.
• Powerup or Warm reset.
The following wake-up events cause only the L2 memory system to exit WFI low-power state:
• If the device is configured to have an ACE interface, deassertion of ACINACTM to service an
external snoop request on the AXI master snoop interface.
• If the device is configured to have a CHI interface:
— Deassertion of SINACT to service an external snoop request.
— Activation of TX or RX links.
• If ACP is present, deassertion of AINACTS to service an ACP transaction on the slave interface.
When the core exits from WFI low-power state STANDBYWFI for that core is deasserted. When the L2
memory system logic exits from WFI low-power state, STANDBYWFIL2 is deasserted.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-44
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

The following figure shows the L2 WFI timing for a 4-core configuration.

CLK

STANDBYWFI[3:0]

ACINACTM

AINACTS

STANDBYWFIL2

CLKEN

Internal L2 clock

nIRQ

Figure 2-12 L2 Wait For Interrupt timing

L2 hardware cache flush

The processor provides an efficient way to fully clean and invalidate the L2 cache in preparation for
powering it down without requiring the waking of a core to perform the clean and invalidate through
software.
Use of L2 hardware cache flush can only occur if specific requirements are met and the following
sequence applied:
1. Disable L2 prefetches by writing a one to bit [38] and zeros to bits [36:35] of the CPU Extended
Control Register.
2. Disable the load-store hardware prefetcher by writing a one to bit [56] of the CPU Auxiliary Control
Register.
3. Execute an ISB instruction to ensure the CPU Extended Control Register and CPU Auxiliary Control
Register writes are complete.
4. Execute a DSB instruction to ensure completion of any prior prefetch requests.
5. All cores are in the WFI low-power state, so all the core STANDBYWFI outputs are asserted.
6. When ACP is present and all outstanding ACP transactions are complete, the SoC asserts the
AINACTS signal to idle the ACP. This is necessary to prevent ACP transactions from allocating new
entries in the L2 cache while the hardware cache flush is occurring. When AINACTS has been
asserted, the SoC must not assert ARVALIDS, AWVALIDS, or WVALIDS.
7. The SoC can now assert the L2FLUSHREQ input.
8. The L2 performs a series of internal clean and invalidate operations to each set and way of the L2
cache. Any dirty cache lines are written back to the system using WriteBack or WriteNoSnoop
operations. Clean cache lines can cause Evict or WriteEvict transactions if the L2 is configured.
9. When the L2 completes the clean and invalidate sequence, it asserts the L2FLUSHDONE signal.
The SoC can now deassert L2FLUSHREQ signal and then the L2 deasserts L2FLUSHDONE.
10. When all outstanding snoop transactions are completed, the SoC can assert the ACINACTM signal
in an ACE implementation or the SINACT signal in a CHI implementation. In response, the L2
asserts the STANDBYWFIL2 signal.
It is possible to terminate the L2 hardware cache flush by deasserting the L2FLUSHREQ signal before
the L2FLUSHDONE signal is asserted. This causes the L2 to abort the hardware cache flush. This
feature can be used when the SoC does not power down the core and must wake up the core quickly.
The following figure shows the L2 hardware cache flush timing.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-45
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

CLK

STANDBYWFI[N:0]

AINACTS

L2FLUSHREQ

L2 flush

L2FLUSHDONE

ACINACTM

STANDBYWFIL2

Figure 2-13 L2 hardware cache flush timing

Related information
4.3.67 CPU Extended Control Register, EL1 on page 4-206.

Processor dynamic retention

When a core is in WFI low-power state or WFE low-power state, the clocks to the core are stopped.
During these low-power states, the core might start the clocks for short periods of time to allow it to
handle snoops or other short events but it remains in the low-power state.
Whenever the clocks to a core are stopped, it is possible for an external power controller to place the core
in a retention state to reduce leakage power consumption without state loss.
Each core in the processor has a CPU Q-channel interface that allows an external power controller to
place the core into a retention state. This interface consists of four pins:
• CPUQACTIVE.
• CPUQREQn.
• CPUQACCEPTn.
• CPUQDENY.
The operational relationship of these signals are:
• CPUQREQn can only go LOW, if CPUQACCEPTn is HIGH and CPUQDENY is LOW.
• After CPUQREQn goes LOW, it must remain LOW until either CPUQACCEPTn goes LOW or
CPUQDENY goes HIGH.
• CPUQREQn can then go HIGH, and must remain HIGH until both CPUQACCEPTn is HIGH and
CPUQDENY is LOW.
• Each CPUQREQn request is followed by the assertion of either CPUQACCEPTn or CPUQDENY,
but not both. CPUQACCEPTn cannot be asserted LOW at the same time as CPUQDENY is
asserted HIGH.
A typical sequence of the external power controller successfully placing the core in retention state is:
1. The core executes a WFI instruction. The clocks in the core are stopped and STANDBYWFI is
asserted. After the programmed number of Generic Timer CNTVALUEB ticks specified by
CPUECTLR[2:0] field has elapsed, the CPUQACTIVE for that core is deasserted. This hints that
retention is possible for that core.
2. The external power controller asserts CPUQREQn to indicate that it wants to put that core into
retention state.
3. While the core is still in WFI low-power state and the clocks are stopped, the core accepts the
retention request by asserting CPUQACCEPTn.
4. While CPUQREQn and CPUQACCEPTn are both asserted, the core is in quiescent state and the
external power controller can safely put the core into retention state.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-46
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

5. During retention, if a snoop occurs to access the cache of the quiescent core, the CPUQACTIVE
signal is asserted to request exit from retention.
6. The external power controller brings the core out of retention and deasserts CPUQREQn.
7. The core deasserts CPUQACCEPTn to complete the handshake.
8. The clocks in the core are restarted temporarily to allow the snoop request to the core to proceed.
9. After the snoop access is complete, the core deasserts CPUQACTIVE.
10. CPUQREQn and CPUQACCEPTn are then asserted. The core has reentered quiescent state and the
external power controller can put the core into retention state again.
11. When the core is ready to exit WFI low-power state, CPUQACTIVE is asserted.
12. CPUQREQn is then deasserted, the core exits WFI low-power state, and CPUQACCEPTn is
deasserted.
The following figure shows a typical sequence where the external power controller successfully places
the core in retention state.

CLK

CPUQACTIVE

CPUQREQn

CPUQACCEPTn

CPUQDENY

STANDBYWFI
retention retention

Figure 2-14 Successful retention timing

The core enters WFI low-power state and deasserts CPUQACTIVE. The external power controller
asserts CPUQREQn. If the core cannot safely enter quiescent state, it asserts CPUQDENY instead of
CPUQACCEPTn. When this occurs, the external power controller cannot put that core into retention
state. The external power controller must then deassert CPUQREQn, then the core deasserts
CPUQDENY.
The following figure shows a sequence where the external power controller attempts to put a core in
retention state but the core denies the request.

CLK

CPUQACTIVE

CPUQREQn

CPUQACCEPTn

CPUQDENY

STANDBYWFI

Figure 2-15 Denied retention timing

Guidelines on the use of core dynamic retention

As cores generally only stay in WFE low-power state for a short period of time, ARM recommends that
you only take a core into retention when it is in WFI low-power state.
If the L1 data cache of a core that is in WFI low-power state contains data that is likely to be the target of
frequent snoops from other cores, entering quiescent state and retention is likely to be inefficient.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-47
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

When using the core retention feature, you must consider the following points:
• During core reset, CPUQREQn must be deasserted HIGH while CPUQACCEPTn is asserted
LOW.
• The Processor dynamic retention control field in the CPU Extended Control Register, CPUECTLR,
must be set to a nonzero value to enable this feature. If this field is 0b000, all assertions of
CPUQREQn LOW receive CPUQDENY responses.
• If the core dynamic retention feature is not used, CPUQREQn must be tied HIGH and the
CPUECTLR retention control field set to disabled.

Note
If you use the core dynamic retention feature then the CPU Auxiliary Control Register,
CPUACTLR[30:29] bits must be zero.

Related information
4.3.66 CPU Auxiliary Control Register, EL1 on page 4-194.
4.3.67 CPU Extended Control Register, EL1 on page 4-206.

L2 RAMs dynamic retention

L2 RAM dynamic retention mode provides a way of saving power in an idle processor while allowing
quick wake-up to service a snoop from ACE or CHI. The core supports dynamic retention of the L2
Data, Dirty, Tag, Inclusion PLRU, and Snoop Tag RAMs.
The processor has an L2 Q-channel interface that allows an external power controller to place the L2
RAMs into a retention state.
L2 RAM dynamic retention mode is entered and exited using the following sequence of events:
1. All cores are in WFI or WFE low-power state and therefore, all the cores STANDBYWFI or
STANDBYWFE outputs are asserted.
2. When all pending L2 activity is complete, and the L2 remains idle for the programmed number of
Generic Timer CNTVALUEB ticks, as specified by L2ECTLR[2:0] field, the L2 deasserts
L2QACTIVE.
3. The external power controller asserts L2QREQn to indicate that it wants to put the L2 RAMs into
retention state.
4. If the L2 is still idle, it accepts the retention request by asserting L2QACCEPTn.
5. While L2QREQn and L2QACCEPTn are both asserted, the power controller can safely put the L2
RAMs into retention state.
6. If the L2 detects that one or more cores have exited WFI low-power state, the ACP becomes active or
a snoop request must be serviced, the L2 asserts L2QACTIVE to request exit from retention.
7. The power controller brings the L2 RAMs out of retention and deasserts L2QREQn.
8. The L2 deasserts L2QACCEPTn to complete the handshake.
The following figure shows the L2 dynamic retention timing.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-48
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

retention

STANDBYWFI[N:0]

L2 idle

L2QACTIVE

L2QREQn

L2QACCEPTn

L2QDENY

Figure 2-16 L2 dynamic retention timing

If the L2 exits idle in step 4 on page 2-48, it asserts L2QDENY instead of L2QACCEPTn. In response,
the power controller must deassert L2QREQn, causing the L2 to deassert L2QDENY.
The L2 dynamic retention control field in the L2 Extended Control Register, L2ECTLR, must be set to a
nonzero value to enable this feature. If this field is 0b000, all assertions of L2QREQn LOW receive
L2QDENY responses.
If the L2 dynamic retention feature is not used, L2QREQn must be tied HIGH and the L2ECTLR
retention control field set to disabled.
Note
If you use the L2 dynamic retention feature then the L2 Auxiliary Control Register, L2ACTLR[28:27]
bits must be zero.

Related information
4.3.59 L2 Extended Control Register, EL1 on page 4-172.
L2 Auxiliary Control Register, EL1.

Advanced SIMD and FP clock gating

The processor supports dynamic high-level clock gating of the Advanced SIMD and FP unit to reduce
dynamic power dissipation.
The clock to the Advanced SIMD and FP unit is enabled when an Advanced SIMD or FP instruction is
detected in the pipeline, and is disabled otherwise.
You can set bit[29] of the CPU Auxiliary Control Register, CPUACTLR_EL1, to 1 to disable dynamic
clock gating of the Advanced SIMD or FP unit.

Related information
4.3.66 CPU Auxiliary Control Register, EL1 on page 4-194.

L2 control and tag banks clock gating

The processor supports dynamic high-level clock gating of the shared L2 control logic and the two L2
tag banks to reduce dynamic power dissipation.
The L2 tag bank clocks are only enabled when a corresponding access to the L2 tag bank is detected in
the pipeline.
The L2 control logic is disabled after 256 consecutive idle cycles. It is then enabled when an L2 access is
detected, with an additional 4-cycle penalty for the wake up before the access is serviced.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-49
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

You can set bit[28] of the L2 Auxiliary Control Register, L2ACTLR_EL1, to 1 to disable dynamic clock
gating of the L2 tag banks.
You can set bit[27] of the L2 Auxiliary Control Register, L2ACTLR_EL1, to 1 to disable dynamic clock
gating of the L2 control logic.

Related information
L2 Auxiliary Control Register, EL1.

Regional clock gating

In addition to extensive local clock gating to register flops, you can configure the processor to include
Regional Clock Gates (RCGs) that can perform additional clock gating of logic blocks such as the
register banks to reduce dynamic power dissipation.
You can set bit[63] of the CPUACTLR_EL1 to 1 to disable regional clock gating for each processor.
You can set bit[26] of the L2ACTLR_EL1 to 1 to disable regional clock gating in the L2, GIC, and
Timer.

Related information
4.3.66 CPU Auxiliary Control Register, EL1 on page 4-194.
L2 Auxiliary Control Register, EL1.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-50
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

2.4.2 Power domains

The processor supports the following power domains:
• Each core in the device.
• The L2 cache and Snoop Tag RAMs.
• A domain for:
— The L2 control.
— The GIC CPU interface.
— The Generic Timer logic.
• The PCLKDBG domain for:
— The Debug APB interface.
— The CTI logic.
— The CTM logic.

Note
• The design does not support a separate power domain for the L1 cache and branch prediction RAMs
within the core. It does not support L1 cache retention when the core is powered down.
• For L2 RAMs dynamic retention, the L2 Data, Dirty, Tag, Inclusion PLRU, and Snoop Tag RAMs are
retained. For L2 cache Dormant mode, the L2 Data, Dirty, Tag, and Inclusion PLRU RAMs are
retained.

The following figure shows the supported power domains in the processor and the placeholders where
you can insert clamps for a core.
Processor Non-processor
Processor power domain L2 control
power domain
Instruction Data L2 cache L2 Snoop Tag
L2 TLB RAM RAM
cache cache
RAM
RAM RAM
Clamp

L2 cache control ACE or CHI

Debug, Clamp and
Integer core miscellaneous ACP
ETM

Clamp

Advanced SIMD and FP PCLKDBG

power domain Clamp
AXI4 Stream,
GIC,
Interrupts,
Timers
Debug, CTI, CTM Timer interrupts

APB, CTI ATB

Figure 2-17 Power domains

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-51
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

2.4.3 Power modes

The power domains can be controlled independently to give different combinations of powered-up and
powered-down domains. However, only some powered-up and powered-down domain combinations are
valid and supported.
The following table shows the valid powered-up and powered-down domain combinations for the
different possible modes of operation. The table uses the following terms:
Off Block is powered down.
WFx Block is in WFI or WFE low-power state.
Ret Logic or RAM retention power only.
On Block is powered up.

Table 2-3 Valid power modes

Mode Cored (CLK) Debug APB, CTI, and CTM L2 RAMse (CLK) L2 control, GIC,
(PCLKDBG) Timer (CLK)

L2 control powerup and L2 RAMs On | WFx | Ret | Off f Off | On On On

powerup

L2 control powerup and L2 RAMs WFx | Ret | Off g Off | On Ret On

retained

L2 cache Dormant mode All Off Off | On On | Ret Off

PCLKDBG powerup All Off On Off Off

Processor shutdown All Off Off Off Off

There are specific requirements that you must meet to power up and power down each power domain
within the core. The supported powerup and powerdown sequences are:
• Individual core powerdown on page 2-52.
• Processor powerdown without system driven L2 flush on page 2-53.
• Processor powerdown with system driven L2 flush on page 2-55.
• Dormant mode on page 2-55.
• Debug powerdown on page 2-57.
• External debug over powerdown on page 2-57.

Note
• The powerup and powerdown sequences in the following sections are the only power sequences that
ARM recommends. Any deviation from these sequences can lead to unpredictable results.
• The powerup and powerdown sequences require that you isolate the powerup domain before power is
removed from the powerdown domain. You must clamp the outputs of the powerdown domain to
benign values to prevent data corruption or unpredictable behavior in the powerup domain.

Individual core powerdown

If an individual core is not required, you can reduce leakage power by turning off the power to the core.
The core refers to all core logic, including Advanced SIMD and FP unit, L1 RAMs, Debug, ETM,
breakpoint and watchpoint logic.

d Core, which includes the Advanced SIMD and FP, Debug, ETM, breakpoint and watchpoint (CLK) logic.
e For L2 RAMs dynamic retention, the L2 Data, Dirty, Tag, Inclusion PLRU, and Snoop Tag RAMs are retained.
For L2 cache Dormant mode, the L2 Data, Dirty, Tag, and Inclusion PLRU RAMs are retained.
f This power mode requires all the cores to be in one of On, WFI, WFE, Retention, or Off state. Each core can be in a different one of these states.
g This power mode requires all the cores to be in one of WFI, WFE, Retention, or Off state. Each core can be in a different one of these states.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-52
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

To enable the core to be powered down, the implementation must place the core on a separately
controlled power supply. In addition, you must clamp the outputs of the core to benign values while the
core is powered down.
To power down the core power domain, apply the following sequence:
1. Clear the appropriate System Control Register C bit, data or unified cache enable, to prevent
additional data cache allocation.
2. Disable L2 prefetches by writing a one to bit [38] and zeros to bits [36:35] of the CPU Extended
Control Register.
3. Disable the Load-store hardware prefetcher by writing a one to bit [56] of the CPU Auxiliary Control
Register.
4. Execute an ISB instruction to ensure the CPU Extended Control Register and CPU Auxiliary Control
Register writes are complete.
5. Execute a DSB instruction to ensure completion of any prior prefetch requests.
6. Clean and invalidate all data from the L1 data cache. The L2 duplicate Snoop Tag RAM for this core
is now empty. This prevents any new data cache snoops or data cache maintenance operations from
other processors in the processor being issued to this core.
7. Clear the CPUECTLR.SMPEN bit. Clearing the SMPEN bit enables the core to be taken out of
coherency by preventing the core from receiving instruction cache, TLB, or BTB maintenance
operations broadcast by other processors in the multiprocessor.
8. Ensure that the system does not send interrupts to the core that is being powered down.
9. Set the DBGOSDLR.DLK, Double lock control bit, that forces the debug interfaces to be quiescent.
10. Execute an ISB instruction to ensure that all of the System register changes from the previous steps
have been committed.
11. Execute a DSB instruction to ensure that all instruction cache, TLB, and branch predictor maintenance
operations issued by any core in the processor before the SMPEN bit was cleared have completed.
12. Execute a WFI instruction and wait until the STANDBYWFI output asserts to indicate that the core is
idle and in the WFI low-power state.
13. Activate the core output clamps.
14. Remove power from the core power domain.
To power up the core power domain, apply the following sequence:
1. Assert nCPUPORESET.
2. Apply power to the core power domain while keeping nCPUPORESET asserted. When power is
restored, continue to hold nCPUPORESET for 16 CLK cycles to allow the reset to propagate.
3. Release the core output clamps.
4. Deassert nCPUPORESET.

Processor powerdown without system driven L2 flush

The Cortex-A72 processor supports processor powerdown where all the processor power domains are
shut down and all state is lost. In this section, a lead core is defined as the last core to powerdown, or the
first core to powerup.
To power down the processor, apply the following sequence:
1. Ensure all non-lead processors are in shutdown mode, see Individual core powerdown on page 2-52.
2. For the lead processor, follow steps 1 on page 2-53 to 5 on page 2-53 in Individual core powerdown
on page 2-52.
3. When ACP is present and all outstanding ACP transactions are complete, the SoC can assert
AINACTS to idle the ACP. When AINACTS has been asserted, the SoC must not assert
ARVALIDS, AWVALIDS, or WVALIDS.
4. Clean and invalidate all data from the L2 data cache.
5. If the core implements:
An ACE interface
When all outstanding snoop transactions are complete, the SoC can assert ACINACTM.
A CHI interface
When all outstanding snoop transactions are complete, the SoC can assert SINACT.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-53
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

6. Ensure system interrupts to the processor are disabled.

7. Follow steps 7 on page 2-53 to 14 on page 2-53 in Individual core powerdown on page 2-52.
8. Wait until STANDBYWFIL2 asserts to indicate that the L2 memory system is idle.
9. Activate the output clamps of the processor in the SoC.
10. Remove power from the L2 control and L2 RAM power domains.
To power up the Cortex-A72 processor, apply the following sequence:
1. For each core in the MPCore device, assert nCPUPORESET LOW.
2. For the lead core in the MPCore device, assert nPRESETDBG and nL2RESET LOW, and hold
L2RSTDISABLE LOW.
3. Apply power to the processor, L2 control, L2 RAM and debug power domains while keeping the
signals described in steps 1 on page 2-53 and 2 on page 2-53 LOW.
4. Release the output clamps of the processor in the SoC.
5. Continue a normal powerup reset sequence.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-54
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

Processor powerdown with system driven L2 flush

The Cortex-A72 processor supports processor powerdown where all the processor power domains are
shut down and all state is lost. In this section, a lead processor is defined as the last core to powerdown,
or the first core to powerup.
To power down the processor, apply the following sequence:
1. Ensure all non-lead cores are in shutdown mode, see Individual core powerdown on page 2-52.
2. For the lead core, follow steps 1 on page 2-53 to 5 on page 2-53 in Individual core powerdown
on page 2-52.
3. When ACP is present and all outstanding ACP transactions are complete, the SoC can assert
AINACTS to idle the ACP. When AINACTS has been asserted, the SoC must not assert
ARVALIDS, AWVALIDS, or WVALIDS.
4. Ensure system interrupts to the processor are disabled.
5. Follow steps 7 on page 2-53 to 14 on page 2-53 in Individual core powerdown on page 2-52.
6. The SoC can now assert the L2FLUSHREQ input.
7. The L2 performs a series of internal clean and invalidate operations to each set and way of the L2
cache.
8. When the L2 completes the clean and invalidate sequence, it asserts the L2FLUSHDONE signal.
The SoC can now deassert L2FLUSHREQ signal and then the L2 deasserts L2FLUSHDONE.
9. If the core implements:
An ACE interface
When all outstanding snoop transactions are complete, the SoC can assert ACINACTM to
idle the AXI master interface. When ACINACTM has been asserted, the SoC must not assert
ACVALIDM..
A CHI interface
When all outstanding snoop transactions are complete, the SoC can assert SINACT.
10. Wait until STANDBYWFIL2 asserts to indicate that the L2 memory system is idle.
11. Activate the output clamps of the processor in the SoC.
12. Remove power from the L2 control and L2 RAM power domains.
To power up the Cortex-A72 processor, apply the following sequence:
1. For each core in the MPCore device, assert nCPUPORESET LOW.
2. For the lead core in the MPCore device, assert nPRESETDBG and nL2RESET LOW, and hold
L2RSTDISABLE LOW.
3. Apply power to the core, L2 control, L2 RAM and debug power domains while keeping the signals
described in steps 1 on page 2-55 and 2 on page 2-55 LOW.
4. Release the output clamps of the processor in the SoC.
5. Continue a normal powerup reset sequence.

Dormant mode
The Cortex-A72 processor supports Dormant mode, where all the processors, debug PCLKDBG, and L2
control logic are powered down while the L2 cache RAMs are powered up and retain state.
This reduces the energy cost of writing dirty lines back to memory and improves response time on
powerup. In Dormant mode, the L2 cache is not kept hardware coherent with other masters in the system.
The RAM blocks that remain powered up and retained during Dormant mode are:
• L2 Tag RAMs.
• L2 Dirty RAMs.
• L2 Data RAMs.
• L2 Inclusion PLRU RAMs.
To support Dormant mode, the L2 cache RAMs must be implemented in a separate power domain. In
addition, you must clamp all inputs to the L2 cache RAMs to benign values, to avoid corrupting data
when the processors and L2 control power domains enter and exit powerdown state.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-55
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

Before entering Dormant mode, the architectural state of the processor, excluding the contents of the L2
cache RAMs that remain powered up, must be saved to external memory.
To exit from Dormant mode to Run mode, the SoC must perform a full powerup reset sequence. The SoC
must assert the reset signals until power is restored. After power is restored, the processor exits the
powerup reset sequence, and the architectural state must be restored.
To enter Dormant mode, apply the following sequence:
1. Clear the appropriate System Control Register C bit, data or unified cache enable, to prevent
additional data cache allocation.
2. Clean and invalidate all data from the L1 data cache. The L2 duplicate Snoop Tag RAM for this core
is now empty. This prevents any new data cache snoops or data cache maintenance operations from
other processors in the processor being issued to this core.
3. Clear the CPUECTLR.SMPEN bit. Clearing the SMPEN bit enables the core to be taken out of
coherency by preventing the core from receiving instruction cache, TLB, or BTB maintenance
operations broadcast by other processors in the MPCore device.
4. Ensure that the system does not send interrupts to the core that is being powered down.
5. Save architectural state, if required. These state saving operations must ensure that the following
occur:
• All ARM registers, including the core state, are saved.
• All System registers are saved.
• All debug related state is saved.
6. Set the DBGOSDLR.DLK, Double lock control bit, that forces the debug interfaces to be quiescent.
7. Execute an ISB instruction to ensure that all of the System register changes from the previous steps
have been committed.
8. Execute a DSB instruction to ensure that all instruction cache, TLB, and branch predictor maintenance
operations issued by any core in the processor before the SMPEN bit was cleared have completed. In
addition, this ensures that all state saving has completed.
9. Execute a WFI instruction and wait until the STANDBYWFI output is asserted, to indicate that the
core is in idle and low-power state.
10. Repeat the previous steps for all processors, and wait for all STANDBYWFI outputs to assert.
11. If the processor implements:
An ACE interface
When all outstanding snoop transactions are complete, the SoC asserts ACINACTM. When
ACINACTM has been asserted, the SoC must not assert ACVALIDM.
A CHI interface
When all outstanding snoop transactions are complete, the SoC asserts SINACT.
When ACP is present and all outstanding ACP transactions are complete, the SoC asserts AINACTS.
When AINACTS has been asserted, the SoC must not assert ARVALIDS, AWVALIDS, or
WVALIDS.
When the L2 completes the outstanding transactions for the AXI, or CHI, interface then
STANDBYWFIL2 asserts to indicate that the L2 memory system is idle.
12. When all of the core STANDBYWFI signals and the STANDBYWFIL2 are asserted, the processor
is ready to enter Dormant mode.
13. Activate the L2 cache RAM input clamps.
14. Remove power from the cores, debug PCLKDBG, and L2 control power domains.
To exit Dormant mode, apply the following sequence:
1. Apply a normal powerup reset sequence. You must apply resets to the cores, debug PCLKDBG, and
the L2 memory system logic until power is restored. During this reset sequence, L2RSTDISABLE
must be held HIGH to disable the L2 cache hardware reset mechanism.
2. When power is restored, release the L2 cache RAM input clamps.
3. Continue a normal powerup reset sequence with L2RSTDISABLE held HIGH. The
L2RSTDISABLE must be held HIGH for a minimum of 32 CLK cycles after the deasserting edge
of nL2RESET.
4. The architectural state must be restored, if required.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-56
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

Debug powerdown
If the Cortex-A72 processor runs in an environment where debug facilities are not required for any of its
cores then you can reduce leakage power by turning off the power to the debug unit in the PCLKDBG
domain.
To enable the debug unit in the PCLKDBG domain to be powered down, the implementation must place
the debug unit on a separately controlled power supply. In addition, you must clamp the outputs of the
debug unit to benign values while the debug unit is powered down.
To power down the debug PCLKDBG power domain, apply the following sequence:
1. Activate the debug output clamps.
2. Remove power from the debug PCLKDBG domain.
Note
If the debug output clamps are released without following the specified debug powerup sequence, the
results are unpredictable.

To power up the debug PCLKDBG power domain, apply the following sequence:
1. Assert nPRESETDBG.
2. Apply power to the debug PCLKDBG power domain while keeping nPRESETDBG asserted.
3. Release the debug output clamps.
4. If the SoC uses the debug hardware, deassert nPRESETDBG.

External debug over powerdown

The Cortex-A72 processor provides support for external debug over powerdown. If any or all of the
cores are powered down, the SoC can still use the debug facilities if the debug PCLKDBG domain is
powered up.
To enable external debug over powerdown, the implementation must place the core and the debug
PCLKDBG unit on separately controlled power supplies. If the core is powered down while the debug
PCLKDBG unit is powered up, you must clamp all outputs from the core power domain to the debug
power domain to benign values.
To power down the core power domain for external debug over powerdown support, apply the following
additional step to the core powerdown sequence, as described in Individual core powerdown
on page 2-52, after STANDBYWFI is asserted in step 12 on page 2-53, and before core clamps are
activated in step 13 on page 2-53:
• Deassert DBGPWRDUP to indicate that the core debug resources are not available for APB
accesses.
When power is removed from the core power domain, keep the debug PCLKDBG unit powered up.
To power up the core power domain after external debug over powerdown support is no longer required,
apply the following additional step to the core powerup sequence, as described in Individual core
powerdown on page 2-52, after nCPUPORESET is deasserted in step 5 on page 2-53.
• Assert DBGPWRDUP to indicate that processor debug resources are available.

Related information
Dormant mode on page 2-55.

2.4.4 Using SMPEN as a power mode indicator

You can use the SMPEN output to differentiate between a retention and powerdown opportunity.
However, this is not a requirement and this communication might occur because of other methods such
as message passing between the core and external power controller.
If SMPEN is HIGH, the core is still in coherency with the other cores in the processor and therefore only
retention is appropriate. If SMPEN is LOW it indicates the core can be powered down.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-57
reserved.
Non-Confidential
 2 Functional Description
2.4 Power management

If the SMPEN is sampled LOW when the CPU Q-Channel handshake has completed the transition to
retention, the core can be returned to the active state using the Q-Channel, then if CPUQACTIVE is still
LOW, the power controller can start a powerdown transition of the core.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 2-58
reserved.
Non-Confidential
Chapter 3
Programmers Model

This chapter describes the processor registers and provides information for programming the processor.

It contains the following sections:

• 3.1 About the programmers model on page 3-60.
• 3.2 ARMv8-A architecture concepts on page 3-61.
• 3.3 ThumbEE instruction set on page 3-69.
• 3.4 Jazelle implementation on page 3-70.
• 3.5 Memory model on page 3-72.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 3-59
reserved.
Non-Confidential
 3 Programmers Model
3.1 About the programmers model

3.1 About the programmers model

The Cortex-A72 processor implements the ARMv8-A architecture. This includes:
• Support for all the Exception levels, EL3-EL0.
• Support for both Execution states, AArch64 and AArch32, at each Exception level.
• The following instruction sets:
AArch64 Execution state
The A64 instruction set.
AArch32 Execution state
The T32 and A32 instruction sets.
The processor supports the following features:
• A32, T32, and A64 Advanced Single Instruction Multiple Data (SIMD) instructions.
• A32, T32, and A64 Floating-point instructions.
• A32, T32, and A64 optional Cryptography Extension instructions.
• Generic Timer.
The processor does not support the T32EE (ThumbEE) instruction set.
Note
The optional Cryptography engine is not included in the base product of the processor. ARM requires
licensees to have contractual rights to obtain the Cortex-A72 processor Cryptography engine.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 3-60
reserved.
Non-Confidential
 3 Programmers Model
3.2 ARMv8-A architecture concepts

3.2 ARMv8-A architecture concepts

Introduces both the ARMv8-A architectural concepts and the associated terminology.
The following sections describe the ARMv8-A architectural concepts. Each section introduces the
corresponding terms that are used to describe the architecture.
Note
A thorough understanding of the terminology defined in this section is a prerequisite for reading the
remainder of this manual.

This section contains the following subsections:

• 3.2.1 Execution state on page 3-61.
• 3.2.2 Exception levels on page 3-62.
• 3.2.3 Security state on page 3-63.
• 3.2.4 Rules for changing Exception state on page 3-64.
• 3.2.5 Stack Pointer selection on page 3-64.
• 3.2.6 ARMv8 security model on page 3-65.
• 3.2.7 Instruction set state on page 3-67.
• 3.2.8 AArch32 execution modes on page 3-67.

3.2.1 Execution state

The Execution state defines the processor execution environment, including:

• Supported register widths.
• Supported instruction sets.
• Significant aspects of:
— The execution model.
— The Virtual Memory System Architecture (VMSA).
— The programmers model.
The Execution states are:
AArch64
The 64-bit Execution state. This Execution state:
• Features 31 64-bit general-purpose registers, with a 64-bit Program Counter (PC), Stack
Pointer (SP), and Exception Link Registers (ELRs).
• Provides a single instruction set, A64.
• Defines the ARMv8 exception model, with four Exception levels, EL0-EL3, that provide an
execution privilege hierarchy.
• Features 48-bit Virtual Address (VA), held in 64-bit registers. The Cortex-A72 processor
VMSA maps these to 44-bit Physical Address (PA) maps.
• Defines a number of elements that hold the processor state (PSTATE). The A64 instruction
set includes instructions that operate directly on various PSTATE elements.
• Names each System register using a suffix that indicates the lowest Exception level that the
register can be accessed.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 3-61
reserved.
Non-Confidential
 3 Programmers Model
3.2 ARMv8-A architecture concepts

AArch32
The 32-bit Execution state. This Execution state is backwards-compatible with implementations
of the ARMv7-A architecture profile that include the Security Extensions and the Virtualization
Extensions. This Execution state:
• Features 13 32-bit general purpose registers, and a 32-bit PC, SP, and Link Register (LR).
Some of these registers have multiple Banked instances for use in different processor modes.
• Provides 32 64-bit registers for Advanced SIMD and Floating-point support.
• Provides two instruction sets, A32 and T32.
• Provides an exception model that maps the ARMv7 exception model onto the ARMv8
exception model and Exception levels. For exceptions taken to an Exception level that is
using AArch32, this supports the ARMv7 exception model use of processor modes.
• Features 32-bit VAs. The VMSA maps these to 40-bit PAs.
• Collects processor state into the Current Processor State Register (CPSR).
The processor can move between Execution states only on a change of Exception level, and subject to
the rules given in 3.2.4 Rules for changing Exception state on page 3-64. This means different software
layers, such as an application, an operating system kernel, and a hypervisor, executing at different
Exception levels, can execute in different Execution states.

Related information
3.2.7 Instruction set state on page 3-67.

3.2.2 Exception levels

The ARMv8 exception model defines Exception levels EL0-EL3, where:

• EL0 has the lowest software execution privilege, and execution at EL0 is called unprivileged
execution.
• Increased values of n, from 1 to 3, indicate increased software execution privilege.
• EL2 provides support for processor virtualization.
• EL3 provides support for two security states.
The Cortex-A72 processor implements all the Exception levels, EL0-EL3, and supports both Execution
states, AArch64 and AArch32, at each Exception level.
Execution can move between Exception levels only on taking an exception, or on returning from an
exception:
• On taking an exception, the Exception level either increases or remains the same. The Exception level
cannot decrease on taking an exception.
• On returning from an exception, the Exception level either decreases or remains the same. The
Exception level cannot increase on returning from an exception.
The Exception level that execution changes to, or remains in, on taking an exception, is called the target
Exception level of the exception and:
• Every exception type has a target Exception level that is either:
— Implicit in the nature of the exception.
— Defined by configuration bits in the System registers.
• An exception cannot target the EL0 Exception level.
Exception levels, and privilege levels, are defined within a particular Security state, and 3.2.6 ARMv8
security model on page 3-65 describes the permitted combinations of Security state and Exception
level.

Exception terminology

This section defines terms used to describe the navigation between Exception levels.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 3-62
reserved.
Non-Confidential
 3 Programmers Model
3.2 ARMv8-A architecture concepts

Terminology for taking an exception

An exception is generated when the processor first responds to an exceptional condition. The processor
state at this time is the state the exception is taken from. The processor state immediately after taking the
exception is the state the exception is taken to.

Terminology for returning from an exception

To return from an exception, the processor must execute an exception return instruction.The processor
state when an exception return instruction is committed for execution is the state the exception returns
from. The processor state immediately after the execution of that instruction is the state the exception
returns to.

Exception level terminology

An Exception level, ELn, with a larger value of n than another Exception level, is described as being a
higher Exception level than the other Exception level. For example, EL3 is a higher Exception level than
EL1.
An Exception level with a smaller value of n than another Exception level is described as being a lower
Exception level than the other Exception level. For example, EL0 is a lower Exception level than EL1.
An Exception level is described as:
Using AArch64
When execution in that Exception level is in AArch64 Execution state.
Using AArch32
When execution in that Exception level is in AArch32 Execution state.

Typical Exception level usage model

The architecture does not specify how software can use the different Exception levels but the following
is a common usage model for the Exception levels:
EL0 Applications.
EL1 OS kernel and associated functions that are typically described as privileged.
EL2 Hypervisor.
EL3 Secure monitor.

Related information
3.2.3 Security state on page 3-63.

3.2.3 Security state

An ARMv8 implementation that includes the EL3 Exception level provides the following Security states,
each with an associated memory address space:
Secure state
In Secure state, the processor:
• Can access both the Secure and the Non-secure memory address space.
• When executing at EL3, can access all the system control resources.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 3-63
reserved.
Non-Confidential
 3 Programmers Model
3.2 ARMv8-A architecture concepts

Non-secure state
In Non-secure state, the processor:
• Can access only the Non-secure memory address space.
• Cannot access the Secure system control resources.
The AArch32 Security state model is unchanged from the model for an ARMv7-A architecture profile
implementation that includes the Security Extensions and the Virtualization Extensions. When the
implementation uses the AArch32 state for all Exception levels, many System registers are Banked to
provide Secure and Non-secure instances, and:
• The Secure instance is accessible only at EL3.
• The Non-secure instance is accessible at EL1 or higher.
• The two instances of a Banked register have the same name.
The 3.2.6 ARMv8 security model on page 3-65 describes how the Security state interacts with other
aspects of the ARMv8 architectural state.

3.2.4 Rules for changing Exception state

This introduction to moving between Execution states does not consider exceptions caused by debug
events.
The Execution state, AArch64 or AArch32, can change only on a change of Exception level, meaning it
can change only on either:
• Taking an exception to a higher Exception level.
• Returning from an exception to a lower Exception level.

Note
The Execution state cannot change if, on taking an exception or on returning from an exception, the
Exception level remains the same.

On taking an exception to a higher Exception level, the Execution state:

• Can either:
— Remain the same.
— Change from AArch32 state to AArch64 state.
• Cannot change from AArch64 state to AArch32 state.
On returning from an exception to a lower Exception level, the Execution state:
• Can either:
— Remain the same.
— Change from AArch64 state to AArch32 state.
• Cannot change from AArch32 state to AArch64 state.
On powerup and on reset, the processor enters EL3, the highest Exception level. The Execution state for
this Exception level is controlled by the configuration input signal, AA64nAA32. For the other
Exception levels the Execution state is determined as follows:
• For an exception return to EL0, the EL0 Execution state is specified as part of the exception return,
subject to the rules given in this section.
• Otherwise, the Execution state is determined by one or more System register configuration bits, that
can be set only in a higher Exception level.

3.2.5 Stack Pointer selection

Stack Pointer behavior depends on the Execution state, as follows:
AArch64
In EL0, the Stack Pointer (SP) maps to the SP_EL0 Stack Pointer register.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 3-64
reserved.
Non-Confidential
 3 Programmers Model
3.2 ARMv8-A architecture concepts

Taking an exception selects the default Stack Pointer for the target Exception level, meaning SP
maps to the SP_ELx Stack Pointer register, where x is the Exception level.
Software executing in the target Exception level can execute an MSR SPSel, #Imm1 instruction
to select whether to use the default SP_ELx Stack Pointer or the SP_EL0 Stack Pointer.
The selected Stack Pointer can be indicated by a suffix to the Exception level:
t Indicates use of the SP0 Stack Pointer.
h Indicates use of the SPx Stack Pointer.

Note
The t and h suffixes are based on the terminology of thread and handler, introduced in ARMv7-
M.

The following table shows the set of AArch64 Stack Pointer options.

Table 3-1 AArch64 Stack Pointer options

Exception level AArch64 Stack Pointer options

EL0 EL0t

EL1 EL1t, EL1h

EL2 EL2t, EL2h

EL3 EL3t, EL3h

AArch32
In AArch32 state, each mode that can be the target of an exception has its own Banked copy of
the Stack Pointer. For example, the Banked Stack Pointer for Hyp mode is called SP_hyp.
Software executing in one of these modes uses the Banked Stack Pointer for that mode.
The modes that have Banked copies of the Stack Pointer are FIQ mode, IRQ mode, Supervisor
mode, Abort mode, Undefined mode, Hyp mode, and Monitor mode. Software executing in
User mode or System mode uses the User mode Stack Pointer, SP_usr.

Related information
3.2.8 AArch32 execution modes on page 3-67.

3.2.6 ARMv8 security model

The Cortex-A72 processor implements all of the Exception levels. This means:
• EL3 exists only in Secure state and a change from Secure state to Non-secure state is made only by
an exception return from EL3.
• EL2 exists only in Non-secure state.
To provide compatibility with ARMv7, the Exception levels available in Secure state are modified when
EL3 is using AArch32. The following sections describe the security model:
• Security model when EL3 is using AArch64 on page 3-66.
• Security model when EL3 is using AArch32 on page 3-66.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 3-65
reserved.
Non-Confidential
 3 Programmers Model
3.2 ARMv8-A architecture concepts

Security model when EL3 is using AArch64

When EL3 is using AArch64, The following figure shows the security model, and the expected use of
the different Exception levels. This figure shows how instances of EL0 and EL1 are present in both
Security states. The figure also shows the expected software usage of the Exception levels.
Non-secure state Secure state

AArch32 or AArch32 or AArch32 or AArch32 or AArch32 or AArch32 or

AArch64† AArch64† AArch64† AArch64† AArch64† AArch64†
EL0 App1 App2 App1 App2 Secure App1 Secure App2

AArch32 or AArch64‡ AArch32 or AArch64‡ AArch32 or AArch64

EL1 Guest OS1 Guest OS2 Secure OS

AArch32 or AArch64

EL2 Hypervisor

AArch64

EL3 Secure monitor

† AArch64 permitted only if EL1 is using AArch64

‡ AArch64 permitted only if EL2 is using AArch64

Figure 3-1 ARMv8 security model when EL3 is using AArch64

Security model when EL3 is using AArch32

To provide software compatibility with VMSAv7 implementations that include the Security Extensions,
in Secure AArch32 state, all modes other than User mode must have the same execution privilege. This
means that, in an implementation where EL3 is using AArch32, the security model is as shown in
following figure. This figure also shows the expected use of the different Exception levels and processor
modes.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 3-66
reserved.
Non-Confidential
 3 Programmers Model
3.2 ARMv8-A architecture concepts

Non-secure state Secure state

AArch32 AArch32 AArch32 AArch32 AArch32 AArch32

App1 App2 App1 App2 Secure App1 Secure App2
EL0
Modes: Modes: Modes: Modes: Modes: Modes:
User User User User User User

AArch32 AArch32
Guest OS1 Guest OS2
EL1
Modes: Modes:
System, FIQ, IRQ, System, FIQ, IRQ,
Supervisor, Abort, Undefined Supervisor, Abort, Undefined

AArch32
Hypervisor
EL2
Modes:
Hyp

AArch32
Secure monitor Secure OS
EL3
Modes: Modes:
System, FIQ, IRQ,
Monitor
Supervisor, Abort, Undefined

Figure 3-2 ARMv8 security model when EL3 is using AArch32

For more information about the AArch32 processor modes see 3.2.8 AArch32 execution modes
on page 3-67.

3.2.7 Instruction set state

The processor instruction set state determines the instruction set that the processor executes. The possible
instruction sets depend on the Execution state:
AArch64
AArch64 state supports only a single instruction set, called A64. This is a fixed-width
instruction set that uses 32-bit instruction encoding.
AArch32
AArch32 state supports the following instruction sets:
A32
This is a fixed-length instruction set that uses 32-bit instruction encodings. It is
compatible with the ARMv7 ARM instruction set.
T32
This is a variable-length instruction set that uses both 16-bit and 32-bit instruction
encodings. It is compatible with the ARMv7 Thumb instruction set.

3.2.8 AArch32 execution modes

ARMv7 and earlier versions of the ARM architecture, define a set of named processor modes, including
modes that correspond to different exception types. For compatibility, AArch32 state retains these
processor modes.
The following table shows the AArch32 processor modes, and the Exception level of each mode.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 3-67
reserved.
Non-Confidential
 3 Programmers Model
3.2 ARMv8-A architecture concepts

Table 3-2 AArch32 processor modes and associated Exception levels

AArch32 processor mode EL3 using Security state Exception level

User AArch32 or AArch64 Non-secure or Secure EL0

System, FIQ, IRQ, Supervisor, AArch64 Non-secure or Secure EL1

Abort, Undefined AArch32 Non-secure EL1

Secure EL3

Hyp AArch32 or AArch64 Non-secure only EL2

Monitor AArch32 Secure only EL3

When the EL3 using column of The following table shows:

AArch64 The row refers to information shown in figure Figure 3-1 ARMv8 security model when EL3
is using AArch64 on page 3-66.
AArch32 The row refers to information shown in figure Figure 3-2 ARMv8 security model when EL3
is using AArch32 on page 3-67.

A processor mode name does not indicate the current Security state. To distinguish between a mode in
Secure state and the equivalent mode in Non-secure state, the mode name is qualified as Secure or Non-
secure. For example, a description of AArch32 operation in EL1 might relate to the Secure FIQ mode, or
to the Non-secure FIQ mode.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 3-68
reserved.
Non-Confidential
 3 Programmers Model
3.3 ThumbEE instruction set

3.3 ThumbEE instruction set

The Thumb Execution Environment (ThumbEE) extension is a variant of the Thumb instruction set that is
designed as a target for dynamically generated code.
The processor does not implement the T32EE (ThumbEE) instruction set.
In AArch32 state, access to the ThumbEE Configuration Register, TEECR, and ThumbEE Handler Base
Register, TEEHBR, results in an Undefined Instruction trap.
In AArch64 state, these registers are named TEECR32_EL1 and TEEHBR32_EL1 respectively, and
access to these registers results in an Undefined Instruction trap.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 3-69
reserved.
Non-Confidential
 3 Programmers Model
3.4 Jazelle implementation

3.4 Jazelle implementation

AArch32 state supports a trivial Jazelle implementation.
This means:
• Jazelle state is not supported.
• The BXJ instruction behaves as a BX instruction.
See the ARM® Architecture Reference Manual ARMv8 for more information.

3.4.1 Register summary

The following table gives a summary of the processor Jazelle registers that are accessed through the
CP14 coprocessor in the AArch32 state. These registers are not implemented in the AArch64 state.

Table 3-3 Summary of Jazelle registers

CRn op1 CRm op2 Name Reset Description

c0 7 c0 0 JIDR 0x00000000 Jazelle Identity Register

c1 7 c0 0 JOSCR 0x00000000 Jazelle OS Control Register

c2 7 c0 0 JMCR 0x00000000 Jazelle Main Configuration Register

Related information
Jazelle Identity Register on page 3-70.
Jazelle OS Control Register on page 3-71.
Jazelle Main Configuration Register on page 3-71.

3.4.2 Register description

This section describes the processor Jazelle Extension registers. The following table provides cross-
references to individual registers.

Jazelle Identity Register

The JIDR characteristics are:

Purpose
Enables software to determine the implementation of the Jazelle Extension provided by the
processor.
Usage constraints
The JIDR is:
• A read-only register
• Accessible from all Exception levels in AArch32.
Configurations
Available in all configurations.
Attributes
See the register summary in table Table 3-3 Summary of Jazelle registers on page 3-70.
The JIDR is a 32-bit register with all bits[31:0] as RES0. Writes are ignored, and all bits read as zero.
To access the JIDR in the AArch32 state, read the register with:
MRC p14, 7, <Rd>, c0, c0, 0; Read Jazelle Identity Register

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 3-70
reserved.
Non-Confidential
 3 Programmers Model
3.4 Jazelle implementation

Jazelle OS Control Register

The JOSCR characteristics are:

Purpose
Provides operating system control of the use of the Jazelle Extension.
Usage constraints
The JOSCR is:
• A read/write register
• Accessible only from EL1 or higher.
Configurations
Available in all configurations.
Attributes
See the register summary in table Table 3-3 Summary of Jazelle registers on page 3-70.
The JOSCR is a 32-bit register with all bits[31:0] as RES0. Writes are ignored, and all bits read as zero.
To access the JOSCR in the AArch32 state, read or write the register with:
MRC p14, 7, <Rd>, c1, c0, 0; Read Jazelle OS Control Register
MCR p14, 7, <Rd>, c1, c0, 0; Write Jazelle OS Control Register

Jazelle Main Configuration Register

The JMCR characteristics are:
Purpose
Provides control of the Jazelle Extension features.
Usage constraints
The JMCR is:
• A read/write register
• Accessible only from EL1 or higher.
Configurations
Available in all configurations.
Attributes
See the register summary in table Table 3-3 Summary of Jazelle registers on page 3-70.
The JMCR is a 32-bit register with all bits[31:0] as RES0. Writes are ignored, and all bits read as zero.
To access the JMCR in the AArch32 state, read or write the register with:
MRC p14, 7, <Rd>, c2, c0, 0; Read Jazelle Main Configuration Register
MCR p14, 7, <Rd>, c2, c0, 0; Write Jazelle Main Configuration Register

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 3-71
reserved.
Non-Confidential
 3 Programmers Model
3.5 Memory model

3.5 Memory model

The Cortex-A72 processor views memory as a linear collection of bytes numbered in ascending order
from zero.
For example, bytes 0-3 hold the first stored word, and bytes 4-7 hold the second stored word.
The processor can store words in memory as either:
• Big-endian format.
• Little-endian format.
See the ARM® Architecture Reference Manual ARMv8 for more information about big-endian and little-
endian memory systems.
Note
Instructions are always treated as little-endian.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 3-72
reserved.
Non-Confidential
Chapter 4
System Control

This chapter describes the System registers, their structure, operation, and how to use them.

It contains the following sections:

• 4.1 About system control on page 4-74.
• 4.2 AArch64 register summary on page 4-75.
• 4.3 AArch64 register descriptions on page 4-88.
• 4.4 AArch32 register summary on page 4-214.
• 4.5 AArch32 register descriptions on page 4-239.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-73
reserved.
Non-Confidential
 4 System Control
4.1 About system control

4.1 About system control

The System registers control and provide status information for the functions implemented in the
processor.
The main functions of the System registers are:
• Overall system control and configuration.
• Memory Management Unit (MMU) configuration and management.
• Cache configuration and management.
• System performance monitoring.
• Generic Interrupt Controller (GIC) configuration and management.
The System registers are accessible in AArch32 and AArch64 states. The Execution states are described
in the 3.2 ARMv8-A architecture concepts on page 3-61. The System register access in AArch64 state is
characterized by three possible scenarios. These scenarios are:
• The AArch64 register is unique and described for the AArch64 state in 4.3 AArch64 register
descriptions on page 4-88.
• The AArch64 register is architecturally mapped to an AArch32 register but has different bit or bit
field assignments. There a separate descriptions for each Execution state in 4.3 AArch64 register
descriptions on page 4-88 and 4.5 AArch32 register descriptions on page 4-239.
• The AArch64 register is architecturally mapped to an AArch32 register with the same bit and bit field
assignments. There is one description for both Execution states in the 4.3 AArch64 register
descriptions on page 4-88 and cross-referenced from the 4.4 AArch32 register summary
on page 4-214.
The System registers accessed in AArch32 state are described in the 4.5 AArch32 register descriptions
on page 4-239.
Some of the System registers can be accessed through the memory-mapped or external debug interfaces.
Bits in the System registers that are described in the ARMv7 architecture are redefined in the ARMv8-A
architecture:
• UNK/SBZP, RAZ/SBZP, and RAZ/WI are redefined as RES0.
• UNK/SBOP and RAO/SBOP are redefined as RES1.
RES0and RES1 are described in the ARM® Architecture Reference Manual ARMv8, for ARMv8-A
architecture profile.
This section contains the following subsections:
• 4.1.1 Registers affected by CP15SDISABLE on page 4-74.

4.1.1 Registers affected by CP15SDISABLE

The CP15SDISABLE signal disables write access to certain secure copies of System registers when EL3
is using AArch32. For a list of registers affected by CP15SDISABLE, see the ARM® Architecture
Reference Manual ARMv8.
The Cortex-A72 processor does not have any IMPLEMENTATION DEFINED registers that are affected by
CP15SDISABLE.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-74
reserved.
Non-Confidential
 4 System Control
4.2 AArch64 register summary

4.2 AArch64 register summary

This section gives a summary of the System registers in AArch64 state.
For more information about using the System registers, see the ARM® Architecture Reference Manual
ARMv8.
This section contains the following subsections:
• 4.2.1 AArch64 identification registers on page 4-75.
• 4.2.2 AArch64 exception handling registers on page 4-77.
• 4.2.3 AArch64 virtual memory control registers on page 4-78.
• 4.2.4 AArch64 other System registers on page 4-79.
• 4.2.5 AArch64 cache maintenance operations on page 4-79.
• 4.2.6 AArch64 TLB maintenance operations on page 4-80.
• 4.2.7 AArch64 address translation operations on page 4-80.
• 4.2.8 AArch64 miscellaneous operations on page 4-81.
• 4.2.9 AArch64 Performance Monitors registers on page 4-82.
• 4.2.10 AArch64 reset registers on page 4-82.
• 4.2.11 Security registers on page 4-82.
• 4.2.12 AArch64 virtualization registers on page 4-83.
• 4.2.13 AArch64 EL2 TLB maintenance operations on page 4-85.
• 4.2.14 Generic Timer registers on page 4-85.
• 4.2.15 AArch64 IMPLEMENTATION DEFINED registers on page 4-86.

4.2.1 AArch64 identification registers

The following table shows the identification registers in AArch64 state. Bits[63:32] are reset to
0x00000000 for all 64-bit registers in the table.

Table 4-1 AArch64 identification registers

Name Type Reset Width Description

MIDR_EL1 RO 0x410FD083 32 4.3.1 Main ID Register,

EL1 on page 4-89

MPIDR_EL1 RO 0x80000003h 64 4.3.2 Multiprocessor

Affinity Register, EL1
on page 4-90

REVIDR_EL1 RO 0x00000000 32 4.3.3 Revision ID

Register, EL1
on page 4-92

ID_PFR0_EL1 RO 0x00000131 32 4.3.4 AArch32 Processor

Feature Register 0, EL1
on page 4-93

ID_PFR1_EL1 RO 0x00011011i 32 4.3.5 AArch32 Processor

Feature Register 1, EL1
on page 4-94

ID_DFR0_EL1 RO 0x03010066 32 4.3.6 AArch32 Debug

Feature Register 0, EL1
on page 4-95

h The reset value depends on the primary inputs, CLUSTERIDAFF1 and CLUSTERIDAFF2, and the number of cores that the processor implements. The value
shown is for a four-core implementation, with CLUSTERIDAFF1 and CLUSTERIDAFF2 set to zero.
i The reset value depends on the primary input GICCDISABLE. The value shown assumes the GICCDISABLE signal is tied HIGH.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-75
reserved.
Non-Confidential
 4 System Control
4.2 AArch64 register summary

Table 4-1 AArch64 identification registers (continued)

Name Type Reset Width Description

ID_AFR0_EL1 RO 0x00000000 32 4.3.7 AArch32 Auxiliary

Feature Register 0, EL1
on page 4-96

ID_MMFR0_EL1 RO 0x10201105 32 4.3.8 AArch32 Memory

Model Feature Register
0, EL1 on page 4-96

ID_MMFR1_EL1 RO 0x40000000 32 4.3.9 AArch32 Memory

Model Feature Register
1, EL1 on page 4-98

ID_MMFR2_EL1 RO 0x01260000 32 4.3.10 AArch32 Memory

Model Feature Register
2, EL1 on page 4-99

ID_MMFR3_EL1 RO 0x02102211 32 4.3.11 AArch32 Memory

Model Feature Register
3, EL1 on page 4-101

ID_ISAR0_EL1 RO 0x02101110 32 4.3.12 AArch32

Instruction Set Attribute
Register 0, EL1
on page 4-103

ID_ISAR1_EL1 RO 0x13112111 32 4.3.13 AArch32

Instruction Set Attribute
Register 1, EL1
on page 4-104

ID_ISAR2_EL1 RO 0x21232042 32 4.3.14 AArch32

Instruction Set Attribute
Register 2, EL1
on page 4-106

ID_ISAR3_EL1 RO 0x01112131 32 4.3.15 AArch32

Instruction Set Attribute
Register 3, EL1
on page 4-107

ID_ISAR4_EL1 RO 0x00011142 32 4.3.16 AArch32

Instruction Set Attribute
Register 4, EL1
on page 4-108

ID_ISAR5_EL1 RO 0x00010001j 32 4.3.17 AArch32

Instruction Set Attribute
Register 5, EL1
on page 4-110

ID_AA64PFR0_EL1 RO 0x00002222 64 4.3.18 AArch64

Processor Feature
Register 0, EL1
on page 4-111

j The reset value is 0x00011121 if the Cryptography engine is implemented.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-76
reserved.
Non-Confidential
 4 System Control
4.2 AArch64 register summary

Table 4-1 AArch64 identification registers (continued)

Name Type Reset Width Description

ID_AA64DFR0_EL1 RO 0x10305106 64 4.3.19 AArch64 Debug

Feature Register 0, EL1
on page 4-113

ID_AA64ISAR0_EL1 RO 0x00010000k 64 4.3.20 AArch64

Instruction Set Attribute
Register 0, EL1
on page 4-114

ID_AA64MMFR0_EL1 RO 0x00001124 64 4.3.21 AArch64 Memory

Model Feature Register
0, EL1 on page 4-115

CCSIDR_EL1 RO UNK 32 4.3.22 Cache Size ID

Register, EL1
on page 4-117

CLIDR_EL1 RO 0x0A200023 32 4.3.23 Cache Level ID

Register, EL1
on page 4-118

AIDR_EL1 - 0x00000000 32 4.3.24 Auxiliary ID

Register, EL1
on page 4-120

CSSELR_EL1 RW UNK 32 4.3.25 Cache Size

Selection Register, EL1
on page 4-120

CTR_EL0 RO 0x8444C004 32 4.3.26 Cache Type

Register, EL0
on page 4-121

DCZID_EL0 RO 0x00000004 32 4.3.27 Data Cache Zero

ID, EL0 on page 4-122

VPIDR_EL2 RW -l 32 4.3.28 Virtualization

Processor ID Register,
EL2 on page 4-123

VMPIDR_EL2 RO -m 64 4.3.2 Multiprocessor

Affinity Register, EL1
on page 4-90

4.2.2 AArch64 exception handling registers

The following table shows the fault handling registers in AArch64 state. Bits[63:32] are reset to
0x00000000 for all 64-bit registers in the following table.

k The reset value is 0x00011120 if the Cryptography engine is implemented.

l The reset value is the value of the Main ID Register.
m The reset value is the value of the Multiprocessor Affinity Register.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-77
reserved.
Non-Confidential
 4 System Control
4.2 AArch64 register summary

Table 4-2 AArch64 exception handling registers

Name Type Reset Width Description

AFSR0_EL1 RW RES0 32 4.3.48 Auxiliary Fault Status Register 0, EL1 and EL3 on page 4-156

AFSR1_EL1 RW RES0 32 4.3.49 Auxiliary Fault Status Register 1, EL1 and EL3 on page 4-157

ESR_EL1 RW UNK 32 4.3.50 Exception Syndrome Register, EL1 and EL3 on page 4-157

IFSR32_EL2 RW UNK 32 4.3.51 Instruction Fault Status Register, EL2 on page 4-159

AFSR0_EL2 RW RES0 32 4.3.52 Auxiliary Fault Status Register 0, EL2 and Hyp Auxiliary Data Fault Status
Register on page 4-163

AFSR1_EL2 RW RES0 32 4.3.53 Auxiliary Fault Status Register 1, EL2 and Hyp Auxiliary Instruction Fault
Status Register on page 4-163

ESR_EL2 RW UNK 32 4.3.54 Exception Syndrome Register, EL2 on page 4-163

AFSR0_EL3 RW RES0 32 4.3.55 Physical Address Register, EL1 on page 4-165

AFSR1_EL3 RW RES0 32 4.3.49 Auxiliary Fault Status Register 1, EL1 and EL3 on page 4-157

ESR_EL3 RW UNK 32 4.3.50 Exception Syndrome Register, EL1 and EL3 on page 4-157

FAR_EL1 RW UNK 64 Fault Address Register, EL1 

FAR_EL2 RW UNK 64 Fault Address Register, EL2 n

HPFAR_EL2 RW 0x00000000 64 Hyp IPA Fault Address Register, EL2 n

FAR_EL3 RW UNK 64 Fault Address Register, EL3 n

VBAR_EL1 RW UNK 64 Vector Base Address Register, EL1 n

ISR_EL1 RO UNK 32 Interrupt Status Register n

VBAR_EL2 RW UNK 64 Vector Base Address Register, EL2 n

4.2.3 AArch64 virtual memory control registers

The following table shows the virtual memory control registers in AArch64 state. Bits[63:32] are reset to
0x00000000 for all 64-bit registers in the following table.

Table 4-3 AArch64 virtual memory control registers

Name Type Reset Width Description

SCTLR_EL1 RW 0x00C50838 32 4.3.30 System Control Register, EL1 on page 4-125

SCTLR_EL2 RW 0x30C50838 32 System Control Register, EL2 

SCTLR_EL3 RW 0x00C50838 o 32 4.3.38 System Control Register, EL3 on page 4-141

TTBR0_EL1 RW UNK 64 Translation Table Base Address Register 0, EL1 p

TTBR1_EL1 RW UNK 64 Translation Table Base Address Register 1, EL1 p

TCR_EL1 RW UNK 64 4.3.41 Translation Control Register, EL1 on page 4-146

TTBR0_EL2 RW UNK 64 Translation Table Base Address Register 0, EL2 p

TCR_EL2 RW UNK 32 4.3.42 Translation Control Register, EL2 on page 4-149

n See the ARM® Architecture Reference Manual ARMv8 for more information.
o The reset value depends on primary input CFGTE. Table 4-3 AArch64 virtual memory control registers on page 4-78 assumes this signal is LOW.
p See the ARM® Architecture Reference Manual ARMv8 for more information.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-78
reserved.
Non-Confidential
 4 System Control
4.2 AArch64 register summary

Table 4-3 AArch64 virtual memory control registers (continued)

Name Type Reset Width Description

VTTBR_EL2 RW UNK 64 Virtualization Translation Table Base Address Register, EL2 p

VTCR_EL2 RW UNK 32 4.3.43 Virtualization Translation Control Register, EL2 on page 4-151

TTBR0_EL3 RW UNK 64 Translation Table Base Address Register 0, EL3 p

TCR_EL3 RW UNK 32 4.3.47 Translation Control Register, EL3 on page 4-155

MAIR_EL1 RW UNK 64 Memory Attribute Indirection Register, EL1 p

AMAIR_EL1 RW RES0 64 4.3.56 Auxiliary Memory Attribute Indirection Register, EL1 and EL3
on page 4-168

MAIR_EL2 RW UNK 64 Memory Attribute Indirection Register, EL2 p

AMAIR_EL2 RW RES0 64 4.3.57 Auxiliary Memory Attribute Indirection Register, EL2 on page 4-168

MAIR_EL3 RW UNK 64 Memory Attribute Indirection Register, EL3 p

AMAIR_EL3 RW RES0 64 4.3.56 Auxiliary Memory Attribute Indirection Register, EL1 and EL3
on page 4-168

CONTEXTIDR_EL1 RW UNK 32 Context ID Register, EL1 p

4.2.4 AArch64 other System registers

The following table shows the other System registers in AArch64 state.

Table 4-4 AArch64 other System registers

Name Type Reset Width Description

ACTLR_EL1 RW 0x00000000 32 4.3.31 Auxiliary Control Register, EL1 on page 4-129

CPACR_EL1 RW 0x00000000 32 4.3.32 Architectural Feature Access Control Register, EL1 on page 4-129

ACTLR_EL2 RW 0x00000000 32 4.3.33 Auxiliary Control Register, EL2 on page 4-130

4.2.5 AArch64 cache maintenance operations

The following table shows the System instructions for cache and maintenance operations in AArch64
state. See the ARM® Architecture Reference Manual ARMv8 for more information about these operations.

Table 4-5 AArch64 cache maintenance operations

Name Description
IC IALLUIS Instruction cache invalidate all to PoUq Inner Shareable

IC IALLU Instruction cache invalidate all to PoU

IC IVAU Instruction cache invalidate by virtual address (VA) to PoU
DC IVAC Data cache invalidate by VA to PoCr
DC ISW Data cache invalidate by set/way

q PoU = Point of Unification. PoU is set by the BROADCASTINNER signal and can be in the L1 data cache or outside of the processor, in which case PoU is
dependent on the external memory system.
r PoC = Point of Coherence. The PoC is always outside of the processor and is dependent on the external memory system.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-79
reserved.
Non-Confidential
 4 System Control
4.2 AArch64 register summary

Table 4-5 AArch64 cache maintenance operations (continued)

Name Description
DC CSW Data cache clean by set/way
DC CISW Data cache clean and invalidate by set/way
DC ZVA Data cache zero by VA
DC CVAC Data cache clean by VA to PoC
DC CVAU Data cache clean by VA to PoU
DC CIVAC Data cache clean and invalidate by VA to PoC

4.2.6 AArch64 TLB maintenance operations

The following table shows the System instructions for TLB maintenance operations in AArch64 state.
See the ARM® Architecture Reference Manual ARMv8 for more information about these operations.

Table 4-6 AArch64 TLB maintenance operations

Name Description
TLBI VMALLE1IS Invalidate all stage 1 translations used at EL1 with the current virtual machine identifier (VMID) in the Inner
Shareable
TLBI VAE1IS Invalidate translation used at EL1 for the specified VA and Address Space Identifier (ASID) and the current
VMID, Inner Shareable
TLBI ASIDE1IS Invalidate all translations used at EL1 with the current VMID and the supplied ASID, Inner Shareable
TLBI VAAE1IS Invalidate all translations used at EL1 for the specified address and current VMID and for all ASID values, Inner
Shareable
TLBI VALE1IS Invalidate all entries from the last level of stage 1 translation table walk used at EL1 with the supplied ASID and
current VMID, Inner Shareable
TLBI VAALE1IS Invalidate all entries from the last level of stage 1 translation table walk used at EL1 for the specified address and
current VMID and for all ASID values, Inner Shareable
TLBI VMALLE1 Invalidate all stage 1 translations used at EL1 with the current VMID
TLBI VAE1 Invalidate translation used at EL1 for the specified VA and ASID and the current VMID
TLBI ASIDE1 Invalidate all translations used at EL1 with the current VMID and the supplied ASID
TLBI VAAE1 Invalidate all translations used at EL1 for the specified address and current VMID and for all ASID values
TLBI VALE1 Invalidate all entries from the last level of stage 1 translation table walk used at EL1 with the supplied ASID and
current VMID
TLBI VAALE1 Invalidate all entries from the last level of stage 1 translation table walk used at EL1 for the specified address and
current VMID and for all ASID values

The Virtualization registers include additional TLB operations for use in Hyp mode. For more
information, see 4.2.13 AArch64 EL2 TLB maintenance operations on page 4-85.

4.2.7 AArch64 address translation operations

The following table shows the address translation register in AArch64 state.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-80
reserved.
Non-Confidential
 4 System Control
4.2 AArch64 register summary

Table 4-7 AArch64 address translation register

Name Type Reset Width Description

PAR_EL1 RW UNK s 64 4.3.55 Physical Address

Register, EL1
on page 4-165

The following table shows the System instructions for address translation operations in AArch64 state.
See the ARM® Architecture Reference Manual ARMv8 for more information.

Table 4-8 AArch64 address translation operations

Name Description
AT S1E1R Stage 1 current state EL1 read 
AT S1E1W Stage 1 current state EL1 write
AT S1E0R Stage 1 current state unprivileged read
AT S1E0W Stage 1 current state unprivileged write
AT S1E2R Stage 1 Hyp mode read
AT S1E2W Stage 1 Hyp mode write
AT S12E1R Stages 1 and 2 Non-secure EL1 read

AT S12E1W Stages 1 and 2 Non-secure EL1 write

AT S12E0R Stages 1 and 2 Non-secure unprivileged read

AT S12E0W Stages 1 and 2 Non-secure unprivileged write

AT S1E3R Stage 1 current state EL3 read

AT S1E3W Stage 1 current state EL3 write

4.2.8 AArch64 miscellaneous operations

The following table shows the miscellaneous operations in AArch64 state. See the ARM® Architecture
Reference Manual ARMv8 for more information about these operations.

Table 4-9 AArch64 miscellaneous System operations

Name Type Reset Width Description

TPIDR_EL0 RW UNK 64 Thread Pointer / ID Register, EL0

TPIDR_EL1 RW UNK 64 Thread Pointer / ID Register, EL1

TPIDRRO_EL0 RW t UNK 64 Thread Pointer / ID Register, Read-Only, EL0

TPIDR_EL2 RW UNK 64 Thread Pointer / ID Register, EL2

TPIDR_EL3 RW UNK 64 Thread Pointer / ID Register, EL3

s Bits[63:32] are reset to 0x00000000.

t RO at EL0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-81
reserved.
Non-Confidential
 4 System Control
4.2 AArch64 register summary

4.2.9 AArch64 Performance Monitors registers

The following table shows the Performance Monitors registers in AArch64 state. Bits[63:32] are reset to
0x00000000 for all 64-bit registers in the following table.

Table 4-10 AArch64 Performance Monitors registers

Name Type Reset Width Description

PMCR_EL0 RW u 0x41023000 32 11.4.1 Performance Monitors Control Register, EL0 on page 11-401

PMCNTENSET_EL0 RW UNK 32 Performance Monitors Count Enable Set Register 

PMCNTENCLR_EL0 RW UNK 32 Performance Monitors Enable Count Clear Register v

PMOVSCLR_EL0 RW UNK 32 Performance Monitors Overflow Flag Status Register v

PMSWINC_EL0 WO - 32 Performance Monitors Software Increment Register v

PMSELR_EL0 RW UNK 32 Performance Monitors Event Counter Selection Register v

PMCEID0_EL0 RO 0x7FFF0F3F 32 11.4.2 Performance Monitors Common Event Identification Register 0, EL0
on page 11-403

PMCEID1_EL0 RO 0x00000000 32 Performance Monitors Common Event Identification Register 1 v

PMCCNTR_EL0 RW UNK 64 Performance Monitors Cycle Count Register v

PMXEVTYPER_EL0 RW UNK 32 Performance Monitors Selected Event Type Register v

PMCCFILTR_EL0 RW 0x00000000 32 Performance Monitors Cycle Count Filter Register v

PMXEVCNTR_EL0 RW UNK 32 Performance Monitors Selected Event Count Register v

PMUSERENR_EL0 RW 0x00000000 32 Performance Monitors User Enable Register v

PMINTENSET_EL1 RW UNK 32 Performance Monitors Interrupt Enable Set Register v

PMINTENCLR_EL1 RW UNK 32 Performance Monitors Interrupt Enable Clear Register v

PMOVSSET_EL0 RW UNK 32 Performance Monitors Overflow Flag Status Set Register v

4.2.10 AArch64 reset registers

The following table shows the reset registers in AArch64 state.

Table 4-11 AArch64 reset registers

Name Type Reset Width Description

RVBAR_EL3 RO - w 64 4.3.60 Reset Vector Base Address, EL3 on page 4-174

RMR_EL3 RW 0x00000000x 32 4.3.61 Reset Management Register, EL3 on page 4-175

4.2.11 Security registers

The following table shows the Security registers in AArch64 state.

u Access permissions also depend on the access condition. See 11.2.5 External register access permissions on page 11-398.
v See the ARM® Architecture Reference Manual ARMv8 for more information.
w The reset value depends on the RVBARADDR signal. Bits[63:32] are reset to 0x00000000.
x For a Cold reset, the AA64nAA32 signal sets the value of bit[0]. The following table assumes this signal is LOW.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-82
reserved.
Non-Confidential
 4 System Control
4.2 AArch64 register summary

Table 4-12 AArch64 security registers

Name Type Reset Width Description

SCR_EL3 RW 0x00000000 32 Secure Configuration Register, EL3 

SDER32_EL3 RW 0x00000000 32 Secure Debug Register, EL3 y

CPTR_EL3 RW 0x00000400 32 4.3.40 Architectural Feature Trap Register, EL3 on page 4-145

MDCR_EL3 RW 0x00000000 32 Monitor Debug Configuration Register, EL3 y

AFSR0_EL3 RW RES0 32 4.3.48 Auxiliary Fault Status Register 0, EL1 and EL3 on page 4-156

AFSR1_EL3 RW RES0 32 4.3.49 Auxiliary Fault Status Register 1, EL1 and EL3 on page 4-157

VBAR_EL3 RW UNKz 64 Vector Base Address Register, EL3 y

4.2.12 AArch64 virtualization registers

The following table shows the virtualization registers in AArch64 state. Bits[63:32] are reset to
0x00000000 for all 64-bit registers in the following table.

Table 4-13 AArch64 virtualization registers

Name Type Reset Width Description

VPIDR_EL2 RW -aa 32 4.3.28 Virtualization

Processor ID Register, EL2
on page 4-123

VMPIDR_EL2 RW -ab 64 4.3.29 Virtualization

Multiprocessor ID
Register, EL2
on page 4-124

SCTLR_EL2 RW 0x30C50838 32 Secure Control Register,

EL3 

ACTLR_EL2 RW 0x00000000 32 4.3.33 Auxiliary Control

Register, EL2
on page 4-130

HCR_EL2 RW 0x00000000 64 4.3.34 Hypervisor

Configuration Register,
EL2 on page 4-131

MDCR_EL2 RW 0x00000006ad 32 Monitor Debug

Configuration Register,
EL2 ac

CPTR_EL2 RW 0x000033FF 32 4.3.35 Architectural

Feature Trap Register, EL2
on page 4-137

y See the ARM® Architecture Reference Manual ARMv8 for more information.
z The reset value of bits[63:32] is 0x00000000.
aa The reset value is the value of the Main ID Register.
ab The reset value is the value of the Multiprocessor Affinity Register.
ac See the ARM® Architecture Reference Manual ARMv8 for more information.
ad The reset value for bit[7] is UNK.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-83
reserved.
Non-Confidential
 4 System Control
4.2 AArch64 register summary

Table 4-13 AArch64 virtualization registers (continued)

Name Type Reset Width Description

HSTR_EL2 RW 0x00000000 32 4.3.36 Hypervisor System

Trap Register
on page 4-139

HACR_EL2 RW 0x00000000 32 4.3.37 Hyp Auxiliary

Configuration Register
on page 4-141

TTBR0_EL2 RW UNK 64 Translation Table Base

Address Register 0, EL3 ac

TCR_EL2 RW UNK 32 4.3.42 Translation Control

Register, EL2
on page 4-149

VTTBR_EL2 RW UNK 64 Virtualization Translation

Table Base Address
Register, EL2 ac

VTCR_EL2 RW UNK 32 4.3.43 Virtualization

Translation Control
Register, EL2
on page 4-151

DACR32_EL2 RW 0x00000000 32 Domain Access Control

Register, EL2 ac

AFSR0_EL2 RW RES0 32 4.3.52 Auxiliary Fault

Status Register 0, EL2 and
Hyp Auxiliary Data Fault
Status Register
on page 4-163

AFSR1_EL2 RW RES0 32 4.3.53 Auxiliary Fault

Status Register 1, EL2 and
Hyp Auxiliary Instruction
Fault Status Register
on page 4-163

ESR_EL2 RW 0x00000000 32 Exception Syndrome

Register, EL2 ac

FAR_EL2 RW UNK 64 Fault Address Register,

EL2 ac

HPFAR_EL2 RW 0x00000000 64 Hyp IPA Fault Address

Register, EL2 ac

MAIR_EL2 RW UNK 64 Memory Attribute

Indirection Register,
EL2 ac

AMAIR_EL2 RW RES0 64 4.3.57 Auxiliary Memory

Attribute Indirection
Register, EL2
on page 4-168

VBAR_EL2 RW UNK 64 Vector Base Address

Register, EL2 ac

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-84
reserved.
Non-Confidential
 4 System Control
4.2 AArch64 register summary

4.2.13 AArch64 EL2 TLB maintenance operations

The following table shows the System instructions for TLB maintenance operations added in AArch64
state. See the ARM® Architecture Reference Manual ARMv8 for more information about these operations.

Table 4-14 AArch64 TLB maintenance operations

Name Description
TLBI IPAS2E1IS Invalidate stage 2 only translations used at EL1 for the specified IPA for the current VMID, Inner Shareable
TLBI IPAS2LE1IS Invalidate entries from the last level of stage 2 only translation used at EL1 for the specified IPA for the
current VMID, Inner Shareable
TLBI ALLE2IS Invalidate all stage 1 translations used at EL2, Inner Shareable
TLBI VAE2IS Invalidate translation used at EL2 for the specified VA and ASID and the current VMID, Inner Shareable
TLBI ALLE1IS Invalidate all stage 1 translations used at EL1, Inner Shareable
TLBI VALE2IS Invalidate all entries from the last level of stage 1 translation table walk used at EL2 with the supplied ASID
and current VMID, Inner Shareable
TLBI VMALLS12E1IS Invalidate all stage 1 and 2 translations used at EL1 with the current VMID, Inner Shareable

TLBI IPAS2E1 Invalidate stage 2 only translations used at EL1 for the specified IPA for the current VMID
TLBI IPAS2LE1 Invalidate entries from the last level of stage 2 only translation used at EL1 for the specified IPA for the
current VMID
TLBI ALLE2 Invalidate all stage 1 translations used at EL2
TLBI VAE2 Invalidate translation used at EL2 for the specified VA and ASID and the current VMID
TLBI ALLE1 Invalidate all stage 1 translations used at EL1
TLBI VALE2 Invalidate all entries from the last level of stage 1 translation table walk used at EL2 with the supplied ASID
and current VMID
TLBI VMALLS12E1 Invalidate all stage 1 and 2 translations used at EL1 with the current VMID
TLBI ALLE3IS Invalidate all stage 1 translations used at EL3, Inner Shareable
TLBI VAE3IS Invalidate translation used at EL3 for the specified VA and ASID and the current VMID, Inner Shareable
TLBI VALE3IS Invalidate all entries from the last level of stage 1 translation table walk used at EL3 with the supplied ASID
and current VMID, Inner Shareable
TLBI ALLE3 Invalidate all stage 1 translations used at EL3
TLBI VAE3 Invalidate translation used at EL3 for the specified VA and ASID and the current VMID
TLBI VALE3 Invalidate all entries from the last level of stage 1 translation table walk used at EL3 with the supplied ASID
and current VMID

4.2.14 Generic Timer registers

See Chapter 9 Generic Timer on page 9-338 for information on the Generic Timer registers.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-85
reserved.
Non-Confidential
 4 System Control
4.2 AArch64 register summary

4.2.15 AArch64 IMPLEMENTATION DEFINED registers

The following table shows the IMPLEMENTATION DEFINED registers in AArch64 state. These registers provide
test features and any required configuration options specific to the Cortex-A72 processor. If a register is
not indicated as mapped to an AArch32 64-bit register, bits[63:32] are 0x00000000.

Table 4-15 AArch64 IMPLEMENTATION DEFINED registers

Name Type Reset Width Description

ACTLR_EL1 RW RES0 32 4.3.31 Auxiliary Control Register, EL1 on page 4-129

ACTLR_EL2 RW 0x00000000 32 4.3.33 Auxiliary Control Register, EL2 on page 4-130

ACTLR_EL3 RW 0x00000000 32 4.3.39 Auxiliary Control Register, EL3 on page 4-143

AFSR0_EL1 RW RES0 32 4.3.48 Auxiliary Fault Status Register 0, EL1 and EL3
on page 4-156

AFSR1_EL1 RW RES0 32 4.3.49 Auxiliary Fault Status Register 1, EL1 and EL3
on page 4-157

AFSR0_EL2 RW RES0 32 4.3.52 Auxiliary Fault Status Register 0, EL2 and Hyp Auxiliary
Data Fault Status Register on page 4-163

AFSR1_EL2 RW RES0 32 4.3.53 Auxiliary Fault Status Register 1, EL2 and Hyp Auxiliary
Instruction Fault Status Register on page 4-163

AFSR0_EL3 RW RES0 32 4.3.48 Auxiliary Fault Status Register 0, EL1 and EL3
on page 4-156

AFSR1_EL3 RW RES0 32 4.3.49 Auxiliary Fault Status Register 1, EL1 and EL3
on page 4-157

AMAIR_EL1 RW RES0 64 4.3.56 Auxiliary Memory Attribute Indirection Register, EL1 and
EL3 on page 4-168

AMAIR_EL2 RW RES0 64 4.3.57 Auxiliary Memory Attribute Indirection Register, EL2

on page 4-168

AMAIR_EL3 RW RES0 64 4.3.57 Auxiliary Memory Attribute Indirection Register, EL2

on page 4-168

L2CTLR_EL1 RW 0x00000000ae 32 4.3.58 L2 Control Register, EL1 on page 4-168

L2ECTLR_EL1 RW 0x00000000 32 4.3.59 L2 Extended Control Register, EL1 on page 4-172

IL1DATA0_EL1 RW UNK 32 4.3.62 Instruction L1 Data n Register, EL1 on page 4-176

IL1DATA1_EL1 UNK

IL1DATA2_EL1 UNK

IL1DATA3_EL1 UNK

DL1DATA0_EL1 RW UNK 32 4.3.63 Data L1 Data n Register, EL1 on page 4-177

DL1DATA1_EL1 UNK

DL1DATA2_EL1 UNK

DL1DATA3_EL1 UNK

DL1DATA4_EL1 UNK

ae The reset value depends on the processor implementation and the state of the L2RSTDISABLE signal.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-86
reserved.
Non-Confidential
 4 System Control
4.2 AArch64 register summary

Table 4-15 AArch64 IMPLEMENTATION DEFINED registers (continued)

Name Type Reset Width Description

RAMINDEX WO - 32 4.3.64 RAM Index operation on page 4-178

L2ACTLR_EL1 RW 0x0000000000000010af 32 L2 Auxiliary Control Register, EL1

CPUACTLR_EL1 RW 0x0000 0000 0000 0000 64 4.3.66 CPU Auxiliary Control Register, EL1 on page 4-194

CPUECTLR_EL1ag RW 0x0000 001B 0000 0000 64 4.3.67 CPU Extended Control Register, EL1 on page 4-206

CPUMERRSR_EL1ag RW UNK 64 4.3.68 CPU Memory Error Syndrome Register, EL1

on page 4-209

L2MERRSR_EL1ag RW UNKah 64 4.3.69 L2 Memory Error Syndrome Register, EL1 on page 4-210

CBAR_EL1 RO UNKai 64 4.3.70 Configuration Base Address Register, EL1 on page 4-212

af This is the reset value for an ACE interface. For a CHI interface the reset value is 0x0000000000004018.
ag Mapped to a 64-bit AArch32 register.
ah Bits[47:40, 39:32, 31] are reset to zero.
ai The reset value depends on the primary input, PERIPHBASE[43:18].

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-87
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

4.3 AArch64 register descriptions

This section describes all the System registers when the processor is in AArch64 state.
Table 4-1 AArch64 identification registers on page 4-75 through Table 4-15 AArch64 implementation
defined registers on page 4-86 provide cross-references to individual registers.
This section contains the following subsections:
• 4.3.1 Main ID Register, EL1 on page 4-89.
• 4.3.2 Multiprocessor Affinity Register, EL1 on page 4-90.
• 4.3.3 Revision ID Register, EL1 on page 4-92.
• 4.3.4 AArch32 Processor Feature Register 0, EL1 on page 4-93.
• 4.3.5 AArch32 Processor Feature Register 1, EL1 on page 4-94.
• 4.3.6 AArch32 Debug Feature Register 0, EL1 on page 4-95.
• 4.3.7 AArch32 Auxiliary Feature Register 0, EL1 on page 4-96.
• 4.3.8 AArch32 Memory Model Feature Register 0, EL1 on page 4-96.
• 4.3.9 AArch32 Memory Model Feature Register 1, EL1 on page 4-98.
• 4.3.10 AArch32 Memory Model Feature Register 2, EL1 on page 4-99.
• 4.3.11 AArch32 Memory Model Feature Register 3, EL1 on page 4-101.
• 4.3.12 AArch32 Instruction Set Attribute Register 0, EL1 on page 4-103.
• 4.3.13 AArch32 Instruction Set Attribute Register 1, EL1 on page 4-104.
• 4.3.14 AArch32 Instruction Set Attribute Register 2, EL1 on page 4-106.
• 4.3.15 AArch32 Instruction Set Attribute Register 3, EL1 on page 4-107.
• 4.3.16 AArch32 Instruction Set Attribute Register 4, EL1 on page 4-108.
• 4.3.17 AArch32 Instruction Set Attribute Register 5, EL1 on page 4-110.
• 4.3.18 AArch64 Processor Feature Register 0, EL1 on page 4-111.
• 4.3.19 AArch64 Debug Feature Register 0, EL1 on page 4-113.
• 4.3.20 AArch64 Instruction Set Attribute Register 0, EL1 on page 4-114.
• 4.3.21 AArch64 Memory Model Feature Register 0, EL1 on page 4-115.
• 4.3.22 Cache Size ID Register, EL1 on page 4-117.
• 4.3.23 Cache Level ID Register, EL1 on page 4-118.
• 4.3.24 Auxiliary ID Register, EL1 on page 4-120.
• 4.3.25 Cache Size Selection Register, EL1 on page 4-120.
• 4.3.26 Cache Type Register, EL0 on page 4-121.
• 4.3.27 Data Cache Zero ID, EL0 on page 4-122.
• 4.3.28 Virtualization Processor ID Register, EL2 on page 4-123.
• 4.3.29 Virtualization Multiprocessor ID Register, EL2 on page 4-124.
• 4.3.30 System Control Register, EL1 on page 4-125.
• 4.3.31 Auxiliary Control Register, EL1 on page 4-129.
• 4.3.32 Architectural Feature Access Control Register, EL1 on page 4-129.
• 4.3.33 Auxiliary Control Register, EL2 on page 4-130.
• 4.3.34 Hypervisor Configuration Register, EL2 on page 4-131.
• 4.3.35 Architectural Feature Trap Register, EL2 on page 4-137.
• 4.3.36 Hypervisor System Trap Register on page 4-139.
• 4.3.37 Hyp Auxiliary Configuration Register on page 4-141.
• 4.3.38 System Control Register, EL3 on page 4-141.
• 4.3.39 Auxiliary Control Register, EL3 on page 4-143.
• 4.3.40 Architectural Feature Trap Register, EL3 on page 4-145.
• 4.3.41 Translation Control Register, EL1 on page 4-146.
• 4.3.42 Translation Control Register, EL2 on page 4-149.
• 4.3.43 Virtualization Translation Control Register, EL2 on page 4-151.
• 4.3.44 Translation Table Base Register 0, EL1 on page 4-152.
• 4.3.45 Translation Table Base Register 0, EL3 on page 4-153.
• 4.3.46 Translation Table Base Register 1, EL1 on page 4-154.
• 4.3.47 Translation Control Register, EL3 on page 4-155.
• 4.3.48 Auxiliary Fault Status Register 0, EL1 and EL3 on page 4-156.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-88
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

• 4.3.49 Auxiliary Fault Status Register 1, EL1 and EL3 on page 4-157.
• 4.3.50 Exception Syndrome Register, EL1 and EL3 on page 4-157.
• 4.3.51 Instruction Fault Status Register, EL2 on page 4-159.
• 4.3.52 Auxiliary Fault Status Register 0, EL2 and Hyp Auxiliary Data Fault Status Register
on page 4-163.
• 4.3.53 Auxiliary Fault Status Register 1, EL2 and Hyp Auxiliary Instruction Fault Status Register
on page 4-163.
• 4.3.54 Exception Syndrome Register, EL2 on page 4-163.
• 4.3.55 Physical Address Register, EL1 on page 4-165.
• 4.3.56 Auxiliary Memory Attribute Indirection Register, EL1 and EL3 on page 4-168.
• 4.3.57 Auxiliary Memory Attribute Indirection Register, EL2 on page 4-168.
• 4.3.58 L2 Control Register, EL1 on page 4-168.
• 4.3.59 L2 Extended Control Register, EL1 on page 4-172.
• 4.3.60 Reset Vector Base Address, EL3 on page 4-174.
• 4.3.61 Reset Management Register, EL3 on page 4-175.
• 4.3.62 Instruction L1 Data n Register, EL1 on page 4-176.
• 4.3.63 Data L1 Data n Register, EL1 on page 4-177.
• 4.3.64 RAM Index operation on page 4-178.
• 4.3.65 L2 Auxiliary Control Register, EL1 on page 4-188.
• 4.3.66 CPU Auxiliary Control Register, EL1 on page 4-194.
• 4.3.67 CPU Extended Control Register, EL1 on page 4-206.
• 4.3.68 CPU Memory Error Syndrome Register, EL1 on page 4-209.
• 4.3.69 L2 Memory Error Syndrome Register, EL1 on page 4-210.
• 4.3.70 Configuration Base Address Register, EL1 on page 4-212.

4.3.1 Main ID Register, EL1

The MIDR_EL1 characteristics are:

Purpose
Provides identification information for the processor, including an implementer code for the
device and a device ID number.
Usage constraints
The accessibility to the MIDR_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

The external debug accessibility to the MIDR by condition code is:

Off DLK OSLK EDAD SLK Default

RO RO RO RO RO RO

Table 10-1 External register access conditions on page 10-347 describes the condition codes.
Configurations
The MIDR_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 MIDR register.
• Architecturally mapped to external MIDR register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the MIDR_EL1 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-89
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

31 24 23 20 19 16 15 4 3 0

Implementer Variant Architecture Primary part number Revision

Figure 4-1 MIDR_EL1 bit assignments

The following table shows the MIDR_EL1 bit assignments.

Table 4-16 MIDR_EL1 bit assignments

Bits Name Function

[31:24] Implementer Indicates the implementer code. This value is:

0x41 ARM Limited.

[23:20] Variant Indicates the variant number of the processor. This is the major revision number n in the rn part of the
rnpn description of the product revision status. This value is:
0 Major revision number.

[19:16] Architecture Indicates the architecture code. This value is:

0xF Defined by CPUID scheme.

[15:4] Primary part number Indicates the primary part number. This value is:
0xD08 Cortex-A72 processor.

[3:0] Revision Indicates the minor revision number of the processor. This is the minor revision number n in the pn
part of the rnpn description of the product revision status. This value is:
3 Minor revision number.

To access the MIDR_EL1 in AArch64 state, read the register with:

MRS <Xt>, MIDR_EL1; Read Main ID Register

To access the MIDR in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c0, 0; Read Main ID Register

The MIDR can be accessed through the memory-mapped interface and the external debug interface,
offset 0xD00.

4.3.2 Multiprocessor Affinity Register, EL1

The MPIDR_EL1 characteristics are:
Purpose
Provides an additional core identification mechanism for scheduling purposes in a cluster
system. EDDEVAFF0 is a read-only copy of MPIDR_EL1[31:0] accessible from the external
debug interface.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-90
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Usage constraints
The accessibility to the MPIDR_EL1 by Exception level is:

EL0 EL1 (NS) EL1 (S) EL2 EL3 (SCR.NS = 1) EL3 (SCR.NS = 0)

- RO RO RO RO RO

The external debug accessibility to the EDDEVAFF0 by condition code is:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 10-1 External register access conditions on page 10-347 describes the condition codes.
Configurations
The MPIDR_EL1[31:0] is:
• Architecturally mapped to the AArch32 MPIDR register. See 4.5.3 Multiprocessor Affinity
Register on page 4-240 for more information.
• Architecturally mapped to external EDDEVAFF0 register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the MPIDR_EL1 bit assignments.
63 32 31 30 29 25 24 23 16 15 8 7 2 1 0

RES0 U RES0 Cluster ID Aff2 Cluster ID Aff1 RES0

RES1 MT CPU ID

Figure 4-2 MPIDR_EL1 bit assignments

The following table shows the MPIDR_EL1 bit assignments.

Table 4-17 MPIDR_EL1 bit assignments

Bits Name Function

[63:32] - Reserved, RES0.

[31] - RES1.

[30] U Indicates a single core system, as distinct from processor 0 in a cluster. This
value is:
0 Core is part of a cluster.

[29:25] - Reserved, RES0.

[24] MT Indicates whether the lowest level of affinity consists of logical cores that are
implemented using a multi-threading type approach. This value is:
0 Performance of cores at the lowest affinity level is largely independent.

[23:16] Cluster ID Aff2 Affinity level 2. Second highest level affinity field.
Indicates the value read in at reset, from the CLUSTERIDAFF2
configuration signal.

[15:8] Cluster ID Aff1 Affinity level 1. Third highest level affinity field.
Indicates the value read in at reset, from the CLUSTERIDAFF1
configuration signal.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-91
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-17 MPIDR_EL1 bit assignments (continued)

Bits Name Function

[7:2] - Reserved, RES0.

[1:0] CPU ID Indicates the core number in the Cortex-A72 processor. The possible values
are:
0x0 A cluster with one processor only.
0x0, 0x1 A cluster with two processors.
0x0, 0x1, 0x2 A cluster with three processors.
0x0, 0x1, 0x2, 0x3 A cluster with four processors.

To access the MPIDR_EL1 in AArch64 state, read the register with:

MRS <Xt>, MPIDR_EL1; Read Multiprocessor Affinity Register

The EDDEVAFF0 can be accessed through the memory-mapped interface and the external debug
interface, offset 0xFA8.

4.3.3 Revision ID Register, EL1

The REVIDR_EL1 characteristics are:

Purpose
Provides implementation-specific minor revision information that can only be interpreted in
conjunction with the MIDR_EL1.
Usage constraints
The accessibility to the REVIDR_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The REVIDR_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 REVIDR register.
The REVIDR_EL1 is a 32-bit register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the REVIDR_EL1 bit assignments.
31 0

ID number

Figure 4-3 REVIDR_EL1 bit assignments

The following table shows the REVIDR_EL1 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-92
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-18 REVIDR_EL1 bit assignments

Bits Name Function

[31:0] ID number Implementation-specific revision information. The reset value is determined by the specific Cortex-A72 processor
implementation.

To access the REVIDR_EL1 in AArch64 state, read the register with:

MRS <Xt>, REVIDR_EL1; Read Revision ID Register

To access the REVIDR in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c0, 6; Read Revision ID Register

4.3.4 AArch32 Processor Feature Register 0, EL1

The ID_PFR0_EL1 characteristics are:
Purpose
Provides information about the instruction sets supported by the processor in AArch32 state.
Usage constraints
The ID_PFR0_EL1 must be interpreted with the ID_PFR1_EL1.
The accessibility to the ID_PFR0_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The ID_PFR0_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 ID_PFR0 register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_PFR0_EL1 bit assignments.
31 16 15 12 11 8 7 4 3 0

RES0 State3 State2 State1 State0

Figure 4-4 ID_PFR0_EL1 bit assignments

The following table shows the ID_PFR0_EL1 bit assignments.

Table 4-19 ID_PFR0_EL1 bit assignments

Bits Name Function

[31:16] - Reserved, RES0.

[15:12] State3 Indicates support for Thumb Execution Environment (ThumbEE) instruction set. This value is:
0x0 Processor does not implement the ThumbEE instruction set.

[11:8] State2 Indicates support for Jazelle extension. This value is:
0x1 Processor supports trivial implementation of Jazelle.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-93
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-19 ID_PFR0_EL1 bit assignments (continued)

Bits Name Function

[7:4] State1 Indicates support for T32 instruction set. This value is:
0x3 Processor supports T32 encoding after the introduction of Thumb-2 technology, and for all 16-bit and 32-bit
T32 basic instructions.

[3:0] State0 Indicates support for A32 instruction set. This value is:
0x1 Processor implements the A32 instruction set.

To access the ID_PFR0_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_PFR0_EL1; Read AArch32 Processor Feature Register 0

To access the ID_PFR0 in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c1, 0; Read AArch32 Processor Feature Register 0

4.3.5 AArch32 Processor Feature Register 1, EL1

The ID_PFR1_EL1 characteristics are:

Purpose
Provides information about the programmers model and extensions support in AArch32.
Usage constraints
The ID_PFR1_EL1 must be interpreted with the ID_PFR0_EL1.
The accessibility to the ID_PFR1_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The ID_PFR1_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 ID_PFR1 register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_PFR1_EL1 bit assignments.
31 28 27 20 19 16 15 12 11 8 7 4 3 0

GIC CP15 RES0 GenTimer MProgMod Security ProgMod

Virtualization

Figure 4-5 ID_PFR1_EL1 bit assignments

The following table shows the ID_PFR1_EL1 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-94
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-20 ID_PFR1_EL1 bit assignments

Bits Name Function

[31:28] GIC CP15 Indicates support for the GIC CP15 interface. The possible values are:
0x0 No GIC CP15 registers are supported. This is the reset value when GICCDISABLE is tied
HIGH.
0x1 GICv3 CP15 registers are supported. This is the reset value when GICCDISABLE is tied
LOW.

[27:20] - Reserved, RES0.

[19:16] GenTimer Indicates support for Generic Timer Extension. This value is:
0x1 Processor supports Generic Timer Extension.

[15:12] Virtualization Indicates support for Virtualization Extensions. This value is:
0x1 Processor supports Virtualization Extensions.

[11:8] MProgMod Indicates support for M-profile programmers model. This value is:
0x0 Processor does not support M-profile programmers model.

[7:4] Security Indicates support for Security Extensions. This value is:
0x1 Processor supports Security Extensions. This includes support for Monitor mode and the
SMC instruction.

[3:0] ProgrMod Indicates support for the standard programmers model for ARMv4 and later. This value is:
0x1 Processor supports the standard programmers model for ARMv4 and later. The model
supports User, FIQ, IRQ, Supervisor, Abort, Undefined, and System modes.

To access the ID_PFR1_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_PFR1_EL1; Read AArch32 Processor Feature Register 1

To access the ID_PFR1 in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c1, 1; Read AArch32 Processor Feature Register 1

4.3.6 AArch32 Debug Feature Register 0, EL1

The ID_DFR0_EL1 characteristics are:

Purpose
Provides top-level information about the debug system in AArch32 state.
Usage constraints
The ID_DFR0_EL1 must be interpreted with the MIDR_EL1.
The accessibility to the ID_DFR0_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The ID_DFR0_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 ID_DFR0 register.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-95
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_DFR0_EL1 bit assignments.
31 28 27 24 23 20 19 16 15 12 11 8 7 4 3 0

RES0 PerfMon MProfDbg MMapTrc CopTrc MMapDbg CopSDbg CopDbg

Figure 4-6 ID_DFR0_EL1 bit assignments

The following table shows the ID_DFR0_EL1 bit assignments.

Table 4-21 ID_DFR0_EL1 bit assignments

Bits Name Function

[31:28] - Reserved, RES0.

[27:24] PerfMon Indicates support for coprocessor-based ARM Performance Monitors Extension. This value is:
0x3 Processor supports Performance Monitors Extension, PMUv3 System registers.

[23:20] MProfDbg Indicates support for memory-mapped debug model for M-profile processors. This value is:
0x0 Processor does not support M-profile Debug architecture, with memory-mapped access.

[19:16] MMapTrc Indicates support for memory-mapped trace model. This value is:
0x1 Processor supports ARM trace architecture, with memory-mapped access.

[15:12] CopTrc Indicates support for coprocessor-based trace model. This value is:
0x0 Processor does not support ARM trace architecture, with CP14 access.

[11:8] MMapDbg Indicates support for memory-mapped debug model. This value is:
0x0 Processor does not support the memory-mapped debug model.

[7:4] CopSDbg Indicates support for coprocessor-based Secure debug model. This value is:
0x6 Processor supports v8-A Debug architecture, with CP14 access.

[3:0] CopDbg Indicates support for coprocessor-based debug model. This value is:
0x6 Processor supports v8-A Debug architecture, with CP14 access.

To access the ID_DFR0_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_DFR0_EL1; Read AArch32 Debug Feature Register 0

To access the ID_DFR0 in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c1, 2; Read AArch32 Debug Feature Register 0

4.3.7 AArch32 Auxiliary Feature Register 0, EL1

The processor does not implement ID_AFR0_EL1. This register is always RES0.

4.3.8 AArch32 Memory Model Feature Register 0, EL1

The ID_MMFR0_EL1 characteristics are:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-96
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Purpose
Provides information about the implemented memory model and memory management support
in AArch32.
Usage constraints
The ID_MMFR0_EL1 must be interpreted with:
• ID_MMFR1_EL1.
• ID_MMFR2_EL1.
• ID_MMFR3_EL1.
The accessibility to the ID_MMFR0_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The ID_MMFR0_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 ID_MMFR0 register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_MMFR0_EL1 bit assignments.
31 28 27 24 23 20 19 16 15 12 11 8 7 4 3 0

InnerShr FCSE AuxReg TCM ShareLvl OuterShr PMSA VMSA

Figure 4-7 ID_MMFR0_EL1 bit assignments

The following table shows the ID_MMFR0_EL1 bit assignments.

Table 4-22 ID_MMFR0_EL1 bit assignments

Bits Name Function

[31:28] InnerShr Indicates the innermost shareability domain implemented. This value is:
0x1 Processor implements hardware coherency support.

[27:24] FCSE Indicates support for Fast Context Switch Extension (FCSE). This value is:
0x0 Processor does not support FCSE.

[23:20] AuxReg Indicates support for Auxiliary registers. This value is:
0x2 Processor supports the ACTLR, AIFSR and ADFSR. See 4.3.39 Auxiliary Control Register,
EL3 on page 4-143,4.3.48 Auxiliary Fault Status Register 0, EL1 and EL3 on page 4-156, and
4.3.49 Auxiliary Fault Status Register 1, EL1 and EL3 on page 4-157.

[19:16] TCM Indicates support for TCMs and associated DMAs. This value is:
0x0 Processor does not support TCM.

[15:12] ShareLvl Indicates the number of shareability levels implemented. This value is:
0x1 Processor implements two levels of shareability.

[11:8] OuterShr Indicates the outermost shareability domain implemented. This value is:
0x1 Processor supports hardware coherency.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-97
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-22 ID_MMFR0_EL1 bit assignments (continued)

Bits Name Function

[7:4] PMSA Indicates support for a Protected Memory System Architecture (PMSA). This value is:
0x0 Processor does not support PMSA.

[3:0] VMSA Indicates support for a Virtual Memory System Architecture (VMSA). This value is:
0x5 Processor supports:
• VMSAv7, with support for remapping and the Access flag
• Privileged Execute Never (PXN) bit in the Short-descriptor translation table format
• The Long-descriptor translation table format.

To access the ID_MMFR0_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_MMFR0_EL1; Read AArch32 Memory Model Feature Register 0

To access the ID_MMFR0 in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c1, 4; Read AArch32 Memory Model Feature Register 0

4.3.9 AArch32 Memory Model Feature Register 1, EL1

The ID_MMFR1_EL1 characteristics are:

Purpose
Provides information about the implemented memory model and memory management support
in AArch32.
Usage constraints
The ID_MMFR1_EL1 must be interpreted with:
• ID_MMFR0_EL1.
• ID_MMFR2_EL1.
• ID_MMFR3_EL1.
The accessibility to the ID_MMFR1_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The ID_MMFR1_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 ID_MMFR1 register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_MMFR1_EL1 bit assignments.
31 28 27 24 23 20 19 16 15 12 11 8 7 4 3 0

BPred L1TstCln L1Uni L1Hvd L1UniSW L1HvdSW L1UniVA L1HvdVA

Figure 4-8 ID_MMFR1_EL1 bit assignments

The following table shows the ID_MMFR1 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-98
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-23 ID_MMFR1_EL1 bit assignments

Bits Name Function

[31:28] BPred Indicates branch predictor management requirements. This value is:
0x4 Branch predictor does not require flushing at any time.

[27:24] L1TstCln Indicates the supported L1 data cache test and clean operations, for Harvard or unified cache implementation.
This value is:
0x0 Not supported.

[23:20] L1Uni Indicates the supported entire L1 cache maintenance operations, for a unified cache implementation. This value
is:
0x0 Not supported.

[19:16] L1Hvd Indicates the supported entire L1 cache maintenance operations, for a Harvard cache implementation. This value
is:
0x0 Not supported.

[15:12] L1UniSW Indicates the supported L1 cache line maintenance operations by set/way, for a unified cache implementation.
This value is:
0x0 Not supported.

[11:8] L1HvdSW Indicates the supported L1 cache line maintenance operations by set/way, for a Harvard cache implementation.
This value is:
0x0 Not supported.

[7:4] L1UniVA Indicates the supported L1 cache line maintenance operations by VA, for a unified cache implementation. This
value is:
0x0 Not supported.

[3:0] L1HvdVA Indicates the supported L1 cache line maintenance operations by VA, for a Harvard cache implementation. This
value is:
0x0 Not supported.

To access the ID_MMFR1_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_MMFR1_EL1; Read AArch32 Memory Model Feature Register 1

To access the ID_MMFR1 in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c1, 5; Read AArch32 Memory Model Feature Register 1

4.3.10 AArch32 Memory Model Feature Register 2, EL1

The ID_MMFR2_EL1 characteristics are:

Purpose
Provides information about the implemented memory model and memory management support
of the processor in AArch32.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-99
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Usage constraints
The ID_MMFR2_EL1 must be interpreted with:
• ID_MMFR0_EL1.
• ID_MMFR1_EL1.
• ID_MMFR3_EL1.
The accessibility to the ID_MMFR2_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The ID_MMFR2_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 ID_MMFR2 register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_MMFR2_EL1 bit assignments.
31 28 27 24 23 20 19 16 15 12 11 8 7 4 3 0

HWAccFlg WFIStall MemBarr UniTLB HvdTLB L1HvdRng L1HvdBG L1HvdFG

Figure 4-9 ID_MMFR2_EL1 bit assignments

The following table shows the ID_MMFR2_EL1 bit assignments.

Table 4-24 ID_MMFR2_EL1 bit assignments

Bits Name Function

[31:28] HWAccFlg Indicates support for Hardware Access flag. This value is:
0x0 Not supported.

[27:24] WFIStall Indicates support for Wait For Interrupt (WFI) stalling. This value is:
0x1 Processor supports WFI stalling.

[23:20] MemBarr Indicates the supported CP15 memory barrier operations. This value is:
0x2 Processor supports:
• Data Synchronization Barrier (DSB).
• Instruction Synchronization Barrier (ISB).
• Data Memory Barrier (DMB).

ARM deprecates the use of these CP15 operations. Instead, use the DMB, DSB, and ISB barrier instructions.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-100
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-24 ID_MMFR2_EL1 bit assignments (continued)

Bits Name Function

[19:16] UniTLB Indicates the supported TLB maintenance operations. This value is:
0x6 Processor supports:
• Invalidate all entries in the TLB.
• Invalidate TLB entry by VA.
• Invalidate TLB entries by ASID match.
• Invalidate instruction TLB and data TLB entries by VA All ASID. This is a shared
unified TLB operation.
• Invalidate Hyp mode unified TLB entry by VA.
• Invalidate entire Non-secure PL1 and PL0 unified TLB.
• Invalidate entire Hyp mode unified TLB.
• Invalidate TLB entry by VA, Last Level.
• Invalidate TLB entry by VA and ASID, Last Level.
• Invalidate Stage 2 TLB only by IPA.
• Invalidate Stage 2 TLB only by IPA, Last Level.

[15:12] HvdTLB Indicates support for Harvard TLB maintenance operations. This value is:
0x0 Not supported.

[11:8] L1HvdRng Indicates support for Harvard L1 cache maintenance range operations. This value is:
0x0 Not supported.

[7:4] L1HvdBG Indicates support for Harvard L1 cache background fetch operations. This value is:
0x0 Not supported.

[3:0] L1HvdFG Indicates support for Harvard L1 cache foreground fetch operations. This value is:
0x0 Not supported.

To access the ID_MMFR2_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_MMFR2_EL1; Read AArch32 Memory Model Feature Register 2

To access the ID_MMFR2 in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c1, 6; Read AArch32 Memory Model Feature Register 2

4.3.11 AArch32 Memory Model Feature Register 3, EL1

The ID_MMFR3_EL1 characteristics are:

Purpose
Provides information about the implemented memory model and memory management support
of the processor in AArch32.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-101
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Usage constraints
The ID_MMFR3_EL1 must be interpreted with:
• ID_MMFR0_EL1.
• ID_MMFR1_EL1.
• ID_MMFR2_EL1.
The accessibility to the ID_MMFR3_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The ID_MMFR3_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 ID_MMFR3 register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_MMFR3_EL1 bit assignments.
31 28 27 24 23 20 19 16 15 12 11 8 7 4 3 0

Supersec CMemSz CohWalk RES0 MaintBcst BPMaint CMaintSW CMaintVA

Figure 4-10 ID_MMFR3_EL1 bit assignments

The following table shows the ID_MMFR3_EL1 bit assignments.

Table 4-25 ID_MMFR3_EL1 bit assignments

Bits Name Function

[31:28] Supersec Indicates support for supersections. This value is:

0x0 Processor supports supersections.

[27:24] CMemSz Indicates the physical memory size supported by the processor caches. This value is:
0x2 Processor caches support 40-bit physical address range.

[23:20] CohWalk Indicates whether translation table updates require a clean to the point of unification. This value is:
0x1 Updates to the translation tables do not require a clean to the point of unification to ensure
visibility by subsequent translation table walks.

[19:16] - Reserved, RES0.

[15:12] MaintBcst Indicates whether cache, TLB and branch predictor operations are broadcast. This value is:
0x2 Cache, TLB and branch predictor operations affect structures according to shareability and
defined behavior of instructions.

[11:8] BPMaint Indicates the supported branch predictor maintenance operations. This value is:
0x2 Processor supports:
• Invalidate all branch predictors.
• Invalidate branch predictors by VA.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-102
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-25 ID_MMFR3_EL1 bit assignments (continued)

Bits Name Function

[7:4] CMaintSW Indicates the supported cache maintenance operations by set/way. This value is:
0x1 Processor supports:
• Invalidate data cache by set/way.
• Clean data cache by set/way.
• Clean and invalidate data cache by set/way.

[3:0] CMaintVA Indicates the supported cache maintenance operations by VA. This value is:
0x1 Processor supports:
• Invalidate data cache by VA.
• Clean data cache by VA.
• Clean and invalidate data cache by VA.
• Invalidate Instruction Cache by VA.
• Invalidate all Instruction Cache entries.

To access the ID_MMFR3_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_MMFR3_EL1; Read AArch32 Memory Model Feature Register 3

To access the ID_MMFR3 in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c1, 7; Read AArch32 Memory Model Feature Register 3

4.3.12 AArch32 Instruction Set Attribute Register 0, EL1

The ID_ISAR0_EL1 characteristics are:

Purpose
Provides information about the instruction set that the processor supports in AArch32.
Usage constraints
The ID_ISAR0_EL1 must be interpreted with:
• ID_ISAR1_EL1.
• ID_ISAR2_EL1.
• ID_ISAR3_EL1.
• ID_ISAR4_EL1.
• ID_ISAR5_EL1.
The accessibility to the ID_ISAR0_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The ID_ISAR0_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 ID_ISAR0 register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_ISAR0_EL1 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-103
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

31 28 27 24 23 20 19 16 15 12 11 8 7 4 3 0

RES0 Divide Debug Coproc CmpBranch BitField BitCount Swap

Figure 4-11 ID_ISAR0_EL1 bit assignments

The following table shows the ID_ISAR0_EL1 bit assignments.

Table 4-26 ID_ISAR0_EL1 bit assignments

Bits Name Function

[31:28] - Reserved, RES0.

[27:24] Divide Returns 0x2 to indicate the processor implements the following divide instructions:
• SDIV and UDIV in the T32 instruction set.
• SDIV and UDIV in the A32 instruction set.

[23:20] Debug Returns 0x1 to indicate the processor implements the BKPT debug instruction.

[19:16] Coproc Returns 0x0 to indicate the processor implements no coprocessor instructions, except for separately attributed
architectures including CP15, CP14, and Advanced SIMD and FP.

[15:12] CmpBranch Returns 0x1 to indicate the processor implements the CBNZ and CBZ, Compare and Branch, instructions in the
T32 instruction set.

[11:8] Bitfield Returns 0x1 to indicate the processor implements the BFC, BFI, SBFX, and UBFX, bit field instructions.

[7:4] BitCount Returns 0x1 to indicate the processor implements the CLZ bit counting instruction.

[3:0] Swap Returns 0x0 to indicate the processor implements no swap instructions in the A32 instruction set.

To access the ID_ISAR0_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_ISAR0_EL1; Read AArch32 Instruction Set Attribute Register 0

To access the ID_ISAR0 in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c2, 0; Read AArch32 Instruction Set Attribute Register 0

4.3.13 AArch32 Instruction Set Attribute Register 1, EL1

The ID_ISAR1_EL1 characteristics are:

Purpose
Provides information about the instruction set that the processor supports in AArch32.
Usage constraints
The ID_ISAR1_EL1 must be interpreted with:
• ID_ISAR0_EL1.
• ID_ISAR2_EL1.
• ID_ISAR3_EL1.
• ID_ISAR4_EL1.
• ID_ISAR5_EL1.
The accessibility to the ID_ISAR1_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-104
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Configurations
The ID_ISAR1_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 ID_ISAR1 register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_ISAR1_EL1 bit assignments.
31 28 27 24 23 20 19 16 15 12 11 8 7 4 3 0

Jazelle Interwork Immediate IfThen Extend Except_AR Except Endian

Figure 4-12 ID_ISAR1_EL1 bit assignments

The following table shows the ID_ISAR1_EL1 bit assignments.

Table 4-27 ID_ISAR1_EL1 bit assignments

Bits Name Function

[31:28] Jazelle Returns 0x1 to indicate the processor implements the BXJ instruction, and the J bit in the PSR.

[27:24] Interwork Returns 0x3 to indicate the processor implements the following interworking instructions:
• BX instruction, and the T bit in the PSR.
• BLX instruction, and PC loads have BX-like behavior.
• Data-processing instructions in the A32 instruction set with the PC as the destination and the S bit clear
have BX-like behavior.

[23:20] Immediate Returns 0x1 to indicate the processor implements the following data-processing instructions with long
immediates:
• MOVT instruction.
• MOV instruction encoding with zero-extended 16-bit immediates.
• Thumb ADD and SUB instruction encoding with zero-extended 12-bit immediates, and other ADD, ADR, and
SUB encoding cross-referenced by the pseudocode for those encodings.

[19:16] IfThen Returns 0x1 to indicate the processor implements the IT instruction and the IT bits in the PSRs, in the T32
instruction set.

[15:12] Extend Returns 0x2 to indicate the processor implements the following Extend instructions:
• SXTB, SXTH, UXTB, and UXTH instructions.
• SXTB16, SXTAB, SXTAB16, SXTAH, UXTB16, UXTAB, UXTAB16, and UXTAH instructions. See the ARM®
Architecture Reference Manual ARMv8 for more information.

[11:8] Except_AR Returns 0x1 to indicate the processor implements the SRS, RFE, and CPS exception-handling instructions.

[7:4] Except Returns 0x1 to indicate the processor implements the LDM (exception return), LDM (user registers), and STM
(user registers) exception-handling instructions in the A32 instruction set.

[3:0] Endian Returns 0x1 to indicate the processor implements the SETEND instruction, and the E bit in the PSRs.

To access the ID_ISAR1_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_ISAR1_EL1; Read AArch32 Instruction Set Attribute Register 1

To access the ID_ISAR1 in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c2, 1; Read AArch32 Instruction Set Attribute Register 1

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-105
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

4.3.14 AArch32 Instruction Set Attribute Register 2, EL1

The ID_ISAR2_EL1 characteristics are:

Purpose
Provides information about the instruction set that the processor supports in AArch32.
Usage constraints
The ID_ISAR2 must be interpreted with:
• ID_ISAR0_EL1.
• ID_ISAR1_EL1.
• ID_ISAR3_EL1.
• ID_ISAR4_EL1.
• ID_ISAR5_EL1.
The accessibility to the ID_ISAR2_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The ID_ISAR2_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 ID_ISAR2 register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_ISAR2_EL1 bit assignments.
31 28 27 24 23 20 19 16 15 12 11 8 7 4 3 0

Reversal PSR_AR MultU MultS Mult MemHint LoadStore

MultiAccessInt

Figure 4-13 ID_ISAR2_EL1 bit assignments

The following table shows the ID_ISAR2_EL1 bit assignments.

Table 4-28 ID_ISAR2_EL1 bit assignments

Bits Name Function

[31:28] Reversal Returns 0x2 to indicate the processor implements the following Reversal instructions:
• REV, REV16, and REVSH.
• RBIT.

[27:24] PSR_AR Returns 0x1 to indicate the processor implements the following instructions that can manipulate the PSR:
• Processor supports MRS and MSR instructions, and the exception return forms of data-processing
instructions. See the ARM® Architecture Reference Manual ARMv8 for more information.

[23:20] MultU Returns 0x2 to indicate the processor implements the UMULL, UMLAL, and UMAAL unsigned multiply
instructions.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-106
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-28 ID_ISAR2_EL1 bit assignments (continued)

Bits Name Function

[19:16] MultS Returns 0x3 to indicate the processor implements the following signed multiply instructions:
• SMULL and SMLAL instructions
• SMLABB, SMLABT, SMLALBB,SMLALBT, SMLALTB, SMLALTT, SMLATB, SMLATT, SMLAWB, SMLAWT,
SMULBB, SMULBT, SMULTB, SMULTT, SMULWB, SMULWT instructions, and the Q bit in the PSRs.
• SMLAD, SMLADX, SMLALD, SMLALDX, SMLSD, SMLSDX, SMLSLD, SMLSLDX, SMMLA, SMMLAR, SMMLS,
SMMLSR, SMMUL, SMMULR, SMUAD, SMUADX, SMUSD, and SMUSDX instructions.

[15:12] Mult Returns 0x2 to indicate the processor implements the MUL, MLA, and MLS multiply instructions.

[11:8] MultiAccessInt Returns 0x0 to indicate no support for interruptible multi-access instructions. This means that the LDM and
STM instructions are not interruptible.

[7:4] MemHint Returns 0x4 to indicate the processor implements the PLD, PLI (NOP), and PLDW memory hint instructions.

[3:0] LoadStore Returns 0x2 to indicate the processor implements the following additional load/store instructions and Load-
Acquire/Store-Release instructions:
• LDRD and STRD load/store instructions.
• STRLB, STRLH, STRL, LDRAB, LDRAH, and LDRA Load-Acquire and Store-Release instructions.

To access the ID_ISAR2_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_ISAR2_EL1; Read AArch32 Instruction Set Attribute Register 2

To access the ID_ISAR2 in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c2, 2; Read AArch32 Instruction Set Attribute Register 2

4.3.15 AArch32 Instruction Set Attribute Register 3, EL1

The ID_ISAR3_EL1 characteristics are:

Purpose
Provides information about the instruction set that the processor supports in AArch32.
Usage constraints
The ID_ISAR3 must be interpreted with:
• ID_ISAR0_EL1.
• ID_ISAR1_EL1.
• ID_ISAR2_EL1.
• ID_ISAR4_EL1.
• ID_ISAR5_EL1.
The accessibility to the ID_ISAR3_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The ID_ISAR3_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 ID_ISAR3 register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_ISAR3_EL1 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-107
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

31 28 27 24 23 20 19 16 15 12 11 8 7 4 3 0

ThumbEE TrueNOP ThumbCopy TabBranch SynchPrim SVC SIMD Saturate

Figure 4-14 ID_ISAR3_EL1 bit assignments

The following table shows the ID_ISAR3_EL1 bit assignments.

Table 4-29 ID_ISAR3_EL1 bit assignments

Bits Name Function

[31:28] ThumbEE Returns 0x0 to indicate no support for Thumb Execution Environment (ThumbEE) extension instructions.

[27:24] TrueNOP Returns 0x1 to indicate the processor implements true NOP instructions in both the A32 and T32 instruction
sets, and additional NOP-compatible hints.

[23:20] ThumbCopy Returns 0x1 to indicate the processor supports T32 instruction set encoding T1 of the MOV (register)
instruction, copying from a low register to a low register.

[19:16] TabBranch Returns 0x1 to indicate the processor implements the TBB and TBH table branch instructions in the T32
instruction set.

[15:12] SynchPrim This field is used with the SynchPrim_frac field of ID_ISAR4 to indicate the supported Synchronization
Primitive instructions. This value is:
0x2 Processor supports:
• LDREX and STREX instructions.
• CLREX, LDREXB, LDREXH, STREXB, and STREXH instructions.
• LDREXD and STREXD instructions.

[11:8] SVC Returns 0x1 to indicate the processor implements the SVC instruction.

[7:4] SIMD Returns 0x3 to indicate the processor implements the following Single Instruction Multiple Data (SIMD)
instructions:
• SSAT and USAT instructions, and the Q bit in the PSRs.
• PKHBT, PKHTB, QADD16, QADD8, QASX, QSUB16, QSUB8, QSAX, SADD16, SADD8, SASX, SEL, SHADD16,
SHADD8, SHASX, SHSUB16, SHSUB8, SHSAX, SSAT16, SSUB16, SSUB8, SSAX, SXTAB16, SXTB16,
UADD16, UADD8, UASX, UHADD16, UHADD8, UHASX, UHSUB16, UHSUB8, UHSAX, UQADD16, UQADD8,
UQASX, UQSUB16, UQSUB8, UQSAX, USAD8, USADA8, USAT16, USUB16, USUB8, USAX, UXTAB16, UXTB16
instructions, and the GE[3:0] bits in the PSRs.
See the ARM® Architecture Reference Manual ARMv8 for more information.

[3:0] Saturate Returns 0x1 to indicate the processor implements the QADD, QDADD, QDSUB, QSUB saturate instructions and the
Q bit in the PSRs.

To access the ID_ISAR3_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_ISAR3_EL1; Read AArch32 Instruction Set Attribute Register 3

To access the ID_ISAR3 in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c2, 3; Read AArch32 Instruction Set Attribute Register 3

4.3.16 AArch32 Instruction Set Attribute Register 4, EL1

The ID_ISAR4_EL1 characteristics are:

Purpose
Provides information about the instruction set that the processor supports in AArch32.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-108
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Usage constraints
The ID_ISAR4_EL1 must be interpreted with:
• ID_ISAR0_EL1.
• ID_ISAR1_EL1.
• ID_ISAR2_EL1.
• ID_ISAR3_EL1.
• ID_ISAR5_EL1.
The accessibility to the ID_ISAR4_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The ID_ISAR4_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 ID_ISAR4 register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_ISAR4_EL1 bit assignments.
31 28 27 24 23 20 19 16 15 12 11 8 7 4 3 0

SWP_frac PSR_M Barrier SMC Writeback WithShifts Unpriv

SynchPrim_frac

Figure 4-15 ID_ISAR4_EL1 bit assignments

The following table shows the ID_ISAR4_EL1 bit assignments.

Table 4-30 ID_ISAR4_EL1 bit assignments

Bits Name Function

[31:28] SWP_frac Returns 0x0 to indicate that SWP or SWPB instructions are not implemented.

[27:24] PSR_M Returns 0x0 to indicate that M-profile instructions, that modify the PSRs, are not implemented.

[23:20] SynchPrim_frac This field is used with the SynchPrim field of ID_ISAR3_EL1 to indicate the supported Synchronization
Primitive instructions. This value is:
0x0 Processor supports:
• LDREX and STREX instructions.
• CLREX, LDREXB, LDREXH, STREXB, and STREXH instructions.
• LDREXD and STREXD instructions.

[19:16] Barrier Returns 0x1 to indicate the processor implements the DMB, DSB, and ISB barrier instructions in the A32
and T32 instruction sets.

[15:12] SMCs Returns 0x1 to indicate the processor implements the SMC instruction.

[11:8] Writeback Returns 0x1 to indicate the processor supports all writeback addressing modes defined in ARMv8
architecture.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-109
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-30 ID_ISAR4_EL1 bit assignments (continued)

Bits Name Function

[7:4] WithShifts Returns 0x4 to indicate the processor supports the following instructions with shifts:
• Shifts of loads and stores over the range LSL 0-3.
• Constant shift options, both on load/store and other instructions.
• Register-controlled shift options.
See the ARM® Architecture Reference Manual ARMv8 for more information.

[3:0] Unpriv Returns 0x2 to indicate the processor implements the following unprivileged instructions:
• LDRBT, LDRT, STRBT, and STRT.
• LDRHT, LDRSBT, LDRSHT, and STRHT.

To access the ID_ISAR4_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_ISAR4_EL1; Read AArch32 Instruction Set Attribute Register 4

To access the ID_ISAR4 in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c2, 4; Read AArch32 Instruction Set Attribute Register 4

4.3.17 AArch32 Instruction Set Attribute Register 5, EL1

The ID_ISAR5_EL1 characteristics are:

Purpose
Provides information about the Cryptography Extension instruction set that the processor can support in
AArch32.
Note
• The optional Cryptography engine is not included in the base product of the processor. ARM requires
licensees to have contractual rights to obtain the Cortex-A72 Cryptography engine.
• The SHA1, SHA2, and AES fields of ID_ISAR5_EL1 are 0x0 if the Cryptography engine is not
included or CRYPTODISABLE is tied HIGH.

Usage constraints
The ID_ISAR5_EL1 must be interpreted with:
• ID_ISAR0_EL1.
• ID_ISAR1_EL1.
• ID_ISAR2_EL1.
• ID_ISAR3_EL1.
• ID_ISAR4_EL1.
The accessibility to the ID_ISAR5_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The ID_ISAR5_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 ID_ISAR5 register.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-110
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_ISAR5_EL1 bit assignments.

31 20 19 16 15 12 11 8 7 4 3 0

Reserved CRC32 SHA2 SHA1 AES SEVL

Figure 4-16 ID_ISAR5_EL1 bit assignments

The following table shows the ID_ISAR5_EL1 bit assignments.

Table 4-31 ID_ISAR5_EL1 bit assignments

Bits Name Function

[31:20] - Reserved, RES0.

[19:16] CRC32 Returns 0x1 to indicate that CRC32 instructions are implemented in AArch32 state.

[15:12] SHA2 Indicates whether SHA2 instructions are implemented in AArch32 state. The possible values are:
0x0 SHA2 instructions are not implemented in AArch32 state.
0x1 SHA256H, SHA256H2, SHA256SU0, and SHA256SU1 instructions are implemented.

All other values are reserved.

[11:8] SHA1 Indicates whether SHA1 instructions are implemented in AArch32 state. The possible values are:
0x0 SHA1 instructions are not implemented in AArch32 state.
0x1 SHA1C, SHA1P, SHA1M, SHA1H, SHA1SU0, and SHA1SU1 instructions are implemented.

All other values are reserved.

[7:4] AES Indicates whether AES instructions are implemented in AArch32 state. The possible values are:
0x0 AES instructions are not implemented in AArch32 state.
0x2 AESE, AESD, AESMC, AESIMC, and PMULL/PMULL2 instructions operating on 64-bit data.

All other values are reserved.

[3:0] SEVL Returns 0x1 to indicate that the SEVL instruction is implemented in AArch32 state.

To access the ID_ISAR5_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_ISAR5_EL1; Read AArch32 Instruction Set Attribute Register 5

To access the ID_ISAR5 in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c2, 5; Read AArch32 Instruction Set Attribute Register 5

4.3.18 AArch64 Processor Feature Register 0, EL1

The ID_AA64PFR0_EL1 characteristics are:
Purpose
Provides information on the exception handling of the processor in AArch64 state.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-111
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Usage constraints
The accessibility to the ID_AA64PFR0_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

The external debug accessibility to ID_AA64PFR0_EL1[63:32] and ID_AA64PFR0_EL1[31:0]

by condition code is:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
The ID_AA64PFR0 is architecturally mapped as follows:
• [63:32] to external ID_AA64PFR0[63:32] register.
• [31:0] to external ID_AA64PFR0[31:0] register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_AA64PFR0_EL1 bit assignments.
63 28 27 24 23 20 19 16 15 12 11 8 7 4 3 0
GIC system
RES0 AdvSIMD FP EL3 EL2 EL1 EL0
registers

Figure 4-17 ID_AA64PFR0_EL1 bit assignments

The following table shows the ID_AA64PFR0_EL1 bit assignments.

Table 4-32 ID_AA64PFR0_EL1 bit assignments

Bits Name Function

[63:28] - Reserved, RES0.

[27:24] GIC system registers Indicates support for the GIC System register interface. The possible values are:
0x0 No GIC System registers are supported. This is the reset value when GICCDISABLE is tied
HIGH.
0x1 GICv3 System registers are supported. This is the reset value when GICCDISABLE is tied
LOW.

[23:20] AdvSIMD Returns 0x0 to indicate support for Advanced SIMD.

[19:16] FP Returns 0x0 to indicate support for Floating-point.

[15:12] EL3 Returns 0x2 to indicate EL3 supports AArch64 state or AArch32 state.

[11:8] EL2 Returns 0x2 to indicate EL2 supports AArch64 state or AArch32 state.

[7:4] EL1 Returns 0x2 to indicate EL1 supports AArch64 state or AArch32 state.

[3:0] EL0 Returns 0x2 to indicate EL0 supports AArch64 state or AArch32 state.

To access the ID_AA64PFR0_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_AA64PFR0_EL1; Read AArch64 Processor Feature Register 0

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-112
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

The ID_AA64PFR0[31:0] can be accessed through the memory-mapped interface and the external debug
interface, offset 0xD20.
The ID_AA64PFR0[63:32] can be accessed through the memory-mapped interface and the external
debug interface, offset 0xD24.

4.3.19 AArch64 Debug Feature Register 0, EL1

The ID_AA64DFR0_EL1 characteristics are:

Purpose
Provides top-level information of the debug system in AArch64 state.
Usage constraints
The accessibility to the ID_AA64DFR0_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

The external debug accessibility to the ID_AA64DFR0[63:32] and the ID_AA64DFR0[31:0] by

condition code is:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
The ID_AA64DFR0_EL1 is architecturally mapped as follows:
• [63:32] to external ID_AA64DFR0[63:32] register.
• [31:0] to external ID_AA64DFR0[31:0] register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_AA64DFR0_EL1 bit assignments.
63 32 31 28 27 24 23 20 19 16 15 12 11 8 7 4 3 0

RES0 CTX_CMPs RES0 WRPs RES0 BRPs PMUVer TraceVer DebugVer

Figure 4-18 ID_AA64DFR0_EL1 bit assignments

The following table shows the ID_AA64DFR0_EL1 bit assignments.

Table 4-33 ID_AA64DFR0_EL1 bit assignments

Bits Name Function

[63:32] - Reserved, RES0

[31:28] CTX_CMPs Returns 0x1 to indicate support for two context-aware breakpoints

[27:24] - Reserved, RES0

[23:20] WRPs Returns 0x3 to indicate support for four watchpoints

[19:16] - Reserved, RES0

[15:12] BRPs Returns 0x5 to indicate support for six breakpoints

[11:8] PMUVer Returns 0x1 to indicate that the Performance Monitors (PMUv3) System registers are implemented

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-113
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-33 ID_AA64DFR0_EL1 bit assignments (continued)

Bits Name Function

[7:4] TraceVer Returns 0x0 to indicate that the Trace System registers are not implemented

[3:0] DebugVer Returns 0x6 to indicate that the v8-A Debug architecture is implemented

To access the ID_AA64DFR0_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_AA64DFR0_EL1; Read AArch64 Debug Feature Register 0

The ID_AA64DFR0[31:0] can be accessed through the memory-mapped interface and the external
debug interface, offset 0xD28.
The ID_AA64DFR0[63:32] can be accessed through the memory-mapped interface and the external
debug interface, offset 0xD2C.

4.3.20 AArch64 Instruction Set Attribute Register 0, EL1

The ID_AA64ISAR0_EL1 characteristics are:

Purpose
Provides information about the Cryptography Extension instruction set that the processor can
support.
Note
• The optional Cryptography engine is not included in the base product of the processor. ARM
requires licensees to have contractual rights to obtain the Cortex-A72 processor
Cryptography engine.
• The SHA1, SHA2, and AES fields of ID_AA64ISAR0_EL1 are 0x0 if the Cryptography
engine is not included or CRYPTODISABLE is HIGH.

Usage constraints
The accessibility to the ID_AA64ISAR0_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

The external debug accessibility to the ID_AA64ISAR0[63:32] and the ID_AA64ISAR0[31:0]

by condition code is:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
The ID_AA64ISAR0_EL1 is architecturally mapped as follows:
• [63:32] to external ID_AA64ISAR0[63:32] register.
• [31:0] to external ID_AA64ISAR0[31:0] register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_AA64ISAR0_EL1 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-114
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

63 20 19 16 15 12 11 8 7 4 3 0

RES0 CRC32 SHA2 SHA1 AES RES0

Figure 4-19 ID_AA64ISAR0_EL1 bit assignments

The following table shows the ID_AA64ISAR0_EL1 bit assignments.

Table 4-34 ID_AA64ISAR0_EL1 bit assignments

Bits Name Function

[63:20] - Reserved, RES0.

[19:16] CRC32 Returns 0x1 to indicate that CRC32 instructions are implemented in AArch64 state.

[15:12] SHA2 Indicates whether SHA2 instructions are implemented in AArch64 state. The possible values are:
0x0 No SHA2 instructions implemented.
0x1 SHA256H, SHA256H2, SHA256U0, and SHA256U1 instructions implemented.

[11:8] SHA1 Indicates whether SHA1 instructions are implemented in AArch64 state. The possible values are:
0x0 No SHA1 instructions implemented.
0x1 SHA1C, SHA1P, SHA1M, SHA1SU0, and SHA1SU1 instructions implemented.

[7:4] AES Indicates whether AES instructions are implemented in AArch64 state. The possible values are:
0x0 No AES instructions implemented.
0x2 AESE, AESD, AESMC, AESIMC and PMULL/PMULL2 instructions implemented.

[3:0] - Reserved, RES0.

To access the ID_AA64ISAR0_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_AA64ISAR0_EL1; Read AArch64 Instruction Set Attribute Register 0

The ID_AA64ISAR0[31:0] can be accessed through the memory-mapped interface and the external
debug interface, offset 0xD30.
The ID_AA64ISAR0[63:32] can be accessed through the memory-mapped interface and the external
debug interface, offset 0xD34.

4.3.21 AArch64 Memory Model Feature Register 0, EL1

The ID_AA64MMFR0_EL1 characteristics are:

Purpose
Provides information about the implemented memory model and memory management support
in AArch64 state.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-115
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Usage constraints
The accessibility of the ID_AA64MMFR0_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

The external debug accessibility to the ID_AA64MMFR0[63:32] and the

ID_AA64MMFR0[31:0] by condition code is:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 10-1 External register access conditions on page 10-347 describes the condition codes.
Configurations
The ID_AA64MMFR0_EL1 is architecturally mapped as follows:
• [63:32] to external ID_AA64MMFR0[63:32] register.
• [31:0] to external ID_AA64MMFR0[31:0] register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the ID_AA64MMFR0_EL1 bit assignments.
63 32 31 28 27 24 23 20 19 16 15 12 11 8 7 4 3 0

RES0 4KB 64KB 16KB RES0 SNSMem BigEnd ASIDBits PARange

Figure 4-20 ID_AA64MMFR0_EL1 bit assignments

The following table shows the ID_AA64MMFR0_EL1 bit assignments.

Table 4-35 ID_AA64MMFR0_EL1 bit assignments

Bits Name Function

[63:32] - Reserved, RES0.

[31:28] 4KB Returns 0x0 to indicate that the 4KB granule is supported.

[27:24] 64KB Returns 0x0 to indicate that the 64KB granule is supported.

[23:20] 16KB Returns 0x0 to indicate that the 16KB granule is not supported.

[19:16] - Reserved, RES0.

[15:12] SNSMem Returns 0x1 to indicate that the processor supports a distinction between Secure and Non-secure memory.

[11:8] BigEnd Returns 0x1 to indicate that the processor supports a mixed-endian configuration. The SCTLR_ELx.EE and
SCTLR_EL1.E0E bits can be configured.

[7:4] ASIDBits Returns 0x2 to indicate that the processor supports 16 ASID bits.

[3:0] PARange Returns 0x4 to indicate that the processor supports a 44-bit physical address range, that is, 16TByte.

To access the ID_AA64MMFR0_EL1 in AArch64 state, read the register with:

MRS <Xt>, ID_AA64MMFR0_EL1; Read AArch64 Memory Model Feature Register 0

The ID_AA64MMFR0[31:0] can be accessed through the memory-mapped interface and the external
debug interface, offset 0xD38.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-116
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

The ID_AA64MMFR0[63:32] can be accessed through the memory-mapped interface and the external
debug interface, offset 0xD3C.

4.3.22 Cache Size ID Register, EL1

The CCSIDR_EL1 characteristics are:

Purpose
Provides information about the architecture of the caches. There is one Cache Size ID Register
for each cache that the processor can access. CSSELR_EL1 selects which Cache Size ID
Register is accessible.
Usage constraints
The accessibility to the CCSIDR_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

If CSSELR_EL1 indicates a cache that is not implemented, reading the Cache Size ID Register
returns an UNKNOWN value.
Configurations
The CCSIDR_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 CCSIDR register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the CCSIDR_EL1 bit assignments.
31 30 29 28 27 13 12 3 2 0

NumSets Associativity LineSize

WT WA
WB RA

Figure 4-21 CCSIDR_EL1 bit assignments

The following table shows the CCSIDR_EL1 bit assignments.

Table 4-36 CCSIDR_EL1 bit assignments

Bits Name Function

[31] WT Returns 0b0 to indicate that the cache level does not support Write-Through.

[30] WB Indicates support for Write-Back. The possible values are:

0 Cache level does not support Write-Back.
1 Cache level supports Write-Back.

[29] RA Returns 0b1 to indicate that the cache level supports Read-Allocation.

[28] WA Indicates support for Write-Allocation. The possible values are:

0 Cache level does not support Write-Allocation.
1 Cache level supports Write-Allocation.

[27:13] NumSets Indicates the (number of sets in cache) – 1. Therefore, a value of 0 indicates 1 set in the cache. The number of
sets does not have to be a power of 2.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-117
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-36 CCSIDR_EL1 bit assignments (continued)

Bits Name Function

[12:3] Associativity Indicates the associativity of the selected cache level. The possible values are:
0b0000000001 2-way.
0b0000000010 3-way.
0b0000001111 16-way.

[2:0] LineSize Returns 0b010 to indicate that the cache line size is 64 bytes.

The following table shows the individual bit field and complete register encoding for the CCSIDR_EL1.
The CSSELR_EL1 determines which Cache Size ID Register to select.

Table 4-37 Encoding of the Cache Size ID Register

CSSELR_EL1 Size Complete register encoding Register bit field encoding

WT WB RA WA NumSets Associativity LineSize

0x0 32KB 0x701FE00A 0 1 1 1 0x0FF 0x1 0x2

0x1 48KB 0x201FE012 0 0 1 0 0x0FF 0x2 0x2

0x2 512KB 0x703FE07A 0 1 1 1 0x1FF 0xF 0x2

1024KB 0x707FE07A 0 1 1 1 0x3FF 0xF 0x2

2048KB 0x70FFE07A 0 1 1 1 0x7FF 0xF 0x2

4096KB 0x71FFE07A 0 1 1 1 0xFFF 0xF 0x2

0x3-0xF - - Reserved

To access the CCSIDR_EL1 in AArch64 state, read the register with:

MRS <Xt>, CCSIDR_EL1; Read Cache Size ID Register

To access the CCSIDR in AArch32 state, read the CP15 register with:
MRC p15, 1, <Rt>, c0, c0, 0; Read Cache Size ID Register

4.3.23 Cache Level ID Register, EL1

The CLIDR_EL1 characteristics are:

Purpose
Identifies:
• The type of cache, or caches, implemented at each level, up to a maximum of seven levels
• The Level of Coherency and Level of Unification for the cache hierarchy.
Usage constraints
The accessibility to the CLIDR_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-118
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Configurations
The CLIDR_EL1 is:
• Common to Secure and Non-secure states.
• A 64-bit register in the AArch64 state.
• Architecturally mapped to the AArch32 CLIDR register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the CLIDR_EL1 bit assignments.
63 30 29 27 26 24 23 21 20 18 17 15 14 12 11 9 8 6 5 3 2 0

RES0 LoUU LoC LoUIS Ctype7 Ctype6 Ctype5 Ctype4 Ctype3 Ctype2 Ctype1

Figure 4-22 CLIDR_EL1 bit assignments

The following table shows the CLIDR_EL1 bit assignments.

Table 4-38 CLIDR_EL1 bit assignments

Bits Name Function

[63:30] - Reserved, RES0.

[29:27] LoUU Indicates the Level of Unification Uniprocessor for the cache hierarchy. This value is:
0b001 L1 cache is the last level of cache that must be cleaned or invalidated when cleaning or
invalidating to the point of unification for the processor.

[26:24] LoC Indicates the Level of Coherency for the cache hierarchy. This value is:
0b010 L3 cache.

[23:21] LoUIS Indicates the Level of Unification Inner Shareable for the cache hierarchy. This value is:
0b001 L2 cache.

[20:18] Ctype7 Indicates the type of cache implemented at level 7. This value is:
0b000 No cache.

[17:15] Ctype6 Indicates the type of cache implemented at level 6. This value is:
0b000 No cache.

[14:12] Ctype5 Indicates the type of cache implemented at level 5. This value is:
0b000 No cache.

[11:9] Ctype4 Indicates the type of cache implemented at level 4. This value is:
0b000 No cache.

[8:6] Ctype3 Indicates the type of cache implemented at level 3. This value is:
0b000 No cache.

[5:3] Ctype2 Indicates the type of cache implemented at level 2. This value is:
0b100 Unified cache.

[2:0] Ctype1 Indicates the type of cache implemented at level 1. This value is:
0b011 Separate instruction and data caches.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-119
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

To access the CLIDR_EL1 in AArch64 state, read the register with:

MRS <Xt>, CLIDR_EL1; Read Cache Level ID Register

To access the CLIDR in AArch32 state, read the CP15 register with:
MRC p15, 1, <Rt>, c0, c0, 1; Read Cache Level ID Register

4.3.24 Auxiliary ID Register, EL1

The processor does not implement AIDR_EL1. This register is always RES0.

4.3.25 Cache Size Selection Register, EL1

The CSSELR_EL1 characteristics are:

Purpose
Selects the current CCSIDR_EL1, by specifying:
• The required cache level.
• The cache type, either instruction or data cache.
Usage constraints
The accessibility to the CSSELR_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

If the CSSELR_EL1 level field is programmed to a cache level that is not implemented, then a
read of CSSELR_EL1 returns an UNKNOWN value in CSSELR_EL1.Level.
Configurations
The CSSELR_EL1 is:
• Banked for the Secure and Non-secure states.
• Architecturally mapped to the Non-secure AArch32 CSSELR register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the CSSELR_EL1 bit assignments.
31 4 3 1 0

RES0 Level

InD

Figure 4-23 CSSELR_EL1 bit assignments

The following table shows the CSSELR_EL1 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-120
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-39 CSSELR_EL1 bit assignments

Bits Name Function

[31:4] - Reserved, RES0.

[3:1] Level Cache level of required cache:

0b000 Level 1.
0b001 Level 2.

All other values are reserved.

[0] InD Instruction not Data bit:

0 Data or unified cache.
1 Instruction cache.

To access the CSSELR_EL1 in AArch64 state, read or write the register with:
MRS <Xt>, CSSELR_EL1; Read Cache Size Selection Register
MSR CSSELR_EL1, <Xt>; Write Cache Size Selection Register

To access the CSSELR in AArch32 state, read or write the CP15 register with:
MRC p15, 2, <Rt>, c0, c0, 0; Read Cache Size Selection Register
MCR p15, 2, <Rt>, c0, c0, 0; Write Cache Size Selection Register

Related information
4.3.22 Cache Size ID Register, EL1 on page 4-117.

4.3.26 Cache Type Register, EL0

The CTR_EL0 characteristics are:

Purpose
Provides information about the architecture of the caches.
Usage constraints
The accessibility to the CTR_EL0 in AArch64 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

Config RO RO RO RO RO

The accessibility to the CTR in AArch32 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The CTR_EL0 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch32 CTR register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the CTR_EL0 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-121
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

31 30 28 27 24 23 20 19 16 15 14 13 4 3 0

RES0 CWG ERG DminLine L1Ip RES0 IminLine

RES1

Figure 4-24 CTR_EL0 bit assignments

The following table shows the CTR_EL0 bit assignments.

Table 4-40 CTR_EL0 bit assignments

Bits Name Function

[31] - Reserved, RES1.

[30:28] - Reserved, RES0.

[27:24] CWG Cache Writeback Granule. Log2 of the number of words of the maximum size of memory that can be overwritten
as a result of the eviction of a cache entry that has had a memory location in it modified. This value is:
0x4 Cache writeback granule size is 16 words.

[23:20] ERG Exclusives Reservation Granule. Log2 of the number of words of the maximum size of the reservation granule
that has been implemented for the Load-Exclusive and Store-Exclusive instructions. This value is:
0x4 Exclusive reservation granule size is 16 words.

[19:16] DminLine Log2 of the number of words in the smallest cache line of all the data and unified caches that the processor
controls. This value is:
0x4 Smallest data cache line size is 16 words.

[15:14] L1lp Level 1 Instruction Cache policy. Indicates the indexing and tagging policy for the L1 Instruction Cache. This
value is:
0b11 Physical index, physical tag (PIPT).

[13:4] - Reserved, RES0.

[3:0] IminLine Log2 of the number of words in the smallest cache line of all the Instruction Caches that the processor controls.
This values is:
0x4 Smallest Instruction Cache line size is 16 words.

To access the CTR_EL0 in AArch64 state, read the register with:

MRS <Xt>, CTR_EL0; Read Cache Type Register

To access the CTR in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c0, 1; Read Cache Type Register

4.3.27 Data Cache Zero ID, EL0

The DCZID_EL0 characteristics are:

Purpose
Indicates the block size written with byte values of 0 by the DC ZVA, Data Cache Zero by
Address, system instruction.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-122
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Usage constraints
The accessibility of the DCZID_EL0 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

RO RO RO RO RO RO

Configurations
The DCZID_EL0 is a 32-bit register.
Attributes
See the register summary in Table 4-1 AArch64 identification registers on page 4-75.
The following figure shows the DCZID_EL0 bit assignments.

63 5 4 3 0

RES0 BlockSize

DZP

Figure 4-25 DCZID_EL0 bit assignments

The following table shows the DCZID_EL0 bit assignments.

Table 4-41 DCZID_EL0 bit assignments

Bits Name Function

[63:5] - Reserved, RES0.

[4] DZP Prohibit the DC ZVA instruction. The possible values are:
0 DC ZVA instruction permitted. This is the reset value.
1 DC ZVA instruction prohibited.

[3:0] BS Returns 0x4 to indicate that the block size is 16 words.

To access the DCZID_EL0 in AArch64 state, read or write the register with:
MRS <Xt>, DCZID_EL0; Read Data Cache Zero ID Register

4.3.28 Virtualization Processor ID Register, EL2

The VPIDR_EL2 characteristics are:

Purpose
Holds the value of the Virtualization Processor ID. A Non-secure read of the MIDR from EL1
returns the value of this register.
Usage constraints
The accessibility to the VPIDR_EL2 in AArch64 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW RW

The accessibility to the VPIDR in AArch32 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW -

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-123
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Configurations
The VPIDR_EL2 is:
• A Banked EL2 register.
• Architecturally mapped to the AArch32 VPIDR register.
Attributes
See the register summary in Table 4-13 AArch64 virtualization registers on page 4-83.
The following figure shows the VPIDR_EL2 bit assignments.
31 0

VPIDR

Figure 4-26 VPIDR_EL2 bit assignments

The following table shows the VPIDR_EL2 bit assignments.

Table 4-42 VPIDR_EL2 bit assignments

Bits Name Function

[31:0] VPIDR MIDR value returned by Non-secure EL1 reads of the MIDR. For information on the subdivision of this value, see
4.3.1 Main ID Register, EL1 on page 4-89.

To access the VPIDR_EL2 in AArch64 state, read or write the register with:
MRS <Xt>, VPIDR_EL1; Read Virtualization Processor ID Register
MSR VPIDR_EL1, <Xt>; Write Virtualization Processor ID Register

To access the VPIDR, read or write the CP15 register with:

MRC p15, 4, <Rt>, c0, c0, 0; Read Virtualization Processor ID Register
MCR p15, 4, <Rt>, c0, c0, 0; Write Virtualization Processor ID Register

Related information
4.3.1 Main ID Register, EL1 on page 4-89.

4.3.29 Virtualization Multiprocessor ID Register, EL2

The VMPIDR_EL2 characteristics are:

Purpose
Holds the value of the Virtualization Multiprocessor ID. This is the value returned by Non-
secure EL1 reads of MPIDR_EL1.
Usage constraints
The accessibility of the VMPIDR_EL2 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW RW

Configurations
The VMPIDR_EL2 is Architecturally mapped to the Non-secure AArch32 VMPIDR register.
Attributes
See the register summary in Table 4-13 AArch64 virtualization registers on page 4-83.
The following figure shows the VMPIDR_EL2 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-124
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

63 32 31 0

Reserved VMPIDR_EL2

Figure 4-27 VMPIDR_EL2 bit assignments

The following table shows the VMPIDR_EL2 bit assignments.

Table 4-43 VMPIDR_EL2 bit assignments

Bits Name Function

[63:32] - Reserved, RES0.

[31:0] VMPIDR_EL2 MPIDR value returned by Non-secure EL1 reads of the MPIDR_EL1. For information on the subdivision of
this value, see 4.5.3 Multiprocessor Affinity Register on page 4-240.

To access the VMPIDR_EL2 in AArch64 state, read or write the register with:
MRS <Xt>, VMPIDR_EL1; Read Virtualization Multiprocessor ID Register
MSR VMPIDR_EL1, <Xt>; Write Virtualization Multiprocessor ID Register

Related information
4.5.4 Virtualization Multiprocessor ID Register on page 4-241.

4.3.30 System Control Register, EL1

The SCTLR_EL1 characteristics are:
Purpose
Provides top-level control of the system, including its memory system at EL1 in AArch64 state.
Usage constraints
The accessibility of the SCTLR_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Configurations
The SCTLR_EL1 is:
• A 32-bit register in AArch64 state.
• Architecturally mapped to the Non-secure AArch32 SCTLR register.
Attributes
See the register summary in Table 4-3 AArch64 virtual memory control registers on page 4-78.
The following figure shows the SCTLR_EL1 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-125
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

RES0 RES1 RES1 I C A M

RES0 RES0 SA
UCI RES1 SA0
EE WXN CP15BEN
E0E nTWE THEE
RES0 ITD
nTWI SED
UCT UMA
DZE RES0
RES0 RES1

Figure 4-28 SCTLR_EL1 bit assignments

The following table shows the SCTLR_EL1bit assignments.

Table 4-44 SCTLR_EL1 bit assignments

Bits Name Function

[31:30] - Reserved, RES0.

[29:28] - Reserved, RES1.

[27] - Reserved, RES0.

[26] UCI Enables EL0 access to the DC CVAU, DC CIVAC, DC CVAC and IC IVAU instructions in AArch64 state. The
values are:
0 EL0 access disabled. This is the reset value.
1 EL0 access enabled.

[25] EE Exception endianness. Indicates the endianness of the translation table data for the translation table lookups. The
EE bit is permitted to be cached in a TLB. The values are:
0 Little-endian.
1 Big-endian.

[24] E0E Endianness of explicit data access at EL0. The values are:
0 Explicit data accesses at EL0 are little-endian. This is reset value.
1 Explicit data accesses at EL0 are big-endian.

[23:22] - Reserved, RES1.

[21] - Reserved, RES0.

[20] - Reserved, RES1.

[19] WXN Write permission implies Execute Never (XN). You can use this bit to require all memory regions with write
permissions are treated as XN. The WXN bit is permitted to be cached in a TLB. The values are:
0 Regions with write permission are not forced to be XN. This is the reset value.
1 Regions with write permissions are forced to be XN.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-126
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-44 SCTLR_EL1 bit assignments (continued)

Bits Name Function

[18] nTWE WFE non-trapping. The values are:

0 A WFE instruction executed at EL0 that causes suspended execution as if the event register is
not set and there is no pending WFE wake-up event, is treated as an exception with error code
of 0x1.
1 WFE instructions executed as normal. This is the reset value.

Conditional WFE instructions that fail their condition do not cause an exception if this bit is 0.

[17] - Reserved, RES0.

[16] nTWI WFI non-trapping. The values are:

0 A WFI instruction executed at EL0 that causes suspended execution as if there is no pending
WFI wake-up event, is treated as an exception with error code of 0x1.
1 WFI instructions executed as normal. This is the reset value.

Conditional WFI instructions that fail their condition do not cause an exception if this bit is 0.

[15] UCT Enables EL0 access to the CTR_EL0 register in AArch64 state. The values are:
0 Disables EL0 access to the CTR_EL0 register. This is the reset value.
1 Enables EL0 access to the CTR_EL0 register.

[14] DZE Enables access to the DC ZVA instruction at EL0. The values are:
0 Disables execution access to the DC ZVA instruction at EL0. Access is treated as UNDEFINED.
This is the reset value.
1 Enables execution access to the DC ZVA instruction at EL0.

[13] - Reserved, RES0.

[12] I Instruction cache enable. The values are:

0 Instruction caches disabled. This is the reset value.
1 Instruction caches enabled.

[11] - Reserved, RES1.

[10] - Reserved, RES0.

[9] UMA User Mask Access. Controls access to interrupt masks from EL0, when EL0 is using AArch64. The values are:
0 Disables access to the interrupt masks from EL0.
1 Enables access to the interrupt masks from EL0.

[8] SED SETEND instruction disable. The values are:

0 The SETEND instruction is enabled. This is the reset value.
1 The SETEND instruction is UNALLOCATED.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-127
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-44 SCTLR_EL1 bit assignments (continued)

Bits Name Function

[7] ITD IT instruction disable. The values are:

0 The IT instruction functionality is enabled. This is the reset value.
1 All encodings of the IT instruction are UNDEFINED when either:
• hw[3:0] are not equal to 0b1000.
• IT instructions with a subsequent 32-bit instruction.
• Subsequent PC reading or writing instruction.

[6] THEE ThumbEE enable:

0 ThumbEE is not implemented.

[5] CP15BEN AArch32 CP15 barrier enable. The values are:

0 CP15 barrier operations disabled. Their encodings are UNDEFINED.
1 CP15 barrier operations enabled. This is the reset value.

[4] SA0 Enable EL0 Stack Alignment check. When set, use of the Stack Pointer as the base address in a load/store
instruction at EL0 must align to a 16-byte boundary, or a Stack Alignment Fault exception is raised. The values
are:
0 Disable EL0 Stack Alignment check.
1 Enable EL0 Stack Alignment check. This is the reset value.

[3] SA Enable Stack Alignment check. When set, use of the Stack Pointer as the base address in a load/store instruction
at the Exception level of this register must align to a 16-byte boundary, or a Stack Alignment Fault exception is
raised. The values are:
0 Disable Stack Alignment check.
1 Enable Stack Alignment check. This is the reset value.

[2] C Cache enable. The values are:

0 Data and unified caches disabled. This is the reset value.
1 Data and unified caches enabled.

[1] A Alignment check enable. The values are:

0 Alignment fault checking disabled. This is the reset value.
1 Alignment fault checking enabled.

[0] M MMU enable. The values are:

0 EL1 and EL0 stage 1 MMU disabled. This is the reset value.
1 EL1 and EL0 stage 1 MMU enabled.

To access SCTLR_EL1 in AArch64 state, read or write the register with:

MRS <Xt>, SCTLR_EL1; Read EL1 System Control Register
MSR SCTLR_EL1, <Xt>; Write EL1 System Control Register

Related information
4.5.5 System Control Register on page 4-242.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-128
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

4.3.31 Auxiliary Control Register, EL1

The processor does not implement the ACTLR_EL1 register. This register is always RES0.

4.3.32 Architectural Feature Access Control Register, EL1

The CPACR_EL1 characteristics are:

Purpose
Controls access to trace functionality and access to registers associated with Floating-point and
Advanced SIMD execution.
Usage constraints
The accessibility of the CPACR_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Configurations
The CPACR_EL1 is:
• A32-bit register in AArch64 state.
• Architecturally mapped to the Non-secure AArch32 CPACR register.
Attributes
See the register summary in Table 4-4 AArch64 other System registers on page 4-79.
The following figure shows the CPACR_EL1 bit assignments.
31 29 28 27 22 21 20 19 0

RES0 RES0 FPEN RES0

TTA

Figure 4-29 CPACR_EL1 bit assignments

The following table shows the CPACR_EL1 bit assignments.

Table 4-45 CPACR_EL1 bit assignments

Bits Name Function

[31:29] - Reserved, RES0.

[28] TTA Traps trace functionality to EL1 when executing from EL0 or EL1. The value is:
0 System register access to trace functionality is not supported. This bit is RES0.

[27:22] - Reserved, RES0.

[21:20] FPEN Traps instructions that access registers associated with floating-point and SIMD execution to trap to EL1 when
executed from EL0 or EL1. The possible values are:
0b00 Trap any instruction in EL0 or EL1 that use registers associated with floating-point and Advanced SIMD
0b10 execution. The reset value is 0b00.
0b01 Trap any instruction in EL0 that use registers associated with floating-point and Advanced SIMD execution.
Instructions in EL1 are not trapped.
0b11 No instructions are trapped.

[19:0] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-129
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

To access the CPACR_EL1 in AArch64 state, read or write the register with:
MRS <Xt>, CPACR_EL1; Read EL1 Architectural Feature Access Control Register
MSR CPACR_EL1, <Xt>; Write EL1 Architectural Feature Access Control Register

Related information
4.5.6 Architectural Feature Access Control Register on page 4-247.

4.3.33 Auxiliary Control Register, EL2

The ACTLR_EL2 characteristics are:

Purpose
Controls access to IMPLEMENTATION DEFINED registers in Non-secure EL1.
Usage constraints
The accessibility to ACTLR_EL2 in AArch64 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW RW

The accessibility to the HACTLR in AArch32 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW -

Configurations
The ACTLR_EL2 is:
• A Banked EL2 register.
• Architecturally mapped to the AArch32 HACTLR register.
Attributes
See the register summary in Table 4-4 AArch64 other System registers on page 4-79.
The following figure shows the ACTLR_EL2 bit assignments.
31 7 6 5 4 3 2 1 0

Reserved

L2ACTLR access control

L2ECTLR access control
L2CTLR access control
Reserved
CPUECTLR access control
CPUACTLR access control

Figure 4-30 ACTLR_EL2 bit assignments

The following table shows the ACTLR_EL2 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-130
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-46 ACTLR_EL2 bit assignments

Bits Name Function

[31:7] - Reserved, RES0.

[6] L2ACTLR access control L2ACTLR access control. The possible values are:
0 The register is not accessible from Non-secure EL1.
1 The register is accessible from Non-secure EL1.

[5] L2ECTLR access control L2ECTLR access control. The possible values are:
0 The register is not accessible from Non-secure EL1.
1 The register is accessible from Non-secure EL1.

[4] L2CTLR access control L2CTLR access control. The possible values are:
0 The register is not accessible from Non-secure EL1.
1 The register is accessible from Non-secure EL1.

[3:2] - Reserved, RES0.

[1] CPUECTLR access control CPUECTLR access control. The possible values are:
0 The register is not accessible from Non-secure EL1.
1 The register is accessible from Non-secure EL1.

[0] CPUACTLR access control CPUACTLR access control. The possible values are:
0 The register is not accessible from Non-secure EL1.
1 The register is accessible from Non-secure EL1.

To access the ACTLR_EL2 in AArch64 state, read or write the register with:
MRS <Xt>, ACTLR_EL2; Read EL2 Auxiliary Control Register
MSR ACTLR_EL2, <Xt>; Write EL2 Auxiliary Control Register

To access the HACTLR in AArch32 state, read or write the CP15 register with:
MRC p15, 4, <Rt>, c1, c0, 1; Read Hypervisor Auxiliary Control Register
MCR p15, 4, <Rt>, c1, c0, 1; Write Hypervisor Auxiliary Control Register

4.3.34 Hypervisor Configuration Register, EL2

The HCR_EL2 characteristics are:

Purpose
Provides configuration control for virtualization, including whether various Non-secure
operations are trapped to EL2.
Usage constraints
The accessibility of the HCR_EL2 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW RW

Configurations
The HCR_EL2 is architecturally mapped as follows:
• [63:32] to the AArch32 HCR2 register.
• [31:0] to the AArch32 HCR register.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-131
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Attributes
See the register summary in Table 4-13 AArch64 virtualization registers on page 4-83.
The following figure shows the HCR_EL2 bit assignments.
63 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

Reserved

ID VM
CD SWIO
RW PTW
TRVM FMO
HCD IMO
TDZ AMO
TGE VF
TVM VI
TTLB VSE
TPU FB
TPC BSU
TSW DC
TACR TWI
TIDCP TWE
TSC TID0
TID3 TID1
TID2

Figure 4-31 HCR_EL2 bit assignments

The following table shows the HCR_EL2 bit assignments.

Table 4-47 HCR_EL2 bit assignments

Bits Name Function

[63:34] - Reserved, RES0.

[33] ID Disables stage 2 Instruction Cache. When HCR_EL2.VM is 1, this forces all stage 2 translations for instruction
accesses to Normal memory to be Non-cacheable for the EL1/EL0 translation regimes. The values are:
0 Has no effect on stage 2 EL1/EL0 translation regime for instruction accesses. This is the reset
value.
1 Forces all stage 2 translations for instruction accesses to Normal memory to be Non-cacheable
for the EL1/EL0 translation regime.

[32] CD Disables stage 2 data cache. When HCR_EL2.VM is 1, this forces all stage 2 translations for data accesses and
translation table walks to Normal memory to be Non-cacheable for the EL1/EL0 translation regimes. The values are:
0 Has no effect on stage 2 EL1/EL0 translation regime for data access or translation table walks.
This is the reset value.
1 Forces all stage 2 translations for data accesses and translation table walks to Normal memory to
be Non-cacheable for the EL1/EL0 translation regime.

[31] RW Register width control for lower Exception levels. The values are:
0 Lower levels are all AArch32. This is the reset value.
1 EL1 is AArch64. EL0 is determined by the register width described in the current processing
state when executing at EL0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-132
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-47 HCR_EL2 bit assignments (continued)

Bits Name Function

[30] TRVM Trap Read of Virtual Memory controls. When 1, this causes reads to the EL1 virtual memory control registers from
EL1 to be trapped to EL2. This covers the following registers:
AArch32 SCTLR, TTBR0, TTBR1, TTBCR, DACR, DFSR, IFSR, DFAR, IFAR, ADFSR, AIFSR,
PRRR/MAIR0, NMRR/MAIR1, AMAIR0, AMAIR1, and CONTEXTIDR.
AArch64 SCTLR_EL1, TTBR0_EL1, TTBR1_EL1, TCR_EL1, ESR_EL1, FAR_EL1, AFSR0_EL1,
AFSR1_EL1, MAIR_EL1, AMAIR_EL1, and CONTEXTIDR_EL1.

The reset value is 0.

[29] HCD Disables Hyp call. The processor implements EL3. This bit is RES0.

[28] TDZ Traps DC ZVA instruction. The values are:

0 DC ZVA instruction is not trapped.
1 DC ZVA instruction is trapped to EL2 when executed in Non-secure EL1 or EL0.

[27] TGE Traps general exceptions. If this bit is set, and SCR_EL3.NS is set, then:
• All EL1 exceptions are routed to EL2.
• For EL1, the SCTLR_EL1.M bit is treated as 0 regardless of its actual state other than the purpose of reading the
bit.
• The HCR_EL2.FMO, HCR_EL2.IMO, and HCR_AMO bits are treated as 1 regardless of their actual state other
than for the purpose of reading the bits.
• All virtual interrupts are disabled.
• Any IMPLEMENTATION DEFINED mechanisms for signaling virtual interrupts are disabled.
• An exception return to EL1 is treated as an illegal exception return.

[26] TVM Trap Virtual Memory controls. When 1, this causes writes to the EL1 virtual memory control registers from EL1 to
be trapped to EL2. This covers the following registers:
AArch32 SCTLR, TTBR0, TTBR1, TTBCR, DACR, DFSR, IFSR, DFAR, IFAR, ADFSR, AIFSR,
PRRR/MAIR0, NMRR/MAIR1, AMAIR0, AMAIR1, and CONTEXTIDR.
AArch64 SCTLR_EL1, TTBR0_EL1, TTBR1_EL1, TCR_EL1, ESR_EL1, FAR_EL1, AFSR0_EL1,
AFSR1_EL1, MAIR_EL1, AMAIR_EL1, and CONTEXTIDR_EL1.

The reset value is 0.

[25] TTLB Trap TLB maintenance instructions. When 1, this causes TLB maintenance instructions executed from EL1 that are
not UNDEFINED to be trapped to EL2. This covers the following instructions:
AArch32 TLBIALLIS, TLBIMVAIS, TLBIASIDIS, TLBIMVAAIS, ITLBIALL, DTLBIALL, TLBIALL,
ITLBIMVA, DTLBIMVA, TLBIMVA, ITLBIASID, DTLBIASID, TLBIASID, TLBIMVAA,
TLBIMVALIS, TLBIMVAALIS, TLBIMVAL, and TLBIMVAAL.
AArch64 TLBI VAMLLE1, TLBI VAE1, TLBI ASIDE1, TLBI VAAE1, TLBI VALE1, TLBI VAALE1,
TLBI VMALLE1IS, TLBI VAE1IS, TLBI ASIDE1IS, TLBI VAAE1IS, TLBI VALE1IS, and
TLBI VAALE1IS.

The reset value is 0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-133
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-47 HCR_EL2 bit assignments (continued)

Bits Name Function

[24] TPU Trap Cache maintenance instructions to Point of Unification. When 1, this causes Cache maintenance instructions to
the point of unification executed from EL1 or EL0 that are not UNDEFINED to be trapped to EL2. This covers the
following instructions:
AArch32 ICIMVAU, ICIALLU, ICIALLUIS, and DCCMVAU.
AArch64 IC IVAU, IC IALLU, IC IALLUIS, and DC CVAU.

The reset value is 0.

[23] TPC Trap Data/Unified Cache maintenance operations to point of coherency. When 1, this causes Data or Unified Cache
maintenance instructions by address to the point of coherency executed from EL1 or EL0 that are not UNDEFINED to
be trapped to EL2. This covers the following instructions:
AArch32 DCIMVAC, DCCIMVAC, and DCCMVAC.
AArch64 DC IVAC, DC CIVAC, and DC CVCA.

The reset value is 0.

[22] TSW Trap Data/Unified Cache maintenance operations by Set/Way. When 1, this causes Data or Unified Cache
maintenance instructions by set/way executed from EL1 that are not UNDEFINED to be trapped to EL2. This covers
the following instructions:
AArch32 DCISW, DCCSW, and DCCISW.
AArch64 DC ISW, DC CSW, and DC CISW.

The reset value is 0.

[21] TACR Traps Auxiliary Control registers. The values are:

0 Accesses to the Auxiliary Control registers are not trapped.
1 Accesses to the ACTLR in AArch32 state or the ACTLR_EL1 in AArch64 state from EL1 are
trapped to EL2.

[20] TIDCP Trap Implementation Dependent functionality. When 1, this causes accesses to the following instruction set space
executed from EL1 to be trapped to EL2:
AArch32 All CP15 MCR and MRC instructions as follows:
• CRn is 9, op1 is 0 to 7, CRm is c0, c1, c2, c5, c6, c7, or c8, and op2 is 0 to 7.
• CRn is 10, op1 is 0 to 7, CRm is c0, c1, c4, or c8, and op2 is 0 to 7.
• CRn is 11, op1 is 0 to 7, CRm is c0 to c8, or c15, and op2 is 0 to 7.
AArch64 Reserved control space for IMPLEMENTATION DEFINED functionality.

Accesses from EL0 are UNDEFINED. The reset value is 0.

[19] TSC Traps SMC instruction. The values are:

0 SMC instruction is not trapped.
1 SMC instruction executed in EL1 is trapped to EL2 for AArch32 and AArch64 states.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-134
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-47 HCR_EL2 bit assignments (continued)

Bits Name Function

[18] TID3 Trap ID Group 3. When 1, this causes reads to the following registers executed from EL1 to be trapped to EL2:
AArch32 ID_PFR0, ID_PFR1, ID_DFR0, ID_AFR0, ID_MMFR0, ID_MMFR1, ID_MMFR2,
ID_MMFR3, ID_ISAR0, ID_ISAR1, ID_ISAR2, ID_ISAR3, ID_ISAR4, ID_ISAR5, MVFR0,
MVFR1, and MVFR2 and MRC instructions to the following locations:
• op1 is 0, CRn is 0, CRm is c3, c4, c5, c6, or c7, and op2 is 0 or 1.
• op1 is 0, CRn is 0, CRm is c3, and op2 is 2.
• op1 is 0, CRn is 0, CRm is 5, and op2 is 4 or 5.
AArch64 ID_PFR0_EL1, ID_PFR1_EL1, ID_DFR0_EL1, ID_AFR0_EL1, ID_MMFR0_EL1,
ID_MMFR1_EL1, ID_MMFR2_EL1, ID_MMFR3_EL1, ID_ISAR0_EL1, ID_ISAR1_EL1,
ID_ISAR2_EL1, ID_ISAR3_EL1, ID_ISAR4_EL1, ID_ISAR5_EL1, MVFR0_EL1,
MVFR1_EL1, MVFR2_EL1, ID_AA64PFRn_EL1, ID_AA64DFRn_EL1,
ID_AA64ISARn_EL1, ID_AA64MMFRn_EL1, and ID_AA64AFRn_EL1.

The reset value is 0.

[17] TID2 Trap ID Group 2. When 1, this causes reads or writes to CSSELR/CSSELR_EL1, to the following registers executed
from EL1 or EL0 that are UNDEFINED to be trapped to EL2:
AArch32 CTR, CCSIDR, CLIDR, and CSSELR.
AArch64 CTR_EL0, CCSIDR_EL1, CLIDR_EL1, and CSSELR_EL1.

The reset value is 0.

[16] TID1 Trap ID Group 1. When 1, this causes reads to the following registers executed from EL1 to be trapped to EL2:
AArch32 TCMTR, TLBTR, AIDR, and REVIDR.
AArch64 AIDR_EL1, and REVIDR_EL1.

The reset value is 0.

[15] TID0 Trap ID Group 0. When 1, this causes reads to the following registers executed from EL1 or EL0 that are UNDEFINED
to be trapped to EL2:
AArch32 FPSID and JIDR.
AArch64 None.

The reset value is 0.

[14] TWE Traps WFE instruction if it would cause suspension of execution. For example, if there is no pending WFE event:
0 WFE instruction is not trapped.
1 WFE instruction executed in EL1 or EL0 is trapped to EL2 for AArch32 and AArch64 states.

[13] TWI Traps WFI instruction if it would cause suspension of execution. For example, if there is no pending WFI event:
0 WFI instruction is not trapped.
1 WFI instruction executed in EL1 or EL0 is trapped to EL2 for AArch32 and AArch64 states.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-135
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-47 HCR_EL2 bit assignments (continued)

Bits Name Function

[12] DC Default Cacheable. When this bit is set to 1 the memory type and attributes determined by stage 1 translation is
Normal, Non-shareable, Inner Write-Back Write-Allocate, Outer Write-Back Write-Allocate.
When executing in Non-secure EL0 or EL1 and the HCR_EL2.DC bit is set, the behavior of processor is consistent
with the behavior when:
• The SCTLR_EL1.M bit is clear, regardless of the actual value of the SCTLR.M bit.
— An explicit read of the SCTLR_EL1.M bit returns its actual value.
• The HCR_EL2.VM bit is set, regardless of the actual value of the HCR_EL2.VM bit.
— An explicit read of the HCR_EL2.VM bit returns its actual value.
The reset value is 0.

[11:10] BSU Barrier shareability upgrade. Determines the minimum shareability domain that is supplied to any barrier executed
from EL1 or EL0. The values are:
0b00 No effect.
0b01 Inner Shareable.
0b10 Outer Shareable.
0b11 Full system.

This value is combined with the specified level of the barrier held in its instruction, according to the algorithm for
combining shareability attributes.

[9] FB Force broadcast. When 1, this causes the following instructions to be broadcast within the Inner Shareable domain
when executed from Non-secure EL1:
AArch32 ITLBIALL, DTLBIALL, TLBIALL, ITLBIMVA, DTLBIMVA, TLBIMVA, ITLBIASID,
DTLBIASID, TLBIASID, TLBIMVAA, BPIALL, and ICIALLU.
AArch64 TLBI VMALLE1, TLBI VAE1, TLBI ASIDE1, TLBI VAAE1, TLBI VALE1, TLBI VAALE1,
and IC IALLU.

The reset value is 0.

[8] VSE Virtual System Error/Asynchronous Abort. The values are:

0 Virtual System Error/Asynchronous Abort is not pending by this mechanism.
1 Virtual System Error/Asynchronous Abort is pending by this mechanism.

The virtual System Error/Asynchronous Abort is only enabled when the HCR_EL2.AMO bit is set.

[7] VI Virtual IRQ interrupt. The values are:

0 Virtual IRQ is not pending by this mechanism.
1 Virtual IRQ is pending by this mechanism.

The virtual IRQ is only enabled when the HCR_EL2.IMO bit is set.

[6] VF Virtual FIQ interrupt. The values are:

0 Virtual FIQ is not pending by this mechanism.
1 Virtual FIQ is pending by this mechanism.

The virtual FIQ is only enabled when the HCR_EL2.FMO bit is set.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-136
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-47 HCR_EL2 bit assignments (continued)

Bits Name Function

[5] AMO Asynchronous abort and error interrupt routing. The values are:
0 Asynchronous external Aborts and SError Interrupts while executing at Exception levels lower
than EL2 are not taken at EL2. Virtual System Error/Asynchronous Abort is disabled.
1 Asynchronous external Aborts and SError Interrupts while executing at EL2 or lower are taken
in EL2 unless routed by SCTLR_EL3.EA bit to EL3. Virtual System Error/Asynchronous Abort
is enabled.

[4] IMO Physical IRQ routing. The values are:

0 Physical IRQ while executing at Exception levels lower than EL2 are not taken at EL2. Virtual
IRQ interrupt is disabled.
1 Physical IRQ while executing at EL2 or lower are taken in EL2 unless routed by
SCTLR_EL3.IRQ bit to EL3. Virtual IRQ interrupt is enabled.

[3] FMO Physical FIQ routing. The values are:

0 Physical FIQ while executing at Exception levels lower than EL2 are not taken at EL2. Virtual
FIQ interrupt is disabled.
1 Physical FIQ while executing at EL2 or lower are taken in EL2 unless routed by
SCTLR_EL3.FIQ bit to EL3. Virtual FIQ interrupt is enabled.

[2] PTW Protected Table Walk. When this bit is set, if stage 2 translation of a translation table access, made as part of a stage
1 translation table walk at EL0 or EL1, maps to Strongly-ordered or Device memory, the access is faulted as a stage
2 Permission fault.

[1] SWIO Set/Way Invalidation Override. EL1 execution of the data cache invalidate by set/way instruction is treated as data
cache clean and invalidate by set/way. When this bit is set:
• DCISW is treated as DCCISW when in AArch32 state
• DC ISW is treated as DC CISW when in AArch64 state.

[0] VM Enables second stage of translation. The values are:

0 Disables second stage translation.
1 Enables second stage translation for execution in EL1 and EL0.

To access the HCR_EL2 in AArch64 state, read or write the register with:
MRS <Xt>, HCR_EL2; Read EL2 Hypervisor Configuration Register
MRS HCR_EL2, <Xt>; Write EL2 Hypervisor Configuration Register

Related information
4.5.11 Hyp Configuration Register 2 on page 4-259.
4.5.10 Hyp Configuration Register on page 4-254.

4.3.35 Architectural Feature Trap Register, EL2

The CPTR_EL2 characteristics are:

Purpose
Controls trapping to EL2 for accesses to the CPACR, Trace functionality and registers
associated with floating-point and Advanced SIMD execution. Controls EL2 access to this
functionality.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-137
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Usage constraints
The accessibility of the CPTR_EL2 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW RW

Configurations
The CPTR_EL2 is:
• A 32-bit register in AArch64 state.
• Architecturally mapped to the AArch32 HCPTR register.
Attributes
See the register summary in Table 4-13 AArch64 virtualization registers on page 4-83.
The following figure shows the CPTR_EL2 bit assignments.
31 30 21 20 19 14 13 12 11 10 9 0

RES0 RES0 RES1 RES1

TCPAC TTA RES0 TFP

Figure 4-32 CPTR_EL2 bit assignments

The following table shows the CPTR_EL2 bit assignments.

Table 4-48 CPTR_EL2 bit assignments

Bits Name Function

[63:32] - Reserved, RES0.

[31] TCPAC Traps direct access to CPACR from EL1 to EL2. The possible values are:
0 Access to CPACR is not trapped. This is the reset value.
1 Access to CPACR is trapped.

[30:21] - Reserved, RES0.

[20] TTA This bit is RES0. The processor does not support System register access to trace functionality.

[19:14] - Reserved, RES0.

[13:12] - Reserved, RES1.

[11] - Reserved, RES0.

[10] TFP Traps instructions that access registers associated with floating-point and SIMD execution from a lower Exception
level to EL2, unless trapped to EL1. The possible values are:
0 Instructions are not trapped. This is the reset value.
1 Instructions are trapped.

[9:0] - Reserved, RES1.

To access the CPTR_EL2 in AArch64 state, read or write the register with:
MRS <Xt>, CPTR_EL2; Read EL2 Architectural Feature Trap Register
MSR CPTR_EL2, <Xt>; Write EL2 Architectural Feature Trap Register

Related information
4.5.13 Hyp Architectural Feature Trap Register on page 4-263.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-138
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

4.3.36 Hypervisor System Trap Register

The HSTR_EL2 characteristics are:

Purpose
Controls trapping to Hyp mode of Non-secure accesses, at EL1 or lower, of use of Jazelle or the
CP15 primary coprocessor registers, c0, c1, c2, c3, c5, c6, c7, c8, c9, c10, c11, c12, c13, or c15
in AArch32 state.
Usage constraints
The accessibility to the HSTR_EL2 in AArch64 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW RW

The accessibility to the HSTR in AArch32 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW -

Configurations
The HSTR_EL2 is:
• A Banked EL2 register.
• Architecturally mapped to AArch32 HSTR register.
Attributes
See the register summary in Table 4-13 AArch64 virtualization registers on page 4-83.
The following figure shows the HSTR_EL2 bit assignments.
31 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

Reserved

TTEE T0
T15 T1
Reserved T2
T13 T3
T12 Reserved
T11 T5
T10 T6
T9 T7
T8

Figure 4-33 HSTR_EL2 bit assignments

The following table shows the HSTR_EL2 bit assignments.

Table 4-49 HSTR_EL2 bit assignments

Bits Name Function

[31:17] - Reserved, RES0.

[16] TEEE Trap ThumbEE. This value is:

0 ThumbEE is not supported.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-139
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-49 HSTR_EL2 bit assignments (continued)

Bits Name Function

[15] T15 Trap coprocessor primary register CRn = 15. The possible values are:
0 Has no effect on Non-secure accesses to CP15 coprocessor registers. This is the reset value.
1 Trap valid Non-secure accesses to coprocessor primary register CRn = c15 in AArch32 state to Hyp mode.

[14] - Reserved, RES0.

[13] T13 Trap coprocessor primary register CRn = 13. The possible values are:
0 Has no effect on Non-secure accesses to CP15 coprocessor registers. This is the reset value.
1 Trap valid Non-secure accesses to coprocessor primary register CRn = c13 in AArch32 state to Hyp mode.

[12] T12 Trap coprocessor primary register CRn = 12. The possible values are:
0 Has no effect on Non-secure accesses to CP15 coprocessor registers. This is the reset value.
1 Trap valid Non-secure accesses to coprocessor primary register CRn = c12 in AArch32 state to Hyp mode.

[11] T11 Trap coprocessor primary register CRn = 11. The possible values are:
0 Has no effect on Non-secure accesses to CP15 coprocessor registers. This is the reset value.
1 Trap valid Non-secure accesses to coprocessor primary register CRn = c11 in AArch32 state to Hyp mode.

[10] T10 Trap coprocessor primary register CRn = 10. The possible values are:
0 Has no effect on Non-secure accesses to CP15 coprocessor registers. This is the reset value.
1 Trap valid Non-secure EL0 or EL1 accesses to coprocessor primary register CRn = c10 in AArch32 state to Hyp
mode.

[9] T9 Trap coprocessor primary register CRn = 9. The possible values are:
0 Has no effect on Non-secure accesses to CP15 coprocessor registers. This is the reset value.
1 Trap valid Non-secure EL0 or EL1 accesses to coprocessor primary register CRn = c9 in AArch32 state to Hyp
mode.

[8] T8 Trap coprocessor primary register CRn = 8. The possible values are:
0 Has no effect on Non-secure accesses to CP15 coprocessor registers. This is the reset value.
1 Trap valid Non-secure EL0 or EL1 accesses to coprocessor primary register CRn = c8 in AArch32 state to Hyp
mode.

[7] T7 Trap coprocessor primary register CRn = 7. The possible values are:
0 Has no effect on Non-secure accesses to CP15 coprocessor registers. This is the reset value.
1 Trap valid Non-secure EL0 or EL1 accesses to coprocessor primary register CRn = c7 in AArch32 state to Hyp
mode.

[6] T6 Trap coprocessor primary register CRn = 6. The possible values are:
0 Has no effect on Non-secure accesses to CP15 coprocessor registers. This is the reset value.
1 Trap valid Non-secure EL0 or EL1 accesses to coprocessor primary register CRn = c6 in AArch32 state to Hyp
mode.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-140
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-49 HSTR_EL2 bit assignments (continued)

Bits Name Function

[5] T5 Trap coprocessor primary register CRn = 5. The possible values are:
0 Has no effect on Non-secure accesses to CP15 coprocessor registers. This is the reset value.
1 Trap valid Non-secure EL0 or EL1 accesses to coprocessor primary register CRn = c5 in AArch32 state to Hyp
mode.

[4] - Reserved, RES0.

[3] T3 Trap coprocessor primary register CRn = 3. The possible values are:
0 Has no effect on Non-secure accesses to CP15 coprocessor registers. This is the reset value.
1 Trap valid Non-secure EL0 or EL1 accesses to coprocessor primary register CRn = c3 in AArch32 state to Hyp
mode.

[2] T2 Trap coprocessor primary register CRn = 2. The possible values are:
0 Has no effect on Non-secure accesses to CP15 coprocessor registers. This is the reset value.
1 Trap valid Non-secure EL0 or EL1 accesses to coprocessor primary register CRn = c2 in AArch32 state to Hyp
mode.

[1] T1 Trap coprocessor primary register CRn = 1. The possible values are:
0 Has no effect on Non-secure accesses to CP15 coprocessor registers. This is the reset value.
1 Trap valid Non-secure EL0 or EL1 accesses to coprocessor primary register CRn = c1 in AArch32 state to Hyp
mode.

[0] T0 Trap coprocessor primary register CRn = 0. The possible values are:
0 Has no effect on Non-secure accesses to CP15 coprocessor registers. This is the reset value.
1 Trap valid Non-secure EL0 or EL1 accesses to coprocessor primary register CRn = c0 in AArch32 state to Hyp
mode.

To access the HSTR_EL2 in AArch64 state, read or write the register with:
MRS <Xt>, HSTR_EL2; Read Hyp System Trap Register
MSR HSTR_EL2, <Xt>; Write Hyp System Trap Register

To access the HSTR in AArch32 state, read or write the CP15 register with:
MRC p15, 4, <Rt>, c1, c1, 3; Read Hyp System Trap Register
MCR p15, 4, <Rt>, c1, c1, 3; Write Hyp System Trap Register

4.3.37 Hyp Auxiliary Configuration Register

The processor does not implement HACR_EL2 in AArch64 state. This register is RES0 in EL2 and EL3.
The processor does not implement HACR in AArch32 state. This register is RES0 in Hyp mode and in
Monitor mode when SCR.NS is 1.

4.3.38 System Control Register, EL3

The SCTLR_EL3 characteristics are:

Purpose
Provides top-level control of the system, including its memory system at EL3 in AArch64 state.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-141
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Usage constraints
The accessibility of the SCTLR_EL3 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - - RW RW

Configurations
The SCTLR_EL3 is:
• A 32-bit register in AArch64 state.
• Architecturally mapped to Secure AArch32 SCTLR register.
Attributes
See the register summary in Table 4-3 AArch64 virtual memory control registers on page 4-78.
The following figure shows the SCTLR_EL3 bit assignments.
31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 13 12 11 10 6 5 4 3 2 1 0

RES0 RES1 RES0 RES1 RES0 RES0 I RES0 RES1 C A M

EE WXN RES1 RES1 SA

RES0 RES1 RES0

Figure 4-34 SCTLR_EL3 bit assignments

The following table shows the SCTLR_EL3 bit assignments.

Table 4-50 SCTLR_EL3 bit assignments

Bits Name Function

[63:30] - Reserved, RES0.

[29:28] - Reserved, RES1.

[27:26] - Reserved, RES0.

[25] EE Exception endianness. The values are:

0 Little-endian.
1 Big-endian.

The reset value depends on the primary input CFGEND.

[24] - Reserved, RES0.

[23:22] - Reserved, RES1.

[21:20] - Reserved, RES0.

[19] WXN Force treatment of all memory regions with write permissions as XN. The values are:
0 Regions with write permissions are not forced to XN. This is the reset value.
1 Regions with write permissions are forced to XN.

[18] - Reserved, RES1.

[17] - Reserved, RES0.

[16] - Reserved, RES1.

[15:13] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-142
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-50 SCTLR_EL3 bit assignments (continued)

Bits Name Function

[12] I Global instruction cache enable. The values are:

0 Instruction caches disabled.
1 Instruction caches enabled.

[11] - Reserved, RES1.

[10:6] - Reserved, RES0.

[5:4] - Reserved, RES1.

[3] SA Enables Stack Alignment check. The values are:

0 Disables Stack Alignment check. This is the reset value
1 Enables Stack Alignment check.

[2] C Global enable for data and unified caches. The values are:
0 Disables data and unified caches. This is the reset value.
1 Enables data and unified caches.

[1] A Enable Alignment fault check. The values are:

0 Disables Alignment fault checking. This is the reset value.
1 Enables Alignment fault checking.

[0] M Global enable for the EL1 and EL0 stage 1 MMU. The values are:
0 Disables EL1 and EL0 stage 1 MMU. This is the reset value.
1 Enables EL1 and EL0 stage 1 MMU.

To access the SCTLR_EL3 in AArch64 state, read or write the register with:
MRS <Xt>, SCTLR_EL3; Read EL3 System Control Register
MSR SCTLR_EL3, <Xt>; Write EL3 System Control Register

Related information
4.5.5 System Control Register on page 4-242.

4.3.39 Auxiliary Control Register, EL3

The ACTLR_EL3 characteristics are:

Purpose
Enables access to the control registers for the L2 cache and the processor control registers.
ACTLR_EL3 is used in conjunction with the ACTLR_EL2 register.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-143
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Usage constraints
The accessibility to the ACTLR_EL3 in AArch64 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - - RW RW

The accessibility to the ACTLR in AArch32 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Configurations
The ACTLR_EL3 is:
• A Banked register.
• Mapped to the Secure AArch32 ACTLR register.
Attributes
See the register summary in Table 4-4 AArch64 other System registers on page 4-79.
The following figure shows the ACTLR_EL3 bit assignments.
31 7 6 5 4 3 2 1 0

RES0 RES0

L2ACTLR
L2ECTLR
L2CTLR
CPUECTLR
CPUACTLR

Figure 4-35 ACTLR_EL3 bit assignments

The following table shows the ACTLR_EL3 bit assignments.

Table 4-51 ACTLR_EL3 bit assignments

Bits Name Function

[31:7] - Reserved, RES0.

[6] L2ACTLR L2 Auxiliary Control Register. The possible values are:

0 The register is not accessible from a lower Exception level. This is the reset value.
1 The register is accessible from a lower Exception level.

[5] L2ECTLR L2 Extended Control Register. The possible values are:

0 The register is not accessible from a lower Exception level. This is the reset value.
1 The register is accessible from a lower Exception level.

[4] L2CTLR L2 Control Register. The possible values are:

0 The register is not accessible from a lower Exception level. This is the reset value.
1 The register is accessible from a lower Exception level.

[3:2] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-144
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-51 ACTLR_EL3 bit assignments (continued)

Bits Name Function

[1] CPUECTLR CPU Extended Control Register. The possible values are:
0 The register is not accessible from a lower Exception level. This is the reset value.
1 The register is accessible from a lower Exception level.

[0] CPUACTLR CPU Auxiliary Control Register. The possible values are:
0 The register is not accessible from a lower Exception level. This is the reset value.
1 The register is accessible from a lower Exception level.

To access the ACTLR_EL3 in AArch64 state, read or write the register with:
MRS <Xt>, ACTLR_EL3; Read Auxiliary Control Register
MSR ACTLR_EL3, <Xt>; Write Auxiliary Control Register

To access the ACTLR in AArch32 state, read or write the CP15 register with:
MRC p15, 0, <Rt>, c1, c0, 1; Read Auxiliary Control Register
MCR p15, 0, <Rt>, c1, c0, 1; Write Auxiliary Control Register

Related information
4.3.33 Auxiliary Control Register, EL2 on page 4-130.

4.3.40 Architectural Feature Trap Register, EL3

The CPTR_EL3 characteristics are:

Purpose
Controls trapping to EL3 for accesses to the CPACR_EL1 register, trace functionality and
registers associated with floating-point and SIMD execution. Also controls EL3 access to this
functionality.
Usage constraints
The accessibility of the CPTR_EL3 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - - RW RW

Configurations
The CPTR_EL3 is a 32-bit register.
Attributes
See the register summary in Table 4-12 AArch64 security registers on page 4-83.
The following figure shows the CPTR_EL3 bit assignments.
31 30 21 20 19 11 10 9 0

RES0 RES0 RES0

TCPAC TTA TFP

Figure 4-36 CPTR_EL3 bit assignments

The following table shows the CPTR_EL3 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-145
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-52 CPTR_EL3 bit assignments

Bits Name Function

[63:32] - Reserved, RES0.

[31] TCPAC Traps direct access to CPACR_EL1 from EL1 to EL3. The possible values are:
0 Access to CPACR_EL1 is not trapped. This is the reset value.
1 Access to CPACR_EL1 is trapped.

[30:21] - Reserved, RES0.

[20] TTA This bit is RES0. The processor does not support System register access to trace functionality.

[19:11] - Reserved, RES0.

[10] TFP Traps instructions that access registers associated with floating-point and Advanced SIMD execution from a lower
Exception level to EL3, unless trapped to EL1. The possible values are:
0 Instructions that access registers associated with floating-point and Advanced SIMD execution are not trapped.
1 Instructions that access registers associated with floating-point and Advanced SIMD execution are trapped.
This is the reset value.

[9:0] - Reserved, RES0.

To access the CPTR_EL3 in AArch64 state, read or write the register with:
MRS <Xt>, CPTR_EL3; Read EL3 Architectural Feature Trap Register
MSR CPTR_EL3, <Xt>; Write EL3 Architectural Feature Trap Register

4.3.41 Translation Control Register, EL1

The TCR_EL1 characteristics are:
Purpose
Controls which Translation Base Register defines the base address register for a translation table
walk required for stage 1 translation of a memory access from EL0 or EL1. Also controls the
translation table format and holds cacheability and shareability information.
Usage constraints
The accessibility of the TCR_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Configurations
TCR_EL1[31:0] is architecturally mapped to the Non-secure AArch32 TTBCR register.
Attributes
See the register summary in Table 4-3 AArch64 virtual memory control registers on page 4-78.
The following figure shows the TCR_EL1 bit assignments.
63 39 38 37 36 35 34 32 31 30 29 28 27 26 25 24 23 22 21 16 15 14 13 12 11 10 9 8 7 6 5 0

RES0 AS IPS SH1 A1 T1SZ SH0 T0SZ

RES0 TG1 EPD1 TG0 RES0

TBI0 RES1 IRGN1 RES0 EPD0
TBI1 ORGN1 IRGN0
ORGN0

Figure 4-37 TCR_EL1 bit assignments

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-146
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

The following table shows the TCR_EL1 bit assignments.

Table 4-53 TCR_EL1 bit assignments

Bits Name Function

[63:39] - Reserved, RES0.

[38] TBI1 Top Byte Ignored. Indicates whether the top byte of the input address is used for address match for the TTBR1
region. The values are:
0 Top byte used in the address calculation.
1 Top byte ignored in the address calculation.

[37] TBI0 Top Byte Ignored. Indicates whether the top byte of the input address is used for address match for the TTBR0
region. The values are:
0 Top byte used in the address calculation.
1 Top byte ignored in the address calculation.

[36] AS ASID size. The values are:

0 8-bit.
1 16-bit.

[35] - Reserved, RES0.

[34:32] IPS Intermediate Physical Address Size. The possible values are:
0b000 32-bit, 4GBytes.
0b001 36-bit, 64GBytes.
0b010 40-bit, 1TByte.
0b011 42-bit, 4TBytes.
0b100 44-bit, 16TBytes.
0b101 48-bit, 256TBytes.

[31] - Reserved, RES1.

[30] TG1 TTBR1_EL1 granule size. The values are:

0 4KB.
1 64KB.

[29:28] SH1 Shareability attribute for memory associated with translation table walks using TTBR1. The values are:
0b00 Non-shareable.
0b01 Reserved.
0b10 Outer Shareable.
0b11 Inner Shareable.

[27:26] ORGN1 Outer cacheability attribute for memory associated with translation table walks using TTBR1. The values are:
0b00 Normal memory, Outer Non-cacheable.
0b01 Normal memory, Outer Write-Back Write-Allocate Cacheable.
0b10 Normal memory, Outer Write-Through Cacheable.
0b11 Normal memory, Outer Write-Back no Write-Allocate Cacheable.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-147
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-53 TCR_EL1 bit assignments (continued)

Bits Name Function

[25:24] IRGN1 Inner cacheability attribute for memory associated with translation table walks using TTBR1. The values are:
0b00 Normal memory, Inner Non-cacheable.
0b01 Normal memory, Inner Write-Back Write-Allocate Cacheable.
0b10 Normal memory, Inner Write-Through Cacheable.
0b11 Normal memory, Inner Write-Back no Write-Allocate Cacheable.

[23] EPD1 Translation table walk disable for translations using TTBR1. Controls if a translation table walk is performed on a
TLB miss for an address that is translated using TTBR1. The values are:
0 Perform translation table walk using TTBR1.
1 A TLB miss on an address translated from TTBR1 generates a Translation fault. No translation
table walk is performed.

[22] A1 Selects whether TTBR0 or TTBR1 defines the ASID. The values are:
0 TTBR0.ASID defines the ASID.
1 TTBR1.ASID defines the ASID.

[21:16] T1SZ Size offset of the memory region addressed by TTBR1. The region size is 2(64–TSIZE) bytes.

[15] - Reserved, RES0.

[14] TG0 TTBR0_EL1 granule size. The values are:

0 4KB.
1 64KB.

[13:12] SH0 Shareability attribute for memory associated with translation table walks using TTBR0. The values are:
0b00 Non-shareable.
0b01 Reserved.
0b10 Outer Shareable.
0b11 Inner Shareable.

[11:10] ORGN0 Outer cacheability attribute for memory associated with translation table walks using TTBR0. The values are:
0b00 Normal memory, Outer Non-cacheable.
0b01 Normal memory, Outer Write-Back Write-Allocate Cacheable.
0b10 Normal memory, Outer Write-Through Cacheable.
0b11 Normal memory, Outer Write-Back no Write-Allocate Cacheable.

[9:8] IRGN0 Inner cacheability attribute for memory associated with translation table walks using TTBR0. The values are:
0b00 Normal memory, Inner Non-cacheable.
0b01 Normal memory, Inner Write-Back Write-Allocate Cacheable.
0b10 Normal memory, Inner Write-Through Cacheable.
0b11 Normal memory, Inner Write-Back no Write-Allocate Cacheable.

[7:6] - Reserved, RES0.

[5:0] T0SZ Size offset of the memory region addressed by TTBR0. The region size is 2(64–TSIZE) bytes.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-148
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

To access the TCR_EL1 in AArch64 state, read or write the register with:
MRS <Xt>, TCR_EL1; Read EL1 Translation Control Register
MSR TCR_EL1, <Xt>; Write EL1 Translation Control Register

Related information
4.5.15 Translation Table Base Control Register on page 4-265.

4.3.42 Translation Control Register, EL2

The TCR_EL2 characteristics are:
Purpose
Controls translation table walks required for stage 1 translation of a memory access from EL2
and holds cacheability and shareability information.
Usage constraints
The accessibility of the TCR_EL2 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW RW

Configurations
The TCR_EL2 is architecturally mapped to the AArch32 HCTR register.
Attributes
See the register summary in Table 4-3 AArch64 virtual memory control registers on page 4-78.
The following figure shows the TCR_EL2 bit assignments.
31 30 24 23 22 21 20 19 18 16 15 14 13 12 11 10 9 8 7 6 5 0

RES0 RES0 PS SH0 RES0 T0SZ

RES1 RES1 RES0 TG0 IRGN0

TBI RES0 ORGN0

Figure 4-38 TCR_EL2 bit assignments

The following table shows the TCR_EL2 bit assignments.

Table 4-54 TCR_EL2 bit assignments

Bits Name Function

[31] - Reserved, RES1.

[30:24] - Reserved, RES0.

[23] - Reserved, RES1.

[22:21] - Reserved, RES0.

[20] TBI Top Byte Ignored. Indicates whether the top byte of the input address is used for address match. The values are:
0 Top byte used in the address calculation.
1 Top byte ignored in the address calculation.

[19] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-149
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-54 TCR_EL2 bit assignments (continued)

Bits Name Function

[18:16] PS Physical Address size. The possible values are:

0b000 32-bit, 4GBytes.
0b001 36-bit, 64GBytes.
0b010 40-bit, 1TByte.
0b011 42-bit, 4TBytes.
0b100 44-bit, 16TBytes.
0b101 48-bit, 256TBytes.

All other values are reserved.

[15] - Reserved, RES0.

[14] TG0 TTBR0_EL2 granule size. The values are:

0 4KByte.
1 64KByte.

[13:12] SH0 Shareability attribute for memory associated with translation table walks using TTBR0. The values are:
0b00 Non-shareable.
0b01 Reserved.
0b10 Outer Shareable.
0b11 Inner Shareable.

[11:10] ORGN0 Outer cacheability attribute for memory associated with translation table walks using TTBR0. The values are:
0b00 Normal memory, Outer Non-cacheable.
0b01 Normal memory, Outer Write-Back Write-Allocate Cacheable.
0b10 Normal memory, Outer Write-Through Cacheable.
0b11 Normal memory, Outer Write-Back no Write-Allocate Cacheable.

[9:8] IRGN0 Inner cacheability attribute for memory associated with translation table walks using TTBR0. The values are:
0b00 Normal memory, Inner Non-cacheable.
0b01 Normal memory, Inner Write-Back Write-Allocate Cacheable.
0b10 Normal memory, Inner Write-Through Cacheable.
0b11 Normal memory, Inner Write-Back no Write-Allocate Cacheable.

[7:6] - Reserved, RES0.

[5:0] T0SZ Size offset of the memory region addressed by TTBR0. The region size is 2(64–TSIZE) bytes.

To access the TCR_EL2 in AArch64 state, read or write the register with:
MRS <Xt>, TCR_EL2; Read EL2 Translation Control Register
MSR TCR_EL2, <Xt>; Write EL2 Translation Control Register

Related information
4.5.16 Hyp Translation Control Register on page 4-266.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-150
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

4.3.43 Virtualization Translation Control Register, EL2

The VTCR_EL2 characteristics are:

Purpose
Controls the translation table walks required for the stage 2 translation of memory accesses from
Non-secure EL0 and EL1, and holds cacheability and shareability information for the accesses.
Usage constraints
The accessibility to the VTCR_EL2 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW RW

Configurations
The VTCR_EL2 is:
• A32-bit register in AArch64 state.
• Architecturally mapped to the AArch32 VTCR register.
Attributes
See the register summary in Table 4-3 AArch64 virtual memory control registers on page 4-78.
The following figure shows the VTCR_EL2 bit assignments.
31 30 19 18 16 15 14 13 12 11 10 9 8 7 6 5 0

RES0 PS SH0 SL0 T0SZ

RES1 TG0 IRGN0

RES0 ORGN0

Figure 4-39 VTCR_EL2 bit assignments

The following table shows the VTCR_EL2 bit assignments.

Table 4-55 VTCR_EL2 bit assignments

Bits Name Function

[31] - Reserved, RES1.

[30:19] - Reserved, RES0.

[18:16] PS Physical Address Size. The possible values are:

0b000 32-bit, 4GBytes.
0b001 36-bit, 64GBytes.
0b010 40-bit, 1TByte.
0b011 42-bit, 4TBytes.
0b100 44-bit, 16TBytes.
0b101 48-bit, 256TBytes.

All other values are reserved.

[15] - Reserved, RES0.

[14] TG0 Granule size for the corresponding TTBR0_ELx.

0 4KB.
1 64KB.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-151
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-55 VTCR_EL2 bit assignments (continued)

Bits Name Function

[13:12] SH0 Shareability attribute for memory associated with translation table walks using TTBR0:
0b00 Non-shareable.
0b01 Reserved.
0b11 Outer Shareable.
0b11 Inner Shareable.

[11:10] ORGN0 Outer cacheability attribute for memory associated with translation table walks using TTBR0.
0b00 Normal memory, Outer Non-cacheable.
0b01 Normal memory, Outer Write-Back Write-Allocate Cacheable.
0b11 Normal memory, Outer Write-Through Cacheable.
0b11 Normal memory, Outer Write-Back no Write-Allocate Cacheable.

[9:8] IRGN0 Inner cacheability attribute for memory associated with translation table walks using TTBR0.
0b00 Normal memory, Inner Non-cacheable.
0b01 Normal memory, Inner Write-Back Write-Allocate Cacheable.
0b11 Normal memory, Inner Write-Through Cacheable.
0b11 Normal memory, Inner Write-Back no Write-Allocate Cacheable.

[7:6] SL0 Starting level of the VTCR_EL2 addressed region.

[5:0] T0SZ The size offset of the memory region addressed by TTBR0. The region size is 2(64–T0SZ) bytes.

To access the VTCR_EL2 in AArch64 state, read or write the register with:
MRS <Xt>, VTCR_EL2; Read EL2 Virtualization Translation Control Register
MSR VTCR_EL2, <Xt>; Write EL2 Virtualization Translation Control Register

4.3.44 Translation Table Base Register 0, EL1

The TTBR0_EL1 characteristics are:

Purpose
Holds the base address of translation table 0, and information about the memory it occupies.
This is one of the translation tables for the stage 1 translation of memory accesses at EL1 if the
highest Exception level is in AArch64 state.
Usage constraints
The TTBR0_EL1 is used in conjunction with TCR_EL1.
The accessibility to the TTBR0_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Configurations
TTBR0_EL1 is architecturally mapped to the Non-secure AArch32 register TTBR0.
Attributes
See the register summary in Table 4-3 AArch64 virtual memory control registers on page 4-78.
The following figure shows the TTBR0_EL1 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-152
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

63 48 47 10 9 0

ASID BADDR Reserved

Figure 4-40 TTBR0_EL1 bit assignments

The following table shows the TTBR0_EL1 bit assignments.

Table 4-56 TTBR0_EL1 bit assignments

Bits Name Function

[63:48] ASID An ASID for the translation table base address.

The TCR_EL1.A1 field selects either the TTBR0.ASID or the TTBR1.ASID.
The TCR_EL1.AS bit selects whether all 16-bits [63:48] or the lower 8-bits [55:48] indicate the current ASID.

[47:10] BADDR Translation table base address. Defining the translation table base address width.

[9:0] - Reserved, RES0.

To access the TTBR0_EL1 in AArch64 state, read or write the register with:
MRS <Xt>, TTBR0_EL1; Read EL1 Translation Table Base Register 0
MSR TTBR0_EL1, <Xt>; Write EL1 Translation Table Base Register 0

4.3.45 Translation Table Base Register 0, EL3

The TTBR0_EL3 characteristics are:

Purpose
Holds the base address of the translation table for the stage 1 translation of memory accesses
from EL3.
Usage constraints
The accessibility to the TTBR0_EL3 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - - RW RW

Configurations
TTBR0_EL3 is mapped to the Secure AArch32 TTBR0 register.
Attributes
See the register summary in Table 4-3 AArch64 virtual memory control registers on page 4-78.
The following figure shows the TTBR0_EL3 bit assignments.
63 48 47 10 9 0

Reserved BADDR Reserved

Figure 4-41 TTBR0_EL3 bit assignments

The following table shows the TTBR0_EL3 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-153
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-57 TTBR0_EL3 bit assignments

Bits Name Function

[63:48] - Reserved, RES0.

[47:10] BADDR Translation table base address. Defining the translation table base address width.

[9:0] - Reserved, UNK/RES0.

To access the TTBR0_EL3 in AArch64 state, read or write the register with:
MRS <Xt>, TTBR0_EL3; Read EL3 Translation Table Base Register 0
MSR TTBR0_EL3, <Xt>; Write EL3 Translation Table Base Register 0

4.3.46 Translation Table Base Register 1, EL1

The TTBR1_EL1 characteristics are:

Purpose
Holds the base address of translation table 1, and information about the memory it occupies.
This is one of the translation tables for the stage 1 translation of memory accesses at EL0 and
EL1. This is one of the translation tables for the stage 1 translation of memory accesses at EL1
if the highest Exception level is in AArch64 state.
Usage constraints
The TTBR1_EL1 is used in conjunction with TCR_EL1.
The accessibility to the TTBR1_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Configurations
TTBR1_EL1 is architecturally mapped to the Non-secure AArch32 register TTBR1.
Attributes
See the register summary in Table 4-3 AArch64 virtual memory control registers on page 4-78.
The following figure shows the TTBR1_EL1 bit assignments.
63 48 47 10 9 0

ASID BADDR Reserved

Figure 4-42 TTBR1_EL1 bit assignments

The following table shows the TTBR0_EL1 bit assignments.

Table 4-58 TTBR1_EL1 bit assignments

Bits Name Function

[63:48] ASID An ASID for the translation table base address.

The TCR_EL1.A1 field selects either the TTBR0.ASID or the TTBR1.ASID.
The TCR_EL1.AS bit selects whether all 16-bits [63:48] or the lower 8-bits [55:48] indicate the current ASID.

[47:10] BADDR Translation table base address. Defining the translation table base address width.

[9:0] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-154
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

To access the TTBR0_EL1 in AArch64 state, read or write the register with:
MRS <Xt>, TTBR1_EL1; Read EL1 Translation Table Base Register 1
MSR TTBR1_EL1, <Xt>; Write EL1 Translation Table Base Register 1

4.3.47 Translation Control Register, EL3

The TCR_EL3 characteristics are:
Purpose
Controls translation table walks required for stage 1 translation of memory accesses from EL3
and holds cacheability and shareability information for the accesses.
Usage constraints
The accessibility of the TCR_EL3 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - - RW RW

Configurations
The TCR_EL3 is architecturally mapped to the Secure AArch32 TTBCR register.
Attributes
See the register summary in Table 4-3 AArch64 virtual memory control registers on page 4-78.
The following figure shows the TCR_EL3 bit assignments.
31 30 24 23 22 21 20 19 18 16 15 14 13 12 11 10 9 8 7 6 5 0

RES0 RES0 PS SH0 RES0 T0SZ

RES1 RES1 RES0 TG0 IRGN0

TBI RES0 ORGN0

Figure 4-43 TCR_EL3 bit assignments

The following table shows the TCR_EL3 bit assignments.

Table 4-59 TCR_EL3 bit assignments

Bits Name Function

[31] - Reserved, RES1.

[30:24] - Reserved, RES0.

[23] - Reserved, RES1.

[22:21] - Reserved, RES0.

[20] TBI Top Byte Ignored. Indicates whether the top byte of the input address is used for address match. The values are:
0 Top byte used in the address calculation.
1 Top byte ignored in the address calculation.

[19] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-155
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-59 TCR_EL3 bit assignments (continued)

Bits Name Function

[18:16] PS Physical Address size. The possible values are:

0b000 32 bits, 4GBytes.
0b001 36 bits, 64GBytes.
0b010 40 bits, 1TByte.
0b011 42 bits, 4TBytes.
0b100 44 bits, 16TBytes.
0b101 48 bits, 256TBytes.

[15] - Reserved, RES0.

[14] TGO TTBR0_EL3 granule size. The values are:

0 4KByte.
1 64KByte.

[13:12] SH0 Shareability attribute for memory associated with translation table walks using TTBR0. The values are:
0b00 Non-shareable.
0b01 Reserved.
0b10 Outer Shareable.
0b11 Inner Shareable.

[11:10] ORGN0 Outer cacheability attribute for memory associated with translation table walks using TTBR0. The values are:
0b00 Normal memory, Outer Non-cacheable.
0b01 Normal memory, Outer Write-Back Write-Allocate Cacheable.
0b10 Normal memory, Outer Write-Through Cacheable.
0b11 Normal memory, Outer Write-Back no Write-Allocate Cacheable.

[9:8] IRGN0 Inner cacheability attribute for memory associated with translation table walks using TTBR0. The values are:
0b00 Normal memory, Inner Non-cacheable.
0b01 Normal memory, Inner Write-Back Write-Allocate Cacheable.
0b10 Normal memory, Inner Write-Through Cacheable.
0b11 Normal memory, Inner Write-Back no Write-Allocate Cacheable.

[7:6] - Reserved, RES0.

[5:0] T0SZ Size offset of the memory region addressed by TTBR0. The region size is 2(64–TSIZE) bytes.

To access the TCR_EL3 in AArch64 state, read or write the register with:
MRS <Xt>, TCR_EL3; Read EL3 Translation Control Register
MRS TCR_EL3, <Xt>; Read EL3 Translation Control Register

Related information
4.5.15 Translation Table Base Control Register on page 4-265.

4.3.48 Auxiliary Fault Status Register 0, EL1 and EL3

The processor implements AFSR0_EL1, AFSR0_EL3, and ADFSR as RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-156
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

4.3.49 Auxiliary Fault Status Register 1, EL1 and EL3

The processor implements AFSR1_EL1, AFSR1_EL3, and AIFSR as RES0.

4.3.50 Exception Syndrome Register, EL1 and EL3

The ESR_EL1 and ESR_EL3 characteristics are:
Purpose
ESR_EL1 holds syndrome information for an exception taken to EL1.
ESR_EL3 holds syndrome information for an exception taken to EL3.
Usage constraints
The accessibility to the ESR_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

The accessibility to the ESR_EL3 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - - RW RW

Configurations
The ESR_EL1 is architecturally mapped to the Non-secure AArch32 DFSR register.
The ESR_EL3 is mapped to the Secure AArch32 DFSR register.
Attributes
See the register summary in Table 4-2 AArch64 exception handling registers on page 4-78.

EC==0b100000 and EC==0b100001, Instruction Aborts

This section describes the IMPLEMENTATION DEFINED behavior of the EA bit for Instruction Abort
exceptions.
The following figure shows the ESR_EL1 and ESR_EL3 bit assignments for the Instruction Abort
exception classes, that is, when EC==0b100000 or EC==0b100001.
31 26 25 24 10 9 8 7 6 5 0

EC Reserved IFSC

IL Reserved
S1TPW
Reserved
EA

Figure 4-44 ESR_EL1 and ESR_EL3 bit assignments

The following table shows the ESR_EL1 and ESR_EL3 bit assignments for the Instruction Abort
exception class.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-157
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-60 ESR_EL1 and ESR_EL3 bit assignments

Bits Name Function

[31:26] EC Exception Class:

0b100000 Instruction Abort that caused entry from a lower Exception level in AArch32 or AArch64.
0b100001 Instruction Abort that caused entry from a current Exception level in AArch64.

[25] IL Instruction Length for synchronous exceptions.

[24:10] - Reserved, RES0.

[9] EA External abort type. This bit indicates whether an AXI decode or slave error caused an abort. The possible values
are:
0 External abort marked as DECERR.
1 External abort marked as SLVERR.

For aborts other than external aborts this bit always returns 0.

[8] - Reserved, RES0.

[7] S1PTW When 1, indicates the instruction fault came from a second stage fault during a first stage translation table walk.

[6] - Reserved, RES0.

[5:0] IFSC Instruction Fault Status Code. This field indicates the type of exception generated. The possible values are:
0b000000 Address size fault in TTBR0 or TTBR1.
0b000101 Translation fault, 1st level.
00b00110 Translation fault, 2nd level.
00b00111 Translation fault, 3rd level.
0b001001 Access flag fault, 1st level.
0b001010 Access flag fault, 2nd level.
0b001011 Access flag fault, 3rd level.
0b001101 Permission fault, 1st level.
0b001110 Permission fault, 2nd level.
0b001111 Permission fault, 3rd level.
0b010000 Synchronous external abort.
0b011000 Synchronous parity error on memory access.
0b010101 Synchronous external abort on translation table walk, 1st level.
0b010110 Synchronous external abort on translation table walk, 2nd level.
0b010111 Synchronous external abort on translation table walk, 3rd level.
0b011101 Synchronous parity error on memory access on translation table walk, 1st level.
0b011110 Synchronous parity error on memory access on translation table walk, 2nd level.
0b011111 Synchronous parity error on memory access on translation table walk, 3rd level.
0b100001 Alignment fault.
0b100010 Debug event.

All other values are reserved.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-158
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

The lookup level associated with a fault is:

• For a fault generated on a translation table walk, the lookup level of the walk being performed.
• For a Translation fault, the lookup level of the translation table that gave the fault. If a fault occurs
because an MMU is disabled, or because the input address is outside the range specified by the
appropriate base address register or registers, the fault is reported as a First level fault.
• For an Access flag fault, the lookup level of the translation table that gave the fault.
• For a Permission fault, including a Permission fault cased by hierarchical permissions, the lookup
level of the final level of translation table accessed for the translation. That is, the lookup level of the
translation table that returned a Block or Page descriptor.
All exception classes except the Instruction Abort are architecturally defined in the ARM® Architecture
Reference Manual ARMv8. The SError Interrupt exception classes are architecturally defined in the
ARM® Generic Interrupt Controller Architecture Specification, GICv3 with the exception of four bits.
The following changes are Cortex-A72 implementation-defined and only apply to SError Interrupt
exception classes.

Table 4-61 ESR_EL1 and ESR_EL3 Cortex-A72 implementation-defined SError Interrupt exception classes bit assignments

Bits Name Function

[15] Unattributable System Error 0b1 Unattributable - cannot be attributed to the processing element counting the
event.
0b0 Attributable - can be attributed to the processing element counting the event.

[14] Uncontainable System Error 0b1 Uncontainable – an event which cannot be contained to a particular code
sequence.
0b0 Containable - an Attributable event which can be contained to a particular code
sequence.

[1:0] System Error Source 0b00 Decode error

0b01 ECC error
0b10 Slave error
0b11 Reserved

4.3.51 Instruction Fault Status Register, EL2

The IFSR32_EL2 characteristics are:

Purpose
Holds status information about the last instruction fault.
Usage constraints
The accessibility to the IFSR32_EL2 in AArch64 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW RW

The accessibility to the IFSR in AArch32 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-159
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Configurations
The IFSR32_EL2 is:
• Banked for Secure and Non-secure states.
• Mapped to the Non-secure AArch32 IFSR register.
Attributes
See the register summary in Table 4-2 AArch64 exception handling registers on page 4-78.
There are two formats for this register. The value of TTBCR.EAE selects which format of the register is
used. The two formats are:
• IFSR32_EL2 format when using the Short-descriptor translation table format.
• IFSR32_EL2 format when using the Long-descriptor translation table format.

IFSR32_EL2 format when using the Short-descriptor translation table format

The following figure shows the IFSR32_EL2 bit assignments when using the Short-descriptor translation
table format.
31 13 12 11 10 9 8 4 3 0

RES0 RES0 FS[3:0]

ExT LPAE
RES0 FS[4]

Figure 4-45 IFSR32_EL2 bit assignments for Short-descriptor translation table format

The following table shows the IFSR32_EL2 bit assignments when using the Short-descriptor translation
table format.

Table 4-62 IFSR32_EL2 bit assignments for Short-descriptor translation table format

Bits Name Function

[31:13] - Reserved, RES0.

[12] ExT External abort type. This field indicates whether an AXI decode or slave error caused an abort. The possible values
are:
0 External abort marked as DECERR.
1 External abort marked as SLVERR.

For aborts other than external aborts this bit always returns 0.

[11] - Reserved, RES0.

[10] FS[4] MSB of the Fault Status field. See bits[3:0] in this table.

[9] LPAE Large physical address extension. The value of the format descriptor is:
0 Short-descriptor translation table formats.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-160
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-62 IFSR32_EL2 bit assignments for Short-descriptor translation table format (continued)

Bits Name Function

[8:4] - Reserved, RES0.

[3:0] FS[3:0] Fault Status bits. This field indicates the type of exception generated. The possible values are:
0b00001 Alignment fault.
0b01100 Synchronous external abort on translation table walk, 1st level.
0b01110 Synchronous external abort on translation table walk, 2nd level.
0b11100 Synchronous parity error on translation table walk, 1st level.
0b11110 Synchronous parity error on translation table walk, 2nd level.
0b00101 Translation fault, 1st level.
0b00111 Translation fault, 2nd level.
0b00011 Access flag fault, 1st level.
0b00110 Access flag fault, 2nd level.
0b01001 Domain fault, 1st level.
0b01011 Domain fault, 2nd level.
0b01101 Permission fault, 1st level.
0b01111 Permission fault, 2nd level.
0b00010 Debug event.
0b01000 Synchronous external abort, non-translation.
0b11001 Synchronous parity error on memory access.

All other values are reserved.

IFSR32_EL2 format when using the Long-descriptor translation table format

The following figure shows the IFSR32_EL2 bit assignments when using the Long-descriptor translation
table format.
31 13 12 11 10 9 8 6 5 0

RES0 RES0 RES0 Status

ExT LPAE

Figure 4-46 IFSR32_EL2 bit assignments for Long-descriptor translation table format

The following table shows the IFSR32_EL2 bit assignments when using the Long-descriptor translation
table format.

Table 4-63 IFSR32_EL2 bit assignments for Long-descriptor translation table format

Bits Name Function

[31:13] - Reserved, res0.

[12] ExT External abort type. This field indicates whether an AXI decode or slave error caused an abort. The possible values
are:
0 External abort marked as DECERR.
1 External abort marked as SLVERR.

For aborts other than external aborts this bit always returns 0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-161
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-63 IFSR32_EL2 bit assignments for Long-descriptor translation table format (continued)

Bits Name Function

[11:10] - Reserved, res0.

[9] LPAE Large physical address extension. The value of the format descriptor is:
1 Long-descriptor translation table formats.

[8:6] - Reserved, res0.

[5:0] Status Fault Status bits. This field indicates the type of exception generated. The possible values are:
0b0000LL Address size fault, LL bits indicate level.
0b0001LL Translation fault, LL bits indicate level.
0b0010LL Access flag fault, LL bits indicate level.
0b0011LL Permission fault, LL bits indicate level.
0b010000 Synchronous external abort.
0b011000 Synchronous parity error on memory access.
0b0101LL Synchronous external abort on translation table walk, LL bits indicate level.
0b0111LL Synchronous parity error on memory access on translation table walk, LL bits indicate level.
0b100001 Alignment fault.
0b100010 Debug event.

All other values are reserved.

The following table shows how the LL bits in the Status field encode the lookup level associated with the
MMU fault.

Table 4-64 Encodings of LL bits associated with the MMU fault

LL bits Meaning

00 Level 0

01 First level

10 Second level

11 Third level

Note
If a Data Abort exception is generated by an Instruction Cache maintenance operation, the fault is
reported as a Cache Maintenance fault in the DFSR or HSR with the appropriate Fault Status code. For
such exceptions reported in the DFSR, the corresponding IFSR is UNKNOWN.

To access the IFSR32_EL2 in AArch64 state, read or write the register with:
MRS <Xt>, IFSR32_EL2; Read EL2 Instruction Fault Status Register
MSR IFSR32_EL2, <Xt>; Write EL2 Instruction Fault Status Register

To access the IFSR in AArch32 state, read or write the CP15 register with:
MRC p15, 0, <Rt>, c5, c0, 1; Read Instruction Fault Status Register
MCR p15, 0, <Rt>, c5, c0, 1; Write Instruction Fault Status Register

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-162
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Related information
IFSR32_EL2 format when using the Short-descriptor translation table format on page 4-160.
IFSR32_EL2 format when using the Long-descriptor translation table format on page 4-161.

4.3.52 Auxiliary Fault Status Register 0, EL2 and Hyp Auxiliary Data Fault Status Register
The processor implements AFSR0_EL2 and HADFSR as RES0.

4.3.53 Auxiliary Fault Status Register 1, EL2 and Hyp Auxiliary Instruction Fault Status Register
The processor implements AFSR1_EL2 and HAIFSR as RES0.

4.3.54 Exception Syndrome Register, EL2

The ESR_EL2 characteristics are:

Purpose Holds syndrome information for an exception taken to EL2.
Usage The accessibility to the ESR_EL2 in AArch64 state by Exception level is:
constraints
EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW RW

The accessibility to the HSR in AArch32 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW -

Configurations The ESR_EL2 is:

• A Banked EL2 register.
• Architecturally mapped to the AArch32 HSR register.
Attributes See the register summary in Table 4-2 AArch64 exception handling registers
on page 4-78.

The following figure shows the ESR_EL2 bit assignments.

31 26 25 24 0

EC IL ISS

Figure 4-47 ESR_EL2 bit assignments

The following table shows the ESR_EL2 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-163
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-65 ESR_EL2 bit assignments

Bits Name Function

[31:26] EC Exception class. The exception class for the exception that is taken in Hyp mode.
When zero, this field indicates that the reason for the exception is not known. In this case, the other fields in this
register are UNKNOWN. Otherwise, the field holds the exception class for the exception. See the ARM® Architecture
Reference Manual ARMv8 for more information.

[25] IL Instruction length. Indicates the size of the instruction that has been trapped to Hyp mode. The values are:
0 16-bit instruction.
1 32-bit instruction.

This field is not valid for:

• Instruction Aborts.
• Data Aborts that do not have ISS information, or for which the ISS is not valid.
In these cases the field is RES0.

[24:0] ISS Instruction specific syndrome. The interpretation of this field depends on the value of the EC field. See Encoding of
ISS[24:20] when HSR[31:30] is 0b00 on page 4-164>.

All exception classes except the Instruction Abort are architecturally defined in the ARM® Architecture
Reference Manual ARMv8. The SError Interrupt exception classes are architecturally defined in the
ARM® Generic Interrupt Controller Architecture Specification, GICv3 with the exception of four bits.
The following changes are Cortex-A72 implementation-defined and only apply to SError Interrupt
exception classes.

Table 4-66 ESR_EL2 Cortex-A72 implementation-defined SError Interrupt exception classes bit assignments

Bits Name Function

[15] Unattributable System Error 0b1 Unattributable - cannot be attributed to the processing element counting the event.
0b0 Attributable - can be attributed to the processing element counting the event.

[14] Uncontainable System Error 0b1 Uncontainable – an event which cannot be contained to a particular code sequence.
0b0 Containable - an Attributable event which can be contained to a particular code sequence.

[1:0] System Error Source 0b00 Decode error

0b01 ECC error
0b10 Slave error
0b11 Reserved

Encoding of ISS[24:20] when HSR[31:30] is 0b00

For EC values that are nonzero and have the two most-significant bits 0b00, ISS[24:20] provides the
condition field for the trapped instruction, together with a valid flag for this field. The encoding of this
part of the ISS field is:
CV, ISS[24] Condition valid. Possible values of this bit are:
0 The COND field is not valid.
1 The COND field is valid.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-164
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

When an instruction is trapped, CV is set to 1.

COND, The Condition field for the trapped instruction. This field is valid only when CV is set to
ISS[23:20] 1.
If CV is set to 0, this field is UNK/RES0.
When an instruction is trapped, the COND field is 0xE.
To access the ESR_EL2 in AArch64 state, read or write the register with:
MRS <Xt>, ESR_EL2; Read EL2 Exception Syndrome Register
MSR ESR_EL2, <Xt>; Write EL2 Exception Syndrome Register

To access the HSR in AArch32 state, read or write the CP15 register with:
MRC p15, 4, <Rt>, c5, c1, 0; Read Hyp Syndrome Register
MCR p15, 4, <Rt>, c5, c1, 0; Write Hyp Syndrome Register

4.3.55 Physical Address Register, EL1

The PAR_EL1 characteristics are:

Purpose
The Physical Address returned from an address translation.
Usage constraints
The accessibility of the PAR_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Configurations
The architectural mapping of the PAR_EL1 is to the Non-secure AArch32 PAR register.
Attributes
See the register summary in Table 4-8 AArch64 address translation operations on page 4-81.
The following figure shows the PAR_EL1 bit assignments when the Virtual Address to Physical Address
conversion completes successfully.
63 60 59 56 55 44 43 12 11 10 9 8 7 6 1 0

AttrH AttrL RES0 PA SHA RES0 F

RES1 NS
RES0

Figure 4-48 PAR_EL1 pass bit assignments

The following table shows the PAR_EL1 bit assignments when the Virtual Address to Physical Address
conversion completes successfully.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-165
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-67 PAR_EL1 pass bit assignments

Bits Name Function

[63:60] AttrH Defines Device memory or Normal memory plus Outer cacheability. Must be used in conjunction with AttrL. The
possible values are:
0x0 Device memory, see AttrL.
0x4 Normal memory, Outer Non-cacheable.
0x8 Normal memory, Outer Write-Through Cacheable.
0x9 Normal memory, Outer Write-Through Cacheable, Outer Write-Allocate.
0xA Normal memory, Outer Write-Through Cacheable, Outer Read-Allocate.
0xB Normal memory, Outer Write-Through Cacheable, Outer Write-Allocate, Outer Read-Allocate.
0xC Normal memory, Outer Write-Back Cacheable.
0xD Normal memory, Outer Write-Back Cacheable, Outer Write-Allocate.
0xE Normal memory, Outer Write-Back Cacheable, Outer Read-Allocate.
0xF Normal memory, Outer Write-Back Cacheable, Outer Write-Allocate, Outer Read-Allocate.

All other values are reserved.

[59:56] AttrL Defines Device memory or Normal memory plus Inner cacheability. Must be interpreted in conjunction with AttrH.
The possible values are:
0x0 Device-nGnRnE memory if AttrH is 0x0. Otherwise this value is reserved.
0x4 Device memory if AttrH is 0x0. Otherwise, Normal memory, Inner Non-cacheable.
0x8 Reserved if AttrH is 0x0. Otherwise, Normal memory, Inner Write-Through Cacheable.
0x9 Reserved if AttrH is 0x0. Otherwise, Normal memory, Inner Write-Through Cacheable, Inner Write-Allocate.
0xA Reserved if AttrH is 0x0. Otherwise, Normal memory, Inner Write-Through Cacheable, Inner Read-Allocate.
0xB Reserved if AttrH is 0x0. Otherwise, Normal memory, Inner Write-Through Cacheable, Inner Write-Allocate,
Inner Read-Allocate.
0xC Reserved if AttrH is 0x0. Otherwise, Normal memory, Inner Write-Back Cacheable.
0xD Reserved if AttrH is 0x0. Otherwise, Normal memory, Inner Write-Back Cacheable, Inner Write-Allocate.
0xE Reserved if AttrH is 0x0. Otherwise, Normal memory, Inner Write-Back Cacheable, Inner Read-Allocate.
0xF Reserved if AttrH is 0x0. Otherwise, Normal memory, Inner Write-Through Cacheable, Inner Write-Allocate,
Inner Read-Allocate.

All other values are reserved.

[55:44] - Reserved, RES0.

[43:12] PA Physical address. The Physical Address corresponding to the supplied Virtual Address. Returns address bits[31:12].

[11] - Reserved, RES1.

[10] - Reserved, RES0.

[9] NS Non-secure. The NS attribute for a translation table entry read from Secure state.
This bit is UNKNOWN for a translation table entry from Non-secure state.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-166
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-67 PAR_EL1 pass bit assignments (continued)

Bits Name Function

[8:7] SHA Shareability attribute for the Physical Address returned from a translation table entry. The values are:
0b00 Non-shareable.
0b01 Reserved.
0b10 Outer Shareable.
0b11 Inner Shareable.

Note
The SHA bit takes the value of 0b10 for:
• Any type of device memory.
• Normal memory with both Inner Non-cacheable and Outer-cacheable attributes.

[6:1] - Reserved, RES0.

[0] F Pass/Fail bit. Indicates whether the conversion completed successfully. This value is:
0 Virtual Address to Physical Address conversion completed successfully.

The following figure shows the PAR_EL1 bit assignments when the Virtual Address to Physical Address
conversion aborts.
63 12 11 10 9 8 7 6 1 0

RES0 S FST F

RES1 RES0
RES0 PTW

Figure 4-49 PAR_EL1 fail bit assignments

The following table shows the PAR_EL1 bit assignments when the Virtual Address to Physical Address
conversion aborts.

Table 4-68 PAR_EL1 fail bit assignments

Bits Name Function

[63:12] - Reserved, RES0.

[11] - Reserved, RES1.

[10] - Reserved, RES0.

[9] S Stage of fault. Indicates the state where the translation aborted. The values are:
0 Translation aborted because of a fault in stage 1 translation.
1 Translation aborted because of a fault in stage 2 translation.

[8] PTW Indicates a stage 2 fault during a stage 1 table walk. The values are:
0 No stage 2 fault during a stage 1 table walk.
1 Translation aborted because of a stage 2 fault during a stage 1 table walk.

[7] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-167
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-68 PAR_EL1 fail bit assignments (continued)

Bits Name Function

[6:1] FST Fault status code, as shown in the Data Abort ESR encoding. See the ARM® Architecture Reference Manual ARMv8
for more information.

[0] F Pass/Fail bit. Indicates whether the conversion completed successfully. The value is:
1 Virtual Address to Physical Address conversion aborted.

To access the PAR_EL1 in AArch64 state, read or write the register with:
MRS <Xt>, PAR_EL1; Read EL1 Physical Address Register
MSR PAR_EL1, <Xt>; Write EL1 Physical Address Register

Related information
4.5.18 Physical Address Register on page 4-271.

4.3.56 Auxiliary Memory Attribute Indirection Register, EL1 and EL3

The processor does not set any IMPLEMENTATION DEFINED attributes in the Auxiliary Memory Attribute
Indirection Registers. AMAIR_EL1 and AMAIR_EL3 are RES0.
AMAIR_EL1[31:0] is architecturally mapped to the Non-secure AArch32 AMAIR0 register.
AMAIR_EL1[63:32] is architecturally mapped to the Non-secure AArch32 AMAIR1 register.
AMAIR_EL3[31:0] is architecturally mapped to the Secure AArch32 AMAIR0 register.
AMAIR_EL3[63:32] is architecturally mapped to the Secure AArch32 AMAIR1 register.
The Non-secure and Secure AArch32 AMAIR0 and AMAIR1 registers are RES0.

4.3.57 Auxiliary Memory Attribute Indirection Register, EL2

The processor does not set any IMPLEMENTATION DEFINED attributes in the Auxiliary Memory Attribute
Indirection Register, EL2. AMAIR_EL2 is RES0.
AMAIR_EL2[31:0] is architecturally mapped to the AArch32 HAMAIR0 register.
AMAIR_EL2[63:32] is architecturally mapped to the AArch32 HAMAIR1 register.
The AArch32 HMAIR0 and HAMAIR1 registers are RES0.

4.3.58 L2 Control Register, EL1

The L2CTLR_EL1 characteristics are:
Purpose
Provides IMPLEMENTATION DEFINED control options for the L2 memory system and ECC/parity
support. There is one L2 Control Register for the Cortex-A72 processor.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-168
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Usage constraints
The accessibility to the L2CTLR_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RWaj RWaj RWak RW RW

Note
The L2CTLR_EL1 must be set statically and not dynamically changed.

The L2 Control Register can only be written when the L2 memory system is idle. ARM
recommends that you write to this register after a powerup reset before the MMU is enabled and
before any ACE, CHI, or ACP traffic begins.
If the register must be modified after a powerup reset sequence, you must idle the L2 memory
system with the following sequence:
1. Disable the MMU from each processor followed by an ISB to ensure the MMU disable
operation is complete, then execute a DSB to drain previous memory transactions.
2. Ensure that the system has no outstanding AC channel or CHI RXRSP coherence requests to
the processor.
3. Ensure that the system has no outstanding ACP requests to the processor.
When the L2 is idle, the processor can update the L2 Control Register followed by an ISB. After
the L2 Control Register is updated, you can enable the MMUs and normal ACE or CHI and
ACP traffic can resume.
Configurations
The L2CTLR_EL1 is:
• Common to the Secure and Non-secure states.
• A 32-bit register in AArch64 state.
• Architecturally mapped to the AArch32 L2CTLR register.
Attributes
See the register summary in Table 4-15 AArch64 implementation defined registers
on page 4-86.
The following figure shows the L2CTLR_EL1 bit assignments.

aj Write access if ACTLR_EL3.L2CTLR is 1 and ACTLR_EL2.L2CTLR is 1, or, ACTLR_EL3.L2CTLR is 1 and the Secure SCR.NS is 0.
ak Write access if ACTLR_EL3.L2CTLR is 1.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-169
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

31 30 26 25 24 23 22 21 20 19 14 13 12 11 10 9 8 6 5 4 3 2 0

RES0 RES0 RES0

L2 Data RAM latency

L2 Data RAM setup
L2 Tag RAM latency
L2 Tag RAM setup
L2 Data RAM slice
L2 Tag RAM slice
L2 arbitration slice
Data inline ECC enable, only applies if ECC is enabled
ECC and parity enable
L1 Cache ECC and Parity protection
RES0
Number of processors
L2RSTDISABLE monitor

Figure 4-50 L2CTRL_EL1 bit assignments

The following table shows the L2CTLR_EL1 bit assignments.

Table 4-69 L2CTLR_EL1 bit assignments

Bits Name Function

[31] L2RSTDISABLE Monitors the L2 hardware reset disable signal, L2RSTDISABLE. The values are:
monitor
0 L2 valid RAM contents are reset by hardware.
1 L2 valid RAM contents are not reset by hardware.

This bit is read-only. The primary input L2RSTDISABLE controls the reset value.

[30:26] - Reserved, RES0.

[25:24] Number of processors Number of processors present. These bits are read-only and set to the number of processors
present in the implementation. The values are:
0b00 One processor, CPU0.
0b01 Two processors, CPU0 and CPU1.
0b10 Three processors, CPU0, CPU1, and CPU2.
0b11 Four processors, CPU0, CPU1, CPU2, and CPU3.

[23] L2 Cache protection This bit is read-only and is set if the cluster implementation supports L2 cache ECC protection.
The L2 cache ECC protection is a configurable implementation option in Cortex-A72 cluster. The
values are:
0
L2 cache ECC is not supported.
1
L2 cache ECC is supported.

[22] L1 Cache ECC and Parity This bit is read-only and is set if the processor implementation supports L1 cache ECC and parity
protection protection. The L1 cache ECC and parity protection is a configurable implementation option in
Cortex-A72 processor. The values are:
0 L1 data cache ECC and L1 instruction cache parity is not supported.
1 L1 data cache ECC and L1 instruction cache parity is supported.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-170
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-69 L2CTLR_EL1 bit assignments (continued)

Bits Name Function

[21] ECC and parity enable ECC and parity enable. The values are:
0
Disables ECC and parity. This is the reset value.
1
Enables ECC and parity.
If Cortex-A72 is implemented with L1 Cache ECC and parity protection, L2CTLR[21] can be
programmed to enable or disable both L1 ECC and parity and L2 ECC protection.
If Cortex-A72 is implemented with no L1 Cache ECC and parity protection but with L2 ECC
protection, L2CTLR[21] can be programmed to enable or disable only L2 ECC protection.
If Cortex-A72 is implemented with neither L1 ECC and parity nor L2 ECC protection,
L2CTLR[21] is RAZ/WI.

[20] Data inline ECC enable, Force inline ECC for Instruction Fetch (IF) and Load/Store (LS) read requests that hit the L2
only applies if ECC is cache increasing the L2 hit latency by 2 cycles. Avoids requirement of flushing requests
enabled associated with L2 cache single-bit ECC errors. The possible values are:
0 Performance optimization reducing L2 hit latency by 2 cycles allowing uncorrected data for
IF and LS read requests that hit the L2 cache. This is the reset value.
1 Forward only corrected data for L2 cache hits avoiding flushing request for single-bit ECC
errors.

[19:14] - Reserved, RES0.

[13] L2 arbitration slice L2 arbitration slice. This is a read-only bit that is set if the L2 arbitration slice is present in the
implementation. The values are:
0 L2 arbitration slice is not present.
1 One L2 arbitration slice is present.

[12] L2 Tag RAM slice L2 Tag RAM slice. This is a read-only bit that is set if the Tag RAM slice is present in the
implementation. The values are:
0 L2 Tag RAM slice is not present.
1 One L2 Tag RAM slice is present.

[11:10] L2 Data RAM slice L2 Data RAM slice. These are read-only bits that are set to the number of Data RAM slices
present in the implementation. The values are:
0b00 L2 Data RAM slices are not present.
0b01 One L2 Data RAM slice is present.
0b10 Two L2 Data RAM slices are present.
0b11 Invalid value.

[9] L2 Tag RAM setup L2 Tag RAM setup. The values are:
0 0 cycle. This the reset value.
1 1 cycle.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-171
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-69 L2CTLR_EL1 bit assignments (continued)

Bits Name Function

[8:6] L2 Tag RAM latency L2 Tag RAM latency. The L2 Tag RAM programmable setup and latency bits only affect the L2
Tag RAM. See 7.2.5 Register slice support for large cache sizes on page 7-301 for more
information. The possible values are:
0b000 2 cycles. This is the reset value.
0b001 2 cycles.
0b010 3 cycles.
0b011 4 cycles.
0b1xx 5 cycles.

[5] L2 Data RAM setup L2 Data RAM setup. The values are:
0 0 cycle. This the reset value.
1 1 cycle.

[4:3] - Reserved, RES0.

[2:0] L2 Data RAM latency L2 Data RAM latency.al The L2 Data RAM programmable setup & latency bits affect only the L2
Data RAM. See 7.2.5 Register slice support for large cache sizes on page 7-301 for more
information. The values are:
0b000
2 cycles. This is the reset value.
0b001
2 cycles.
0b010
3 cycles.
0b011
4 cycles.
0b100
5 cycles.
0b101
6 cycles.
0b11x
6 cycles.

To access the L2CTLR_EL1 in AArch64 state, read or write the register with:
MRS <Xt>, S3_1_c11_c0_2; Read L2 Control Register
MSR S3_1_c11_c0_2, <Xt>; Write L2 Control Register

To access the L2CTLR in AArch32 state, read or write the CP15 register with:
MRC p15, 1, <Rt>, c9, c0, 2; Read L2 Control Register
MCR p15, 1, <Rt>, c9, c0, 2; Write L2 Control Register

4.3.59 L2 Extended Control Register, EL1

The L2ECTLR_EL1 characteristics are:

Purpose
Provides additional IMPLEMENTATION DEFINED control options for the L2 memory system. There is
one L2 Extended Control Register for the Cortex-A72 processor.

al Slice and Set-up have priority over programmed latency in determining total adjusted pipeline depth.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-172
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Usage constraints
The accessibility to the L2ECTLR_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RWam RWam RWan RW RW

The L2ECTLR_EL1 can be written dynamically.

Configurations
The L2ECTLR_EL1 is:
• Common to the Secure and Non-secure states.
• A 32-bit register in AArch64 state.
• Architecturally mapped to the AArch32 L2ECTLR register.
Attributes
See the register summary in Table 4-15 AArch64 implementation defined registers
on page 4-86.
The following figure shows the L2ECTLR bit assignments.
31 30 29 28 3 2 0

RES0

AXI or CHI imprecise bus error L2 dynamic retention control

L2 internal asynchronous error
RES0

Figure 4-51 L2ECTLR_EL1 bit assignments

The following table shows the L2ECTLR_EL1 bit assignments.

Table 4-70 L2ECTLR_EL1 bit assignments

Bits Name Function

[31] - Reserved, RES0.

[30] L2 internal L2 internal asynchronous error caused by L2 RAM double-bit ECC error. The possible values are:
asynchronous error
0 No pending asynchronous error. This is the reset value.
1 An asynchronous error has occurred.

A write of 0 clears this bit. A write of 1 is ignored.

[29] AXI or CHI AXI or CHI asynchronous error indication. The possible values are:
asynchronous error
0 No pending asynchronous error. This is the reset value.
1 An asynchronous error has occurred.

A write of 0 clears this bit. A write of 1 is ignored.

am Write access if ACTLR_EL3.L2ECTLR is 1 and ACTLR_EL2.L2ECTLR is 1, or ACTLR_EL3.L2ECTLR is 1 and the Secure SCR.NS is 0.
an Write access if ACTLR_EL3.L2ECTLR is 1.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-173
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-70 L2ECTLR_EL1 bit assignments (continued)

Bits Name Function

[28:3] - Reserved, RES0.

[2:0] L2 dynamic retention L2 dynamic retention control. The possible values are:
control
0b000 L2 dynamic retention disabled. This is the reset value.
0b001 2 Generic Timer ticks required before retention entry.
0b010 8 Generic Timer ticks required before retention entry.
0b011 32 Generic Timer ticks required before retention entry.
0b100 64 Generic Timer ticks required before retention entry.
0b101 128 Generic Timer ticks required before retention entry.
0b110 256 Generic Timer ticks required before retention entry.
0b111 512 Generic Timer ticks required before retention entry.

To access the L2ECTLR_EL1 in AArch32 state, read or write the CP15 register with:
MRS <Xt>, S3_1_c11_c0_3; Read L2 Extended Control Register
MSR S3_1_c11_c0_3, <Xt>; Write L2 Extended Control Register

To access the L2ECTLR in AArch32 state, read or write the CP15 register with:
MRC p15, 1, <Rt>, c9, c0, 3; Read L2 Extended Control Register
MCR p15, 1, <Rt>, c9, c0, 3; Write L2 Extended Control Register

Related information
L2 RAMs dynamic retention on page 2-48.

4.3.60 Reset Vector Base Address, EL3

The RVBAR_EL3 characteristics are:

Purpose
Defines the address that execution starts from after reset when executing in the AArch64 state.
RVBAR_EL3 is part of the reset management registers functional group.
Usage constraints
The accessibility of the RVBAR_EL3 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - - RO RO

Configurations
Only implemented if the highest Exception level implemented is EL3.
Attributes
See the register summary in Table 4-11 AArch64 reset registers on page 4-82.
The following figure shows the RVBAR_EL3 bit assignments.
63 44 43 2 1 0

RES0 Reset Vector Base Address RES0

Figure 4-52 RVBAR_EL3 bit assignments

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-174
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

The following table shows the RVBAR_EL3 bit assignments.

Table 4-71 RVBAR_EL3 bit assignments

Bits Name Function

[63:44] - Reserved, RES0.

[43:2] Reset Vector Base Address Reset Vector Base Address when executing in the AArch64 state. The reset address for processor
n is set by the RVBARADDRn[43:2] input signals.

[1:0] - Reserved, RES0.

To access the RVBAR_EL3 in AArch64 state, read the register with:

MRS <Xt>, RVBAR_EL3; Read RVBAR_EL3 Reset Vector Base Address Register

4.3.61 Reset Management Register, EL3

The RMR_EL3 characteristics are:

Purpose
Controls the Execution state that the processor boots into and allows request of a Warm reset.
Usage constraints
The accessibility to the RMR_EL3 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - - RW RW

Configurations
The RMR_EL3 is
• Common to the Secure and Non-secure states.
• Architecturally mapped to the AArch32 RMR register.
Attributes
Write access to RMR_EL3 is disabled when the CP15SDISABLE signal is HIGH and EL3 is
using AArch32.
See the register summary in Table 4-11 AArch64 reset registers on page 4-82.
The following figure shows the RMR_EL3 bit assignments.
31 2 1 0

RES0

RR
AA64

Figure 4-53 RMR_EL3 bit assignments

The following table shows the RMR_EL3 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-175
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-72 RMR_EL3 bit assignments

Bits Name Function

[31:2] - Reserved, RES0.

[1] RR Reset Request. The values are:

0 This is the reset value. It is set to zero by either a Cold or Warm reset.
1 Requests a Warm reset.

[0] AA64ao Determines the Execution state at processor boot time. The values are:
0 AArch32 state.
1 AArch64 state.

If software requests a Warm reset by setting RR=1 then it can use the AA64 bit to change Execution state.

To access the RMR_EL3 in AArch64 state, read or write the register with:
MRS <Xt>, RMR_EL3; Read EL3 Reset Management Register
MSR RMR_EL3, <Xt>; Write EL3 Reset Management Register

To access the RMR, in AArch32 state, read or write the CP15 register with:
MRC p15, 0, <Rt>, c12, c0, 2; Read Reset Management Register
MCR p15, 0, <Rt>, c12, c0, 2; Write Reset Management Register

4.3.62 Instruction L1 Data n Register, EL1

The IL1DATAn_EL1, where n is from 0 to 3, characteristics are:

Purpose
Holds the instruction side L1 array information returned by the RAMINDEX system operation.
Note
Because all of the I-side arrays are greater than 32-bit wide, the processor contains multiple
IL1DATA registers, to hold the array information.

Usage constraints
The accessibility to the IL1DATAn_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Configurations
The IL1DATAn_EL1 is:
• Common to the Secure and Non-secure states.
• A 32-bit register in AArch64 state.
• Architecturally mapped to the AArch32 IL1DATAn registers.
Attributes
See the register summary in Table 4-15 AArch64 implementation defined registers
on page 4-86.
The following figure shows the IL1DATAn_EL1 bit assignments.

ao For a Cold reset, the value of this bit is set by the AA64nAA32 signal.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-176
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

31 0

Data

Figure 4-54 IL1DATA

The following table shows the IL1DATAn_EL1 bit assignments.

Table 4-73 IL1DATAn_EL1 bit assignments

Bits Name Function

[31:0] Data Holds the instruction side L1 array information

To access the IL1DATAn_EL1 in AArch64 state, read or write the registers with:
MRS <Xt>, s3_0_c15_c0_n; Read EL1 Instruction L1 Data n Register
MSR s3_0_c15_c0_n, <Xt>; Write EL1 Instruction L1 Data n Register

n is 0, 1, 2, or 3 for Opcode2 of IL1DATAn_EL1 registers.

To access the IL1DATAn in AArch32 state, read or write the CP15 registers with:
MRC p15, 0, <Rt>, c15, c0, n; Read Instruction L1 Data n Register
MCR p15, 0, <Rt>, c15, c0, n; Write Instruction L1 Data n Register

n is 0, 1, 2, or 3 for Opcode2 of IL1DATAn registers.

Related information
4.3.64 RAM Index operation on page 4-178.

4.3.63 Data L1 Data n Register, EL1

The DL1DATAn_EL1, where n is from 0 to 4, characteristics are:

Purpose
Holds the data side L1 or L2 array information returned by the RAMINDEX write operation.
Note
Because the Data, Tag, and TLB arrays are greater than 32-bit wide, the processor contains
multiple DL1DATA registers, to hold the array information.

Usage constraints
The accessibility to the DL1DATAn_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Configurations
The DL1DATAn_EL1 is:
• Common to the Secure and Non-secure states.
• A 32-bit register in AArch64 state.
• Architecturally mapped to the AArch32 DL1DATAn registers.
Attributes
See the register summary in Table 4-15 AArch64 implementation defined registers
on page 4-86.
The following figure shows the DL1DATAn_EL1 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-177
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

31 0

Data

Figure 4-55 DL1DATA

The following table shows the DL1DATAn_EL1 bit assignments.

Table 4-74 DL1DATAn_EL1 bit assignments

Bits Name Function

[31:0] Data Holds the data side L1 or L2 array information

To access the DL1DATAn_EL1 in AArch64 state, read or write the registers with:
MRS <Xt>, s3_0_c15_c1_n; Read EL1 Data L1 Data n Register
MSR s3_0_c15_c1_n, <Xt>; Write EL1 Data L1 Data n Register

n is 0, 1, 2, 3, or 4 for Opcode2 of the DL1DATAn_EL1 registers.

To access the DL1DATAn in AArch32 state, read or write the CP15 registers with:
MRC p15, 0, <Rt>, c15, c1, n; Read Data L1 Data n Register
MCR p15, 0, <Rt>, c15, c1, n; Write Data L1 Data n Register

n is 0, 1, 2, 3, or 4 for Opcode2 of the DL1DATAn registers.

Related information
4.3.64 RAM Index operation on page 4-178.

4.3.64 RAM Index operation

The RAMINDEX characteristics are:

Purpose
Read the instruction side L1 array contents into the IL1DATAn register or read the data side L1
or L2 array contents into the DL1DATAn register.
Usage constraints
The accessibility to the RAMINDEX by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- WO WO WO WO WO

Configurations
The RAMINDEX operates in the Secure and Non-secure states.
The RAMINDEX command takes one argument or source register. You must write an ARM
core register with the bit pattern described in the following figure for each RAM listed in the
following table.
A 32-bit register in AArch64 state.
Attributes
See the register summary in Table 4-15 AArch64 implementation defined registers
on page 4-86.
The following figure shows the RAMINDEX bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-178
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

31 24 23 22 21 18 17 0

RAMID RES0 Way Index

Figure 4-56 RAMINDEX bit assignments

The following table shows the RAMINDEX bit assignments.

Table 4-75 RAMINDEX bit assignments

Bits Name Function

[31:24] RAMID RAM identifier. This field indicates which RAM is being accessed. The possible values are ap:
0x00 L1-I Tag RAM, see L1-I Tag RAM on page 4-180.
0x01 L1-I Data RAM, see L1-I Data RAM on page 4-180.
0x02 L1-I BTB RAM, see L1-I BTB RAM on page 4-181.
0x03 L1-I GHB RAM, see L1-I GHB RAM on page 4-181.
0x04 L1-I TLB array, see L1-I TLB array on page 4-181.
0x05 L1-I indirect predictor RAM, see L1-I indirect predictor RAM on page 4-182.
0x08 L1-D Tag RAM, see L1-D Tag RAM on page 4-183.
0x09 L1-D Data RAM, see L1-D Data RAM on page 4-183.
0x0A L1-D TLB array, see L1-D TLB array on page 4-183.
0x10 L2 Tag RAM, see L2 Tag RAM on page 4-184.
0x11 L2 Data RAM, see L2 Data RAM on page 4-185.
0x12 L2 Snoop Tag RAM, see L2 Snoop Tag RAM on page 4-186.
0x13 L2 Data ECC RAM, see L2 Data ECC RAM on page 4-186.
0x14 L2 Dirty RAM, see L2 Dirty RAM on page 4-187.
0x18 L2 TLB RAM, see L2 TLB RAM on page 4-187.

All other values are reserved.

[23:22] - Reserved, RES0.

[21:18] Way Indicates the way of the RAM that is being accessed.

[17:0] Index Indicates the index address of the RAM that is being accessed.

Note
• Executing a RAMINDEX operation with a reserved value of RAMID, Way, or Index results in the
corruption of the IL1DATAn or DL1DATAn register contents.
• In Non-secure EL1 and EL2, the RAMINDEX operation returns the contents of the RAM only if the
entry is marked valid and Non-secure. Entries that are marked invalid or Secure update the
IL1DATAn or DL1DATAn registers with 0x0 values.
• In Secure EL1 or EL3, the RAMINDEX operation returns the contents of the RAM, regardless of
whether the entry is marked valid or invalid, and Secure or Non-secure.
• When the RAMID field is set to L1-I BTB RAM in Non-secure EL1 and EL2, the RAMINDEX
operation always returns zero.
• The L1-I, L1-D, L2 TLB, and L2 Snoop Tag RAMs can only be accessed by the processor where the
RAM resides or that owns the RAM.
• The L2 Tag, Data, and Dirty RAMs can be accessed by any processor.

ap All other values reserved.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-179
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

L1-I Tag RAM

The following figure shows the RAMINDEX register bit assignments for accessing L1-I Tag RAM.
31 24 23 20 19 18 17 14 13 6 5 0

RAMID = 0x00 Reserved Way Reserved Virtual address [13:6] Reserved

Figure 4-57 RAMINDEX bit assignments for L1-I Tag RAM

The RAMINDEX address bits for accessing L1-I Tag RAM are:
Way[1:0] Way select.
Note
The instruction cache is 3-way set-associative. Setting the way field to a value of 3, reads
way 2 of the cache.

VA[13:7] Row select.

VA[6] Bank select.

The data returned from accessing L1-I Tag RAM are:

ILDATA1[1] Valid bit.
ILDATA1[0] Non-secure identifier for the physical address.
ILDATA0 Physical address tag [43:12].

L1-I Data RAM

The following figure shows the RAMINDEX bit assignments for accessing L1-I Data RAM.
31 24 23 20 19 18 17 14 13 3 2 0

RAMID = 0x01 Reserved Way Reserved Virtual address [13:3]

Reserved

Figure 4-58 RAMINDEX bit assignments for L1-I Data RAM

The RAMINDEX address bits for accessing L1-I Data RAM are:
Way[1:0] Way select.
Note
The instruction cache is 3-way set-associative. Setting the Way field to 3, reads way 2 of the
cache.

VA[13:6] Set select.

VA[5:4] Bank select.
VA[3] Upper or lower doubleword within the quadword.

The data returned from accessing L1-I Data RAM are:

ILDATA1[31:0] Data word 1.
ILDATA0[31:0] Data word 0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-180
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

L1-I BTB RAM

The following figure shows the RAMINDEX bit assignments for accessing L1-I BTB RAM.
31 24 23 15 14 4 3 0

RAMID = 0x02 Reserved Virtual address [14:4] Reserved

Figure 4-59 RAMINDEX bit assignments for L1-I BTB RAM

The RAMINDEX address bits for accessing L1-I BTB RAM are:
VA[14:6] Row select.
VA[5:4] Bank select.

ARM does not disclose the format of the returned data.

L1-I GHB RAM

The following figure shows the RAMINDEX bit assignments for accessing L1-I GHB RAM.

31 24 23 16 15 4 3 0

RAMID = 0x03 Reserved Index[15:4] Reserved

Figure 4-60 RAMINDEX bit assignments for L1-I GHB RAM

The RAMINDEX address bits for accessing L1-I GHB RAM are:
Index[15:5]
Row select.
Index[4]
Bank select.
ARM does not disclose the format of the returned data.

L1-I TLB array

The following figure shows the RAMINDEX bit assignments for accessing L1-I TLB array.
31 24 23 6 5 0

RAMID = 0x04 Reserved TLB entry

Figure 4-61 RAMINDEX bit assignments for L1-I TLB array

The RAMINDEX address bits for accessing L1-I TLB array are:
TLB entry Selects one of the 48 entries.

The data returned from accessing L1-I TLB array are:

ILDATA3[27] Valid bit.
ILDATA3[26:25] Shareability attribute:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-181
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

0b00
Non-Shareable.
0b01
Reserved.
0b10
Outer Shareable.
0b11
Inner Shareable.
ILDATA3[15:14] VA memory space ID:
0b00
Secure EL1.
0b01
EL3, AArch64 only.
0b10
Non-secure EL1.
0b11
Non-secure EL2.
ILDATA3[13:6] Virtual Machine ID (VMID).
{ILDATA3[5:0], ILDATA2[31:22]} Address Space ID (ASID).
ILDATA2[21:14] Memory Attribute Indirection Register.
ILDATA2[11:10] Page size:
0b00
4KB.
0b01
64KB.
0b10
1MB.
0b11
Reserved.
ILDATA2[9:6] Domain ID.
ILDATA2[5] Non-secure identifier for the physical address.
{ILDATA2[4:0], ILDATA1[31:5]} Physical address [43:12].
{ILDATA1[4:0], ILDATA0[31:0]} Virtual address [48:12].

L1-I indirect predictor RAM

The following figure shows the RAMINDEX bit assignments for accessing L1-I indirect predictor RAM.
31 24 23 19 18 17 8 7 0

RAMID = 0x05 Reserved Reserved Index[7:0]

Way

Figure 4-62 RAMINDEX bit assignments for L1-I indirect predictor RAM

The RAMINDEX address bits for accessing L1-I indirect predictor RAM are:
Way Way select.
Index[7:0] Indirect predictor entry.

ARM does not disclose the format of the returned data.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-182
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

L1-D Tag RAM

The following figure shows the RAMINDEX bit assignments for accessing L1-D Tag RAM.
31 24 23 19 18 17 14 13 6 5 0

RAMID = 0x08 Reserved Reserved Physical address [13:6] Reserved

Way

Figure 4-63 RAMINDEX bit assignments for L1-D Tag RAM

The RAMINDEX address bits for accessing L1-D Tag RAM are:
Way Way select.
PA[13:8] Row select.
PA[7:6] Bank select.

The data returned from accessing L1-D Tag RAM are:

DL1DATA1[1:0] MESI state:
0b00 Invalid.
0b01 Exclusive.
0b10 Shared.
0b11 Modified.

DL1DATA0[30] Non-secure identifier for the physical address.

DL1DATA0[29:0] Physical address tag [43:14].

L1-D Data RAM

The following figure shows the RAMINDEX bit assignments for accessing L1-D Data RAM.
31 24 23 19 18 17 14 13 3 2 0

RAMID = 0x09 Reserved Unused Physical address [13:3]

Way Reserved

Figure 4-64 RAMINDEX bit assignments for L1-D Data RAM

The RAMINDEX address bits for accessing L1-D Data RAM are:
Way Way select.
PA[13:6] Set select.
PA[5:4] Bank select.
PA[3] Upper or lower doubleword within the quadword.

The data returned from accessing L1-D Data RAM are:

DL1DATA1[31:0] Data word 1.
DL1DATA0[31:0] Data word 0.

L1-D TLB array

The following figure shows the RAMINDEX bit assignments for accessing L1-D TLB array.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-183
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

31 24 23 5 4 0

RAMID = 0x0A Reserved TLB entry

Figure 4-65 RAMINDEX bit assignments for L1-D TLB array

The RAMINDEX address bits for accessing L1-D TLB array are:
TLB entry Selects one of the 32 entries.

The data returned from accessing L1-D TLB array are:

DL1DATA3[12] Valid bit.
DL1DATA3[11:10] VA memory space ID:
0b00
Secure EL1.
0b01
EL3, AArch64 only.
0b10
Non-secure EL1.
0b11
Non-secure EL2.
DL1DATA3[1:0] Shareability attribute:
0b00
Non-Shareable.
0b01
Reserved.
0b10
Outer Shareable.
0b11
Inner Shareable.
DL1DATA2[31:24] Memory Attribute Indirection Register.
DL1DATA2[23:22] Page size:
0b00
4KB.
0b01
64KB.
0b10
1MB.
0b11
Reserved.
DL1DATA2[21:18] Domain ID.
DL1DATA2[5] Non-secure identifier for the physical address.
{DL1DATA2[4:0], DL1DATA1[31:5]} Physical address [43:12].
{DL1DATA1[4:0], DL1DATA0[31:0]} Virtual address [48:12].

L2 Tag RAM

The following figure shows the RAMINDEX bit assignments for accessing L2 Tag RAM.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-184
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

31 24 23 22 21 18 17 16 6 5 0

RAMID = 0x10 Way Physical address [16:6] Reserved

Reserved Reserved

Figure 4-66 RAMINDEX bit assignments for L2 Tag RAM

The RAMINDEX address bits for accessing L2 Tag RAM are:

Way[3:0] Way select.
PA[16:7] Row select.
PA[6] Tag bank select.

The data returned from accessing L2 Tag RAM are:

DL1DATA0[31] Non-secure identifier for the physical address.
DL1DATA0[30:2] Physical address tag [43:15].
DL1DATA0[1:0] MOESI state:
0b00
Invalid.
0b01
Exclusive or Modified.
0b10
Reserved.
0b11
Shared or Owned.

Note
The Dirty bit in the L2 Dirty RAM must be used to differentiate between the
Exclusive, Modified, Shared, and Owned states.

L2 Data RAM

The following figure shows the RAMINDEX bit assignments for accessing L2 Data RAM.
31 24 23 22 21 18 17 16 4 3 0

RAMID = 0x11 Way Physical address [16:4] Reserved

Reserved Reserved

Figure 4-67 RAMINDEX bit assignments for L2 Data RAM

The RAMINDEX address bits for accessing L2 Data RAM are:

Way[3:0] Way select.
PA[16:7] Row select.
PA[6] Tag bank select.
PA[5:4] Data bank select.

The data returned from accessing L2 Data RAM are:

DL1DATA3 Data[127:96].
DL1DATA2 Data[95:64].
DL1DATA1 Data[63:32].

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-185
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

DL1DATA0 Data[31:0].

L2 Snoop Tag RAM

The following figure shows the RAMINDEX bit assignments for accessing L2 Snoop Tag RAM.
31 24 23 21 20 19 18 17 14 13 6 5 0
CPU
RAMID = 0x12 Reserved Physical address [13:6] Reserved
ID

Way
Reserved

Figure 4-68 RAMINDEX bit assignments for L2 Snoop Tag RAM

The RAMINDEX address bits for accessing L2 Snoop Tag RAM are:
CPUID[1:0] Processor ID of the executing processor that has access to the L2 Snoop Tag RAM.
Way Way select.
PA[13:7] Row select.
PA[6] Bank select.

The data returned from accessing L2 Snoop Tag RAM are:

DL1DATA1[0] Non-secure identifier for the physical address.
DL1DATA0[31:2] Physical address tag [43:14].
DL1DATA0[1:0] MESI state:
0b00
Invalid.
0b01
Exclusive or Modified.
0b10
Reserved.
0b11
Shared.

L2 Data ECC RAM

The following figure shows the RAMINDEX bit assignments for accessing L2 Data ECC RAM.
31 24 23 22 21 18 17 16 4 3 0

RAMID = 0x13 Way Physical address [16:4] Reserved

Reserved Reserved

Figure 4-69 RAMINDEX bit assignments for L2 Data ECC RAM

The RAMINDEX address bits for accessing L2 Data ECC RAM are:
Way[3:0] Way select.
PA[16:7] Row select.
PA[6] Tag bank select.
PA[5:4] Data bank select.

ARM does not disclose the format of the returned data.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-186
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

L2 Dirty RAM

The following figure shows the RAMINDEX bit assignments for accessing L2 Dirty RAM.
31 24 23 22 21 18 17 16 6 5 0

RAMID = 0x14 Way Physical address [16:6] Reserved

Reserved Reserved

Figure 4-70 RAMINDEX bit assignments for L2 Dirty RAM

The RAMINDEX address bits for accessing L2 Dirty RAM are:

Way[3:0] Way select.
PA[16:7] Row select.
PA[6] Tag bank select.

The data returned from accessing L2 Dirty RAM are:

DL1DATA0[7] Outer Shareable page attribute.
DL1DATA0[6] Read Allocate page attribute.
DL1DATA0[5] Write Allocate page attribute.
DL1DATA0[4] Inner Shareable page attribute.
DL1DATA0[0] Dirty bit indicator.

L2 TLB RAM

The following figure shows the RAMINDEX bit assignments for accessing L2 TLB RAM.
31 24 23 20 19 18 17 8 7 0

RAMID = 0x18 Reserved Way Reserved TLB entry

Figure 4-71 RAMINDEX bit assignments for L2 TLB RAM

The RAMINDEX address bits for accessing L2 TLB RAM are:

Way Way select.
TLB entry Selects one of the 256 entries in each way.

The data returned from accessing L2 TLB RAM are:

DL1DATA3[31] Valid bit for EL3 AArch64 only.
DL1DATA3[30] Valid bit for EL2.
DL1DATA3[29] Valid bit for Secure EL1.
DL1DATA3[28] Valid bit for Non-secure EL1.

Note
Only a single bit in DLDATA3[31:28] is set to 1.

DL1DATA3[27:20] VMID.
DL1DATA3[19:4] ASID.
{DL1DATA3[3:0], Virtual address [48:19].
DL1DATA2[31:6]}

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-187
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

DL1DATA2[5] Non-secure identifier for the physical address.

{DL1DATA2[4:0], Physical address [43:12].
DL1DATA1[31:5]}
{DL1DATA1[1:0], Fully resolved page size:
DL1DATA0[31]} 0b000 4KB.
0b001 64KB.
0b010 1MB.
0b011 2MB.
0b100 16MB.
0b101 512MB.
0b110 1GB.

DL1DATA0[13:10] Domain ID.

DL1DATA0[9:8] Shareability attribute:
0b00 Non-Shareable.
0b01 Reserved.
0b10 Outer Shareable.
0b11 Inner Shareable.

DL1DATA0[7:0] Memory Attribute Indirection Register.

For example, to read an entry in the instruction side TLB in AArch64 state:
LDR X0, =0x0000000001000D80
SYS #0, c15, c4, #0, X0
DSB SY
ISB
MRS X1, S3_0_c15_c0_0 ; Move ILData0 register to X1
MRS X2, S3_0_c15_c0_1 ; Move ILData1 register to X2
MRS X3, S3_0_c15_c0_2 ; Move ILData2 register to X3
MRS X4, S3_0_c15_c0_3 ; Move ILData3 register to X4

To complete the RAMINDEX operation in AArch64 state, use the following instruction:
SYS #0, c15, c4, #0, X0 ; Execute RAMINDEX operation

For example, to read one entry in the instruction side L1 data array in AArch32 state:
LDR R0, =0x01000D80;
MCR p15, 0, R0, c15, c4, 0; Read I-L1 TLB data into IL1DATA0-2
DSB
ISB
MRC p15, 0, R1, c15, c0, 0; Move IL1DATA0 Register to R1
MRC p15, 0, R2, c15, c0, 1; Move IL1DATA1 Register to R2
MRC p15, 0, R3, c15, c0, 2; Move IL1DATA2 Register to R3

To complete the RAMINDEX operation in AArch32 state, use the following instruction:
MCR p15, 0, <Rt>, c15, c4, 0; Execute RAMINDEX operation

4.3.65 L2 Auxiliary Control Register, EL1

The L2ACTLR_EL1 characteristics are:
Purpose Provides IMPLEMENTATION DEFINED configuration and control options for the L2 memory
system. There is one L2 Auxiliary Control Register for the Cortex-A72 processor.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-188
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Usage The accessibility to the L2ACTLR by Exception level is:

constraints
EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RWaq RWaq RWar RW RW

Note
The L2ACTLR_EL1 must be set statically and not dynamically changed.

The L2 Auxiliary Control Register can only be written when the L2 memory system is
idle. ARM recommends that you write to this register after a powerup reset, before the
MMU is enabled, and before any ACE, CHI, or ACP traffic begins.
If the register must be modified after a powerup reset sequence, you must to idle the
L2 memory system with the following sequence:
1. Disable the MMU from each core followed by an ISB to ensure the MMU disable
operation is complete, then execute a DSB to drain previous memory transactions.
2. Ensure that the system has no outstanding ACE AC channel or CHI RXRSP
coherence requests to the Cortex-A72 processor.
3. Ensure that the system has no outstanding ACP requests to the Cortex-A72
processor.
When the L2 is idle, the processor can update the L2 Auxiliary Control Register
followed by an ISB. After the L2 Auxiliary Control Register is updated, you can
enable the MMUs and normal ACE or CHI and ACP traffic can resume.
Configurations The L2ACTLR_EL1 is:
• Common to the Secure and Non-secure states.
• A 32 bit register in AArch64 state.
• Architecturally mapped to the AArch32 L2ACTLR register.
Attributes See the register summary in Table 4-15 AArch64 implementation defined registers
on page 4-86.

The following figure shows the L2ACTLR_EL1 bit assignments.

aq Write access if ACTLR_EL3.L2ACTLR is 1 and ACTLR_EL2.L2ACTLR is 1, or ACTLR_EL3.L2ACTLR is 1 and the Secure SCR.NS is 0.
ar Write access if ACTLR_EL3.L2ACTLR is 1.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-189
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

L2 PLRU insertion point

L2 replacement policy
Force L2 tag banks clock enable active
Force L2 logic clock enable active
Force L2, GIC, Timer RCG enables active
Enable single issue across all tbnks when
L2 arbitration replay threshold is reached
L2 PLRU mode
Disable ACP MakeUnique and WriteLineUnique transactions

Disable dynamic throttling of LS prefetch requests

RES0

Disable limit on NC/SO/Dev stores in Address Sequence Queue (ASQ)

Disable L2 round-robin arbitration that only clocks through
paths with an active requestor waiting to be arbitrated
Enable replay threshold single issue
Disable fast forwarding of data from ACE or CHI to LS and IF
Enable UniqueClean evictions with data
Disable clean evict optimization
Disable prefetch set hazard optimization
Disable DSB with no DVM-sync
Disable Non-secure debug array access
Disable set/way hazard optimization on back to back read requests from same CPU
Disable DVM/CMO message broadcast
Enable hazard-detect timeout
Disable ACE shareable or CHI snoopable transactions from master
Disable set/way hazard optimization for WBNA/WT memory
Disable WriteUnique and WriteLineUnique transactions from master
Disable clean/evict push to external
Limit to one request per tag bank
Enable arbitration replay threshold timeout
Disable hardware prefetch forwarding

Figure 4-72 L2ACTLR_EL1 bit assignments

The following table shows the L2ACTLR_EL1 bit assignments.

Table 4-76 L2ACTLR_EL1 bit assignments

Bits Name Function

[31:30]as L2 PLRU insertion point Selects the L2 PLRU insertion point:

00 Insert new lines midway between MRU and LRU positions. This is the reset value.
01 Insert new lines at MRU position.
10 Insert new lines at 3/4 LRU position.
11 Insert new lines at LRU position.

[29]as L2 replacement policy Selects the L2 cache replacement policy:

0 Pseudo Least Recently Used (PLRU). This is the reset value.
1 Pseudo random

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-190
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-76 L2ACTLR_EL1 bit assignments (continued)

Bits Name Function

[28]as Force L2 tag bank clock enable Forces L2 tag bank clock enable active:
active
0 Does not prevent the clock generator from stopping the L2 tag bank clock. This is
the reset value.
1 Prevents the clock generator from stopping the L2 tag bank clock.

This bit applies to each of the two L2 cache tag bank clocks.
See L2 control and tag banks clock gating on page 2-49.
If the L2 dynamic retention feature is used then this bit must be zero. See L2 RAMs
dynamic retention on page 2-48.

[27]as Force L2 logic clock enable Forces L2 logic clock enable active:
active
0 Does not prevent the clock generator from stopping the L2 logic clock. This is the
reset value.
1 Prevents the clock generator from stopping the L2 logic clock.

See L2 control and tag banks clock gating on page 2-49.

If the L2 dynamic retention feature is used then this bit must be zero. See L2 RAMs
dynamic retention on page 2-48.

[26]as Force L2, GIC, Timer RCG Forces L2, GIC CPU interface, and Timer Regional Clock Gate (RCG) enables active:
enables active
0 Enables L2, GIC CPU interface, and Timer RCGs for additional clock gating and
potentially reduce dynamic power dissipation. This is the reset value.
1 Forces L2, GIC CPU interface, and Timer RCG enables HIGH.

Setting this bit to 1 has no effect if the processor is configured to not include RCGs. See
Regional clock gating on page 2-50.

[25]as Enable single issue across all Enables single issue across all tag banks when the L2 arbitration replay threshold is
tbnks when L2 arbitration replay reached, so that only one request can be active across both tag banks at any given time:
threshold is reached
0 Disables single issue across the tag banks when the L2 arbitration replay threshold is
reached. This is the reset value.
1 Enables single issue across the tag banks when the L2 arbitration replay threshold is
reached.

[24]as L2 PLRU mode Disables PLRU dynamic insertion and update policy:
0 Enables PLRU dynamic insertion and update policy. This is the reset value.
1 Disables PLRU dynamic insertion and update policy.

[23]as Disable ACP MakeUnique and Disables ACP MakeUnique and WriteLineUnique transactions:
WriteLineUnique transactions
0 Enables MakeUnique and WriteLineUnique transactions for full cache line ACP
writes with all byte enables asserted. This is the reset value.
1 Disables MakeUnique and WriteLineUnique transactions for full cache line ACP
writes with all byte enables asserted.

as This bit is provided for debugging and characterization purpose only. For normal operation, ARM recommends that you do not change the value of this bit from its
reset value.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-191
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-76 L2ACTLR_EL1 bit assignments (continued)

Bits Name Function

[22]as Disable dynamic throttling of Disables dynamic throttling of load/store prefetch requests:
load/store prefetch requests
0 Enables dynamic throttling of load/store prefetch requests. This is the reset value.
1 Disables dynamic throttling of load/store prefetch requests.

[21:20] - Reserved, RES0.

[19:18] Disable limit on NC/SO/Dev Disables limit on NC/SO/Dev stores in ASQ:

stores in Address Sequence
00 NC/SO/Dev stores limited to 12 entries in the ASQ. This is the reset value.
Queue (ASQ)
01 NC/SO/Dev stores limited to 10 entries in the ASQ.
10 NC/SO/Dev stores limited to 8 entries in the ASQ.
11 There is no limit on NC/SO/Dev stores in the ASQ.

[17]as Disable L2 round-robin Disable L2 round-robin arbitration that only clocks through paths with an active
arbitration that only clocks requestor waiting to be arbitrated:
through paths with an active
0 Enables L2 round-robin arbitration that only clocks through paths with an active
requestor waiting to be arbitrated
requestor waiting to be arbitrated. This is the reset value.
1 Disables L2 round-robin arbitration that only clocks through paths with an active
requestor waiting to be arbitrated.

[16]as Enable replay threshold single Enables replay threshold single issue:
issue
0 Disables replay threshold single issue. This is the reset value.
1 Enables replay threshold single issue. If there are 32 consecutive transactions on a
tag bank replay, then single issue is forced until a transaction successfully passes
hazard checking.

[15]as Disable fast forwarding of data Disables fast forwarding of data from ACE or CHI to LS and IF:
from ACE or CHI to LS and IF
0 Enables fast forwarding of data from ACE or CHI to LS and IF. This is the reset
value.
1 Disables fast forwarding of data from ACE or CHI to LS and IF.

[14] Enable UniqueClean evictions Enables UniqueClean evictions with data:

with data
0 Disables UniqueClean evictions with data. This is the reset value if the
multiprocessor implements the ACE interface.
1 Enables UniqueClean evictions with data. This is the reset value if the
multiprocessor implements the CHI interface.

[13]as Disable clean evict optimization Disables clean evict optimization:

0 Enables clean evict optimization. This is the reset value.
1 Disables clean evict optimization.

[12] Disable prefetch set hazard Disables set hazard optimization against prefetch entries.
optimization
0 Enables set hazard optimization, preventing a set hazard for prefetch entries. This is
the reset value.
1 Disables set hazard optimization against prefetch entries.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-192
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-76 L2ACTLR_EL1 bit assignments (continued)

Bits Name Function

[11]as Disable DSB with no DVM Disables Data Synchronization Barrier (DSB) with no Distributed Virtual Memory
synchronization (DVM) synchronization:
0 Enables DSB with no DVM synchronization. This is the reset value.
A DSB does not cause a DVM Sync message to occur. However, if a TLB
maintenance operation, cache maintenance operation, or branch predictor
maintenance operation occurs after the previous DSB then a DVM Sync message is
generated regardless of the setting of this bit.
1 Disables DSB with no DVM synchronization. Therefore, a DSB always causes a
DVM Sync message to occur.

[10] Disable Non-secure debug array Disables Non-secure debug array read:
read
0 Enables Non-secure debug array read access to Non-secure memory. This is the
reset value.
1 Disables Non-secure debug array read access.

[9] Disable set/way hazard Disables set/way hazard optimization on back to back reads from the same CPU
optimization on back to back targeting the same set:
read requests from same CPU
0 Enables set/way hazard optimization. This is the reset value.
1 Disables set/way hazard optimization.

[8]as Disable DVM and cache Disables DVM transactions and cache maintenance operation message broadcast:
maintenance operation message
0 Enables DVM and cache maintenance operation message broadcast. This is the reset
broadcast
value.
1 Disables DVM and cache maintenance operation message broadcast.

[7]as Enable hazard detect timeout Enables hazard detect timeout:

0 Disables hazard detect timeout. This is the reset value.
1 Enables hazard detect timeout.

[6]as Disable ACE shareable or CHI Disables shareable or snoopable transactions from master:
snoopable transactions from
0 Enables ACE shareable or CHI snoopable transactions from master. This is the reset
master
value.
1 Disables ACE shareable or CHI snoopable transactions from master.

[5]as Disable set/way hazard Disables set/way hazard optimization for WBNA/WT memory:
optimization for WBNA/WT
0 Enables optimization removing set/way hazard for WBNA/WT memory. This is the
memory
reset value.
1 Disables optimization for forcing set/way hazard for WBNA/WT memory.

[4] Disable WriteUnique and Disables WriteUnique and WriteLineUnique transactions from master:
WriteLineUnique transactions
0 Enables WriteUnique and WriteLineUnique transactions from master.
from master
1 Disables WriteUnique and WriteLineUnique transactions from master. This is the
reset value.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-193
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-76 L2ACTLR_EL1 bit assignments (continued)

Bits Name Function

[3] Disable clean/evict push to Disables clean/evict push to external:

external
0 Enables clean/evict to be pushed out to external. This is the reset value if the
processor implements the ACE interface.
1 Disables clean/evict from being pushed to external. This is the reset value if the
processor implements the CHI interface.

[2]as Limit to one request per tag bank Limit to one request per tag bank:
0 Normal behavior permitting parallel requests to the tag banks. This is the reset
value.
1 Limits to one request per tag bank.

[1]as Enable arbitration replay Enables arbitration replay threshold timeout:

threshold timeout
0 Disables arbitration replay threshold timeout. This is the reset value.
1 Enables arbitration replay threshold timeout.

[0]as Disable hardware prefetch Disables hardware prefetch forwarding:

forwarding
0 Enables hardware prefetch forwarding. This is the reset value.
1 Disables hardware prefetch forwarding.

To access the L2ACTLR_EL1 in AArch64 state, read or write the register with:
MRS <Xt>, s3_1_c15_c0_0; Read EL1 L2 Auxiliary Control Register
MSR s3_1_c15_c0_0, <Xt>; Write EL1 L2 Auxiliary Control Register

To access the L2ACTLR in AArch32 state, read or write the CP15 register with:
MRC p15, 1, <Rt>, c15, c0, 0; Read L2 Auxiliary Control Register
MCR p15, 1, <Rt>, c15, c0, 0; Write L2 Auxiliary Control Register

4.3.66 CPU Auxiliary Control Register, EL1

The CPUACTLR_EL1 characteristics are:

Purpose
Provides IMPLEMENTATION DEFINED configuration and control options for the processor. There is
one 64-bit CPU Auxiliary Control Register for each core in the cluster.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-194
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Usage constraints
The accessibility to the CPUACTLR_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RWat RWat RWau RW RW

The CPU Auxiliary Control Register can only be written when the system is idle. ARM
recommends that you write to this register after a powerup reset, before the MMU is enabled,
and before any ACE or ACP traffic begins.
Note
Setting many of these bits can cause significantly lower performance on your code. Therefore, it
is suggested that you do not modify this register unless directed by ARM.

Configurations
CPUACTLR_EL1 is:
• Common to the Secure and Non-secure states.
• A 64-bit read/write register.
• Architecturally mapped to the AArch32 CPUACTLR register.
Attributes
See the register summary in Table 4-15 AArch64 implementation defined registers
on page 4-86.
The following figure shows the CPUACTLR_EL1[63:32] bit assignments.

at Write access if ACTLR_EL3.CPUACTLR is 1 and ACTLR_EL2.CPUACTLR is 1, or ACTLR_EL3.CPUACTLR is 1 and SCR.NS is 0.

au Write access if ACTLR_EL3.CPUACTLR is 1.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-195
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

63 62 61 60 59 58 57 56 55 54 53 52 51 50 49 48 47 46 45 44 43 42 41 40 39 38 37 36 35 34 33 32

Disable Instruction Cache prefetch

Disable main prediction suppression
at target fetch of microBTB
Disable static branch predictor
Disable ISB optimization
Force implicit DSB on an ISB
Disable instruction group split
Force FPSCR write flush
Disable instruction merging
RES0
Enable snoop hazard while waiting for the second half of atomic exclusive pair
Disable prefetch requests from ReadUnique transactions
Disable VA based hardware prefetch
Execute Data Cache clean as Data Cache clean/invalidate
Disable L1 Data Cache way tracker
Disable multiple outstanding L1 Data TLB misses and L2 TLB hit under miss
Disable D-side L1/L2 hardware prefetch across 4KB page boundary even if page is 64KB or larger
Disable early speculative read access from LS to L2
Disable non-allocate hint of Write-Back No-Allocate memory type
Disable store streaming on NC/GRE memory type
Disable contention detection and fast exclusive monitor path
Disable over-read from LDNP instruction
Treat DMB/DSB as if their domain field is SY
Treat GRE/nGRE as nGnRE
Disable load pass store
Disable Load-store hardware prefetcher
Treat DMB st/st and ld/all as DMB all/all. Treat DSB st/st and ld/all as DSB all/all
Disable DMB nullification
Disable load pass DMB
Force processor Dispatch idle RCG enables active
Force processor Decode and Integer Execute idle RCG enables active
Force processor non-memory-system RCG enables active
Force processor memory-system RCG enables active

Figure 4-73 CPUACTLR_EL1[63:32] bit assignments

The following table shows the CPUACTLR_EL1[63:32] bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-196
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-77 CPUACTLR_EL1[63:32] bit assignments

Bits Name Function

[63]av Force processor memory- Forces processor memory-system RCG enables active:
system RCG enables active
0
Enables the processor memory-system RCGs for additional clock gating and
potentially reduce dynamic power dissipation. This is the reset value.
1
Forces the processor memory-system RCG enables HIGH.
Setting this bit to 1 has no effect if the processor is configured to not include RCGs. See
Regional clock gating on page 2-50.

[62]av Force processor non- Forces processor non-memory-system RCG enables active:
memory-system RCG
0
enables active
Enables the processor non-memory-system RCGs for additional clock gating and
potentially reduce dynamic power dissipation. This is the reset value.
1
Forces the processor non-memory-system RCG enables HIGH.
Setting this bit to 1 has no effect if the processor is configured to not include RCGs. See
Regional clock gating on page 2-50.

[61]av Force processor Decode and Forces processor Decode and Integer Execute idle RCG enables active:
Integer Execute idle RCG
0
enables active
Enables the processor Decode and Integer Execute idle RCGs for additional clock
gating and potentially reduce dynamic power dissipation. This is the reset value.
1
Forces the processor Decode and Integer Execute idle RCG enables HIGH.
Setting this bit to 1 has no effect if the processor is configured to not include RCGs. See
Regional clock gating on page 2-50.

[60]av Force processor Dispatch Forces processor Dispatch idle RCG enables active:
idle RCG enables active
0
Enables the processor Dispatch idle RCGs for additional clock gating and potentially
reduce dynamic power dissipation. This is the reset value.
1
Forces the processor Dispatch idle RCG enables HIGH.
Setting this bit to 1 has no effect if the processor is configured to not include RCGs. See
Regional clock gating on page 2-50.

[59]av Disable load pass DMB Disables load pass DMB. This does not include the implicit barrier from Load-Acquire and Load-
Acquire Exclusive. The possible values are:
0
Enables load pass DMB. This is the reset value.
1
Disables load pass DMB.

av This bit is used internally for debugging and characterization purposes only. For normal operation, ARM recommends that you do not change the value of this bit
from its reset value.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-197
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-77 CPUACTLR_EL1[63:32] bit assignments (continued)

Bits Name Function

[58]av Disable DMB nullification Disables DMB nullification. This includes the implicit barrier from Store-Release and Store-
Release Exclusive:
0
Enables DMB nullification. This is the reset value.
1
Disables DMB nullification.

[57]av Treat DMB st/st and DMB Treats DMB st/st and DMB ld/all as DMB all/all. Treat DSB st/st and DSB ld/all
ld/all as DMB all/all. as DSB all/all. This does not include the implicit barrier from Load-Acquire/Store-Release.
The possible values are:
Treat DSB st/st and DSB
ld/all as DSB all/all. 0
Normal behavior. This the reset value.
1
• Treat DMB st/st and DMB ld/all as DMB all/all.
• Treat DSB st/st and DSB ld/all as DSB all/all.

[56]av Disable Load–store Disables Load–store hardware prefetcher:

hardware prefetcher
0
Enables Load–store hardware prefetcher. This the reset value.
1
Disables Load–store hardware prefetcher.

[55]av Disable load pass store Disables load pass store:

0
Enables load pass store. This the reset value.
1
Disables load pass store.

[54]av Treat GRE/nGRE as nGnRE Treat GRE and nGRE as nGnRE:

0
Enables optimization for GRE and nGRE load/store. This is the reset value.
1
Treats GRE and nGRE as nGnRE. Disables optimization for GRE and nGRE load/
store.

[53]av Treat DMB and DSB as if Treats DMB and DSB as if their domain field is SY. The possible values are:
their domain field is SY
0
Normal behavior. This is the reset value.
1
Treat DMB NSH, DMB ISH, and DMB OSH as DMB SY.
Treat DSB NSH, DSB ISH, and DSB OSH as DSB SY.

[52]av Disable over-read from Disables over-read from LDNP instruction:

LDNP instruction
0
Enables the over-read from LDNP instruction. This is the reset value.
1
Disables the over-read from LDNP instruction.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-198
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-77 CPUACTLR_EL1[63:32] bit assignments (continued)

Bits Name Function

[51]av Enable contention detection Enables contention detection and fast exclusive monitor path:
and fast exclusive monitor
0
path
Disables contention detection and fast exclusive monitor path. This is the reset value.
1
Enables contention detection and fast exclusive monitor path.

[50]av Disable store streaming on Disables store streaming on NC/GRE memory type:
NC/GRE memory type
0
Enables store streaming on NC/GRE memory type. This is the reset value.
1
Disables store streaming on NC/GRE memory type.

[49]av Disable non-allocate hint of Disables non-allocate hint of Write-Back No-Allocate memory type:
Write-Back No-Allocate
0
(WBNA) memory type
Enables non-allocate hint of WBNA memory type. This is the reset value.
1
Disables non-allocate hint of WBNA memory type.

[48]av Disable early speculative Disables early speculative read access from LS to L2:
read access from LS to L2
0
Enables speculative early read access from LS to L2. This is the reset value.
1
Disables speculative early read access from LS to L2.

[47]av Disable D-side L1/L2 Disables L1 and L2 hardware prefetch across 4KB page boundary even if page is 64KB or
hardware prefetch across larger:
4KB page boundary even if
0
page is 64KB or larger.
Enables D-side L1/L2 hardware prefetch across 4KB page boundary if the page is
64KB or larger. This is the reset value.
1
Disables D-sideL1/L2 hardware prefetch across 4KB page boundary even if the page
is 64KB or larger.

[46]av Disable multiple Disables multiple outstanding L1 Data TLB misses and L2 TLB hit under miss:
outstanding L1 Data TLB
0
misses and L2 TLB hit
Enables multiple outstanding L1 Data TLB misses and L2 TLB hit under miss. This is
under miss
the reset value.
1
Disables multiple outstanding L1 Data TLB misses and L2 TLB hit under miss.

[45]av Disable L1-DCache way Disables L1-DCache way tracker:

tracker
0
Enables L1-DCache way tracker. This is the reset value.
1
Disables L1-DCache way tracker.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-199
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-77 CPUACTLR_EL1[63:32] bit assignments (continued)

Bits Name Function

[44]av Enable data cache clean as Enables data cache clean as data cache clean and invalidate:
data cache clean/invalidate
0
Normal behavior, executes data cache clean as data cache clean.
This is the reset value.
1
Executes data cache clean as data cache clean and invalidate.

[43]av Disable VA based hardware Disables the Load/Store hardware prefetcher from using VA to cross page boundaries:
prefetch
0
Enables the Load/Store hardware prefetcher to use VA in generating prefetches that
can cross page boundaries. This is the reset value.
1
Disables the Load/Store hardware prefetcher from using VA in prefetch generation.

[42]av Disable prefetch requests Disables prefetch requests from ReadUnique transactions:
from ReadUnique
0
transactions
Enables prefetch requests to be generated by ReadUnique transactions. This is the
reset value.
1
Disables prefetch requests to be generated by ReadUnique transactions.

[41]av Enable snoop hazard while Enables snoop hazard while waiting for second half of atomic exclusive pair:
waiting for second half of
0
atomic exclusive pair
Disable snoop hazard while waiting for second half of atomic exclusive pair. This is
the reset value.
1
Enable snoop hazard while waiting for second half of the atomic exclusive pair.

[40] - Reserved, RES0.

[39]av Disable instruction merging Disables instruction merging:

0
Enables instruction merging. This is the reset value.
1
Disables instruction merging.

[38]av Force FPSCR write flush Forces FPSCR write flush:

0
Normal behavior for FPSCR writes. This is the reset value.
1
Forces synchronizing flush on all FPSCR writes.

[37]av Disable instruction group Disables instruction group split:

split
0
Enables instruction group split. This is the reset value.
1
Disables instruction group split.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-200
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-77 CPUACTLR_EL1[63:32] bit assignments (continued)

Bits Name Function

[36]av Force implicit DSB on an Forces implicit DSB on ISB event:

ISB event
0
Normal behavior. This is the reset value.
1
Force implicit DSB on an ISB event.

[35]av Disable ISB optimization Disables ISB optimization:

0
Enables ISB optimization. This is the reset value.
1
Disables ISB optimization.

[34]av Disable Static Branch Disables static branch predictor:

Predictor
0
Enables static branch predictor. This is the reset value.
1
Disables static branch predictor.

[33]av Disable main prediction Disables main prediction suppression at target fetch of microBTB:
suppression at target fetch of
0
microBTB
Enables prediction suppression at target fetch of microBTB. This is the reset value.
1
Disables prediction suppression at target fetch of microBTB.

[32]av Disable L1 Instruction Disables L1 Instruction Cache prefetch:

Cache prefetch
0
Enables Instruction Cache prefetch. This is the reset value.
1
Disables Instruction Cache prefetch.

The following figure shows the CPUACTLR_EL1[31:0] bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-201
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

Enable invalidates of BTB

Disable Instruction cache
streaming
RES0
Disable micro-BTB
Disable indirect predictor
Execute PLD instructions as a NOP
RES0
Execute WFE instructions as a NOP
Execute WFI instructions as a NOP
Disable flag renaming optimization
Force serialization after each group
Limit to one instruction per instruction group
Force push of SPRs
Flush after SPR writes
Force limit of one instruction group commit/deallocate per cycle
Force in-order issue in branch execute unit
Enable full Strongly-Ordered & Device loads replay
Disable L2 TLB performance optimization
Disable L2 stage 1 translation table walk L2 PA cache
Disable L2 stage 1 translation table walk cache
Disable L2 translation table walk IPA PA cache
Disable L2 TLB prefetching
Force in-order load issue
Force in-order requests to same set/way
Non-cacheable streaming enhancement
Write streaming no-L1-allocate threshold
Write streaming no-allocate threshold
Force CX clock enable active
Force main clock enable active
Snoop-delayed exclusive handling

Figure 4-74 CPUACTLR_EL1[31:0] bit assignments

The following table shows the CPUACTLR_EL1[31:0] bit assignments.

Table 4-78 CPUACTLR_EL1[31:0] bit assignments

Bits Name Function

[31] Snoop-delayed exclusive Snoop-delayed exclusive handling. The possible values are:
handling
0 Normal exclusive handling behavior. This is the reset value.
1 Modifies exclusive handling behavior by delaying certain snoop requests.

[30]aw Force main clock enable Forces main clock enable active. The possible values are:
active
0 Does not prevent the clock generator from stopping the processor clock. This is the reset
value.
1 Prevents the clock generator from stopping the processor clock.

If the processor dynamic retention feature is used then this bit must be zero. See Processor
dynamic retention on page 2-46.

aw This bit is used internally for debugging and characterization purposes only. For normal operation, ARM recommends that you do not change the value of this bit
from its reset value.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-202
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-78 CPUACTLR_EL1[31:0] bit assignments (continued)

Bits Name Function

[29]aw Force Advanced SIMD Forces Advanced SIMD and Floating-point clock enable active. The possible values are:
and floating-point clock
0 Does not prevent the clock generator from stopping the Advanced SIMD and Floating-point
enable active
clock. This is the reset value.
1 Prevents the clock generator from stopping the Advanced SIMD and Floating-point clock.

See Advanced SIMD and FP clock gating on page 2-49.

If the processor dynamic retention feature is used then this bit must be zero. See Processor
dynamic retention on page 2-46.

[28:27] Write streaming no- Write streaming no-allocate threshold. The possible values are:
allocate threshold
0b00 12th consecutive streaming cache line does not allocate in the L1 or L2 cache. This is the
reset value.
0b01 128th consecutive streaming cache line does not allocate in the L1 or L2 cache.
0b10 512th consecutive streaming cache line does not allocate in the L1 or L2 cache.
0b11 Disables streaming. All Write-Allocate lines allocate in the L1 or L2 cache.

[26:25] Write streaming no-L1- Write streaming no-L1-allocate threshold. The possible values are:
allocate threshold
0b00 4th consecutive streaming cache line does not allocate in the L1 cache. This is the reset
value.
0b01 64th consecutive streaming cache line does not allocate in the L1 cache.
0b10 128th consecutive streaming cache line does not allocate in the L1 cache.
0b11 Disables streaming. All Write-Allocate lines allocate in the L1 cache.

[24] Non-cacheable streaming Non-cacheable streaming enhancement. You can set this bit only if your memory system meets
enhancement the requirement that cache line fill requests from the Cortex-A72 processor are atomic. The
possible values are:
0 Disables higher performance Non-cacheable load forwarding. This is the reset value.
1 Enables higher performance Non-cacheable load forwarding. See 6.4.4 Non-cacheable
streaming enhancement on page 6-292 for more information.

[23]aw Force in-order requests to Forces in-order requests to the same set and way. The possible values are:
the same set and way
0 Does not force in-order requests to the same set and way. This is the reset value.
1 Forces in-order requests to the same set and way.

[22]aw Force in-order load issue Forces in-order load issue. The possible values are:
0 Does not force in-order load issue. This is the reset value.
1 Forces in-order load issue.

[21]aw Disable L2 TLB Disables L2 TLB prefetching. The possible values are:
prefetching
0 Enables L2 TLB prefetching. This is the reset value.
1 Disables L2 TLB prefetching.

[20]aw Disable L2 translation Disables L2 translation table walk Immediate Physical Address (IPA) to Physical Address (PA)
table walk IPA PA cache cache. The possible values are:
0 Enables L2 translation table walk IPA to PA cache. This is the reset value.
1 Disables L2 translation table walk IPA to PA cache.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-203
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-78 CPUACTLR_EL1[31:0] bit assignments (continued)

Bits Name Function

[19]aw Disable L2 stage 1 Disables L2 stage 1 translation table walk cache. The possible values are:
translation table walk
0 Enables L2 stage 1 translation table walk cache. This is the reset value.
cache
1 Disables L2 stage 1 translation table walk cache.

[18]aw Disable L2 stage 1 Disables L2 stage 1 translation table walk L2 PA cache. The possible values are:
translation table walk L2
0 Enables L2 stage 1 translation table walk L2 PA cache. This is the reset value.
PA cache
1 Disables L2 stage 1 translation table walk L2 PA cache.

[17]aw Disable L2 TLB Disables L2 TLB performance optimization. The possible values are
performance optimization
0 Enables L2 TLB optimization. This is the reset value.
1 Disables L2 TLB optimization.

[16]aw Enable full Strongly- Enables full Strongly-ordered or Device load replay. The possible values are:
ordered and Device load
0 Disables full Strongly-ordered or Device load replay. This is the reset value.
replay
1 Enables full Strongly-ordered or Device load replay.

[15]aw Force in-order issue in Forces in-order issue in branch execute unit. The possible values are:
branch execute unit
0 Disables forced in-order issue. This is the reset value.
1 Forces in-order issue.

[14]aw Force limit of one Forces limit of one instruction group to commit and de-allocate per cycle. The possible values
instruction group are:
commit/de-allocate per
0 Normal commit and de-allocate behavior. This is the reset value.
cycle
1 Limits commit and de-allocate to one instruction group per cycle.

[13]aw Flush after Special Flushes after certain SPR writes. The possible values are:
Purpose Register (SPR)
0 Normal behavior for SPR writes. This is the reset value.
writes
1 Flushes after certain SPR writes.

[12]aw Force push of SPRs Forces push of certain SPRs from local dispatch copies to shadow copies. The possible values
are:
0 Normal behavior for SPRs. This is the reset value.
1 Pushes certain SPRs from local dispatch copies to shadow copies.

Note
Setting this bit to 1 forces the processor to behave as if bit[13] is set to 1.

[11]aw Limit to one instruction Limits to one instruction per instruction group. The possible values are:
per instruction group
0 Normal instruction grouping. This is the reset value.
1 Limits to one instruction per instruction group.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-204
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-78 CPUACTLR_EL1[31:0] bit assignments (continued)

Bits Name Function

[10]aw Force serialization after Forces serialization after each instruction group. The possible values are:
each instruction group
0 Disables forced serialization after each instruction group. This is the reset value.
1 Forces serialization after each instruction group.

Note
Setting this bit to 1 forces the processor to behave as if bit[11] is set to 1.

[9]aw Disable flag renaming Disables flag renaming optimization. The possible values are:
optimization
0 Enables normal flag renaming optimization. This is the reset value.
1 Disables normal flag renaming optimization.

[8]aw Execute WFI instruction Executes WFI instruction as a NOP instruction. The possible values are:
as a NOP instruction
0 Executes WFI instruction as defined in the ARM® Architecture Reference Manual ARMv8.
This is the reset value.
1 Executes WFI instruction as a NOP instruction, and does not put the processor in WFI low-
power state.

[7]aw Execute WFE instruction Executes WFE instruction as a NOP instruction. The possible values are:
as a NOP instruction
0 Executes WFE instruction as defined in the ARM® Architecture Reference Manual ARMv8.
This is the reset value.
1 Executes WFE instruction as a NOP instruction, and does not put the processor in WFE low-
power state.

[6] - Reserved, RES0.

[5]aw Execute PLD and PLDW Executes PLD and PLDW instructions as a NOP instruction. The possible values are:
instructions as a NOP
0 Executes PLD and PLDW instructions as defined in the ARM® Architecture Reference Manual
ARMv8. This is the reset value.
1 Executes PLD and PLDW instructions as a NOP instruction.

[4]aw Disable indirect predictor Disables indirect predictor. The possible values are:
0 Enables indirect predictor. This is the reset value.
1 Disables indirect predictor.

[3]aw Disable micro-BTB Disables micro-Branch Target Buffer (BTB). The possible values are:
0 Enables micro-BTB. This is the reset value.
1 Disables micro-BTB.

[2] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-205
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-78 CPUACTLR_EL1[31:0] bit assignments (continued)

Bits Name Function

[1]aw Disable Instruction Cache Disables Instruction Cache miss streaming. The possible values are:
miss streaming
0 Enables Instruction Cache miss streaming. Sequential fetches resulting from Instruction
Cache misses wait until individual packets arrive. This is the reset value.
1 Disables Instruction Cache miss streaming. Sequential fetches resulting from Instruction
Cache misses internally generate misses for each packet.

[0]aw Enable invalidates of Enables invalidate of BTB. The possible values are:
BTB
0 The Invalidate Instruction Cache All and Invalidate Instruction Cache by VA instructions
only invalidates the instruction cache array. This is the reset value.
1 The Invalidate Instruction Cache All and Invalidate Instruction Cache by VA instructions
invalidates the instruction cache array and branch target buffer.

To access the CPUACTLR_EL1 in AArch64 state, read or write the register with:
MRS <Xt>, S3_1_c15_c2_0; Read EL1 CPU Auxiliary Control Register
MSR S3_1_c15_c2_0, <Xt>; Write EL1 CPU Auxiliary Control Register

To access the CPUACTLR in AArch32 state, read or write the CP15 register with:
MRRC p15, 0, <Rt>, <Rt2>, c15; Read CPU Auxiliary Control Register
MCRR p15, 0, <Rt>, <Rt2>, c15; Write CPU Auxiliary Control Register

4.3.67 CPU Extended Control Register, EL1

The CPUECTLR_EL1 characteristics are:
Purpose
Provides additional IMPLEMENTATION DEFINED configuration and control options for the processor.
Usage constraints
The accessibility to the CPUECTLR_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RWax RWax RWay RW RW

The CPUECTLR_EL1 can be written dynamically.

Configurations
The CPUECTLR_EL1 is:
• Common to the Secure and Non-secure states.
• A 64-bit read/write register.
• Architecturally mapped to the AArch32 CPUECTLR register.
Attributes
See the register summary in Table 4-15 AArch64 implementation defined registers
on page 4-86.
The following figure shows the CPUECTLR_EL1 bit assignments.

ax Write access if ACTLR_EL3.CPUECTLR is 1 and ACTLR_EL2.CPUECTLR is 1, or ACTLR_EL3.CPUECTLR is 1 and SCR.NS is 0.

ay Write access if ACTLR_EL3.CPUECTLR is 1.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-206
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

63 39 38 37 36 35 34 33 32 31 7 6 5 3 2 0

RES0 RES0 RES0

Disable table walk descriptor access prefetch SMPEN

RES0 Processor dynamic retention control
L2 instruction fetch prefetch distance
RES0
L2 load/store data prefetch distance

Figure 4-75 CPUECTLR_EL1 bit assignments

The following table shows the CPUECTLR_EL1 bit assignments.

Table 4-79 CPUECTLR_EL1 bit assignments

Bits Name Function

[63:39] - Reserved, RES0.

[38] Disable table Disables table walk descriptor access prefetch. The possible values are:
walk descriptor
0
access prefetch
Enables table walk descriptor access prefetch. This is the reset value.
1
Disables table walk descriptor access prefetch.

[37] - Reserved, RES0.

[36:35] L2 instruction Indicates the L2 instruction fetch prefetch distance. It is the number of requests by which the prefetcher is
fetch prefetch ahead of the demand request stream. It also specifies the maximum number of prefetch requests generated
distance on a demand miss. The possible values are:
0b00
0 requests, disables instruction prefetch.
0b01
1 request.
0b10
2 requests.
0b11
3 requests. This is the reset value.

[34] - Reserved, RES0.

[33:32] L2 load data Indicates the L2 load data prefetch distance. It is the number of requests by which the prefetch request to
prefetch distance the L2, on a load stream, is ahead of the demand request stream. The possible values are:
0b00
16 requests.
0b01
18 requests.
0b10
20 requests.
0b11
22 requests. This is the reset value.

[31:7] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-207
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-79 CPUECTLR_EL1 bit assignments (continued)

Bits Name Function

[6] SMPEN Enables the processor to receive instruction cache and TLB maintenance operations broadcast from other
processors in the cluster.
You must set this bit before enabling the caches and MMU, or performing any cache and TLB
maintenance operations.
You must clear this bit during a processor power down sequence. See 2.4 Power management
on page 2-42.
The possible values are:
0
Disables receiving of instruction cache and TLB maintenance operations. This is the reset value.
1
Enables receiving of instruction cache and TLB maintenance operations.

Note
• Any processor instruction cache and TLB maintenance operations can execute the request, regardless
of the value of the SMPEN bit.
• This bit has no impact on data cache maintenance operations.
• In the Cortex-A72 processor, the L1 data cache and L2 cache are always coherent, for shared or non-
shared data, regardless of the value of the SMPEN bit.

[5:3] - Reserved, RES0.

[2:0] Processor Processor dynamic retention control. The possible values are:
dynamic
0b000
retention control
Processor dynamic retention disabled. This is the reset value.
0b001
2 Generic Timer ticks required before retention entry.
0b010
8 Generic Timer ticks required before retention entry.
0b011
32 Generic Timer ticks required before retention entry.
0b100
64 Generic Timer ticks required before retention entry.
0b101
128 Generic Timer ticks required before retention entry.
0b110
256 Generic Timer ticks required before retention entry.
0b111
512 Generic Timer ticks required before retention entry.
All other values are reserved.

To access the CPUECTLR_EL1 in AArch64 state, read or write the register with:
MRS <Xt>, S3_1_c15_c2_1; Read EL1 CPU Extended Control Register
MSR S3_1_c15_c2_1, <Xt>; Write EL1 CPU Extended Control Register

To access the CPUECTLR in AArch32 state, read or write the CP15 register with:
MRRC p15, 1, <Rt>, <Rt2>, c15; Read CPU Extended Control Register
MCRR p15, 1, <Rt>, <Rt2>, c15; Write CPU Extended Control Register

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-208
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

4.3.68 CPU Memory Error Syndrome Register, EL1

The CPUMERRSR_EL1 characteristics are:

Purpose
Holds the number of memory errors that have occurred in the following L1 and L2 RAMs:
• L1-I Tag RAM.
• L1-I Data RAM.
• L1-D Tag RAM.
• L1-D Data RAM.
• L2 TLB RAM.
A write of any value to the register updates the register to zero.
Usage constraints
The accessibility to the CPUMERRSR_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Configurations
The CPUMERRSR_EL1 is:
• Common to the Secure and Non-secure states.
• A 64-bit read/write register.
• Architecturally mapped to the AArch32 CPUMERRSR register.
Attributes
See the register summary in Table 4-15 AArch64 implementation defined registers
on page 4-86.
The following figure shows the CPUMERRSR_EL1 bit assignments.
63 62 48 47 40 39 32 31 30 24 23 22 18 17 0
Other error Repeat error
Reserved RAMID Bank/Way Index
count count

Fatal Valid Reserved

Figure 4-76 CPUMERRSR_EL1 bit assignments

The following table shows the CPUMERRSR_EL1 bit assignments.

Table 4-80 CPUMERRSR_EL1 bit assignments

Bits Name Function

[63] Fatal Fatal bit. This bit is set to 1 on the first memory error that caused a Data Abort. It is a sticky bit so that after
it is set, it remains set until the register is written. The reset value is 0.

[62:48] - Reserved, RES0.

[47:40] Other error This field is set to 0 on the first memory error and is incremented on any memory error that does not match
count the RAMID, bank, way, or index information in this register while the sticky Valid bit is set. The reset
value is 0.

[39:32] Repeat error This field is set to 0 on the first memory error and is incremented on any memory error that exactly
count matches the RAMID, bank, way or index information in this register while the sticky Valid bit is set. The
reset value is 0.

[31] Valid Valid bit. This bit is set to 1 on the first memory error. It is a sticky bit so that after it is set, it remains set
until the register is written. The reset value is 0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-209
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-80 CPUMERRSR_EL1 bit assignments (continued)

Bits Name Function

[30:24] RAMID RAM Identifier. Indicates the RAM, the first memory error occurred in. The possible values are:
0x00 L1-I Tag RAM.
0x01 L1-I Data RAM.
0x08 L1-D Tag RAM.
0x09 L1-D Data RAM.
0x18 L2 TLB RAM.

[23] - Reserved, RES0.

[22:18] Bank/Way Indicates the bank or way of the RAM where the first memory error occurred.

[17:0] Index Indicates the index address of the first memory error.

Note
• If two or more memory errors in the same RAM occur in the same cycle, only one error is reported.
• If two or more first memory error events from different RAMs occur in the same cycle, one of the
errors is selected arbitrarily, while the Other error count field is only incremented by one.
• If two or more memory error events from different RAMs, that do not match the RAMID, bank, way,
or index information in this register while the sticky Valid bit is set, occur in the same cycle, the
Other error count field is only incremented by one.

To access the CPUMERRSR_EL1 in AArch64 state, read or write the register with:
MRS <Xt>, S3_1_c15_c2_2 ; Read EL1 CPU Memory Error Syndrome Register
MSR S3_1_c15_c2_2 , <Xt>; Write EL1 CPU Memory Error Syndrome Register

To access the CPUMERRSR in AArch32 state, read or write the CP15 register with:
MRRC p15, 2, <Rt>, <Rt2>, c15; Read CPU Memory Error Syndrome Register
MCRR p15, 2, <Rt>, <Rt2>, c15; Write CPU Memory Error Syndrome Register

4.3.69 L2 Memory Error Syndrome Register, EL1

The L2MERRSR_EL1 characteristics are:
Purpose Holds the number of memory errors that have occurred in the following L2 RAMs:
• L2 Tag RAM.
• L2 Data RAM.
• L2 Snoop Tag RAM.
• L2 Dirty RAM.
• L2 Inclusion PLRU RAM.
A write of any value to the register updates the register to zero.
Usage The accessibility to the L2MERRSR_EL1 by Exception level is:
constraints
EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Configurations The L2MERRSR_EL1 is:

• Common to the Secure and Non-secure states.
• A 64-bit read/write register.
• Architecturally mapped to the AArch32 L2MERRSR register.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-210
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Attributes See the register summary in Table 4-15 AArch64 implementation defined registers
on page 4-86.

The following figure shows the L2MERRSR_EL1 bit assignments.

63 62 48 47 40 39 32 31 30 24 23 22 21 18 17 0
Other error Repeat error
RES0 RAMID RES0 CPUID/Way Index
count count

Fatal Valid

Figure 4-77 L2MERRSR_EL1 bit assignments

The following table shows the L2MERRSR_EL1 bit assignments.

Table 4-81 L2MERRSR_EL1 bit assignments

Bits Name Function

[63] Fatal Fatal bit. This bit is set to 1 on the first memory error that caused a Data Abort. It is a sticky bit so that after it
is set, it remains set until the register is written. The reset value is 0.

[62:48] - Reserved, RES0.

[47:40] Other error This field is set to 0 on the first memory error and is incremented on any memory error that does not match
count the RAMID, bank, way, or index information in this register while the sticky Valid bit is set. The reset value is
0.

[39:32] Repeat error This field is set to 0 on the first memory error and is incremented on any memory error that exactly matches
count the RAMID, bank, way or index information in this register while the sticky Valid bit is set. The reset value is
0.

[31] Valid Valid bit. This bit is set to 1 on the first memory error. It is a sticky bit so that after it is set, it remains set until
the register is written. The reset value is 0.

[30:24] RAMID RAM Identifier. Indicates the RAM where the first memory error occurred. The possible values are:
0b001 0000 L2 Tag RAM.
0b001 0001 L2 Data RAM.
0b001 0010 L2 Snoop Tag RAM.
0b001 0100 L2 Dirty RAM.
0b0011000 L2 Inclusion PLRU RAM.

[23:22] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-211
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Table 4-81 L2MERRSR_EL1 bit assignments (continued)

Bits Name Function

[21:18] CPUID/Way Indicates which processor and way of the RAM where the first memory error occurred.
For L2 Tag, Data, and Dirty RAMs, bits[21:18] indicate one of 16 ways, from way 0 to way 15.
For L2 Snoop Tag RAM:
• Bits[20:19] indicate which processor of the L1 Tag RAM.
• Bit[18] indicates which way of the Tag RAM.
The possible values are:
0b0000 CPU0 tag, way 0.
0b0001 CPU0 tag, way 1.
0b0010 CPU1 tag, way 0.
0b0011 CPU1 tag, way 1.
0b0100 CPU2 tag, way 0.
0b0101 CPU2 tag, way 1.
0b0110 CPU3 tag, way 0.
0b0111 CPU3 tag, way 1.

[17:0] Index Indicates the index address of the first memory error.

Note
• If two or more memory errors in the same RAM occur in the same cycle, only one error is reported.
• If two or more first memory error events from different RAMs occur in the same cycle, one of the
errors is selected arbitrarily, while the Other error count field is only incremented by one.
• If two or more memory error events from different RAMs, that do not match the RAMID, bank, way,
or index information in this register while the sticky Valid bit is set, occur in the same cycle, the
Other error count field is only incremented by one.

To access the L2MERRSR_EL1 in AArch64 state, read or write the register with:
MRS <Xt>, S3_1_c15_c2_3 ; Read EL1 L2 Memory Error Syndrome Register
MSR S3_1_c15_c2_3, <Xt> ; Write EL1 L2 Memory Error Syndrome Register

To access the L2MERRSR in AArch32 state, read or write the CP15 register with:
MRRC p15, 3, <Rt>, <Rt2>, c15; Read L2 Memory Error Syndrome Register
MCRR p15, 3, <Rt>, <Rt2>, c15; Write L2 Memory Error Syndrome Register

4.3.70 Configuration Base Address Register, EL1

The CBAR_EL1 characteristics are:

Purpose
Holds the physical base address of the memory-mapped GIC CPU interface registers.
Usage constraints
The accessibility to the CBAR_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-212
reserved.
Non-Confidential
 4 System Control
4.3 AArch64 register descriptions

Configurations
The CBAR_EL1 is:
• Common to the Secure and Non-secure states.
• A 64-bit register in AArch64 state.
Attributes
See the register summary in Table 4-15 AArch64 implementation defined registers
on page 4-86.
The following figure shows the CBAR_EL1 bit assignments.
63 44 43 18 17 0

RES0 PERIPHBASE RES0

Figure 4-78 CBAR_EL1 bit assignments

The following table shows the CBAR_EL1 bit assignments.

Table 4-82 CBAR_EL1 bit assignments

Bits Name Function

[63:44] - Reserved, RES0

[43:18] PERIPHBASE The primary input PERIPHBASE[43:18] determines the reset value

[17:0] - Reserved, RES0

To access the CBAR_EL1 in AArch64 state, read the register with:

MRS <Xt>, s3_1_c15_c3_0; Read EL1 Configuration Base Address Register

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-213
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

4.4 AArch32 register summary

This section gives a summary of the System registers in AArch32 state.
The System registers are a set of registers that you can write to and read from. Some of the registers
permit more than one type of operation.
The registers are accessed by the MCR and MRC instructions for 32-bit registers and the MCRR and MRRC
instructions for 64-bit registers. The following sections describe the System registers grouped by CRn in
the order of op1, CRm, and op2:
• 4.4.1 c0 registers on page 4-215.
• 4.4.2 c1 registers on page 4-216.
• 4.4.3 c2 registers on page 4-217.
• 4.4.4 c3 registers on page 4-218.
• 4.4.5 c5 registers on page 4-218.
• 4.4.6 c6 registers on page 4-218.
• 4.4.7 c7 register on page 4-219.
• 4.4.8 c7 System operations on page 4-219.
• 4.4.9 c8 System operations on page 4-220.
• 4.4.10 c9 registers on page 4-221.
• 4.4.11 c10 registers on page 4-222.
• 4.4.12 c12 registers on page 4-222.
• 4.4.13 c13 registers on page 4-222.
• 4.4.14 c14 registers on page 4-223.
• 4.4.15 c15 registers on page 4-224.
The following subsection describes the 64-bit registers and provides cross-references to individual
register descriptions:
• 4.4.16 64-bit registers on page 4-224.
In addition to listing the System registers by CRn ordering, the following subsections describe the
System registers by functional group:
• 4.4.17 Identification registers on page 4-225.
• 4.4.18 CPUID registers on page 4-227.
• 4.4.19 Virtual memory control registers on page 4-228.
• 4.4.20 Fault and Exception handling registers on page 4-229.
• 4.4.21 Other System registers on page 4-230.
• 4.4.22 Cache maintenance operations on page 4-231.
• 4.4.23 TLB maintenance operations on page 4-231.
• 4.4.24 Address translation operations on page 4-232.
• 4.4.25 Miscellaneous operations on page 4-233.
• 4.4.26 Performance Monitors registers on page 4-233.
• 4.4.27 Security registers on page 4-234.
• 4.4.28 Virtualization registers on page 4-235.
• 4.4.29 Hyp mode TLB maintenance operations on page 4-236.
• 4.4.30 Generic Timer registers on page 4-236.
• 4.4.31 Implementation defined registers on page 4-236.
The following table describes the column headings that the System register summary tables use
throughout AArch32 state. These correspond to fields within the MCR and MRC instruction mnemonics:
MCR p15, op1, Rt, CRn, CRm, op2
MRC p15, op1, Rt, CRn, CRm, op2

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-214
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-83 Column headings definition for System register summary tables

Column name Description

CRn Register number within the System registers

op1 Opcode_1 value for the register

CRm Operational register number within CRn

op2 Opcode_2 value for the register

Name Short form architectural, operation, or code name for the register

Type One of:

• Read-only (RO).
• Write-only (WO).
• Read/write (RW).

Reset Reset value of register

Description Cross-reference to register description

4.4.1 c0 registers

The following table shows the CP15 System registers when CRn is c0 and the processor is in AArch32
state.

Table 4-84 c0 register summary

op1 CRm op2 Name Type Reset Description

0 c0 0 MIDR RO 0x410FD083 Main ID Register. See 4.3.1 Main ID Register, EL1 on page 4-89.

1 CTR RO 0x8444C004 Cache Type Register. See 4.3.26 Cache Type Register, EL0 on page 4-121.

2 TCMTR - 0x00000000 4.5.1 TCM Type Register on page 4-239.

3 TLBTR RO 0x00000000 4.5.2 TLB Type Register on page 4-239.

4, 7 MIDR RO 0x410FD083 Aliases of Main ID Register, see 4.3.1 Main ID Register, EL1
on page 4-89.

5 MPIDR RO 0x80000003az 4.5.3 Multiprocessor Affinity Register on page 4-240.

6 REVIDR RO 0x00000000 Revision ID Register. See 4.3.3 Revision ID Register, EL1 on page 4-92.

c1 0 ID_PFR0 RO 0x00000131 Processor Feature Register 0. See 4.3.4 AArch32 Processor Feature
Register 0, EL1 on page 4-93.

1 ID_PFR1 RO 0x00011011ba Processor Feature Register 1. See 4.3.5 AArch32 Processor Feature
Register 1, EL1 on page 4-94.

2 ID_DFR0 RO 0x03010066 Debug Feature Register 0. See 4.3.6 AArch32 Debug Feature Register 0,
EL1 on page 4-95.

3 ID_AFR0 RO 0x00000000 Auxiliary Feature Register 0. See 4.3.7 AArch32 Auxiliary Feature
Register 0, EL1 on page 4-96.

4 ID_MMFR0 RO 0x10201105 Memory Model Feature Register 0. See 4.3.8 AArch32 Memory Model
Feature Register 0, EL1 on page 4-96.

az The reset value depends on the primary inputs, CLUSTERIDAFF1 and CLUSTERIDAFF2, and the number of cores that the device implements. The value shown
is for a four processor implementation, with CLUSTERIDAFF1 and CLUSTERIDAFF2 set to zero.
ba The reset value depends on the primary input GICCDISABLE. The value shown assumes the GICCDISABLE signal is tied HIGH.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-215
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-84 c0 register summary (continued)

op1 CRm op2 Name Type Reset Description

5 ID_MMFR1 RO 0x40000000 Memory Model Feature Register 1. See 4.3.9 AArch32 Memory Model
Feature Register 1, EL1 on page 4-98.

6 ID_MMFR2 RO 0x01260000 Memory Model Feature Register 2. See 4.3.10 AArch32 Memory Model
Feature Register 2, EL1 on page 4-99.

7 ID_MMFR3 RO 0x02102211 Memory Model Feature Register 3. See 4.3.11 AArch32 Memory Model
Feature Register 3, EL1 on page 4-101.

c2 0 ID_ISAR0 RO 0x02101110 Instruction Set Attribute Register 0. See 4.3.12 AArch32 Instruction Set
Attribute Register 0, EL1 on page 4-103.

1 ID_ISAR1 RO 0x13112111 Instruction Set Attribute Register 1. See 4.3.13 AArch32 Instruction Set
Attribute Register 1, EL1 on page 4-104.

2 ID_ISAR2 RO 0x21232042 Instruction Set Attribute Register 2. See 4.3.14 AArch32 Instruction Set
Attribute Register 2, EL1 on page 4-106.

3 ID_ISAR3 RO 0x01112131 Instruction Set Attribute Register 3. See 4.3.15 AArch32 Instruction Set
Attribute Register 3, EL1 on page 4-107.

4 ID_ISAR4 RO 0x00011142 Instruction Set Attribute Register 4. See 4.3.16 AArch32 Instruction Set
Attribute Register 4, EL1 on page 4-108.

5 ID_ISAR5 RO 0x00010001bb Instruction Set Attribute Register 5. See 4.3.17 AArch32 Instruction Set
Attribute Register 5, EL1 on page 4-110.

1 c0 0 CCSIDR RO UNK Cache Size ID Register. See 4.3.22 Cache Size ID Register, EL1
on page 4-117.

1 CLIDR RO 0x0A200023 Cache Level ID Register. See 4.3.23 Cache Level ID Register, EL1
on page 4-118.

7 AIDR - 0x00000000 Auxiliary ID Register. See 4.3.24 Auxiliary ID Register, EL1

on page 4-120.

2 c0 0 CSSELR RW UNK Cache Size Selection Register. See 4.3.25 Cache Size Selection Register,
EL1 on page 4-120.

4 c0 0 VPIDR RW -bc Virtualization Processor ID Register. See 4.3.28 Virtualization Processor

ID Register, EL2 on page 4-123.

5 VMPIDR RO -bd Virtualization Multiprocessor ID Register. See 4.5.4 Virtualization

Multiprocessor ID Register on page 4-241.

4.4.2 c1 registers

The following table shows the System registers when CRn is c1 and the processor is in AArch32 state.

bb The reset value is 0x00011121 if the Cryptography engine is implemented.

bc The reset value is the value of the Main ID Register.
bd The reset value is the value of the Multiprocessor Affinity Register.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-216
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-85 c1 register summary

op1 CRm op2 Name Type Reset Description

0 c0 0 SCTLR RW 0x00C50838be 4.5.5 System Control Register on page 4-242.

1 ACTLR - 0x00000000 Auxiliary Control Register. See 4.3.39 Auxiliary Control Register, EL3
on page 4-143.

2 CPACR RW 0x00000000 4.5.6 Architectural Feature Access Control Register on page 4-247.

c1 0 SCR RW 0x00000000 4.5.7 Secure Configuration Register on page 4-248.

1 SDER RW 0x00000000 Secure Debug Enable Register. 

2 NSACR RW bg 0x00000000 4.5.8 Non-secure Access Control Register on page 4-251.

c3 1 SDCR RW 0x00000000 4.5.9 Secure Debug Configuration Register on page 4-252.

4 c0 0 HSCTLR RW 0x30C50838 Hyp System Control Register. bf

1 HACTLR RW 0x00000000 Hyp Auxiliary Control Register. See 4.3.33 Auxiliary Control Register, EL2
on page 4-130.

c1 0 HCR RW 0x00000000 4.5.10 Hyp Configuration Register on page 4-254.

1 HDCR RW 0x00000006bh 4.5.12 Hyp Debug Control Register on page 4-261.

2 HCPTR RW 0x000033FF 4.5.13 Hyp Architectural Feature Trap Register on page 4-263.

3 HSTR RW 0x00000000 Hyp System Trap Register. See 4.3.36 Hypervisor System Trap Register
on page 4-139.

4 HCR2 RW 0x00000000 4.5.11 Hyp Configuration Register 2 on page 4-259.

7 HACR RW 0x00000000 4.3.37 Hyp Auxiliary Configuration Register on page 4-141.

4.4.3 c2 registers

The following table shows the System registers when CRn is c2 and the processor is in AArch32 state.

Table 4-86 c2 register summary

op1 CRm op2 Name Type Reset Description

0 c0 0 TTBR0 RW UNK
4.5.14 Translation Table Base Register 0 and Register 1 on page 4-265
1 TTBR1 RW UNK

2 TTBCR RW 0x00000000bi 4.5.15 Translation Table Base Control Register on page 4-265

4 c0 2 HTCR RW UNK 4.5.16 Hyp Translation Control Register on page 4-266

c1 2 VTCR RW UNK Virtualization Translation Control Register, see the ARM® Architecture
Reference Manual ARMv8

be The reset value depends on primary input, CFGEND. The value shown assumes this signal is set to zero.
bf See the ARM® Architecture Reference Manual ARMv8 for more information.
bg RO at EL2 and EL0(NS).
bh The reset value for bit[7] is UNK.
bi The reset value is 0x00000000 for the Secure copy of the register. The reset value for the EAE bit of the Non-secure copy of the register is 0b0. You must program
the Non-secure copy of the register with the required initial value, as part of the processor boot sequence.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-217
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

4.4.4 c3 registers

The following table shows the System registers when CRn is c3 and the processor is in AArch32 state.

Table 4-87 c3 register summary

op1 CRm op2 Name Type Reset Description

0 c0 0 DACR RW UNK Domain Access Control Register, see the ARM® Architecture Reference Manual ARMv8

4.4.5 c5 registers

The following table shows the System registers when CRn is c5 and the processor is in AArch32 state.

Table 4-88 c5 register summary

op1 CRm op2 Name Type Reset Description

0 c0 0 DFSR RW UNK 4.5.17 Data Fault Status Register on page 4-267.

1 IFSR RW UNK Instruction Fault Status Register. See 4.3.51 Instruction Fault Status Register,
EL2 on page 4-159.

c1 0 ADFSR RW 0x00000000 Auxiliary Data Fault Status Register. See 4.3.48 Auxiliary Fault Status Register
0, EL1 and EL3 on page 4-156.

1 AIFSR RW 0x00000000 Auxiliary Instruction Fault Status Register. See 4.3.49 Auxiliary Fault Status
Register 1, EL1 and EL3 on page 4-157.

4 c1 0 HADFSR RW 0x00000000 Hyp Auxiliary Data Fault Status Register. See 4.3.52 Auxiliary Fault Status
Register 0, EL2 and Hyp Auxiliary Data Fault Status Register on page 4-163.

1 HAIFSR RW 0x00000000 Hyp Auxiliary Instruction Fault Status Register. See 4.3.53 Auxiliary Fault
Status Register 1, EL2 and Hyp Auxiliary Instruction Fault Status Register
on page 4-163.

c2 0 HSR RW UNK Hyp Syndrome Register. See 4.3.54 Exception Syndrome Register, EL2
on page 4-163.

4.4.6 c6 registers

The following table shows the System registers when CRn is c6 and the processor is in AArch32 state.
See the ARM® Architecture Reference Manual ARMv8 for more information about these registers.

Table 4-89 c6 register summary

op1 CRm op2 Name Type Reset Description

0 c0 0 DFAR RW UNK Data Fault Address Register

2 IFAR RW UNK Instruction Fault Address Register

4 c0 0 HDFAR RW UNK Hyp Data Fault Address Register

2 HIFAR RW UNK Hyp Instruction Fault Address Register

4 HPFAR RW UNK Hyp IPA Fault Address Register

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-218
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

4.4.7 c7 register

The following table shows the System registers when CRn is c7 and the processor is in AArch32 state.

Table 4-90 c7 register summary

op1 CRm op2 Name Type Reset Description

0 c4 0 PAR RW UNK 4.5.18 Physical Address Register on page 4-271

4.4.8 c7 System operations

The following table shows the System operations when CRn is c7 and the processor is in AArch32 state.
See the ARM® Architecture Reference Manual ARMv8 for more information about these operations.

Table 4-91 c7 System operation summary

op1 CRm op2 Name Description

0 c1 0 ICIALLUIS Invalidate all instruction caches Inner Shareable to PoUbj

6 BPIALLIS Invalidate all entries from branch predictors Inner Shareable

c5 0 ICIALLU Invalidate all Instruction Caches to PoU

1 ICIMVAU Invalidate Instruction Caches by VA to PoU

4 CP15ISB Instruction Synchronization Barrier operation, this operation is deprecated in ARMv8-A

6 BPIALL Invalidate all entries from branch predictors

7 BPIMVA Invalidate VA from branch predictors

c6 1 DCIMVAC Invalidate data cache line by VA to PoCbk

2 DCISW Invalidate data cache line by set/way

c8 0 ATS1CPR Stage 1 current state PL1 read

1 ATS1CPW Stage 1 current state PL1 write

2 ATS1CUR Stage 1 current state unprivileged read

3 ATS1CUW Stage 1 current state unprivileged write

4 ATS12NSOPR Stages 1 and 2 Non-secure only PL1 read

5 ATS12NSOPW Stages 1 and 2 Non-secure only PL1 write

6 ATS12NSOUR Stages 1 and 2 Non-secure only unprivileged read

7 ATS12NSOUW Stages 1 and 2 Non-secure only unprivileged write

c10 1 DCCMVAC Clean data cache line by VA to PoC

2 DCCSW Clean data cache line by set/way

4 CP15DSB Data Synchronization Barrier operation, this operation is deprecated in ARMv8-A

5 CP15DMB Data Memory Barrier operation, this operation is deprecated in ARMv8-A

c11 1 DCCMVAU Clean data cache line by VA to PoU

bj PoU = Point of Unification. PoU is set by the BROADCASTINNER signal and can be in the L1 data cache or outside of the processor, in which case PoU is
dependent on the external memory system.
bk PoC = Point of Coherence. The PoC is always outside of the processor and is dependent on the external memory system.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-219
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-91 c7 System operation summary (continued)

op1 CRm op2 Name Description

c14 1 DCCIMVAC Clean and invalidate data cache line by VA to PoC

2 DCCISW Clean and invalidate data cache line by set/way

4 c8 0 ATS1HR Stage 1 Hyp mode read

1 ATS1HW Stage 1 Hyp mode write

4.4.9 c8 System operations

The following table shows the System operations when CRn is c8 and the processor is in AArch32 state.
See the ARM® Architecture Reference Manual ARMv8 for more information about these operations.

Table 4-92 c8 System operations summary

op1 CRm op2 Name Description

0 c3 0 TLBIALLIS Invalidate entire TLB Inner Shareable

1 TLBIMVAIS Invalidate unified TLB entry by VA and ASID Inner Shareable

2 TLBIASIDIS Invalidate unified TLB by ASID match Inner Shareable

3 TLBIMVAAIS Invalidate unified TLB entry by VA all ASID Inner Shareable

5 TLBIMVALIS Invalidate unified TLB entry by VA Inner Shareable, Last level

7 TLBIMVAALIS Invalidate unified TLB by VA all ASID Inner Shareable, Last level

c5 0 ITLBIALL Invalidate instruction TLB

1 ITLBIMVA Invalidate instruction TLB entry by VA and ASID

2 ITLBIASID Invalidate instruction TLB by ASID match

c6 0 DTLBIALL Invalidate data TLB

1 DTLBIMVA Invalidate data TLB entry by VA and ASID

2 DTLBIASID Invalidate data TLB by ASID match

c7 0 TLBIALL Invalidate unified TLB

1 TLBIMVA Invalidate unified TLB by VA and ASID

2 TLBIASID Invalidate unified TLB by ASID match

3 TLBIMVAA Invalidate unified TLB entries by VA all ASID

5 TLBIMVAL Invalidate last level of stage 1 TLB entry by VA

7 TLBIMVAAL Invalidate last level of stage 1 TLB entry by VA all ASID

4 c0 1 TLBIIPAS2IS TLB Invalidate entry by Intermediate Physical Address, Stage 2, Inner Shareable

5 TLBIIPAS2LIS TLB Invalidate entry by Intermediate Physical Address, Stage 2, Last level, Inner Shareable

c3 0 TLBIALLHIS Invalidate entire Hyp unified TLB Inner Shareable

1 TLBIMVAHIS Invalidate Hyp unified TLB entry by VA Inner Shareable

4 TLBIALLNSNHIS Invalidate entire Non-secure non-Hyp unified TLB Inner Shareable

5 TLBIMVALHIS Invalidate Unified Hyp TLB entry by VA Inner Shareable, Last level

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-220
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-92 c8 System operations summary (continued)

op1 CRm op2 Name Description

c4 1 TLBIIPAS2 TLB Invalidate entry by Intermediate Physical Address, Stage 2

5 TLBIIPAS2L TLB Invalidate entry by Intermediate Physical Address, Stage 2, Last level

c7 0 TLBIALLH Invalidate entire Hyp unified TLB

1 TLBIMVAH Invalidate Hyp unified TLB entry by VA

4 TLBIALLNSNH Invalidate entire Non-secure non-Hyp unified TLB

5 TLBIMVALH Invalidate Unified Hyp TLB entry by VA, Last level

4.4.10 c9 registers

The following table shows the System registers when CRn is c9 and the processor is in AArch32 state.

Table 4-93 c9 register summary

op1 CRm op2 Name Type Reset Description

0 c12 0 PMCR RW 0x41023000 Performance Monitors Control Register. See 11.4.1 Performance
Monitors Control Register, EL0 on page 11-401.

1 PMCNTENSET RW UNK Performance Monitors Count Enable Set Register. 

2 PMCNTENCLR RW UNK Performance Monitors Count Enable Clear Register. bl

3 PMOVSR RW UNK Performance Monitors Overflow Flag Status Register. bl

4 PMSWINC WO - Performance Monitors Software Increment Register. bl

5 PMSELR RW UNK Performance Monitors Event Counter Selection Register. bl

6 PMCEID0 RO 0x7FFF0F3F Performance Monitors Common Event Identification Register 0. See

11.4.2 Performance Monitors Common Event Identification Register
0, EL0 on page 11-403.

7 PMCEID1 RO 0x00000000 Performance Monitors Common Event Identification Register 1. bl

c13 0 PMCCNTR RW UNK Performance Monitors Cycle Counter Register. bl

1 PMXEVTYPER RW UNK Performance Monitors Selected Event Type Register. bl

PMCCFILTR RW 0x00000000 Performance Monitors Cycle Count Filter Register. bl

2 PMXEVCNTR RW UNK Performance Monitors Selected Event Count Register. bl

c14 0 PMUSERENR RW 0x00000000 Performance Monitors User Enable Register. bl

1 PMINTENSET RW UNK Performance Monitors Interrupt Enable Set Register. bl

2 PMINTENCLR RW UNK Performance Monitors Interrupt Event Clear Register. bl

3 PMOVSSET RW UNK Performance Monitors Overflow Flag Status Set Register. bl

1 c0 2 L2CTLR RW 0x00000000bm L2 Control Register. See 4.3.58 L2 Control Register, EL1

on page 4-168.

3 L2ECTLR RW 0x00000000 L2 Extended Control Register. See 4.3.59 L2 Extended Control

Register, EL1 on page 4-172.

bl See the ARM® Architecture Reference Manual ARMv8 for more information.
bm The reset value depends on the processor implementation and the state of the L2RSTDISABLE signal.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-221
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

4.4.11 c10 registers

The following table shows the System registers when CRn is c10 and the processor is in AArch32 state.

Table 4-94 c10 register summary

op1 CRm op2 Name Type Reset Description

0 c2 0 PRRR RW 0x00098AA4 4.5.19 Primary Region Remap Register on page 4-272.

0 MAIR0 RW UNK 4.5.20 Memory Attribute Indirection Register 0 on page 4-273.

1 NMRR RW 0x44E048E0 4.5.21 Normal Memory Remap Register. on page 4-273.

1 MAIR1 RW UNK 4.5.22 Memory Attribute Indirection Register 1 on page 4-274.

c3 0 AMAIR0 RW UNK Auxiliary Memory Attribute Indirection Register 0. See 4.3.56 Auxiliary
Memory Attribute Indirection Register, EL1 and EL3 on page 4-168.

1 AMAIR1 RW UNK Auxiliary Memory Attribute Indirection Register 1. See 4.3.56 Auxiliary
Memory Attribute Indirection Register, EL1 and EL3 on page 4-168.

4 c2 0 HMAIR0 RW UNK Hyp Memory Attribute Indirection Register 0, See 4.3.57 Auxiliary Memory
Attribute Indirection Register, EL2 on page 4-168.

1 HMAIR1 RW UNK Hyp Memory Attribute Indirection Register 1, See 4.3.57 Auxiliary Memory
Attribute Indirection Register, EL2 on page 4-168.

c3 0 HAMAIR0 RW UNK Hyp Auxiliary Memory Attribute Indirection Register 0. See 4.3.57 Auxiliary
Memory Attribute Indirection Register, EL2 on page 4-168.

1 HAMAIR1 RW UNK Hyp Auxiliary Memory Attribute Indirection Register 1. See 4.3.57 Auxiliary
Memory Attribute Indirection Register, EL2 on page 4-168.

4.4.12 c12 registers

The following table shows the System registers when CRn is c12 and the processor is in AArch32 state.

Table 4-95 c12 register summary

op1 CRm op2 Name Type Reset Description

0 c0 0 VBAR RW 0x00000000bn Vector Base Address Register. 

1 MVBAR RW UNK Monitor Vector Base Address Register. bo

2 RMR RW 0x00000000bp Reset Management Register. See 4.3.61 Reset Management Register, EL3
on page 4-175.

c1 0 ISR RO UNK Interrupt Status Register. bo

4 c0 0 HVBAR RW UNK Hyp Vector Base Address Register. bo

4.4.13 c13 registers

The following table shows the System registers when CRn is c13 and the processor is in AArch32 state.

bn The reset value is 0x00000000 for the Secure copy of the register. You must program the Non-secure copy of the register with the required initial value, as part of
the processor boot sequence.
bo See the ARM® Architecture Reference Manual ARMv8 for more information.
bp The reset value of bit[0] depends on the AA64nAA32 signal. The following table assumes this signal is LOW.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-222
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-96 c13 register summary

op1 CRm op2 Name Type Reset Description

0 c0 0 FCSEIDR RW 0x00000000 4.5.23 FCSE Process ID

Register on page 4-274

1 CONTEXTIDR RW UNK Context ID Register 

2 TPIDRURW RW UNK User Read/Write Thread

Pointer ID Register bq

3 TPIDRURO RW br UNK User Read-Only Thread

Pointer ID Register bq

4 TPIDRPRW RW UNK EL1 only Thread Pointer ID

Register bq

4 c0 2 HTPIDR RW UNK Hyp Thread Pointer ID

Register bq

4.4.14 c14 registers

The following table shows the System registers when CRn is C14 and the processor is in AArch32 state.
See the ARM® Architecture Reference Manual ARMv8 for more information about these registers.

Table 4-97 c14 register summary

op1 CRm op2 Name Type Reset Description

0 c0 0 CNTFRQ RW bs UNK Timer Counter Frequency register

c1 0 CNTKCTL RW -bt EL1 Timer Control register

c2 0 CNTP_TVAL RW UNK EL1 Physical Timer TimerValue register

1 CNTP_CTL RW - EL1 Physical Timer Control register

c3 0 CNTV_TVAL RW UNK Virtual Timer TimerValue register

1 CNTV_CTL RW -bu Virtual Timer Control register

c8 0 PMEVCNTR0 RW UNK Performance Monitors Event Count Registers

1 PMEVCNTR1

2 PMEVCNTR2

3 PMEVCNTR3

4 PMEVCNTR4

5 PMEVCNTR5

bq See the ARM® Architecture Reference Manual ARMv8 for more information.
br RO at EL0.
bs Ar EL3(S) only, otherwise it is RO.
bt The reset value for bits[9:8, 2:0] is 0b00000.
bu The reset value for bit[0] is 0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-223
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-97 c14 register summary (continued)

op1 CRm op2 Name Type Reset Description

c12 0 PMEVTYPER0 RW UNK Performance Monitors Event Type Registers

1 PMEVTYPER1

2 PMEVTYPER2

3 PMEVTYPER3

4 PMEVTYPER4

5 PMEVTYPER5

c15 7 PMCCFILTR RW 0x00000000 Performance Monitors Cycle Count Filter Register

4 c1 0 CNTHCTL RW -bv EL2 Timer Control register

c2 0 CNTHP_TVAL RW UNK EL2 Physical Timer TimerValue register

1 CNTHP_CTL RW -bu EL2 Physical Timer Control register

4.4.15 c15 registers

The following table shows the System registers when CRn is c15 and the processor is in AArch32 state.

Table 4-98 c15 register summary

op1 CRm op2 Name Type Reset Description

0 c0 0 IL1DATA0 RW UNK 4.3.62 Instruction L1 Data n Register, EL1 on page 4-176.

1 IL1DATA1

2 IL1DATA2

3 IL1DATA3

c1 0 DL1DATA0 RW UNK 4.3.63 Data L1 Data n Register, EL1 on page 4-177.

1 DL1DATA1

2 DL1DATA2

3 DL1DATA3

4 DL1DATA4

c4 0 RAMINDEXbw WO - 4.3.64 RAM Index operation on page 4-178.

1 c0 0 L2ACTLR RW 0x00000010bx L2 Auxiliary Control Register. See L2 Auxiliary Control Register, EL1.

c3 0 CBAR RO -by 4.5.24 Configuration Base Address Register on page 4-274.

4.4.16 64-bit registers

The following table gives a summary of the 64-bit wide System registers, accessed by the MCRR and MRRC
instructions when the processor is in AArch32 state.

bv The reset value for bit[2] is 0 and for bits[1:0] is 0b11.

bw RAMINDEX is a system operation.
bx The reset value is 0x00000010 for an ACE interface and 0x00004018 for a CHI interface.
by The reset value depends on the primary input, PERIPHBASE[43:18].

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-224
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-99 64-bit register summary

CRn op1 CRm op2 Name Type Reset Description

- 0 c2 - TTBR0 RW UNK Translation Table Base Register 0. 

- 1 c2 - TTBR1 RW UNK Translation Table Base Register 1. bz

- 4 c2 - HTTBR RW UNK Hyp Translation Table Base Register. bz

- 6 c2 - VTTBR RW UNKca Virtualization Translation Table Base Register. bz

- 0 c7 - PAR RW UNK 4.5.18 Physical Address Register on page 4-271.

- 0 c9 - PMCCNTR RW - Performance Monitors Cycle Count Register. bz

- 0 c14 - CNTPCT RO UNK Physical Timer Count register. bz

- 1 c14 - CNTVCT RO UNK Virtual Timer Count register. bz

- 2 c14 - CNTP_CVAL RW UNK EL1 Physical Timer CompareValue register. bz

- 3 c14 - CNTV_CVAL RW UNK Virtual Timer CompareValue register. bz

- 4 c14 - CNTVOFF RW UNK Virtual Timer Offset register. bz

- 6 c14 - CNTHP_CVAL RW UNK EL2 Physical Timer CompareValue register. bz

0 c15 - CPUACTLR RW -cb CPU Auxiliary Control Register. See 4.3.66 CPU Auxiliary Control
Register, EL1 on page 4-194.

1 c15 - CPUECTLR RW -cc CPU Extended Control Register. See 4.3.67 CPU Extended Control
Register, EL1 on page 4-206.

- 2 c15 - CPUMERRSR RW - CPU Memory Error Syndrome Register. See 4.3.68 CPU Memory Error
Syndrome Register, EL1 on page 4-209.

- 3 c15 - L2MERRSR RW -cd L2 Memory Error Syndrome Register. See 4.3.69 L2 Memory Error
Syndrome Register, EL1 on page 4-210.

4.4.17 Identification registers

The following table shows the Identification registers in AArch32 state.

Table 4-100 Identification registers

Name CRn op1 CRm op2 Type Reset Description

MIDR c0 0 c0 0 RO 0x410FD083 Main ID Register. See 4.3.1 Main ID Register, EL1 on page 4-89.

CTR 1 RO 0x8444C004 Cache Type Register. See 4.3.26 Cache Type Register, EL0
on page 4-121.

TCMTR 2 - 0x00000000 4.5.1 TCM Type Register on page 4-239.

TLBTR 3 RO 0x00000000 4.5.2 TLB Type Register on page 4-239.

MPIDR 5 RO 0x80000003ce 4.5.3 Multiprocessor Affinity Register on page 4-240.

bz See the ARM® Architecture Reference Manual ARMv8 for more information.
ca The reset value for bits[55:48] is zero.
cb The reset value is zero.
cc The reset value is 0x0000 001B 0000 0000.
cd The reset value for bits[63,47:40,39:32,31] is zero.
ce The reset value depends on the primary inputs, CLUSTERIDAFF1, and the number of processors that the MPCore device implements. The value shown is for a
four processor implementation, with CLUSTERIDAFF1 set to zero.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-225
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-100 Identification registers (continued)

Name CRn op1 CRm op2 Type Reset Description

REVIDR 6 RO 0x00000000 Revision ID Register. See 4.3.3 Revision ID Register, EL1

on page 4-92.

MIDR 4, 7 RO 0x410FD083 Aliases of Main ID Register, see 4.3.1 Main ID Register, EL1
on page 4-89.

ID_PFR0 c1 0 RO 0x00000131 Processor Feature Register 0. See 4.3.4 AArch32 Processor Feature
Register 0, EL1 on page 4-93.

ID_PFR1 1 RO 0x00011011cf Processor Feature Register 1. See 4.3.5 AArch32 Processor Feature
Register 1, EL1 on page 4-94.

ID_DFR0 2 RO 0x03010066 Debug Feature Register 0. See 4.3.6 AArch32 Debug Feature
Register 0, EL1 on page 4-95.

ID_AFR0 3 RO 0x00000000 Auxiliary Feature Register 0. See 4.3.7 AArch32 Auxiliary Feature
Register 0, EL1 on page 4-96.

ID_MMFR0 4 RO 0x10201105 Memory Model Feature Register 0. See 4.3.8 AArch32 Memory
Model Feature Register 0, EL1 on page 4-96.

ID_MMFR1 5 RO 0x40000000 Memory Model Feature Register 1. See 4.3.9 AArch32 Memory
Model Feature Register 1, EL1 on page 4-98.

ID_MMFR2 6 RO 0x01260000 Memory Model Feature Register 2. See 4.3.10 AArch32 Memory
Model Feature Register 2, EL1 on page 4-99.

ID_MMFR3 7 RO 0x02102211 Memory Model Feature Register 3. See 4.3.11 AArch32 Memory
Model Feature Register 3, EL1 on page 4-101.

ID_ISAR0 c2 0 RO 0x02101110 Instruction Set Attribute Register 0. See 4.3.12 AArch32 Instruction
Set Attribute Register 0, EL1 on page 4-103

ID_ISAR1 1 RO 0x13112111 Instruction Set Attribute Register 1. See 4.3.13 AArch32 Instruction
Set Attribute Register 1, EL1 on page 4-104.

ID_ISAR2 2 RO 0x21232042 Instruction Set Attribute Register 2. See 4.3.14 AArch32 Instruction
Set Attribute Register 2, EL1 on page 4-106.

ID_ISAR3 3 RO 0x01112131 Instruction Set Attribute Register 3. See 4.3.15 AArch32 Instruction
Set Attribute Register 3, EL1 on page 4-107.

ID_ISAR4 4 RO 0x00011142 Instruction Set Attribute Register 4. See 4.3.16 AArch32 Instruction
Set Attribute Register 4, EL1 on page 4-108.

ID_ISAR5 5 RO 0x00010001cg Instruction Set Attribute Register 5. See 4.3.17 AArch32 Instruction
Set Attribute Register 5, EL1 on page 4-110.

CCSIDR 1 c0 0 RO UNK Cache Size ID Register. See 4.3.22 Cache Size ID Register, EL1
on page 4-117.

CLIDR 1 RO 0x0A200023 Cache Level ID Register. See 4.3.23 Cache Level ID Register, EL1
on page 4-118.

AIDR 7 - 0x00000000 Auxiliary ID Register. See 4.3.24 Auxiliary ID Register, EL1

on page 4-120.

CSSELR 2 c0 0 RW UNK Cache Size Selection Register. See 4.3.25 Cache Size Selection
Register, EL1 on page 4-120.

cf The reset value depends on the primary input GICCDISABLE. The value shown assumes the GICCDISABLE signal is tied HIGH.
cg The reset value is 0x00011121 if the Cryptography engine is implemented.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-226
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-100 Identification registers (continued)

Name CRn op1 CRm op2 Type Reset Description

VPIDR 4 c0 0 RW -ch Virtualization Processor ID Register. See 4.3.28 Virtualization

Processor ID Register, EL2 on page 4-123.

VMPIDR 5 RO -ci Virtualization Multiprocessor ID Register. See 4.5.4 Virtualization

Multiprocessor ID Register on page 4-241.

4.4.18 CPUID registers

The following table shows the CPUID registers in AArch32 state.

Table 4-101 CPUID registers

Name CRn op1 CRm op2 Type Reset Description

ID_PFR0 c0 0 c1 0 RO 0x00000131 Processor Feature Register 0. See 4.3.4 AArch32

Processor Feature Register 0, EL1 on page 4-93.

ID_PFR1 1 RO 0x00011011cj Processor Feature Register 1. See 4.3.5 AArch32

Processor Feature Register 1, EL1 on page 4-94.

ID_DFR0 2 RO 0x03010066 Debug Feature Register 0. See 4.3.6 AArch32 Debug

Feature Register 0, EL1 on page 4-95.

ID_AFR0 3 RO 0x00000000 Auxiliary Feature Register 0. See 4.3.7 AArch32

Auxiliary Feature Register 0, EL1 on page 4-96.

ID_MMFR0 4 RO 0x10201105 Memory Model Feature Register 0. See

4.3.8 AArch32 Memory Model Feature Register 0,
EL1 on page 4-96.

ID_MMFR1 5 RO 0x40000000 Memory Model Feature Register 1. See

4.3.9 AArch32 Memory Model Feature Register 1,
EL1 on page 4-98.

ID_MMFR2 6 RO 0x01260000 Memory Model Feature Register 2. See

4.3.10 AArch32 Memory Model Feature Register 2,
EL1 on page 4-99.

ID_MMFR3 7 RO 0x02102211 Memory Model Feature Register 3. See

4.3.11 AArch32 Memory Model Feature Register 3,
EL1 on page 4-101.

ch The reset value is the value of the Main ID Register. See 4.3.1 Main ID Register, EL1 on page 4-89 for more information.
ci The reset value is the value of the Multiprocessor Affinity Register.
cj The reset value depends on the primary input GICCDISABLE. The value shown assumes the GICCDISABLE signal is tied HIGH.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-227
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-101 CPUID registers (continued)

Name CRn op1 CRm op2 Type Reset Description

ID_ISAR0 c0 0 c2 0 RO 0x02101110 Instruction Set Attribute Register 0. See

4.3.12 AArch32 Instruction Set Attribute Register 0,
EL1 on page 4-103.

ID_ISAR1 1 RO 0x13112111 Instruction Set Attribute Register 1. See

4.3.13 AArch32 Instruction Set Attribute Register 1,
EL1 on page 4-104.

ID_ISAR2 2 RO 0x21232042 Instruction Set Attribute Register 2. See

4.3.14 AArch32 Instruction Set Attribute Register 2,
EL1 on page 4-106.

ID_ISAR3 3 RO 0x01112131 Instruction Set Attribute Register 3. See

4.3.15 AArch32 Instruction Set Attribute Register 3,
EL1 on page 4-107.

ID_ISAR4 4 RO 0x00011142 Instruction Set Attribute Register 4. See

4.3.16 AArch32 Instruction Set Attribute Register 4,
EL1 on page 4-108.

ID_ISAR5 5 RO 0x00000001ck Instruction Set Attribute Register 5. See

4.3.17 AArch32 Instruction Set Attribute Register 5,
EL1 on page 4-110.

4.4.19 Virtual memory control registers

The following table shows the Virtual memory control registers in AArch32 state.

Table 4-102 Virtual memory control registers

Name CRn op1 CRm op2 Type Reset Width Description

SCTLR c1 0 c0 0 RW 0x00C50838cl 32-bit 4.5.5 System Control Register on page 4-242.

HSCTLR c1 4 c0 0 RW 0x30C50838 32-bit Hyp System Control Register. 

TTBR0 c2 0 c0 0 RW UNK 32-bit

Translation Table Base Register 0. cm
- 0 c2 - 64-bit

TTBR1 c2 0 c0 1 RW UNK 32-bit

Translation Table Base Register 1. cm
- 1 c2 - 64-bit

TTBCR c2 0 c0 2 RW 0x00000000cn 32-bit 4.5.15 Translation Table Base Control Register

on page 4-265.

HTCR c2 4 c0 2 RW UNK 32-bit 4.5.16 Hyp Translation Control Register on page 4-266.

VTCR c1 2 RW UNK 32-bit Virtualization Translation Control Register. cm

DACR c3 0 c0 0 RW UNK 32-bit Domain Access Control Register. cm

ck The reset value is 0x00001121 if the Cryptography engine is implemented.

cl The reset value depends on primary inputs, CFGTE, CFGEND, and VINITHI. Table 4-102 Virtual memory control registers on page 4-228 assumes these signals
are LOW.
cm See the ARM® Architecture Reference Manual ARMv8 for more information.
cn The reset value is 0x00000000 for the Secure copy of the register. The reset value for the EAE bit of the Non-secure copy of the register is 0x0. You must program
the Non-secure copy of the register with the required initial value, as part of the processor boot sequence.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-228
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-102 Virtual memory control registers (continued)

Name CRn op1 CRm op2 Type Reset Width Description

PRRR c10 0 c2 0 RW 0x00098AA4 32-bit 4.5.19 Primary Region Remap Register on page 4-272.

MAIR0 0 RW UNK 32-bit 4.5.20 Memory Attribute Indirection Register 0

on page 4-273.

NMRR 1 RW 0x44E048E0 32-bit 4.5.21 Normal Memory Remap Register. on page 4-273.

MAIR1 1 RW UNK 32-bit 4.5.22 Memory Attribute Indirection Register 1

on page 4-274.

AMAIR0 c3 0 RW UNK 32-bit Auxiliary Memory Attribute Indirection Register 0. See

4.3.56 Auxiliary Memory Attribute Indirection Register,
EL1 and EL3 on page 4-168.

AMAIR1 1 RW UNK 32-bit Auxiliary Memory Attribute Indirection Register 1. See

4.3.56 Auxiliary Memory Attribute Indirection Register,
EL1 and EL3 on page 4-168.

HMAIR0 4 c2 0 RW UNK 32-bit Hyp Memory Attribute Indirection Register 0. cm

HMAIR1 1 RW UNK 32-bit Hyp Memory Attribute Indirection Register 1. cm

HAMAIR0 c3 0 RW UNK 32-bit Hyp Auxiliary Memory Attribute Indirection Register 0.

See 4.3.57 Auxiliary Memory Attribute Indirection
Register, EL2 on page 4-168.

HAMAIR1 1 RW UNK 32-bit Hyp Auxiliary Memory Attribute Indirection Register 1.

See 4.3.57 Auxiliary Memory Attribute Indirection
Register, EL2 on page 4-168.

CONTEXTIDR c13 0 c0 1 RW UNK 32-bit Context ID Register. cm

4.4.20 Fault and Exception handling registers

The following table shows the Fault handling registers in AArch32 state.

Table 4-103 Fault and Exception handling registers

Name CRn op1 CRm op2 Type Reset Description

DFSR c5 0 c0 0 RW UNK 4.5.17 Data Fault Status Register on page 4-267.

IFSR 1 RW UNK Instruction Fault Status Register. See 4.3.51 Instruction Fault Status
Register, EL2 on page 4-159.

ADFSR c1 0 RW UNK Auxiliary Data Fault Status Register. See 4.3.48 Auxiliary Fault Status
Register 0, EL1 and EL3 on page 4-156.

AIFSR 1 RW UNK Auxiliary Instruction Fault Status Register. See 4.3.49 Auxiliary Fault
Status Register 1, EL1 and EL3 on page 4-157.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-229
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-103 Fault and Exception handling registers (continued)

Name CRn op1 CRm op2 Type Reset Description

HADFSR 4 c1 0 RW UNK Hyp Auxiliary Data Fault Status Register. See 4.3.52 Auxiliary Fault
Status Register 0, EL2 and Hyp Auxiliary Data Fault Status Register
on page 4-163.

HAIFSR 1 RW UNK Hyp Auxiliary Instruction Fault Status Register. See 4.3.53 Auxiliary
Fault Status Register 1, EL2 and Hyp Auxiliary Instruction Fault
Status Register on page 4-163.

HSR c2 0 RW UNK Hyp Syndrome Register. See 4.3.54 Exception Syndrome Register, EL2
on page 4-163.

DFAR c6 0 c0 0 RW UNK Data Fault Address Register. 

IFAR 2 RW UNK Instruction Fault Address Register. co

HDFAR 4 c0 0 RW UNK Hyp Data Fault Address Register. co

HIFAR 2 RW UNK Hyp Instruction Fault Address Register. co

HPFAR 4 RW UNK Hyp IPA Fault Address Register. co

VBAR c12 0 c0 0 RW 0x00000000cp Vector Base Address Register. co

MVBAR 1 RW UNK Monitor Vector Base Address Register. co

RMR 2 RW 0x00000000cq Reset Management Register. See 4.3.53 Auxiliary Fault Status Register
1, EL2 and Hyp Auxiliary Instruction Fault Status Register
on page 4-163.

ISR c1 0 RO UNK Instruction Status Register. co

HVBAR 4 c0 2 RW UNK Hyp Vector Base Address Register. co

The Virtualization registers include additional fault handling registers. For more information see
4.4.28 Virtualization registers on page 4-235.

4.4.21 Other System registers

The following table shows the other System registers in AArch32 state.

Table 4-104 Other System registers

Name CRn op1 CRm op2 Type Reset Description

ACTLR c1 0 c0 1 - 0x00000000 Auxiliary Control Register. See 4.3.39 Auxiliary Control Register, EL3
on page 4-143.

CPACR 2 RW 0x00000000 4.5.6 Architectural Feature Access Control Register on page 4-247.

HACTLR 4 c0 1 RW 0x00000000 Hyp Auxiliary Control Register. See 4.3.33 Auxiliary Control Register,
EL2 on page 4-130.

FCSEIDR c13 0 c0 0 RW 0x00000000 4.5.23 FCSE Process ID Register on page 4-274.

co See the ARM® Architecture Reference Manual ARMv8 for more information.
cp The reset value is 0x00000000 for the Secure copy of the register. You must program the Non-secure copy of the register with the required initial value, as part of
the processor boot sequence.
cq The reset value of bit[0] depends on the AA64nAA32 signal. Table 4-103 Fault and Exception handling registers on page 4-229 assumes this signal is LOW.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-230
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

4.4.22 Cache maintenance operations

The following table shows the System instructions for cache and branch predictor maintenance
operations in AArch32 state. See the ARM® Architecture Reference Manual ARMv8 for more information
about these operations.

Table 4-105 Cache and branch predictor maintenance operations

Name CRn op1 CRm op2 Description

ICIALLUIS c7 0 c1 0 Instruction Cache invalidate all to PoUcr Inner Shareable
BPIALLIS 6 Branch predictor invalidate all Inner Shareable
ICIALLU c5 0 Instruction Cache invalidate all to PoU
ICIMVAU 1 Instruction Cache invalidate by VA to PoU
BPIALL 6 Branch predictor invalidate all
BPIMVA 7 Branch predictor invalidate by VA
DCIMVAC c6 1 Data cache invalidate by VA to PoCcs
DCISW 2 Data cache invalidate by set/way
DCCMVAC c10 1 Data cache clean by VA to PoC
DCCSW 2 Data cache clean by set/way
DCCMVAU c11 1 Data cache clean by VA to PoU
DCCIMVAC c14 1 Data cache clean and invalidate by VA to PoC
DCCISW 2 Data cache clean and invalidate by set/way

4.4.23 TLB maintenance operations

The following table shows the System instructions for TLB maintenance operations in AArch32 state.
See the ARM® Architecture Reference Manual ARMv8 for more information about these operations.

Table 4-106 TLB maintenance operations

Name CRn op1 CRm op2 Description

TLBIALLIS c8 0 c3 0 Invalidate entire unified TLB Inner Shareable
TLBIMVAIS 1 Invalidate unified TLB by VA and ASID Inner Shareable
TLBIASIDIS 2 Invalidate unified TLB by ASID Inner Shareable
TLBIMVAAIS 3 Invalidate unified TLB by VA all ASID Inner Shareable
TLBIMVALIS 5 Invalidate unified TLB entry by VA Inner Shareable, Last level
TLBIMVAALIS 7 Invalidate unified TLB by VA all ASID Inner Shareable, Last level
ITLBIALL c5 0 Invalidate entire instruction TLB
ITLBIMVA 1 Invalidate instruction TLB entry by VA and ASID
ITLBIASID 2 Invalidate instruction TLB by ASID

cr PoU = Point of Unification. PoU is set by the BROADCASTINNER signal and can be in the L1 data cache or outside of the processor, in which case PoU is
dependent on the external memory system.
cs PoC = Point of Coherence. The PoC is always outside of the processor and is dependent on the external memory system.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-231
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-106 TLB maintenance operations (continued)

Name CRn op1 CRm op2 Description

DTLBIALL c6 0 Invalidate entire data TLB
DTLBIMVA 1 Invalidate data TLB entry by VA and ASID
DTLBIASID 2 Invalidate data TLB by ASID
TLBIALL c7 0 Invalidate entire unified TLB
TLBIMVA 1 Invalidate unified TLB by VA and ASID
TLBIASID 2 Invalidate unified TLB by ASID
TLBIMVAA 3 Invalidate unified TLB by VA all ASID
TLBIMVAL 5 Invalidate unified TLB entry by VA, Last level
TLBIMVAAL 7 Invalidate unified TLB by VA all ASID, Last level

The Virtualization registers include additional TLB operations for use in Hyp mode.

Related information
4.2.13 AArch64 EL2 TLB maintenance operations on page 4-85.

4.4.24 Address translation operations

The following table shows the address translation register in AArch32 state.

Table 4-107 Address translation register

Name CRn op1 CRm op2 Reset Width Description

PAR c7 0 c4 0 UNK 32-bit 4.5.18 Physical Address Register on page 4-271

- 0 c7 - 64-bit

The following table shows the System instructions for address translation operations in AArch32 state.

Table 4-108 Address translation operations

Name CRn op1 CRm op2 Reset Width Description

ATS1CPR c7 0 c8 0 UNK 32-bit Stage 1 current state EL1 read 
ATS1CPW 1 UNK 32-bit Stage 1 current state EL1 write ct
ATS1CUR 2 UNK 32-bit Stage 1 current state unprivileged read ct
ATS1CUW 3 UNK 32-bit Stage 1 current state unprivileged write ct
ATS12NSOPR 4 UNK 32-bit Stages 1 and 2 Non-secure EL1 read ct
ATS12NSOPW 5 UNK 32-bit Stages 1 and 2 Non-secure EL1 write ct
ATS12NSOUR 6 UNK 32-bit Stages 1 and 2 Non-secure unprivileged read ct
ATS12NSOUW 7 UNK 32-bit Stages 1 and 2 Non-secure unprivileged write ct

ct See the ARM® Architecture Reference Manual ARMv8 for more information.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-232
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-108 Address translation operations (continued)

Name CRn op1 CRm op2 Reset Width Description

ATS1HR 4 c8 0 UNK 32-bit Stage 1 Hyp mode read ct
ATS1HW 1 UNK 32-bit Stage 1 Hyp mode write ct

4.4.25 Miscellaneous operations

The following table shows the System instructions and the registers for miscellaneous operations in
AArch32 state.

Table 4-109 Miscellaneous System operations

Name CRn op1 CRm op2 Type Reset Description

CP15ISB c7 4 c5 4 - UNK Instruction Synchronization Barrier operation, this operation is deprecated
in ARMv8-A
CP15DSB c10 4 - UNK Data Synchronization Barrier operation, this operation is deprecated in
ARMv8-A
CP15DMB 5 - UNK Data Memory Barrier operation, this operation is deprecated in ARMv8-A

TPIDRURW c13 0 c0 2 RW UNK User Read/Write Thread ID Register 

TPIDRURO 3 RW cv UNK EL1 only Thread ID Register cu

TPIDRPRW 4 RW UNK Hyp Software Thread ID Register cu

HTPIDR 4 c0 2 RW UNK User Read-Only Thread ID Register cu

4.4.26 Performance Monitors registers

The following table shows the Performance Monitors registers in AArch32 state.

cu See the ARM® Architecture Reference Manual ARMv8 for more information.
cv RO at EL0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-233
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-110 Performance Monitors registers

Name CRn op1 CRm op2 Type Reset Description

PMCR c9 0 c12 0 RW 0x41023000 Performance Monitors Control Register. See 11.4.1 Performance
Monitors Control Register, EL0 on page 11-401.

PMCNTENSET 1 RW UNK Performance Monitors Count Enable Set Register. 

PMCNTENCLR 2 RW UNK Performance Monitors Count Enable Clear Register. cw

PMOVSR 3 RW UNK Performance Monitors Overflow Flag Status Register. cw

PMSWINC 4 WO - Performance Monitors Software Increment Register. cw

PMSELR 5 RW UNK Performance Monitors Event Counter Selection Register. cw

PMCEID0 6 RO 0x7FFF0F3F Performance Monitors Common Event Identification Register 0.

See 11.4.2 Performance Monitors Common Event Identification
Register 0, EL0 on page 11-403.

PMCEID1 7 RO UNK Performance Monitors Common Event Identification Register

1. cw

PMCCNTR c13 0 RW UNK Performance Monitors Cycle Count Register. cw

PMXEVTYPER 1 RW UNK Performance Monitors Selected Event Type Register. cw

PMCCFILTR RW 0x00000000 Performance Monitors Cycle Count Filter Register. cw

PMXEVCNTR 2 RW UNK Performance Monitors Selected Event Count Register. cw

PMUSERENR c14 0 RW 0x00000000 Performance Monitors User Enable Register. cw

PMINTENSET 1 RW UNK Performance Monitors Interrupt Enable Set Register. cw

PMINTENCLR 2 RW UNK Performance Monitors Interrupt Enable Clear Register. cw

PMOVSSET 3 RW UNK Performance Monitors Overflow Flag Status Set Register. cw

4.4.27 Security registers

The following table shows the 32-bit wide Security registers in AArch32 state.

Table 4-111 Security registers

Name CRn op1 CRm op2 Type Reset Description

SCR c1 0 c1 0 RW 0x00000000 4.5.7 Secure Configuration Register on page 4-248

SDER 1 RW 0x00000000 Secure Debug Enable Register 

NSACR 2 RW 0x00000000 4.5.8 Non-secure Access Control Register on page 4-251

SDCR c3 1 RW 0x00000000 4.5.9 Secure Debug Configuration Register on page 4-252

VBAR c12 0 c0 0 RW 0x00000000cy Vector Base Address Register cx

MVBAR 1 RW UNK Monitor Vector Base Address Register cx

ISR c1 0 RO UNK Interrupt Status Register cx

cw See the ARM® Architecture Reference Manual ARMv8 for more information.
cx See the ARM® Architecture Reference Manual ARMv8 for more information.
cy The reset value is 0x00000000 for the Secure copy of the register. You must program the Non-secure copy of the register with the required initial value, as part of
the processor boot sequence.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-234
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

4.4.28 Virtualization registers

The following table shows the Virtualization registers in AArch32 state.

Table 4-112 Virtualization registers

Name CRn op1 CRm op2 Type Reset Width Description

VPIDR c0 4 c0 0 RW -cz 32-bit Virtualization Processor ID Register. See

4.3.28 Virtualization Processor ID Register, EL2
on page 4-123.

VMPIDR 5 RO -da 32-bit 4.5.4 Virtualization Multiprocessor ID Register

on page 4-241.

HSCTLR c1 4 c0 0 RW 0x30C50838 32-bit Hyp System Control Register.

HACTLR 1 RW 0x00000000 32-bit Hyp Auxiliary Control Register. See 4.3.33 Auxiliary
Control Register, EL2 on page 4-130.

HCR c1 0 RW 0x00000000 32-bit 4.5.10 Hyp Configuration Register on page 4-254.

HDCR 1 RW 0x00000006dc 32-bit 4.5.12 Hyp Debug Control Register on page 4-261.

HCPTR 2 RW 0x000033FF 32-bit 4.5.13 Hyp Architectural Feature Trap Register

on page 4-263.

HSTR 3 RW 0x00000000 32-bit Hyp System Trap Register. See 4.3.36 Hypervisor System
Trap Register on page 4-139.

HCR2 4 RW 0x00000000 32-bit 4.5.11 Hyp Configuration Register 2 on page 4-259.

HACR 7 RW 0x00000000 32-bit 4.3.37 Hyp Auxiliary Configuration Register on page 4-141.

HTCR c2 4 c0 2 RW UNK 32-bit 4.5.16 Hyp Translation Control Register on page 4-266.

VTCR c1 2 RW UNK 32-bit Virtualization Translation Control Register. db

HTTBR - 4 c2 - RW UNK 64-bit Hyp Translation Table Base Register. db

VTTBR - 6 c2 - RW UNKdd 64-bit Virtualization Translation Table Base Register. db

HADFSR c5 4 c1 0 RW UNK 32-bit Hyp Auxiliary Data Fault Status Register. See
4.3.52 Auxiliary Fault Status Register 0, EL2 and Hyp
Auxiliary Data Fault Status Register on page 4-163.

HAIFSR 1 RW UNK 32-bit Hyp Auxiliary Instruction Fault Status Register. See
4.3.53 Auxiliary Fault Status Register 1, EL2 and Hyp
Auxiliary Instruction Fault Status Register on page 4-163.

HSR c2 0 RW UNK 32-bit Hyp Syndrome Register. See 4.3.54 Exception Syndrome
Register, EL2 on page 4-163.

HDFAR c6 4 c0 0 RW -de 32-bit Hyp Data Fault Address Register. db

HIFAR 2 RW -df 32-bit Hyp Instruction Fault Address Register. db

HPFAR 4 RW UNK 32-bit Hyp IPA Fault Address Register. db

cz The reset value is the value of the Main ID Register.

da The reset value is the value of the Multiprocessor Affinity Register.
db See the ARM® Architecture Reference Manual ARMv8 for more information.
dc The reset value for bit[7] is UNK.
dd The reset value for bits[55:48] is 0b00000000.
de The reset value is the value of the Secure copy of the DFAR register.
df The reset value is the value of the Secure copy of the IFR register.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-235
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-112 Virtualization registers (continued)

Name CRn op1 CRm op2 Type Reset Width Description

HMAIR0 c10 4 c2 0 RW UNK 32-bit Hyp Memory Attribute Indirection Register 0. db

HMAIR1 1 RW UNK 32-bit Hyp Memory Attribute Indirection Register 1. db

HAMAIR0 c3 0 RW UNK 32-bit Hyp Auxiliary Memory Attribute Indirection Register 0. See
4.3.57 Auxiliary Memory Attribute Indirection Register, EL2
on page 4-168.

HAMAIR1 1 RW UNK 32-bit Hyp Auxiliary Memory Attribute Indirection Register 1. See
4.3.57 Auxiliary Memory Attribute Indirection Register, EL2
on page 4-168.

HVBAR c12 4 c0 0 RW UNK 32-bit Hyp Vector Base Address Register. db

4.4.29 Hyp mode TLB maintenance operations

The following table shows the System instructions for TLB maintenance operations added for
Virtualization in AArch32 state. See the ARM® Architecture Reference Manual ARMv8 for more
information about these operations.

Table 4-113 Hyp mode TLB maintenance operations

Name CRn op1 CRm op2 Description

TLBIIPAS2IS c8 4 c0 1 TLB Invalidate entry by Intermediate Physical Address, Stage 2, Inner Shareable
TLBIIPAS2LIS 5 TLB Invalidate entry by Intermediate Physical Address, Stage 2, Last level, Inner
Shareable
TLBIALLHIS c3 0 Invalidate entire Hyp unified TLB Inner Shareable
TLBIMVAHIS 1 Invalidate Hyp unified TLB by VA Inner Shareable
TLBIALLNSNHIS 4 Invalidate entire Non-secure Non-Hyp unified TLB Inner Shareable
TLBIMVALHIS 5 Invalidate Unified Hyp TLB entry by VA Inner Shareable, Last level
TLBIIPAS2 c4 1 TLB Invalidate entry by Intermediate Physical Address, Stage 2
TLBIIPAS2L 5 TLB Invalidate entry by Intermediate Physical Address, Stage 2, Last level
TLBIALLH c8 4 c7 0 Invalidate entire Hyp unified TLB
TLBIMVAH 1 Invalidate Hyp unified TLB by VA
TLBIALLNSNH 4 Invalidate entire Non-secure Non-Hyp unified TLB
TLBIMVALH 5 Invalidate Unified Hyp TLB entry by VA, Last level

4.4.30 Generic Timer registers

See 9.3 Generic Timer register summary on page 9-341 for information on the Generic Timer registers.

4.4.31 Implementation defined registers

The following table shows the IMPLEMENTATION DEFINED registers in AArch32 state. These registers provide
test features and any required configuration options specific to the Cortex-A72 processor.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-236
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-114 IMPLEMENTATION DEFINED registers

Name CRn op1 CRm op2 Type Reset Width Description

AIDR c0 1 c0 7 - 0x00000000 32-bit Auxiliary ID Register. See 4.3.24 Auxiliary ID Register,

EL1 on page 4-120.

ACTLR c1 0 c0 1 - 0x00000000 32-bit Auxiliary Control Register. See 4.3.39 Auxiliary Control
Register, EL3 on page 4-143.

HACTLR 4 c0 1 RW 0x00000000 32-bit Hyp Auxiliary Control Register. See 4.3.33 Auxiliary
Control Register, EL2 on page 4-130.

HADFSR c5 4 c1 0 RW UNK 32-bit Hyp Auxiliary Data Fault Status Register. See
4.3.52 Auxiliary Fault Status Register 0, EL2 and Hyp
Auxiliary Data Fault Status Register on page 4-163.

HAIFSR 1 RW UNK 32-bit Hyp Auxiliary Instruction Fault Status Register. See
4.3.52 Auxiliary Fault Status Register 0, EL2 and Hyp
Auxiliary Data Fault Status Register on page 4-163.

L2CTLR c9 1 c0 2 RW 0x00000000dg 32-bit L2 Control Register. See 4.3.58 L2 Control Register, EL1
on page 4-168.

L2ECTLR 3 RW 0x00000000 32-bit L2 Extended Control Register. See 4.3.59 L2 Extended

Control Register, EL1 on page 4-172.

AMAIR0 c10 0 c3 0 RW UNK 32-bit Auxiliary Memory Attribute Indirection Register 0. See
4.3.56 Auxiliary Memory Attribute Indirection Register,
EL1 and EL3 on page 4-168.

AMAIR1 1 RW UNK 32-bit Auxiliary Memory Attribute Indirection Register 1. See

4.3.56 Auxiliary Memory Attribute Indirection Register,
EL1 and EL3 on page 4-168.

HAMAIR0 0 RW UNK 32-bit Hyp Auxiliary Memory Attribute Indirection Register 0.

See 4.3.57 Auxiliary Memory Attribute Indirection
Register, EL2 on page 4-168.

HAMAIR1 1 RW UNK 32-bit Hyp Auxiliary Memory Attribute Indirection Register 1.

See 4.3.57 Auxiliary Memory Attribute Indirection
Register, EL2 on page 4-168.

IL1DATA0 c15 0 c0 0 RW UNK 32-bit 4.3.62 Instruction L1 Data n Register, EL1

on page 4-176.
IL1DATA1 1

IL1DATA2 2

IL1DATA3 3

DL1DATA0 c1 0 RW UNK 32-bit 4.3.63 Data L1 Data n Register, EL1 on page 4-177.

DL1DATA1 1

DL1DATA2 2

DL1DATA3 3

DL1DATA4 4

RAMINDEX c4 0 WO - 32-bit 4.3.64 RAM Index operation on page 4-178.

dg The reset value depends on the processor implementation and the state of the L2RSTDISABLE signal.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-237
reserved.
Non-Confidential
 4 System Control
4.4 AArch32 register summary

Table 4-114 IMPLEMENTATION DEFINED registers (continued)

Name CRn op1 CRm op2 Type Reset Width Description

L2ACTLR 1 c0 0 RW 0x00000010dh 32-bit L2 Auxiliary Control Register. See L2 Auxiliary Control

Register, EL1.

CBAR 4 c0 0 RO -di 32-bit 4.5.24 Configuration Base Address Register

on page 4-274>.

CPUACTLR - 0 c15 - RW -dj 64-bit CPU Auxiliary Control Register. See 4.3.66 CPU
Auxiliary Control Register, EL1 on page 4-194.

CPUECTLR - 1 - RW -dk 64-bit CPU Extended Control Register. See 4.3.67 CPU
Extended Control Register, EL1 on page 4-206.

CPUMERRSR - 2 - RW - 64-bit CPU Memory Error Syndrome Register. See 4.3.68 CPU
Memory Error Syndrome Register, EL1 on page 4-209.

L2MERRSR - 3 - RW -dl 64-bit L2 Memory Error Syndrome Register. See 4.3.69 L2

Memory Error Syndrome Register, EL1 on page 4-210.

dh The reset value is 0x00000010 for an ACE interface and 0x00004018 for a CHI interface.
di The reset value depends on the primary input, PERIPHBASE[43:18].
dj The reset value is zero.
dk The reset value is 0x0000 001B 0000 0000.
dl The reset value for bits[63,47:40,39:32,31] is zero.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-238
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

4.5 AArch32 register descriptions

This section describes all the System registers in register number order when the processor is in AArch32
state.
Table 4-84 c0 register summary on page 4-215 to Table 4-99 64-bit register summary on page 4-225
provide cross-references to individual registers.
This section contains the following subsections:
• 4.5.1 TCM Type Register on page 4-239.
• 4.5.2 TLB Type Register on page 4-239.
• 4.5.3 Multiprocessor Affinity Register on page 4-240.
• 4.5.4 Virtualization Multiprocessor ID Register on page 4-241.
• 4.5.5 System Control Register on page 4-242.
• 4.5.6 Architectural Feature Access Control Register on page 4-247.
• 4.5.7 Secure Configuration Register on page 4-248.
• 4.5.8 Non-secure Access Control Register on page 4-251.
• 4.5.9 Secure Debug Configuration Register on page 4-252.
• 4.5.10 Hyp Configuration Register on page 4-254.
• 4.5.11 Hyp Configuration Register 2 on page 4-259.
• 4.5.12 Hyp Debug Control Register on page 4-261.
• 4.5.13 Hyp Architectural Feature Trap Register on page 4-263.
• 4.5.14 Translation Table Base Register 0 and Register 1 on page 4-265.
• 4.5.15 Translation Table Base Control Register on page 4-265.
• 4.5.16 Hyp Translation Control Register on page 4-266.
• 4.5.17 Data Fault Status Register on page 4-267.
• 4.5.18 Physical Address Register on page 4-271.
• 4.5.19 Primary Region Remap Register on page 4-272.
• 4.5.20 Memory Attribute Indirection Register 0 on page 4-273.
• 4.5.21 Normal Memory Remap Register. on page 4-273.
• 4.5.22 Memory Attribute Indirection Register 1 on page 4-274.
• 4.5.23 FCSE Process ID Register on page 4-274.
• 4.5.24 Configuration Base Address Register on page 4-274.

4.5.1 TCM Type Register

The processor does not implement instruction or data Tightly Coupled Memory (TCM), so this register is
always RES0.

4.5.2 TLB Type Register

The TLBTR characteristics are:

Purpose
Provides information about the TLB implementation.
Usage constraints
The accessibility to the TLBTR by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The TLBTR is Common to Secure and Non-secure states.
Attributes
See the register summary in Table 4-84 c0 register summary on page 4-215.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-239
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

The following figure shows the TLBTR bit assignments.

31 1 0

RES0 nU

Figure 4-79 TLBTR bit assignments

The following table shows the TLBTR bit assignments.

Table 4-115 TLBTR bit assignments

Bits Name Function

[31:1] - Reserved, RES0.

[0] nU Not Unified. Indicates whether the implementation has a unified TLB. The value is:
0 Processor has a unified TLB.

To access the TLBTR in AArch32 state, read the CP15 register with:
MRC p15, 0, <Rt>, c0, c0, 3; Read TLB Type Register

4.5.3 Multiprocessor Affinity Register

The MPIDR characteristics are:

Purpose
Provides an additional core identification mechanism for scheduling purposes in a cluster.
Usage constraints
The accessibility to the MPIDR by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The MPIDR is:
• Common to Secure and Non-secure states.
• Architecturally mapped to the AArch64 MPIDR_EL1[31:0] register.
Attributes
See the register summary in Table 4-84 c0 register summary on page 4-215.
The following figure shows the MPIDR bit assignments.
31 30 29 25 24 23 16 15 8 7 2 1 0

U Reserved Cluster ID Aff2 Cluster ID Aff1 Reserved

RES1 MT CPU ID

Figure 4-80 MPIDR bit assignments

The following table shows the MPIDR bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-240
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-116 MPIDR bit assignments

Bits Name Function

[31] - Reserved, RES1.

[30] U Indicates a single processor system, as distinct from core 0 in a cluster. This value
is:
0 Processor is part of a multicore system.

[29:25] - Reserved, RES0.

[24] MT Indicates whether the lowest level of affinity consists of logical cores that are
implemented using a multi-threading type approach:
0 Performance of cores at the lowest affinity level is largely independent.
1 Performance of cores at the lowest affinity level is very interdependent.

[23:16] Cluster ID Aff2 Indicates the value read in at reset, from the CLUSTERIDAFF2 configuration
signal. It identifies a Cortex-A72 device in a system with more than one Cortex-
A72 device present.

[15:8] Cluster ID Aff1 Indicates the value read in at reset, from the CLUSTERIDAFF1 configuration
signal. It identifies a Cortex-A72 device in a system with more than one Cortex-
A72 device present.

[7:2] - Reserved, RES0.

[1:0] CPU ID Indicates the core number in the Cortex-A72 device. The possible values are:
0x0 An MPCore device with one core only.
0x0, 0x1 A Cortex-A72 device with two cores.
0x0, 0x1, 0x2 A Cortex-A72 device with three cores.
0x0, 0x1, 0x2, 0x3 A Cortex-A72 device with four cores.

To access the MPIDR in AArch32 state, read the CP15 registers with:
MRC p15, 0, <Rt>, c0, c0, 5; Read Multiprocessor Affinity Register

Related information
4.3.2 Multiprocessor Affinity Register, EL1 on page 4-90.

4.5.4 Virtualization Multiprocessor ID Register

The VMPIDR characteristics are:

Purpose
Holds the value of the Virtualization Multiprocessor ID. This is the value returned by Non-
secure EL1 reads of MPIDR. .
Usage constraints
The accessibility to the VMPIDR in AArch32 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW -

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-241
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Configurations
The VMPIDR is:
• A Banked EL2 register.
• Architecturally mapped to the AArch64 VMPIDR_EL2 register.
Attributes
See the register summary in Table 4-84 c0 register summary on page 4-215.
The following figure shows the VMPIDR bit assignments.
31 0

VMPIDR

Figure 4-81 VMPIDR bit assignments

The following table shows the VMPIDR bit assignments.

Table 4-117 VMPIDR bit assignments

Bits Name Function

[31:0] VMPIDR MPIDR value returned by Non-secure EL1 reads of the MPIDR. For information on the subdivision of this value,
see 4.5.3 Multiprocessor Affinity Register on page 4-240.

To access the VMPIDR, read or write the CP15 register with:

MRC p15, 4, <Rt>, c0, c0, 5; Read Virtualization Multiprocessor ID Register
MCR p15, 4, <Rt>, c0, c0, 5; Write Virtualization Multiprocessor ID Register

Related information
4.5.3 Multiprocessor Affinity Register on page 4-240.
4.3.29 Virtualization Multiprocessor ID Register, EL2 on page 4-124.

4.5.5 System Control Register

The SCTLR characteristics are:
Purpose
Provides the top-level control of the system, including its memory system.
Usage constraints
The accessibility to the SCTLR by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Control bits in the SCTLR that are not applicable to a VMSA implementation read as the value
that most closely reflects that implementation, and ignore writes.
Some bits in the register are read-only. These bits relate to non-configurable features of an
implementation, and are provided for compatibility with other versions of the architecture.
Write access to the Secure copy of SCTLR is disabled when the CP15SDISABLE signal is
HIGH.
Configurations
The SCTLR is Banked for Secure and Non-secure states.
The architectural mapping of the SCTLR is:
• The Non-secure SCTLR is mapped to the AArch64 SCTLR_EL1.
• The Secure SCTLR is mapped to the AArch64 SCTLR_EL3.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-242
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Attributes
See the register summary in Table 4-85 c1 register summary on page 4-217.
The following figure shows the SCTLR bit assignments.
31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

RES0 RES1 RES0 V I RES0 RES1 C A M

TRE RES0 nTWI RES1 CP15BEN

AFE EE RES0 THEE
TE nTWE ITD
RES0 WXN SED
UWXN
RES0

Figure 4-82 SCTLR bit assignments

The following table shows the SCTLR bit assignments.

Table 4-118 SCTLR bit assignments

Bits Name Access Function

[31] - - Reserved, RES0.

[30] TE Banked Thumb Exception enable. This bit controls whether

exceptions are taken in ARM or Thumb state:
0 Exceptions, including reset, taken in ARM state.
1 Exceptions, including reset, taken in Thumb state.

The primary input CFGTE defines the reset value of the TE

bit of the Secure Banked register.

[29] AFE Banked Access flag enable. This bit enables use of the AP[0] bit in
the translation table descriptors as the Access flag. It also
restricts access permissions in the translation table
descriptors to the simplified model as described in the ARM®
Architecture Reference Manual ARMv8. In the translation
table descriptors, AP[0] is:
0 An access permissions bit. The full range of access
permissions is supported. No access flag is implemented.
This is the reset value.
1 The Access flag. Only the simplified model for access
permissions is supported.

When TTBCR.EAE is set to 1, to enable use of the Long-

descriptor translation table format, this bit is UNK/RES1.
This bit is permitted to be cached in a TLB.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-243
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-118 SCTLR bit assignments (continued)

Bits Name Access Function

[28] TRE Banked TEX remap enable. This bit enables remapping of the
TEX[2:1] bits for use as two translation table bits that can be
managed by the operating system. Enabling this remapping
also changes the scheme that describes the memory region
attributes in the VMSA. The possible values are:
0 TEX remap disabled. TEX[2:0] are used, with the C and
B bits, to describe the memory region attributes. This is
the reset value.
1 TEX remap enabled. TEX[2:1] are reassigned for use as
bits managed by the operating system. The TEX[0], C
and B bits describe the memory region attributes, with
the MMU remap registers.

When TTBCR.EAE is set to 1, to enable use of the Long-

descriptor translation table format, this bit is UNK/RES1.
This bit is permitted to be cached in a TLB.

[27:26] - - Reserved, RES0.

[25] EE Banked Exception Endianness. The value of this bit defines the value
of the CPSR.E bit on entry to an exception vector, including
reset. This value also indicates the endianness of the
translation table data for translation table lookups. The
values are:
0 Little endian.
1 Big endian.

The primary input CFGEND defines the reset value of the

EE bit of the Secure Banked register.

[24] - - Reserved, RES0.

[23:22] - - Reserved, RES1.

[21] - - Reserved, RES0.

[20] UWXN Banked Unprivileged write permission implies EL1 Execute Never
(XN). You can use this bit to require all memory regions with
unprivileged write permissions are treated as XN for accesses
from software executing at EL1. Regions with unprivileged
write permission are:
0 Not forced to be XN. This is the reset value.
1 Forced to be XN for accesses from software executing at
EL1.

This bit is permitted to be cached in a TLB.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-244
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-118 SCTLR bit assignments (continued)

Bits Name Access Function

[19] WXN Banked Write permission implies Execute Never (XN). You can use
this bit to require all memory regions with write permissions
are treated as XN. Regions with write permission are:
0 Not forced to be XN. This is the reset value.
1 Forced to be XN.

This bit is permitted to be cached in a TLB.

[18] nTWE Banked WFE trap. The values are:

0 A WFE instruction executed at EL0 that causes suspended
execution as if the event register is not set and there is no
pending WFE wake-up event. It is treated as UNDEFINED.
1 WFE instructions executed as normal. This is the reset
value.

[17] - - Reserved, RES0.

[16] nTWI Banked WFI trap. The values are:

0 A WFI instruction executed at EL0 that causes suspended
execution as if there is not a pending WFI wake-up
event. It is treated as UNDEFINED.
1 WFE instructions executed as normal. This is the reset
value.

[15:14] - - Reserved, RES0.

[13] V Banked Vectors bit. This bit selects the base address of the exception
vectors:
0 Normal exception vectors, base address 0x00000000.
This base address can be remapped.
1 High exception vectors, base address 0xFFFF0000. This
base address is never remapped.

The primary input VINITHI defines the reset value of the V

bit of the Secure Banked register.

[12] I Banked Instruction Cache enable. This is a global enable bit for
Instruction Caches:
0 Instruction Caches disabled. This is the reset value.
1 Instruction Caches enabled.

[11] - - Reserved, RES1.

[10:9] - - Reserved, RES0.

[8] SED Banked SETEND instruction disable. The values are:

0 SETEND instruction is enabled. This is the reset value.
1 SETEND instruction is UNALLOCATED.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-245
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-118 SCTLR bit assignments (continued)

Bits Name Access Function

[7] ITD Banked IT instruction disable. The values are:

0 IT instruction functionality is enabled. This is the reset
value.
1 All encodings of the IT instruction are UNDEFINED when
either:
• hw[3:0] are not equal to 0b1000.
• IT instructions with a subsequent 32-bit instruction.
• Subsequent PC reading or writing instruction.

[6] THEE Banked ThumbEE enable. This value is:

0 ThumbEE is not implemented.

[5] CP15BEN Banked AArch32 CP15 barrier enable. The values are:
0 CP15 barrier operations disabled. Their encodings are
UNDEFINED.

1 CP15 barrier operations enabled. This is the reset value.

[4:3] - - Reserved, RES1.

[2] C Banked Cache enable. This is a global enable bit for data and unified
caches:
0 Data and unified caches disabled. This is the reset value.
1 Data and unified caches enabled.

[1] A Banked Alignment check enable. This is the enable bit for Alignment
fault checking:
0 Alignment fault checking disabled. This is the reset
value.
1 Alignment fault checking enabled.

[0] M Banked MMU enable. This is a global enable bit for the EL1 and
EL0 stage 1 MMU:
0 EL1 and EL0 stage 1 MMU disabled. This is the reset
value.
1 EL1 and EL0 stage 1 MMU enabled.

To access the SCTLR in AArch32 state, read or write the CP15 register with:
MRC p15, 0, <Rt>, c1, c0, 0; Read System Control Register
MCR p15, 0, <Rt>, c1, c0, 0; Write System Control Register

To access the SCTLR_EL1 in AArch64 state, read or write the register with:
MRS <Xt>, SCTLR_EL1; Read System Control Register
MSR SCTLR_EL1, <Xt>; Write System Control Register

To access the SCTLR_EL3 in AArch64 state, read or write the register with:
MRS <Xt>, SCTLR_EL3; Read System Control Register
MSR SCTLR_EL3, <Xt>; Write System Control Register

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-246
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Related information
4.3.30 System Control Register, EL1 on page 4-125.
4.3.38 System Control Register, EL3 on page 4-141.

4.5.6 Architectural Feature Access Control Register

The CPACR characteristics are:

Purpose
Controls access to the CP10 and CP11 coprocessors. It also enables software to check for the
presence of coprocessors CP10 to CP11.
Usage constraints
The accessibility to the CPACR by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

The CPACR has no effect on instructions executed in Hyp mode.

Configurations
The CPACR is:
• Common to the Secure and Non-secure states.
• Architecturally mapped to the AArch64 CPACR_EL1 register.

Note
The NSACR controls Non-secure access to the CPACR fields.

Attributes
See the register summary in Table 4-85 c1 register summary on page 4-217.
The following figure shows the CPACR bit assignments.
31 30 29 28 27 24 23 22 21 20 19 0

RES0 RES0 cp11 cp10 RES0

TRCDIS
ASEDIS

Figure 4-83 CPACR bit assignments

The following table shows the CPACR bit assignments.

Table 4-119 CPACR bit assignments

Bits Name Function

[31] ASEDIS Disables Advanced SIMD functionality:

0 All Advanced SIMD and FP instructions execute normally. This is the reset value.
1 All instruction encodings that are part of Advanced SIMD, but not FP instructions, are UNDEFINED.

[30:29] - Reserved, RES0.

[28] TRCDIS Disable CP14 access to trace registers:

0 CP14 access to trace registers is not supported. This bit is RES0.

[27:24] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-247
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-119 CPACR bit assignments (continued)

Bits Name Function

[23:22] cp11 Defines the access rights for coprocessor 11. The values are:
0b00 Access denied. Any attempt to access the coprocessor generates an Undefined Instruction exception. This
is the reset value.
0b01 Access at EL1 or higher only. Any attempt to access the coprocessor from software executing at EL0
generates an Undefined Instruction exception.
0b10 Reserved.
0b11 Full access. The meaning of full access is defined by the appropriate coprocessor.

If NSACR[11:10] is 0b00 in Non-secure state, these bits are RES0.

[21:20] cp10 Defines the access rights for coprocessor 10. The values are:
0b00 Access denied. Any attempt to access the coprocessor generates an Undefined Instruction exception. This
is the reset value.
0b01 Access at EL1 or higher only. Any attempt to access the coprocessor from software executing at EL0
generates an Undefined Instruction exception.
0b10 Reserved.
0b11 Full access. The meaning of full access is defined by the appropriate coprocessor.

If NSACR[11:10] is 0b00 in Non-secure state, these bits are RES0.

[19:0] - Reserved, RES0.

Note
If the values of the cp11 and cp10 fields are not the same, the behavior is same as if both fields were set
to the value of cp10, in all respects other than the value read back by explicitly reading cp11.

To access the CPACR in AArch32 state, read or write the CP15 register with:
MRC p15, 0, <Rt>, c1, c0, 2; Read Architectural Feature Access Control Register
MCR p15, 0, <Rt>, c1, c0, 2; Write Architectural Feature Access Control Register

Related information
4.3.32 Architectural Feature Access Control Register, EL1 on page 4-129.
4.5.8 Non-secure Access Control Register on page 4-251.

4.5.7 Secure Configuration Register

The SCR characteristics are:

Purpose Defines the configuration of the current Security state. It specifies:
• The Security state of the processor, Secure or Non-secure.
• What mode the processor branches to, if an IRQ, FIQ, or external abort occurs.
• Whether the CPSR.F and CPSR.A bits can be modified when SCR.NS is 1.
If EL3 is using AArch64, accesses to this register from Secure EL1 using AArch32
are trapped to EL3.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-248
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Usage The accessibility to the SCR by Exception level is:

constraints
EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - TRAP - RW RW

Configurations The SCR is a Restricted access register that exists only in the Secure state.
The SCR is mapped to the AArch64 SCR_EL3 register.
Attributes See the register summary in Table 4-85 c1 register summary on page 4-217.

The following figure shows the SCR bit assignments.

31 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

RES0 RES0 EA NS

TWI FW IRQ
TWE AW FIQ
nET
SCD
HCE
SIF

Figure 4-84 SCR bit assignments

The following table shows the SCR bit assignments.

Table 4-120 SCR bit assignments

Bits Name Function

[31:14] - Reserved, RES0.

[13] TWE Trap WFE instructions. The possible values are:

0 WFE instructions are not trapped. This is the reset value.
1 WFE instructions executed in any mode other than Monitor mode that
would cause suspended execution as if the event register is not set, there
is not a pending WFE wake-up event and the instruction does not cause
another exception, is trapped to Monitor mode using the UNDEFINED
exception vector.

[12] TWI Trap WFI instructions. The possible values are:

0 WFI instructions are not trapped. This is the reset value.
1 WFI instructions executed in any mode other than Monitor mode that
would cause suspended execution, as if there is no pending WFI wake-
up event and the instruction does not cause another exception, is trapped
to Monitor mode using the UNDEFINED exception vector.

[11:10] - Reserved, RES0.

[9] SIF Secure Instruction Fetch. When the processor is in Secure state, this bit
disables instruction fetches from Non-secure memory. The possible values
are:
0 Secure state instruction fetches from Non-secure memory permitted.
This is the reset value.
1 Secure state instruction fetches from Non-secure memory not permitted.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-249
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-120 SCR bit assignments (continued)

Bits Name Function

[8] HCE Hyp Call enable. This bit enables the use of HVC instruction. The possible
values are:
0 The HVC instruction is UNDEFINED in any mode. This is the reset value.
1 The HVC instruction enabled in Non-secure EL1 or EL2, and performs a
Hyp Call.

[7] SCD Secure Monitor Call disable. This bit causes the SMC instruction to be
UNDEFINED in all privileged modes. The possible values are:

0 The SMC instruction executes normally from privileged modes, and

performs a Secure Monitor Call. This is the reset value.
1 The SMC instruction is UNDEFINED in any mode.

A trap of the SMC instruction to Hyp mode from Non-secure EL1 takes
priority over the value of this bit. See the ARM® Architecture Reference
Manual ARMv8 for more information.

[6] nET Not Early Termination. This bit disables early termination.
This bit is not implemented, RES0.

[5] AW A bit writable. This bit controls whether CPSR.A can be modified in Non-
secure state. For the Cortex-A72 processor:
• This bit has no effect on whether CPSR.A can be modified in Non-
secure state. The AW bit can be modified in either Security state.
• This bit, with the HCR.AMO bit, determines whether CPSR.A has any
effect on exceptions that are routed to a Non-secure mode.

[4] FW F bit writable. This bit controls whether CPSR.F can be modified in Non-
secure state. For the Cortex-A72 processor:
• This bit has no effect on whether CPSR.F can be modified in Non-
secure state. The FW bit can be modified in either Security state.
• This bit, with the HCR.FMO bit, determines whether CPSR.F has any
effect on exceptions that are routed to a Non-secure mode.

[3] EA External Abort handler. This bit controls which mode takes external aborts.
The possible values are:
0 External aborts taken in Abort mode. This is the reset value.
1 External aborts taken in Monitor mode.

[2] FIQ FIQ handler. This bit controls which mode takes FIQ exceptions. The
possible values are:
0 FIQs taken in FIQ mode. This is the reset value.
1 FIQs taken in Monitor mode.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-250
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-120 SCR bit assignments (continued)

Bits Name Function

[1] IRQ IRQ handler. This bit controls which mode takes IRQ exceptions. The
possible values are:
0 IRQs taken in IRQ mode. This is the reset value.
1 IRQs taken in Monitor mode.

[0] NS Non-secure bit. Except when the processor is in Monitor mode, this bit
determines the Security state of the processor. The possible values are:
0 Secure. This is the reset value.
1 Non-secure.

Note
When the processor is in Monitor mode, it is always in Secure state,
regardless of the value of the NS bit.

To access the SCR in AArch32 state, read or write the CP15 register with:
MRC p15, 0, <Rt>, c1, c1, 0; Read Secure Configuration Register data
MCR p15, 0, <Rt>, c1, c1, 0; Write Secure Configuration Register data

4.5.8 Non-secure Access Control Register

The NSACR characteristics are:

Purpose
Defines the Non-secure access permission to the CP10 and CP11 coprocessors and controls
Non-secure Advanced SIMD functionality.
Usage constraints
The accessibility to the NSACR by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO TRAP RO RO RW

If EL3 is using AArch64, accesses to this register from Secure EL1 using AArch32 are trapped
to EL3.
Configurations
The NSACR:
• Is a Restricted access register that exists only in the Secure state but can be read from the
Non-secure state.
• Functionality is replaced by the behavior in the CPTR_EL3 register in AArch64 state.
When EL3 is using AArch64, reads of the NSACR from Non-secure EL2 or Non-secure EL1
using AArch32, return a fixed value of 0x00000C00.
Attributes
See the register summary in Table 4-85 c1 register summary on page 4-217.
The following figure shows the NSACR bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-251
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

31 21 20 19 16 15 14 12 11 10 9 0

RES0 RES0 RES0 RES0

NSTRCDIS NSASEDIS cp11 cp10

Figure 4-85 NSACR bit assignments

The following table shows the NSACR bit assignments.

Table 4-121 NSACR bit assignments

Bits Name Function

[31:21] - Reserved, RES0.

[20] NSTRCDIS Disable Non-secure access to CP14 trace registers:

0 CP14 access to trace registers is not supported. This bit is RES0.

[19:16] - Reserved, RES0.

[15] NSASEDIS Disables Non-secure Advanced SIMD functionality. The values are:
0 This bit has no effect on the ability to write to the CAPCR.ASEDIS bit. This is the reset value.
1 When executing in Non-secure state, the CPACR.ASEDIS bit is RES1.

[14:12] - Reserved, RES0.

[11] cp11 Non-secure access to coprocessor 11 enable. The values are:

0 Secure access only. Any attempt to access coprocessor 11 in Non-secure state results in an Undefined
Instruction exception. If the processor is in Non-secure state, the corresponding bits in the CPACR ignore
writes and read as 0b00, access denied. This is the reset value.
1 Access from any Security state.

[10] cp10 Non-secure access to coprocessor 10 enable. The values are:

0 Secure access only. Any attempt to access coprocessor 10 in Non-secure state results in an Undefined
Instruction exception. If the processor is in Non-secure state, the corresponding bits in the CPACR ignore
writes and read as 0b00, access denied. This is the reset value.
1 Access from any Security state.

[9:0] - Reserved, RES0.

To access the NSACR in AArch32 state, read or write the CP15 register with:
MRC p15, 0, <Rt>, c1, c1, 2; Read Non-secure Access Control Register data
MCR p15, 0, <Rt>, c1, c1, 2; Write Non-secure Access Control Register data

Related information
4.3.40 Architectural Feature Trap Register, EL3 on page 4-145.

4.5.9 Secure Debug Configuration Register

The SDCR characteristics are:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-252
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Purpose
Controls the trapping to Hyp mode of Secure accesses, at EL1 or lower, to functions provided by
the debug and trace architectures.
If EL3 is using AArch64, accesses to this register from Secure EL1 using AArch32 are trapped
to EL3.
Usage constraints
The accessibility to the SDCR by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - TRAP - RW RW

Configurations
The SDCR is a Restricted access register that only exists in the Secure state.
The SDCR is mapped to the AArch64 MDCR_EL3 register.
Attributes
See the register summary in Table 4-85 c1 register summary on page 4-217.
The following figure shows the SDCR bit assignments.
31 22 21 20 19 18 17 16 15 14 13 0

RES0 RES0 SPD RES0

EPMAD EDAD RES0

SPME

Figure 4-86 SDCR bit assignments

The following table shows the SDCR bit assignments

Table 4-122 SDCR bit assignments

Bits Name Function

[31:22] - Reserved, RES0.

[21] EPMAD Disables access to the performance monitor configuration registers by an external debugger:
0 External debugger access to the performance monitor configuration registers enabled. This is the reset value.
1 External debugger access to the performance monitor configuration registers disabled, unless overridden by
the authentication interface.

Resets to 0 on Warm reset.

[20] EDAD Disables access to the breakpoint and watchpoint registers by an external debugger:
0 External debugger access to the breakpoint and watchpoint registers enabled. This is the reset value.
1 External debugger access to the breakpoint and watchpoint registers disabled, unless overridden by the
authentication interface.

Resets to 0 on Warm reset.

[19:18] - Reserved, RES0.

[17] SPME Enables Secure performance monitor:

0 Performance monitors disabled in Secure state, no events are counted. This is the reset value.
1 Performance monitors enabled in Secure state.

Resets to 0 on Warm reset.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-253
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-122 SDCR bit assignments (continued)

Bits Name Function

[16] - Reserved, RES0.

[15:14] SPDdm AArch32 Secure privileged debug. Enables or disables debug exceptions from Secure state if Secure EL1 is using
AArch32, other than Software breakpoint instructions. The possible values are:
0b00 Legacy mode. Debug exceptions from Secure EL1 are enabled if
AArch32SelfHostedSecurePrivilegedInvasiveDebugEnabled() is true.
0b10 Secure privileged debug disabled. Debug exceptions from Secure EL1 are disabled.
0b11 Secure privileged debug enabled. Debug exceptions from Secure EL1 are enabled.

The value 0b01 is reserved.

Note
If debug exceptions from Secure EL1 are enabled, then debug exceptions from Secure EL0 are also enabled.
Otherwise, debug exceptions from Secure EL0 are enabled only if SDER32_EL3.SUIDEN is 1.
Ignored if Secure EL1 is using AArch64 and in Non-secure state. Debug exceptions from Software breakpoint
instruction debug events are always enabled.

Resets to 0 on Warm reset.

[13:0] - Reserved, RES0.

To access the SDCR in AArch32 state, read or write the CP15 register with:
MRC p15, 0, <Rt>, c1, c3, 1; Read Secure Debug Configuration Register
MCR p15, 0, <Rt>, c1, c3, 1; Write Secure Debug Configuration Register

4.5.10 Hyp Configuration Register

The HCR characteristics are:

Purpose Provides configuration controls for virtualization, including defining whether various
Non-secure operations are trapped to Hyp mode.
Usage The accessibility to the HCR in AArch32 state by Exception level is:
constraints
EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW -

Configurations The HCR is:

• A Banked EL2 register
• Architecturally mapped to the AArch64 HCR_EL2[31:0] register.
Attributes See the register summary in Table 4-85 c1 register summary on page 4-217.

The following figure shows the HCR bit assignments.

dm SPD only applies in Secure state and when either Secure EL1 or EL3 is using AArch32.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-254
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

RES0 BSU

RES0 TGE DC VM
TRVM TVM TWI SWIO
TTLB TWE PTW
TPU TID0 FMO
TPC TID1 IMO
TSW TID2 AMO
TAC TID3 VF
TIDCP VI
TSC VA
FB

Figure 4-87 HCR bit assignments

The following table shows the HCR bit assignments.

Table 4-123 HCR bit assignments

Bits Name Function

[31] - Reserved, RES0.

[30] TRVM Trap Read of Virtual Memory controls. When 1, this causes reads to the
EL1 virtual memory control registers from EL1 to be trapped to EL2.
This covers the following registers:
AArch32 SCTLR, TTBR0, TTBR1, TTBCR, DACR, DFSR, IFSR,
DFAR, IFAR, ADFSR, AIFSR, PRRR/MAIR0, NMRR/
MAIR1, AMAIR0, AMAIR1, and CONTEXTIDR.

The reset value is 0.

[29:28] - Reserved, RES0.

[27] TGE Trap general exceptions. When this bit is set to 1, and the processor is
executing at EL0 in Non-secure state, Undefined Instruction exceptions,
Supervisor Call exceptions, synchronous External aborts and some
Alignment faults are taken in Hyp mode.
The SCTLR.M bit is treated as being 0 regardless of its actual state,
other than for the purpose of reading the bit.
When the processor is executing at EL1 in Non-secure state, and this bit
is set to 1, the Illegal Exception Return mechanism is invoked.
The reset value is 0.

[26] TVM Trap Virtual Memory controls. When 1, this causes writes to the EL1
virtual memory control registers from EL1 to be trapped to EL2. This
covers the following registers:
AArch32 SCTLR, TTBR0, TTBR1, TTBCR, DACR, DFSR, IFSR,
DFAR, IFAR, ADFSR, AIFSR, PRRR/MAIR0, NMRR/
MAIR1, AMAIR0, AMAIR1, and CONTEXTIDR.

The reset value is 0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-255
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-123 HCR bit assignments (continued)

Bits Name Function

[25] TTLB Trap TLB maintenance instructions. When 1, this causes TLB
maintenance instructions executed from EL1 that are not UNDEFINED to
be trapped to EL2. This covers the following instructions:
AArch32 TLBIALLIS, TLBIMVAIS, TLBIASIDIS, TLBIMVAAIS,
ITLBIALL, DTLBIALL, TLBIALL, ITLBIMVA, DTLBIMVA,
TLBIMVA, ITLBIASID, DTLBIASID, TLBIASID,
TLBIMVAA, TLBIMVALIS, TLBIMVAALIS, TLBIMVAL, and
TLBIMVAAL.

The reset value is 0.

[24] TPU Trap Cache maintenance instructions to Point of Unification. When 1,

this causes Cache maintenance instructions to the point of unification
executed from EL1 or EL0 that are not UNDEFINED to be trapped to EL2.
This covers the following instructions:
AArch32 ICIMVAU, ICIALLU, ICIALLUIS, and DCCMVAU.

The reset value is 0.

[23] TPC Trap Data/Unified Cache maintenance operations to Point of Coherency.

When 1, this causes Data or Unified Cache maintenance instructions by
address to the point of coherency executed from EL1 or EL0 that are not
UNDEFINED to be trapped to EL2. This covers the following instructions:

AArch32 DCIMVAC, DCCIMVAC, and DCCMVAC.

The reset value is 0.

[22] TSW Trap Data/Unified Cache maintenance operations by Set/Way. When 1,

this causes Data or Unified Cache maintenance instructions by set/way
executed from EL1 that are not UNDEFINED to be trapped to EL2. This
covers the following instructions:
AArch32 DCISW, DCCSW, and DCCISW.

The reset value is 0.

[21] TAC Trap ACTLR accesses. When this bit is set to 1, any valid Non-secure
access to the ACTLR is trapped to Hyp mode.
The reset value is 0.

[20] TIDCP Trap Implementation Dependent functionality. When 1, this causes

accesses to the following instruction set space executed from EL1 to be
trapped to EL2:
AArch32 All CP15 MCR and MRC instructions as follows:
• CRn is 9, op1 is 0-7, CRm is c0, c1, c2, c5, c6, c7, or
c8, and op2 is 0-7.
• CRn is 10, op1 is 0-7, CRm is c0, c1, c4, or c8, and op2
is 0-7.
• CRn is 11, op1 is 0-7, CRm is c0 to c8, or c15, and op2
is 0-7.

The reset value is 0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-256
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-123 HCR bit assignments (continued)

Bits Name Function

[19] TSC Trap SMC instruction. When this bit is set to 1, any attempt from Non-
secure EL1 to execute an SMC instruction, that passes its condition check
if it is conditional, is trapped to Hyp mode.
The reset value is 0.

[18] TID3 Trap ID Group 3. When 1, this causes reads to the following registers
executed from EL1 to be trapped to EL2:
AArch32 ID_PFR0, ID_PFR1, ID_DFR0, ID_AFR0, ID_MMFR0,
ID_MMFR1, ID_MMFR2, ID_MMFR3, ID_ISAR0,
ID_ISAR1, ID_ISAR2, ID_ISAR3, ID_ISAR4, ID_ISAR5,
MVFR0, MVFR1, and MVFR2 and MRC instructions to the
following locations:
• op1 is 0, CRn is 0, CRm is c3, c4, c5, c6, or c7, and op2
is 0 or 1.
• op1 is 0, CRn is 0, CRm is c3, and op2 is 2.
• op1 is 0, CRn is 0, CRm is 5, and op2 is 4 or 5.

The reset value is 0.

[17] TID2 Trap ID Group 2. When 1, this causes reads (or writes to CSSELR/
CSSELR_EL1) to the following registers executed from EL1 or EL0 if
not UNDEFINED to be trapped to EL2:
AArch32 CTR, CCSIDR, CLIDR, and CSSELR.

The reset value is 0.

[16] TID1 Trap ID Group 1. When 1, this causes reads to the following registers
executed from EL1 to be trapped to EL2:
AArch32 TCMTR, TLBTR, AIDR, and REVIDR.

The reset value is 0.

[15] TID0 Trap ID Group 0. When 1, this causes reads to the following registers
executed from EL1 or EL0 if not UNDEFINED to be trapped to EL2:
AArch32 FPSID and JIDR.

The reset value is 0.

[14] TWE Traps WFE instruction if it would cause suspension of execution. For
example, if there is no pending WFE event:
0 WFE instruction is not trapped. This is the reset value.
1 WFE instruction executed in Non-secure EL1 or EL0 is trapped to
EL2.

[13] TWI Traps WFI instruction if it would cause suspension of execution. For
example, if there is no pending WFI event:
0 WFI instruction is not trapped. This is the reset value.
1 WFI instruction executed in Non-secure EL1 or EL0 is trapped to
EL2.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-257
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-123 HCR bit assignments (continued)

Bits Name Function

[12] DC Default Cacheable. When this bit is set to 1 the memory type and
attributes determined by the stage 1 translation is Normal, Non-
shareable, Inner Write-Back Write-Allocate, Outer Write-Back Write-
Allocate.
When executing in a Non-secure mode other than Hyp mode and the
HCR.DC bit is set, the processor behavior is consistent with the
behavior when:
• The SCTLR.M bit is clear, regardless of the actual value of the
SCTLR.M bit.
— An explicit read of the SCTLR.M bit returns its actual value.
• The HCR.VM bit is set, regardless of the actual value of the
HCR.VM bit.
— An explicit read of the HCR.VM bit returns its actual value.
The reset value is 0.

[11:10] BSU Barrier Shareability upgrade. The value in this field determines the
minimum shareability domain that is applied to any barrier executed
from EL1 or EL0. The values are:
0b00 No effect.
0b01 Inner Shareable.
0b10 Outer Shareable.
0b11 Full System.

The reset value is 0.

[9] FB Force broadcast. When 1, this causes the following instructions to be

broadcast within the Inner Shareable domain when executed from Non-
secure EL1:
AArch32 ITLBIALL, DTLBIALL, TLBIALL, ITLBIMVA, DTLBIMVA,
TLBIMVA, ITLBIASID, DTLBIASID, TLBIASID,
TLBIMVAA, BPIALL, and ICIALLU.

The reset value is 0.

[8] VA Virtual Asynchronous Abort exception. Setting this bit signals a virtual
Asynchronous Abort exception to the Guest OS, when the AMO bit is
set to 1 and the processor is executing in Non-secure state at EL0 or
EL1.
The Guest OS cannot distinguish the virtual exception from the
corresponding physical exception.
The reset value is 0.

[7] VI Virtual IRQ exception. Setting this bit signals a virtual IRQ exception to
the Guest OS, when the IMO bit is set to 1 and the processor is
executing in Non-secure state at EL0 or EL1.
The Guest OS cannot distinguish the virtual exception from the
corresponding physical exception.
The reset value is 0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-258
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-123 HCR bit assignments (continued)

Bits Name Function

[6] VF Virtual FIQ exception. Setting this bit signals a virtual FIQ exception to
the Guest OS, when the FMO bit is set to 1 and the processor is
executing in Non-secure state at EL0 or EL1.
The Guest OS cannot distinguish the virtual exception from the
corresponding physical exception.
The reset value is 0.

[5] AMO Asynchronous Abort Mask Override. When this bit is set to 1, it
overrides the effect of CPSR.A, and enables virtual exception signaling
by the VA bit.
The reset value is 0.

[4] IMO IRQ Mask Override. When this bit is set to 1, it overrides the effect of
CPSR.I, and enables virtual exception signaling by the VI bit.
The reset value is 0.

[3] FMO FIQ Mask Override. When this bit is set to 1, it overrides the effect of
CPSR.F, and enables virtual exception signaling by the VF bit.
The reset value is 0.

[2] PTW Protected Table Walk. When 1, if the stage 2 translation of a translation
table access made as part of a stage 1 translation table walk at Non-
secure EL0 or EL1 maps that translation table access to Device memory,
the access is faulted as a stage 2 Permission fault.
The reset value is 0.

[1] SWIO Set/Way Invalidation Override. When 1, this causes EL1 execution of
the Data Cache Invalidate by Set/Way instruction to be treated as Data
Cache Clean and Invalidate by Set/Way. The affected instructions are:
AArch32 DCISW is executed as DCCISW.

The reset value is 0.

[0] VM Second stage of Translation enable. When 1, this enables the second
stage of translation for execution in EL1 and EL0. This bit is permitted
to be cached in a TLB.
The reset value is 0.

To access the HCR in AArch32 state, read or write the CP15 register with:
MRC p15, 4, <Rt>, c1, c1, 0; Read Hyp Configuration Register
MCR p15, 4, <Rt>, c1, c1, 0; Write Hyp Configuration Register

Related information
4.3.34 Hypervisor Configuration Register, EL2 on page 4-131.

4.5.11 Hyp Configuration Register 2

The HCR2 characteristics are:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-259
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Purpose
Provides additional configuration controls for virtualization.
Usage constraints
The accessibility to the HCR2 in AArch32 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW -

Configurations
The HCR2 is:
• A Banked EL2 register.
• Architecturally mapped to the AArch64 HCR_EL2[63:31] register.
Attributes
See the register summary in Table 4-85 c1 register summary on page 4-217.
The following figure shows the HCR2 bit assignments.
31 2 1 0

RES0 ID

CD

Figure 4-88 HCR2 bit assignments

The following table shows the HCR2 bit assignments.

Table 4-124 HCR2 bit assignments

Bits Name Function

[31:2] - Reserved, RES0.

[1] ID Stage 2 Instruction Cache disable. When HCR_EL2.VM is 1, this forces all stage2 translations for instruction accesses
to Normal memory to be Non-cacheable for the EL1/EL0 translation regime. The values are:
0 No effect on the stage 2 of the EL1/EL0 translation regime for instruction accesses. This is the reset value.
1 Forces all stage 2 translations for instruction accesses to Normal memory to be Non-cacheable for the EL0/EL1
translation regime.

[0] CD Stage 2 Data cache disable. When HCR_EL2.VM is 1, this forces all stage2 translations for data accesses and
translation table walks to Normal memory to be Non-cacheable for the EL1/EL0 translation regime. The values are:
0 No effect on the stage 2 of the EL1/EL0 translation regime for data accesses and translation table walks. This is
the reset value.
1 Forces all stage 2 translations for data accesses and translation table walks to Normal memory to be Non-
cacheable for the EL0/EL1 translation regime.

To access the HCR2 in AArch32 state, read or write the CP15 register with:
MRC p15, 4, <Rt>, c1, c1, 4; Read Hyp Configuration Register 2
MCR p15, 4, <Rt>, c1, c1, 4; Write Configuration Register 2

Related information
4.3.34 Hypervisor Configuration Register, EL2 on page 4-131.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-260
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

4.5.12 Hyp Debug Control Register

The HDCR characteristics are:

Purpose
Controls the trapping to Hyp mode of Non-secure accesses, at EL1 or lower, to functions
provided by the debug and trace architectures.
Usage constraints
The accessibility to the HDCR in AArch32 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW -

Configurations
The HDCR is:
• A Banked EL2 register.
• Architecturally mapped to the AArch64 MDCR_EL2 register.
Attributes
See the register summary in Table 4-85 c1 register summary on page 4-217.
The following figure shows the HDCR bit assignments.
31 12 11 10 9 8 7 6 5 4 0

RES0 HPMN

TDRA TPMCR
TDOSA TPM
TDA HPME
TDE

Figure 4-89 HDCR bit assignments

The following table shows the HDCR bit assignments.

Table 4-125 HDCR bit assignments

Bits Name Function

[31:12] - Reserved, RES0.

[11] TDRA Trap Debug ROM Access. The values are:

0 Has no effect on Debug ROM accesses. This is the reset value.
1 Trap valid Non-secure EL0 or EL1 Debug ROM accesses to Hyp mode.

When this bit is set to 1, any valid Non-secure access to DBGDRAR or DBGDSAR is trapped to Hyp mode.
If bit[8], TDE, is set, or if the HCR.TGE bit is set, the TDRA value is ignored and the processor behaves as if this
bit is set to 1.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-261
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-125 HDCR bit assignments (continued)

Bits Name Function

[10] TDOSA Trap Debug OS-related register Access. The values are:
0 Has no effect on accesses to CP14 Debug registers. This is the reset value.
1 Trap valid EL0 or EL1 Non-secure accesses to CP14 OS-related Debug registers to Hyp mode.

When this bit is set to 1, any valid Non-secure CP14 access to the following OS-related Debug registers is trapped
to Hyp mode:
• DBGOSLSR.
• DBGOSLAR.
• DBGOSDLR.
• DBGPRCR.
If bit[8], TDE, is set, or if the HCR.TGE bit is set, the TDRA value is ignored and the processor behaves as if this
bit is set to 1.

[9] TDA Trap Debug Access. The values are:

0 Has no effect on accesses to CP14 Debug registers. This is the reset value.
1 Trap valid EL0 or EL1 Non-secure accesses to CP14 Debug registers to Hyp mode.

When this bit is set to 1, any valid Non-secure access to the CP14 Debug registers, other than the registers trapped
by the TDRA and TDOSA bits, is trapped to Hyp mode.
If bit[8], TDE, is set, or if the HCR.TGE bit is set, the TDRA value is ignored and the processor behaves as if this
bit is set to 1.

[8] TDE Trap Debug Exceptions. The values are:

0 Has no effect on Debug exceptions. This is the reset value.
1 Trap valid Non-secure Debug exceptions to Hyp mode.

When this bit is set to 1, any Debug exception taken in Non-secure state is trapped to Hyp mode.
When this bit is set to 1, the TRA, TDOSA, and TDA bits are treated as if they are set to 1, irrespective of the value
stored in the register. If the HCR.TGE bit is set to 1, this bit is treated as if it was set to 1, irrespective of the value
stored in the register.

[7] HPME Hypervisor Performance Monitors Enable. The values are:

0 Hyp mode Performance Monitors counters disabled. This is the reset value.
1 Hyp mode Performance Monitors counters enabled.

When this bit is set to 1, access to the Performance Monitors counters that are reserved for use from Hyp mode is
enabled. For more information, see the description of the HPMN field.

[6] TPM Trap Performance Monitors accesses. The values are:

0 Has no effect on Performance Monitors accesses. This is the reset value.
1 Trap valid Non-secure Performance Monitors accesses to Hyp mode.

When this bit is set to 1, any valid Non-secure EL0 or EL1 access to the Performance Monitors registers is trapped
to Hyp mode.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-262
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-125 HDCR bit assignments (continued)

Bits Name Function

[5] TPMCR Trap Performance Monitors Control Register accesses. The values are:
0 Has no effect on PMCR accesses. This is the reset value.
1 Trap valid Non-secure PMCR accesses to Hyp mode.

When this bit is set to 1, any valid Non-secure access to the PMCR is trapped to Hyp mode.

[4:0] HPMN Defines the number of Performance Monitors counters that are accessible from Non-secure EL1, and from Non-
secure EL0 if unprivileged access is enabled.
This field behaves as if it contains an UNKNOWN value of less than or equal to PMCR.N, in all ways other than
when reading back this field if:
• This field is set to 0.
• This field is set to a value greater than PMCR.N.
In Non-secure state, HPMN divides the Performance Monitors counters as follows:
If PMXEVCNTR is accessing Performance Monitors counter n then, in Non-secure state:
• If n is in the range 0 ≤ n < HPMN, the counter is accessible from EL1 and EL2, and from EL0 if unprivileged
access to the counters is enabled.
• If n is in the range HPMN ≤ n < PMCR.N, the counter is accessible only from EL2. The HPME bit enables
access to the counters in this range.
This field resets to 0x6, the value of PMCR.N.

To access the HDCR in AArch32 state, read or write the CP15 register with:
MRC p15, 4, <Rt>, c1, c1, 1; Read Hyp Debug Configuration Register
MCR p15, 4, <Rt>, c1, c1, 1; Write Hyp Debug Configuration Register

To access the MDCR_EL2 in AArch64 state, read or write the register with:
MRS <Xt>, MDCR_EL2; Read Monitor Debug Configuration Register
MSR MDCR_EL2, <Xt>; Write Monitor Debug Configuration Register

4.5.13 Hyp Architectural Feature Trap Register

The HCPTR characteristics are:

Purpose
Controls the trapping to Hyp mode of Non-secure accesses, at EL1 or lower, to coprocessors
other than CP14 and CP15 and to floating-point and Advanced SIMD functionality. The HCPTR
also controls the access to this functionality from Hyp mode.
Usage constraints
The accessibility to the HCPTR by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW -

If a bit in the NSACR prohibits a Non-secure access, then the corresponding bit in the HCPTR
behaves as RES1 for Non-secure accesses. See the bit descriptions for more information.
Configurations
The HCPTR is:
• A Banked EL2 register.
• Architecturally mapped to the AArch64 CPTR_EL2 register.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-263
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Attributes
See the register summary in Table 4-85 c1 register summary on page 4-217.
The following figure shows the HCPTR bit assignments.
31 30 21 20 19 16 15 14 13 12 11 10 9 0

RES0 RES0 RES1 RES1

TCPAC TTA RES0 TCP10

TASE TCP11

Figure 4-90 HCPTR bit assignments

The following table shows the HCPTR bit assignments.

Table 4-126 HCPTR bit assignments

Bits Name Function

[31] TCPAC Trap Coprocessor Access Control Register accesses. When this bit is set to 1, any valid Non-secure EL1 accesses to
the CPACR is trapped to Hyp mode. The values are:
0 Has no effect on CPACR accesses. This is the reset value.
1 Trap valid Non-secure EL1 CPACR accesses to Hyp mode.

[30:21] - Reserved, UNK/RES0.

[20] TTA Trap Trace Access. This value is:

0 CP14 access to the trace registers is not supported.

[19:16] - Reserved, UNK/RES0.

[15] TASE Trap Advanced SIMD use. If NSACR.NSASEDIS is set to 1, this bit behaves as RES1 on Non-secure accesses. The
values are:
0 If the NSACR settings permit Non-secure use of the Advanced SIMD functionality then Hyp mode can access
that functionality, regardless of any settings in the CPACR. This is the reset value.
Note
This bit value has no effect on possible use of the Advanced SIMD functionality from Non-secure EL1 and
EL0.

1 Trap valid Non-secure accesses to Advanced SIMD functionality to Hyp mode.

When this bit is set to 1, any otherwise-valid access to Advanced SIMD functionality from:
• A Non-secure EL1 or EL0 access is trapped to Hyp mode.
• Hyp mode generates an UNDEFINED Instruction exception, taken in Hyp mode.

Note
If TCP10 and TCP11 are set to 1, then all Advanced SIMD use is trapped to Hyp mode, regardless of the value of
this field.

[14] - Reserved, RES0.

[13:12] - Reserved, RES1.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-264
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-126 HCPTR bit assignments (continued)

Bits Name Function

[11] TCP11 Trap coprocessor 11. The values are:

0 If NSACR.CP11 is set to 1, then Hyp mode can access CP11, regardless of the value of CPACR.CP11. This is
the reset value.
Note
This bit value has no effect on possible use of CP11 from Non-secure EL1 and EL0.

1 Trap valid Non-secure accesses to CP11 to Hyp mode.

When TCP11 is set to 1, any otherwise-valid access to CP11 from:
• A Non-secure EL1 or EL0 access is trapped to Hyp mode.
• Hyp mode generates an Undefined Instruction exception, taken in Hyp mode.

[10] TCP10 Trap coprocessor 10. The possible values are:

0 If NSACR.CP10 is set to 1, then Hyp mode can access CP10, regardless of the value of CPACR.CP10. This is
the reset value.
Note
This bit value has no effect on possible use of CP10 from Non-secure EL1 and EL0.

1 Trap valid Non-secure accesses to CP10 to Hyp mode.

When TCP10 is set to 1, any otherwise-valid access to CP10 from:
• A Non-secure EL1 or EL0 access is trapped to Hyp mode.
• Hyp mode generates an Undefined Instruction exception, taken in Hyp mode.

[9:0] - Reserved, RES1.

To access the HCPTR in AArch32 state, read or write the CP15 register with:
MRC p15, 4, <Rt>, c1, c1, 2; Read Hyp Architectural Feature Trap Register
MCR p15, 4, <Rt>, c1, c1, 2; Write Hyp Architectural Feature Trap Register

Related information
4.5.8 Non-secure Access Control Register on page 4-251.
4.3.35 Architectural Feature Trap Register, EL2 on page 4-137.

4.5.14 Translation Table Base Register 0 and Register 1

The processor does not use any IMPLEMENTATION DEFINED bits in the 32-bit TTBR0 and TTBR1 format, so
these bits are RES0.

4.5.15 Translation Table Base Control Register

The TTBCR characteristics are:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-265
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Purpose
Controls which Translation Table Base Register defines the base address for a translation table
walk required for the stage 1 translation of a memory access from any mode other than Hyp
mode in AArch32 state. This register also controls the translation table format and, when using
the Long-descriptor translation table format, holds cacheability and shareability information.
The processor does not use the IMPLEMENTATION DEFINED bit, TTBCR[30], when using the Long-
descriptor translation table format, so this bit is RES0.
Usage constraints
The accessibility to the TTBCR by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Write access to the Secure copy of SCTLR is disabled when the CP15SDISABLE signal is
HIGH.
Configurations
The TTBCR is Banked in the Secure and Non-secure states.
The architectural mapping of the TTBCR is:
• The Non-secure TTBCR is mapped to the AArch64 TCR_EL1[31:0] register.
• The Secure TTBCR is mapped to the AArch64 TCR_EL3[31:0] register
Attributes
See the register summary in Table 4-85 c1 register summary on page 4-217.
See the ARM® Architecture Reference Manual ARMv8 for more information.
To access the TTBCR in AArch32 state, read or write the CP15 register with:
MRC p15, 0, <Rt>, c2, c0, 2; Read Translation Table Base Control Register
MCR p15, 0, <Rt>, c2, c0, 2; Write Translation Table Base Control Register

Related information
4.3.41 Translation Control Register, EL1 on page 4-146.
4.3.47 Translation Control Register, EL3 on page 4-155.

4.5.16 Hyp Translation Control Register

The processor does not use the IMPLEMENTATION DEFINED bit, HTCR[30], so this bit is RES0.
The HTCR characteristics are:
Purpose
Controls translation table walks required for the stage 1 translation of memory accesses from
Hyp mode, and holds cacheability and shareability information for the accesses.
Usage constraints
The accessibility to the HTCR by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW -

Configurations
The HTCR is:
• A Banked EL2 register.
• Architecturally mapped to the AArch64 TCR_EL2.
The TCR_EL2 is a 32-bit register in AArch64 state.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-266
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Attributes
See the register summary in Table 4-85 c1 register summary on page 4-217.
See the ARM® Architecture Reference Manual ARMv8 for more information.
To access the HTCR in AArch32 state, read or write the CP15 register with:
MRC p15, 4, <Rt>, c2, c0, 2; Read Hyp Translation Control Register
MCR p15, 4, <Rt>, c2, c0, 2; Write Hyp Translation Control Register

Related information
4.3.42 Translation Control Register, EL2 on page 4-149.

4.5.17 Data Fault Status Register

The DFSR characteristics are:

Purpose
Holds status information about the last data fault.
Usage constraints
The accessibility to the DFSR by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Configurations
The DFSR is Banked for Secure and Non-secure states.
The architectural mapping of the DFSR is:
• The Non-secure DFSR is mapped to the AArch64 ESR_EL1 register.
• The Secure DFSR is mapped to the AArch64 ESR_EL3 register.
Attributes
See the register summary in Table 4-2 AArch64 exception handling registers on page 4-78.
There are two formats for this register. The value of TTBCR.EAE selects which format of the register is
used. The two formats are:
• DFSR format when using the Short-descriptor translation table format on page 4-268.
• DFSR format when using the Long-descriptor translation table format on page 4-269.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-267
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

DFSR format when using the Short-descriptor translation table format

The following figure shows the DFSR bit assignments when using the Short-descriptor translation table
format.
31 16 15 14 13 12 11 10 9 8 7 4 3 0

RES0 Domain FS[3:0]

UA RES0
UC LPAE
CM FS[4]
ExT WnR

Figure 4-91 DFSR bit assignments for Short-descriptor translation table format

The following table shows the DFSR bit assignments when using the Short-descriptor translation table
format.

Table 4-127 DFSR bit assignments for Short-descriptor translation table format

Bits Name Function

[31:16] - Reserved, RES0.

[15] UA Unattributable fault. This bit is only set for System Errors. For other
faults, it is RES0. The values are:
0 Attributable, can be attributed to the processing element counting
the event.
1 Unattributable, cannot be attributed to any particular processor.

[14] UC Uncontainable fault. This bit is only set for System Errors. For other
faults, it is RES0. The values are:
0 Containable, an attributable event that can be contained to a
particular code sequence.
1 Uncontainable, cannot be contained to a particular code sequence.

Unattributable events are Uncontainable.

[13] CM Cache maintenance fault. For synchronous faults, this bit indicates
whether a cache maintenance operation generated the fault. The values
are:
0 Abort not caused by a cache maintenance operation.
1 Abort caused by a cache maintenance operation.

On an asynchronous fault, this bit is UNKNOWN.

[12] ExT External abort type. This field indicates whether an AXI decode or slave
error caused an abort:
0 External abort marked as DECERR.
1 External abort marked as SLVERR.

For aborts other than external aborts this bit always returns 0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-268
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-127 DFSR bit assignments for Short-descriptor translation table format (continued)

Bits Name Function

[11] WnR Write not Read bit. This field indicates whether a write or a read access
caused the abort:
0 Abort caused by a read access.
1 Abort caused by a write access.

For faults on CP15 cache maintenance operations, including the VA to

PA translation operations, this bit always returns a value of 1.

[10] FS[4] Part of the Fault Status field. See bits[3:0] in this table.

[9] LPAE Large physical address extension. The value of the format descriptor is:
0 Short-descriptor translation table formats.

[8] - Reserved, RES0.

[7:4] Domain The domain of the fault address. Use of the field is deprecated.

[3:0] FS[3:0] Fault Status bits. This field indicates the type of exception generated.
The possible values are:
0b00001 Alignment fault.
0b01100 Synchronous external abort on translation table walk, 1st
level.
0b01110 Synchronous external abort on translation table walk, 2nd
level.
0b11100 Synchronous parity error on translation table walk, 1st
level.
0b11110 Synchronous parity error on translation table walk, 2nd
level.
0b00101 Translation fault, 1st level.
0b00111 Translation fault, 2nd level.
0b00011 Access flag fault, 1st level.
0b00110 Access flag fault, 2nd level.
0b01001 Domain fault, 1st level.
0b01011 Domain fault, 2nd level.
0b01101 Permission fault, 1st level.
0b01111 Permission fault, 2nd level.
0b00010 Debug event.
0b01000 Synchronous external abort, non-translation.
0b11001 Synchronous parity error on memory access.
0b10110 Asynchronous external abort.
0b11000 Asynchronous parity error on memory access.

All other values are reserved.

DFSR format when using the Long-descriptor translation table format

The following figure shows the DFSR bit assignments when using the Long-descriptor translation table
format.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-269
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

31 16 15 14 13 12 11 10 9 8 6 5 0

RES0 RES0 Status

UA LPAE
UC RES0
CM WnR
ExT

Figure 4-92 DFSR bit assignments for Long-descriptor translation table format

The following table shows the DFSR bit assignments when using the Long-descriptor translation table
format.

Table 4-128 DFSR bit assignments for Long-descriptor translation table format

Bits Name Function

[31:16] - Reserved, RES0.

[15] UA Unattributable fault. This bit is only set for System Errors. For other
faults, it is RES0. The values are:
0 Attributable, can be attributed to the processing element counting
the event.
1 Unattributable, cannot be attributed to any particular processor.

[14] UC Uncontainable fault. This bit is only set for System Errors. For other
faults, it is RES0. The values are:
0 Containable, an attributable event that can be contained to a
particular code sequence.
1 Uncontainable, cannot be contained to a particular code sequence.

Unattributable events are Uncontainable.

[13] CM Cache maintenance fault. For synchronous faults, this bit indicates
whether a cache maintenance operation generated the fault:
0 Abort not caused by a cache maintenance operation.
1 Abort caused by a cache maintenance operation.

On an asynchronous fault, this bit is UNKNOWN.

[12] ExT External abort type. This field indicates whether an AXI decode or slave
error caused an abort:
0 External abort marked as DECERR.
1 External abort marked as SLVERR.

For aborts other than external aborts this bit always returns 0.

[11] WnR Write not Read bit. This field indicates whether a write or a read access
caused the abort:
0 Abort caused by a read access.
1 Abort caused by a write access.

For faults on CP15 cache maintenance operations, including the VA to

PA translation operations, this bit always returns a value of 1.

[10] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-270
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Table 4-128 DFSR bit assignments for Long-descriptor translation table format (continued)

Bits Name Function

[9] LPAE Large physical address extension. The value of the format descriptor is:
1 Long-descriptor translation table formats.

[8:6] - Reserved, RES0.

[5:0] Status Fault Status bits. This field indicates the type of exception generated.
The possible values are:
0b0000LL Address size fault, LL bits indicate level.
0b0001LL Translation fault, LL bits indicate level.
0b0010LL Access flag fault, LL bits indicate level.
0b0011LL Permission fault, LL bits indicate level.
0b010000 Synchronous external abort.
0b011000 Synchronous parity error on memory access.
0b010001 Asynchronous external abort.
0b011001 Asynchronous parity error on memory access.
0b0101LL Synchronous external abort on translation table walk, LL
bits indicate level.
0b0111LL Synchronous parity error on memory access on translation
table walk, LL bits indicate level.
0b100001 Alignment fault.
0b100010 Debug event.

All other values are reserved.

The following table shows how the LL bits in the Status field encode the lookup level associated with the
MMU fault.

Table 4-129 Encodings of LL bits associated with the MMU fault

LL bits Meaning

00 Level 0 fault

01 First level

10 Second level

11 Third level

To access the DFSR in AArch32 state, read or write the CP15 register with:
MRC p15, 0, <Rt>, c5, c0, 0; Read Data Fault Status Register
MCR p15, 0, <Rt>, c5, c0, 0; Write Data Fault Status Register

Related information
4.3.50 Exception Syndrome Register, EL1 and EL3 on page 4-157.

4.5.18 Physical Address Register

The PAR characteristics are:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-271
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Purpose
Receives the PA from any address translation operation.
Usage constraints
The accessibility to the PAR by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Configurations
The PAR is Banked for the Secure and Non-secure states.
The Non-secure PAR is architecturally mapped to AArch64 PAR_EL1 register.
The PAR[63:32] is RES0 when using the Short-descriptor translation format.
Attributes
The processor does not use any IMPLEMENTATION DEFINED bits in the 32-bit or 64-bit format PAR or
the PAR_EL1, so these bits are RES0.
See the register summary in Table 4-90 c7 register summary on page 4-219.
See the ARM® Architecture Reference Manual ARMv8 for more information.
To access the PAR in AArch32 state when using the Short-descriptor translation format, read or write the
CP15 register with:
MRC p15, 0, <Rt>, c7, c4, 0; Read Physical Address Register
MCR p15, 0, <Rt>, c7, c4, 0; Write Physical Address Register

To access the PAR in AArch32 state when using the Long-descriptor translation format, read or write the
CP15 register with:
MRRC p15, 0, <Rt>, <Rt2>, c7; Read Physical Address Register
MCRR p15, 0, <Rt>, <Rt2>, c7; Write Physical Address Register

Related information
4.3.55 Physical Address Register, EL1 on page 4-165.

4.5.19 Primary Region Remap Register

The PRRR characteristics are:

Purpose
Controls the top-level mapping of the TEX[0], C, and B memory region attributes.
Usage constraints
The accessibility to the PRRR by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Write access to the Secure copy of PRRR is disabled when the CP15SDISABLE signal is
HIGH.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-272
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

Configurations
The PRRR is:
• Banked for the Secure and Non-secure states.
• Only relevant if the TTBCR.EAE bit is 0.
• Architecturally mapped to the MAIR0 register in AArch32 state.
The Non-secure PRRR is architecturally mapped to the AArch64 MAIR_EL1[31:0] register.
The Secure PRRR is mapped to the AArch64 MAIR_EL3[31:0] register.
Attributes
See the register summary in Table 4-94 c10 register summary on page 4-222.
See the ARM® Architecture Reference Manual ARMv8 for more information.
To access the PRRR in AArch32 state when TTBCR.EAE is 0, read or write the CP15 register with:
MRC p15, 0, <Rt>, c10, c2, 0; Read Primary Region Remap Register
MCR p15, 0, <Rt>, c10, c2, 0; Write Primary Region Remap Register

4.5.20 Memory Attribute Indirection Register 0

The processor does not set any IMPLEMENTATION DEFINED attributes with the Memory Attribute Indirection
Register 0 (MAIR0).

4.5.21 Normal Memory Remap Register.

The NMRR characteristics are:

Purpose
Provides additional mapping controls for memory regions that are mapped as Normal memory
by their entry in the PRRR.
Usage constraints
The accessibility to the NMRR by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

Write access to the Secure copy of NMRR is disabled when the CP15SDISABLE signal is
HIGH.
Configurations
The NMRR is:
• Banked for the Secure and Non-secure states.
• Only relevant if the TTBCR.EAE bit is 0.
• Architecturally mapped on to the MAIR1 register in AArch32 state.
The Non-secure NMRR is architecturally mapped to the AArch64 MAIR_EL1[63:32] register.
The Secure NMRR is mapped to the AArch64 MAIR_EL3[63:32] register.
Attributes
See the register summary in Table 4-94 c10 register summary on page 4-222.
See the ARM® Architecture Reference Manual ARMv8 for more information.
To access the NMRR in AArch32 state, read or write the CP15 register with:
MRC p15, 0, <Rt>, c10, c2, 1; Read Normal Memory Remap Register
MCR p15, 0, <Rt>, c10, c2, 1; Write Normal Memory Remap Register

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-273
reserved.
Non-Confidential
 4 System Control
4.5 AArch32 register descriptions

4.5.22 Memory Attribute Indirection Register 1

The processor does not set any IMPLEMENTATION DEFINED attributes with the Memory Attribute Indirection
Register 1 (MAIR1).

4.5.23 FCSE Process ID Register

The processor does not implement Fast Context Switch Extension (FCSE), so this register is always RES0.

4.5.24 Configuration Base Address Register

The CBAR characteristics are:

Purpose
Holds the physical base address of the memory-mapped GIC CPU interface registers.
Usage constraints
The accessibility to the CBAR by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The CBAR is Common to the Secure and Non-secure states.
Attributes
See the register summary in Table 4-98 c15 register summary on page 4-224.
The following figure shows the CBAR bit assignments.
31 18 17 12 11 0

PERIPHBASE[31:18] RES0 PERIPHBASE[43:32]

Figure 4-93 CBAR bit assignments

The following table shows the CBAR bit assignments.

Table 4-130 CBAR bit assignments

Bits Name Function

[31:18] PERIPHBASE[31:18] The primary input PERIPHBASE[31:18] determines the reset value.

[17:12] - Reserved, RES0.

[11:0] PERIPHBASE[43:32] The primary input PERIPHBASE[43:32] determines the reset value.

To access the CBAR in AArch32 state, read the CP15 register with:
MRC p15, 1, <Rt>, c15, c3, 0; Read Configuration Base Address Register

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 4-274
reserved.
Non-Confidential
Chapter 5
Memory Management Unit

This section describes the Memory Management Unit (MMU).

It contains the following sections:

• 5.1 About the MMU on page 5-276.
• 5.2 TLB organization on page 5-277.
• 5.3 TLB match process on page 5-278.
• 5.4 Memory access sequence on page 5-279.
• 5.5 MMU enabling and disabling on page 5-281.
• 5.6 Intermediate table walk caches on page 5-282.
• 5.7 External aborts on page 5-283.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 5-275
reserved.
Non-Confidential
 5 Memory Management Unit
5.1 About the MMU

5.1 About the MMU

The Cortex-A72 processor is an ARMv8 compliant processor that supports execution in both the
AArch64 and AArch32 states.
In AArch32 state, the ARMv8 address translation system resembles the ARMv7 address translation
system with LPAE and Virtualization Extensions. In AArch64 state, the ARMv8 address translation
system resembles an extension to the Long Descriptor Format address translation system to support the
expanded virtual and physical address spaces. For more information regarding the address translation
formats, see the ARM® Architecture Reference Manual ARMv8. Key differences between the AArch64
and AArch32 address translation systems are that the AArch64 state provides the ability to:
• Select the translation granule to either be 4KB or 64KB. In AArch32, the translation granule is
limited to be 4KB.
• Configure the ASID size to be either 8-bit or 16-bit. In AArch32, the ASID is limited to an 8-bit
value.
The maximum supported physical address size is:
• 44-bit in AArch64 state.
• 40-bit in AArch32 state.
The MMU controls table walk hardware that accesses translation tables in memory. The MMU works
with the L1 and L2 memory system to translate a Virtual Address (VA) to a Physical Address (PA). The
MMU enables fine-grained memory system control through a set of virtual-to-physical address mappings
and memory attributes held in the L1 and L2 Translation Look-aside Buffers (TLBs).
The MMU has the following features:
• 48-entry fully-associative L1 instruction TLB.
• 32-entry fully-associative L1 data TLB for data load and store pipelines.
• 4-way set-associative 1024-entry L2 TLB in each processor.
• Intermediate table walk caches.
• The TLB entries contain a global indicator or an Address Space Identifier (ASID) to permit context
switches without TLB flushes.
• The TLB entries contain a Virtual Machine Identifier (VMID) to permit virtual machine switches
without TLB flushes.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 5-276
reserved.
Non-Confidential
 5 Memory Management Unit
5.2 TLB organization

5.2 TLB organization

The Cortex-A72 processor implements a 2-level TLB structure.
The TLBs, at either the L1 or the L2 level, do not require to be flushed on a context or virtual machine
switch. The MMU does not support the locking of TLB entries at either Level 1 or Level 2.
This section contains the following subsections:
• 5.2.1 L1 instruction TLB on page 5-277.
• 5.2.2 L1 data TLB on page 5-277.
• 5.2.3 L2 TLB on page 5-277.

5.2.1 L1 instruction TLB

The L1 instruction TLB is a 48-entry fully-associative structure. This TLB caches entries of three
different page sizes, natively 4KB, 64KB, and 1MB, of VA to PA mappings. If the page tables map the
memory region to a larger granularity than 1MB, it only allocates one mapping for the particular 1MB
region to which the current access corresponds.
A hit in the instruction TLB provides a single CLK cycle access to the translation, and returns the PA to
the instruction cache for comparison. It also checks the access permissions to signal a Prefetch Abort.

5.2.2 L1 data TLB

The L1 data TLB is a 32-entry fully-associative TLB that is used for data loads and stores. This TLB
caches entries of three different page sizes, natively 4KB, 64KB, and 1MB, of VA to PA mappings.
A hit in the data TLB provides a single CLK cycle access to the translation, and returns the PA to the
data cache for comparison. It also checks the access permissions to signal a Data Abort.

5.2.3 L2 TLB

Misses from the L1 instruction and data TLBs are handled by a unified L2 TLB. This is a 1024-entry 4-
way set-associative structure. The L2 TLB supports the page sizes of 4K, 64K, 1MB and 16MB. It also
supports page sizes of 2MB and 1GB for the long descriptor format translation in AArch32 state and in
AArch64 state when using the 4KB translation granule. In addition, the L2 TLB supports the 512MB
page map size defined for the AArch64 translations that use a 64KB translation granule.
Accesses to the L2 TLB take a variable number of cycles, based on the competing requests from each of
the L1 TLBs, TLB maintenance operations in flight, and the different page size mappings in use.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 5-277
reserved.
Non-Confidential
 5 Memory Management Unit
5.3 TLB match process

5.3 TLB match process

The ARMv8-A architecture provides for multiple VA spaces that are translated differently.
The TLB entries store all the required context information to facilitate a match and avoid the requirement
for a TLB flush on a context or virtual machine switch. Each TLB entry contains a VA, page size, PA,
and a set of memory properties that include the memory type and access permissions. Each entry is
associated with a particular ASID, or as global for all application spaces. The TLB entry also contains a
field to store the VMID in the entry, applicable to accesses made from the Non-secure state. There is also
a memory space identifier that records whether the request occurred at the EL3 Exception level, Non-
secure EL2 Exception level, or Secure and Non-secure EL0 or EL1 Exception levels. A TLB entry match
occurs when the following conditions are met:
• Its VA, moderated by the page size such as the VA bits[48:N], where N is log2 of the page size for
that translation stored in the TLB entry, matches that of the requested address.
• The memory space matches the memory space state of the requests. The memory space can be one of
four values:
— Secure EL3.
— Non-secure EL2.
— Secure EL0 or EL1.
— Non-secure EL0 or EL1.
• The ASID matches the current ASID held in the CONTEXTIDR, TTBR0, or TTBR1 register or the
entry is marked global.
• The VMID matches the current VMID held in the VTTBR register.

Note
• For a request originating from EL2 or EL3, the ASID and VMID match are ignored.
• For a request originating from Secure state, the VMID match is ignored.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 5-278
reserved.
Non-Confidential
 5 Memory Management Unit
5.4 Memory access sequence

5.4 Memory access sequence

When the processor generates a memory access, the MMU:
1. Performs a lookup for the requested VA, current ASID, current VMID, and memory space in the
relevant L1 instruction or data TLB.
2. Performs a lookup for the requested VA, current ASID, current VMID, and memory space in the
unified L2 TLB if there is a miss in the relevant L1 TLB.
3. Performs a hardware translation table walk if there is a miss in the L2 TLB.
When executing in AArch64 at a particular Exception level, you can configure the hardware translation
table walk to use either the 4KB translation granule or the 64KB translation granule. Program the
Translation Granule bit, TG0, in the appropriate translation control register:
• TCR_EL1.
• TCR_EL2.
• TCR_EL3.
• VTCR_EL2.
When executing in AArch32 in a particular mode, you can configure the MMU to perform translation
table walks using either the Short Descriptor Translation Table or the Long Descriptor Translation table
format, by programming the Extended Address Enable bit, EAE, in the appropriate translation table
control register. Only the Long Descriptor Translation format is supported in Hyp mode.
You can configure the MMU to perform translation table walks in Cacheable regions, by programming
the IRGN bits:
AArch32 • Translation table base registers (TTBR0/TTBR1_ELx) when using the Short Descriptor
translation table format.
• TCR_ELx register when using the Long Descriptor translation table format.
AArch64 In the appropriate TCR_ELx register.

For Stage2 translations, the IRGN bits must be programmed in the VTCR_EL2 register.
If the encoding of the IRGN bits is WriteBack, an L2 data cache lookup is performed and data is read
from the data cache. If the encoding of the IRGN bits is Write-Through or Non-cacheable, an access to
external memory is performed.
In the case of an L2TLB miss, the hardware does a translation table walk provided the MMU is enabled,
and the translation using the base register has not been disabled by:
• Setting the PD0 or PD1 bit in the 4.5.15 Translation Table Base Control Register on page 4-265, to
disallow translation using either TTBR0 or TTBR1 respectively, when using AArch32 along with the
Short Descriptor Format.
• Setting of the EPD0 or EPD1 bit in the TCR_EL1 register when using AArch64 or when using the
Long Descriptor format in AArch32.
If the translation table walk is disabled for a particular base register, the processor returns a Translation
Fault. If the TLB finds a matching entry, it uses the information in the entry as follows:
• The access permission bits and the domain, when using the Short Descriptor format in AArch32 state,
determine if the access is permitted. If the matching entry does not pass the permission checks, the
MMU signals a Permission fault. See the ARM® Architecture Reference Manual ARMv8 for:
— A description of the various faults.
— The fault codes.
— Information regarding the registers where the fault codes are set.
• The memory region attributes specified in the TLB entry determine if the access is:
— Secure or Non-secure.
— Inner, Outer or not Cacheable.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 5-279
reserved.
Non-Confidential
 5 Memory Management Unit
5.4 Memory access sequence

— Normal Memory or Device type, Strongly-ordered or Device type when using the Short
Descriptor Format in AArch32.
— One of the four different device memory types defined for ARMv8:
Device-nGnRnE Device non-Gathering, non-Reordering, No Early Write Acknowledgment.
Device-nGnRE Device non-Gathering, non-Reordering, Early Write Acknowledgment.
Device-nGRE Device non-Gathering, Reordering, Early Write Acknowledgment.
Device-GRE Device Gathering, Reordering, Early Write Acknowledgment.
• The TLB translates the VA to a PA for the memory access.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 5-280
reserved.
Non-Confidential
 5 Memory Management Unit
5.5 MMU enabling and disabling

5.5 MMU enabling and disabling

You can enable or disable the MMU. See the ARM Architecture Reference Manual ARMv8 for more
information.
You must set CPUECTLR.SMPEN to 1 before the caches and MMU are enabled, or any instruction
cache or TLB maintenance operations are performed.

Related information
4.3.67 CPU Extended Control Register, EL1 on page 4-206.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 5-281
reserved.
Non-Confidential
 5 Memory Management Unit
5.6 Intermediate table walk caches

5.6 Intermediate table walk caches

The Cortex-A72 processor implements dedicated caches that store intermediate levels of translation table
entries as part of a table walk.
Cached entries are associated with an ASID and a VMID where applicable for Non-secure EL1
translations.
Care is required when using the reserved ASID method for context switch. See the ARM® Architecture
Reference Manual ARMv8 for more information.
The following example shows how to synchronize ASID and TTBR changes using a reserved ASID.

Example 5-1 Using a reserved ASID to synchronize ASID and TTBR changes

In this example, the operating system uses a particular reserved ASID value for the synchronization of
the ASID and the Translation Table Base Register. You can use this approach only when the size of the
mapping for any given Virtual Address is the same in the old and new translation tables. The example
uses the value of 0.
The software uses the following sequences that must be executed from memory marked as global:
Change ASID to 0
ISB
Change Translation Table Base Register
ISB
Change ASID to new value
ISB

If the code relies on only leaf translation table entries that are cached, it can incorrectly assume that
entries tagged with the reserved ASID are not required to be flushed. For example:
• Global leaf entries that remain valid or must be flushed for all ASIDs when modified
• Non-global leaf entries that are not used because the reserved ASID is not set outside the context
switch code.
The incorrect assumption leads to the following failure:
• The context switch code sets the ASID to the reserved value.
• Speculative fetching reads and caches the first level page table entry, using the current TTBR, and
tagging the entry with the reserved ASID. This is a pointer to a second level table.
• Context switch completes.
• Processing continues, and the process with the page tables terminates. The OS frees and reallocates
the page table memory.
• A later context switch sets the ASID to the reserved value
• Speculative fetching makes use of the cached first level page table entry, because it is tagged with the
reserved ASID, and uses it to fetch a second level page table entry. Because the memory is
reallocated and reused, the entry contains random data that can appear to be a valid, global entry. This
second level page table entry is cached.
• Context switch completes, and application execution continues.
• The application references the address range covered by the cached second level page table entry.
Because the entry is marked as global, a match occurs and so data is fetched from a random address.

Note
When you use a reserved ASID, you must invalidate the TLB to deallocate the translation table memory.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 5-282
reserved.
Non-Confidential
 5 Memory Management Unit
5.7 External aborts

5.7 External aborts

External memory errors are defined as those that occur in the memory system rather than those that the
MMU detects.
External memory errors are extremely rare. External errors are caused by errors flagged by the AXI
interfaces or generated because of an uncorrected ECC error in the L1 data cache or L2 cache arrays
when the request is external to the Cortex-A72 MPCore processor. You can configure external aborts to
trap to Monitor mode by setting the EA bit in the Secure Configuration Register to 1.
This section contains the following subsections:
• 5.7.1 External aborts on data read or write on page 5-283.
• 5.7.2 Synchronous and asynchronous aborts on page 5-283.

5.7.1 External aborts on data read or write

Externally generated errors during a data read or write can be asynchronous. This means that the
ELR_EL1, ELR_EL2, ELR_EL3, of r14 entry into the abort handler on such an abort might not hold the
address of the instruction that caused the abort.
The DFAR is UNPREDICTABLE when an asynchronous abort occurs.
For a load multiple or store multiple operation, the address captured in the DFAR is that of the address
that generated the synchronous external abort.

5.7.2 Synchronous and asynchronous aborts

To determine a fault type, check the Execution state. If the abort handler code targeted by the exception
is in AArch64 state, read the appropriate ESR_ELx register. If the abort handler code is an AArch32
hypervisor, see 4.3.54 Exception Syndrome Register, EL2 on page 4-163. If the abort handler code is not
an AArch32 non-hypervisor, see 4.3.51 Instruction Fault Status Register, EL2 on page 4-159 for an
Instruction Abort or the ">4.5.17 Data Fault Status Register on page 4-267 for a Data Abort.

Related information
4.5.7 Secure Configuration Register on page 4-248.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 5-283
reserved.
Non-Confidential
Chapter 6
Level 1 Memory System

This section describes the Level 1 (L1) memory system.

It contains the following sections:

• 6.1 About the L1 memory system on page 6-285.
• 6.2 Cache organization on page 6-286.
• 6.3 L1 instruction memory system on page 6-287.
• 6.4 L1 data memory system on page 6-289.
• 6.5 Program flow prediction on page 6-295.
• 6.6 L1 RAM memories on page 6-297.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 6-284
reserved.
Non-Confidential
 6 Level 1 Memory System
6.1 About the L1 memory system

6.1 About the L1 memory system

The L1 memory system consists of separate instruction and data caches.
The L1 instruction memory system has the following features:
• 48KB 3-way set-associative instruction cache.
• Fixed line length of 64 bytes.
• Parity protection per 16 bits.
• Instruction cache that behaves as Physically-indexed and physically-tagged (PIPT).
• Least Recently Used (LRU) cache replacement policy.
• MBIST support.
The L1 data memory system has the following features:
• 32KB 2-way set-associative data cache.
• Fixed line length of 64 bytes.
• ECC protection per 32 bits.
• Data cache that is PIPT.
• Out-of-order, speculative, non-blocking load requests to Normal memory and non-speculative, non-
blocking load requests to Device memory.
• LRU cache replacement policy.
• Hardware prefetcher that generates prefetches targeting both the L1 data cache and the L2 cache.
• MBIST support.

Note
The Cortex-A72 processor does not support cache lockdown.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 6-285
reserved.
Non-Confidential
 6 Level 1 Memory System
6.2 Cache organization

6.2 Cache organization

You can disable each cache independently.
On a cache miss, critical word-first filling of the cache is performed.

Related information
4.5.5 System Control Register on page 4-242.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 6-286
reserved.
Non-Confidential
 6 Level 1 Memory System
6.3 L1 instruction memory system

6.3 L1 instruction memory system

The instruction cache can source up to 128 bits per fetch depending on alignment.
Sequential cache read operations reduce the number of full cache reads. This has the benefit of reducing
power consumption. If a cache read is sequential to the previous cache read, and the read is within the
same cache line, only the data RAM way that was previously read is accessed.
The L1 instruction cache appears to software as a physically tagged, physically indexed array. Therefore,
the instruction cache is only required to be flushed when writing new data to an instruction address.
This section contains the following subsections:
• 6.3.1 Instruction cache disabled behavior on page 6-287.
• 6.3.2 Instruction cache speculative memory accesses on page 6-287.
• 6.3.3 Fill buffers on page 6-287.
• 6.3.4 Non-cacheable fetching on page 6-288.
• 6.3.5 Parity error handling on page 6-288.
• 6.3.6 Hardware L1 I-cache prefetching on page 6-288.

6.3.1 Instruction cache disabled behavior

The SCTLR.I bit enables or disables the L1 instruction cache. If the I bit is disabled, fetches cannot
access any of the instruction cache arrays. An exception to this rule is the instruction cache system
operations. If the instruction cache is disabled, the instruction cache maintenance operations can still
execute normally.

Related information
4.5.5 System Control Register on page 4-242.

6.3.2 Instruction cache speculative memory accesses

An instruction remains in the pipeline between the fetch and the execute stages. Because there can be
several unresolved branches in the pipeline, instruction fetches are speculative, meaning there is no
guarantee that they are executed. A branch or exceptional instruction in the code stream can cause a
pipeline flush, discarding the currently fetched instructions.
Because of the aggressive prefetching behavior, you must not place read-sensitive devices in the same
page as code. Pages with Device memory type attributes are treated as Non-cacheable Normal Memory.
You must mark pages that contain read-sensitive devices with the TLB Execute Never (XN) attribute bit.
To avoid speculative fetches to read sensitive devices when address translation is disabled, these devices
and code that are fetched must be separated in the physical memory map. See the ARM® Architecture
Reference Manual ARMv8 for more information. To avoid speculative fetches to potential non-code
regions, the static predictor is disabled and branches are forced to resolve in order when address
translation is disabled.

6.3.3 Fill buffers

The instruction cache is fed by three fill buffers that hold instructions returned from the L2 cache on a
linefill operation, or instructions from Non-cacheable regions. The fill buffers are non-blocking. An
instruction cache hit can bypass an in-progress cache miss, even before the critical word is returned. A
line at a given Physical Address remains in a fill buffer until the fill buffer must be reclaimed. At this
time, the fill buffer contents are either transferred to the main instruction cache or discarded if no fetch
has occurred to the address of the line over the lifetime of the line in the fill buffer.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 6-287
reserved.
Non-Confidential
 6 Level 1 Memory System
6.3 L1 instruction memory system

6.3.4 Non-cacheable fetching

Fetches occurring when the instruction cache is disabled, or from a page with attributes not indicating
Inner Cacheable, do not result in the line entering the instruction cache.
Incoming instructions from the L2 cache are stored in the fill buffers until the fetch reaches the end of the
cache line or a nonsequential fetch occurs, whichever occurs first. Therefore, multiple sequential fetches
from the same 64-byte region, corresponding to a cache line, can occur without incurring multiple L2
requests when a region is not given Cacheable attributes.

6.3.5 Parity error handling

The instruction cache implements one parity bit per 16-bits of instruction data. The instruction cache Tag
array is also protected by two parity bits per tag entry. Parity errors invalidate the offending cache line,
and force a fetch from the L2 cache on the next access. No aborts are generated on parity errors that
occur within the instruction cache. The location of a parity error is reported in the CPU Memory Error
Syndrome Register. Because the data cache shares this register, there is no guarantee that this register
contains the location of the last instruction side parity error.

Related information
4.3.68 CPU Memory Error Syndrome Register, EL1 on page 4-209.

6.3.6 Hardware L1 I-cache prefetching

The processor implements speculative prefetching on the instruction side. Following an L1 I-cache miss,
the next sequential line is looked up in the L1 instruction cache. If a miss is indicated, and no pipeline
flushes have occurred, a second L2 request is initiated for the next sequential line. This line is not
committed to the instruction cache unless actually demanded by a fetch. This is the default behavior.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 6-288
reserved.
Non-Confidential
 6 Level 1 Memory System
6.4 L1 data memory system

6.4 L1 data memory system

The L1 data memory system executes all memory operations in the Cortex-A72 processor.
In addition, it handles cache maintenance operations, TLB maintenance operations, and exclusive
operations using the Load-Exclusive, Store-Exclusive and Clear-Exclusive instructions.
The L1 memory system supports out-of-order execution of instructions. Loads can be executed and
return their data while they are still speculative and might be flushed. Stores can be executed, but not
committed to memory, while they are still speculative. Speculative loads can forward data from older
speculative stores.
The L1 memory system is non-blocking and supports hit-under-miss. For Normal memory, up to six 64-
byte cache line requests can be outstanding at a time. While those requests are waiting for memory, loads
to different cache lines can hit the cache and return their data.
The L1 data memory system includes the following:
• L1 data cache.
• Address generation logic.
• The L1 TLB.
• Buffering for stores that have not been written to the cache or memory.
• Fill buffers for processing cache line fills and Non-cacheable reads.
• Coherence logic for handling snoop requests.
• Hardware prefetch logic.
This section contains the following subsections:
• 6.4.1 Behavior for different memory types on page 6-289.
• 6.4.2 Coherence on page 6-291.
• 6.4.3 Cache disabled behavior on page 6-292.
• 6.4.4 Non-cacheable streaming enhancement on page 6-292.
• 6.4.5 Synchronization primitives on page 6-292.
• 6.4.6 Load/Store unprivileged instructions on page 6-293.
• 6.4.7 Preload instruction behavior on page 6-293.
• 6.4.8 Error Correction Code on page 6-293.
• 6.4.9 Load/store hardware prefetcher on page 6-294.

6.4.1 Behavior for different memory types

The L1 data memory system uses memory attributes from the MMU to determine the behaviors of
memory transactions to regions of memory.
The L1 data memory system uses the following memory types:
• Write-Back Read-Write-Allocate on page 6-290.
• Write-Back No-Allocate on page 6-291.
• Write-Through on page 6-291.
• Non-cacheable on page 6-291.
• Device on page 6-291.

Note
Some attribute combinations are only available if the LPAE page table format is used.

Table 6-1 Memory attribute combinations on page 6-290 shows the memory attribute combinations
available.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 6-289
reserved.
Non-Confidential
 6 Level 1 Memory System
6.4 L1 data memory system

Table 6-1 Memory attribute combinations

Outer MemAttr Inner MemAttr Cortex-A72 processor internal memory type

Device nGnRnE Device nGnRnE

Device nGnRE Device nGnRE

Device nGRE Device nGRE

Device GRE Device GRE

Non-cacheable Non-cacheable Non-cacheable

Non-cacheable Write-Through Non-cacheable

Non-cacheable Write-Back Non-cacheable

Write-Through Non-cacheable Non-cacheable

Write-Through Write-Through Non-cacheable

Write-Through Write-Back Non-cacheable

Write-Back Non-cacheable Non-cacheable

Write-Back Write-Through Non-cacheable

Write-Back Write-Back No-Allocate Write-Back No-Allocate

Write-Back Write-Back Read-Allocate Write-Back Read-Write-Allocate

Write-Back Write-Back Write-Allocate Write-Back Read-Write-Allocate

Write-Back Write-Back Read-Write-Allocate Write-Back Read-Write-Allocate

The L1 and L2 data memory system use the internal memory type to determine its behavior in addition to
the value of the ARCACHE, AWCACHE, and TXREQFLIT[MemAttr] signals. The L1 and L2 caches
use allocation hints from the inner memory attributes and the ARCACHE, AWCACHE, and
TXREQ[MemAttr] signals use allocation hints from the outer memory attributes.
Note
The Cortex-A72 processor provides the raw memory attributes from the MMU on external signals.

If any memory instruction crosses a 4KB page boundary between two pages with different memory types
such as Normal and Device memory, the result is unpredictable and an abort might be triggered or
incorrect data delivered.
If any given Physical Address is mapped to Virtual Addresses with different memory types or different
cacheability such as Non-cacheable, Write-Through, or Write-Back, the result is unpredictable. This can
occur if two Virtual Addresses are mapped to the same Physical Address at the same time with different
memory type or cacheability, or if the same Virtual Address has its memory type or cacheability changed
over time without the appropriate cache cleaning or barriers.

Write-Back Read-Write-Allocate

This is expected to be the most common and highest performance memory type. Any read or write to this
memory type searches the cache to determine if the line is resident. If it is, the line is read or updated. A
store that hits a Write-Back cache line does not update main memory.
If the required cache line is not in the cache, one or more cache lines is requested from the L2 cache. The
L2 cache can obtain the lines from its cache, from another coherent L1 cache, or from memory. The line
is then placed in the L1 cache, and the operation completes from the L1 cache.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 6-290
reserved.
Non-Confidential
 6 Level 1 Memory System
6.4 L1 data memory system

Write-Back No-Allocate

Use Write-Back No-Allocate memory to access data that might be in the cache because other virtual
pages that are mapped to the same Physical Address are Write-Back Read-Write-Allocate. Write-Back
No-Allocate memory avoids polluting the caches when accessing large memory structures that are used
only one time. The cache is searched and the correct data is delivered or updated if the data resides in one
of the caches. However, if the request misses the L1 or L2 cache, the line is not allocated into that cache.
For a read that misses all caches, the required data is read to satisfy the memory request, but the line is
not added to the cache. For a write that misses in all caches, the modified bytes are updated in memory.
Note
The No-Allocate allocation hint is only a performance hint. The processor might in some cases, allocate
Write-Back No-Allocate lines into the L1 data cache or the L2.

Write-Through

The Cortex-A72 processor memory system treats all Write-Through pages as Non-cacheable.

Non-cacheable

Normal Non-cacheable memory is not looked up in any cache. The requests are sent directly to memory.
Read requests might over-read in memory, for example, reading 64 bytes of memory for a 4-byte access,
and a single external memory access might satisfy multiple memory requests. Write requests might
merge with other write requests to the same bytes or nearby bytes.

Device

Device memory types are used for communicating with input and output devices and memory-mapped
peripherals. They are not looked up in any cache.
All the memory operations for a single instruction can be sent to the interconnect as multiple naturally
aligned requests.

Related information
Chapter 5 Memory Management Unit on page 5-275.
A.8 ACE and CHI interface signals on page Appx-A-540.

6.4.2 Coherence

All memory requests for pages that are marked as Inner Shareable in the page tables and are Write-Back
Cacheable, regardless of allocation policy, are coherent in all the caches that comprise the inner domain.
At a minimum, this includes the L1 data cache of the executing core, the L2 cache, and all other L1 data
caches in the processor. The inner domain might contain additional caches outside the processor
depending on how the system is configured.
It is unpredictable whether memory requests for pages that are marked as Inner Non-shareable are
coherent with the processor. No code must assume that Non-shareable pages are incoherent among the
caches.
The L1 data cache implements a MESI coherence protocol.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 6-291
reserved.
Non-Confidential
 6 Level 1 Memory System
6.4 L1 data memory system

6.4.3 Cache disabled behavior

When you clear the C bit in the CP15 System Control Register for a given processor, data caching is
disabled and no new cache lines are allocated to the L1 data cache and L2 cache because of data requests
from that processor. This is important when cleaning and invalidating the caches for power down.
Cache lines can be allocated from memory requests of other processors, unless their cache enable bits are
also cleared. The effect on the L1 memory system is that all Write-Back Read-Write-Allocate pages are
treated as Non-cacheable pages.
When you disable the cache, all Write-Back Cacheable requests do not look up the L1 data cache. L1
cache still services the snoops from the L2 cache.

Related information
4.5.5 System Control Register on page 4-242.

6.4.4 Non-cacheable streaming enhancement

You can enable the CPUACTLR[24], Non-cacheable streaming enhancement bit, only if your memory
system meets the requirement that cache line fill requests from the processor are atomic. Specifically, if
the processor requests a cache line fill on the AXI master read address channel, any given write request
from a different master is ordered completely before or after the cache line fill read. This means that after
the memory read for the cache line fill starts, writes from any other master to the same cache line are
stalled until that memory read completes. Setting this bit enables higher performance for applications
with streaming reads from memory types that do not allocate into the cache.
Because it is possible to build an AXI interconnect that does not comply with the specified requirement,
the CPUACTLR[24] bit defaults to disabled.

6.4.5 Synchronization primitives

The L1 memory system supports the Load-Exclusive, Store-Exclusive, and Clear-Exclusive

synchronization primitive instructions. For all Non-shareable memory pages, the synchronization
primitives are supported with a local monitor that is in each L1 memory system. For Shareable memory
pages, the local monitor is used in conjunction with a global monitor. Where the global monitor resides
depends on the memory type and cacheability.

Internal coherent global monitor

If synchronization primitives are used for memory pages that are Shareable Normal Write-Back and the
cache is enabled, SCTLR.C is 1, the external monitor on AXI is not used. Instead, the global monitor
function is handled in the L1 cache using the cache coherence information.

External global monitor

If synchronization primitives are used for memory pages that are Device, or Inner-Shareable Normal
Non-cacheable, a global monitor must be provided in the interconnect. See the ARM® Architecture
Reference Manual ARMv8, for ARMv8-A architecture profile for more information. The memory requests
are sent on the AXI interface as Read-Exclusive or Write-Exclusive. See the ARM® AMBA® AXI and ACE
Protocol Specification for more information.
Note
Use of synchronization primitives on addresses in regions marked as Device memory is UNPREDICTABLE in
the ARMv8-A architecture. Code that makes such accesses is not portable.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 6-292
reserved.
Non-Confidential
 6 Level 1 Memory System
6.4 L1 data memory system

6.4.6 Load/Store unprivileged instructions

The load/store unprivileged instructions are used in privileged modes to emulate User mode instructions
and to enforce User mode permissions. These instructions are for all memory types when enforcing
permission checking against the permissions that the page table specifies. The User mode permissions
from the page table are used instead of the privileged mode permissions.
You can also use these instructions to modify the privileged and user information on the ARPROT and
AWPROT signals on the AXI. This is required if external permission checking hardware exists in the
fabric memory.
The LDRT and STRT instructions for Strongly-ordered and Device pages appear on the AXI with an
AxPROT value that indicates User mode access. However, the same instructions for Normal Memory
might not always result in AXI transactions with an AxPROT value that indicates User mode access.
This is because any Normal Memory page permits speculative prefetching at any time. Those prefetch
requests, either caused by hardware prefetching or speculative prefetching triggered by flushed memory
instructions, can have a value of the AxPROT field that indicates privileged mode access. This reflects
the mode of the processor during the prefetch.
For Normal Write-Through Cacheable or Non-cacheable memory, the processor can still access the
memory speculatively, and can merge multiple stores together before issuing them to the AXI. Because
of this, you must use the LDRT and STRT instructions to present User mode on AxPROT if the LDRT and
STRT instructions are preceded and followed by DMB instructions:
• DMB.
• LDRT or STRT.
• DMB.
The DMB instructions prevent the LDRT or STRT instruction from hitting any previously requested read
data, or from merging with any other requests. The DMB instructions can be DMBSY, DMBISH, DMBISH, and
DMBOSH.

6.4.7 Preload instruction behavior

The multiprocessor supports the PLD, PLDW, and PRFM prefetch hint instructions. For Normal Write-Back
Cacheable memory page, the PLD, PLDW, and PRFM L1 instructions cause the line to be allocated to the L1
data cache of the executing processor. The PLD instruction brings the line into the cache in Exclusive or
Shared state and the PLDW instruction brings the line into the cache in Exclusive state. The preload
instruction cache, PLDI, is treated as a NOP. PLD and PLDW instructions are performance hints instructions
only and might be dropped in some cases.

6.4.8 Error Correction Code

The L1 data cache supports optional single bit correct and double bit detect error correction logic in both
the Tag and Data arrays. The ECC granularity for the Tag array is the tag for a single cache line and the
ECC granularity for the Data array is a 32-bit word.
Because of the ECC granularity in the Data array, a write to the array cannot update a portion of a 4-byte
aligned memory location because there is not enough information to calculate the new ECC value. This
is the case for any store instruction that does not write one or more aligned 4-byte regions of memory. In
this case, the L1 data memory system reads the existing data in the cache, merges in the modified bytes,
and calculates the ECC from the merged value. The L1 memory system attempts to merge multiple stores
together to meet the aligned 4-byte ECC granularity and to avoid the read-modify-write requirement.
Single bit ECC errors in the Tag or cache are corrected in the background. Because the line is removed
from the L1 cache as part of the correction process, no software intervention is required. No exception or
interrupt is generated. The CPU Memory Error Syndrome Register is updated to indicate a nonfatal error.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 6-293
reserved.
Non-Confidential
 6 Level 1 Memory System
6.4 L1 data memory system

Double bit ECC errors in the Tag or cache are detected and an imprecise Data Abort is triggered. The
line that contains the error is evicted from the cache. When a double bit error is reported, you must
assume that data corruption has occurred and handle this appropriately.
For any detected ECC error in the L1 memory system, the CPU Memory Error Syndrome Register is
updated. For the first error reported, the register is updated with information for the RAM, bank, way,
and index that contain the error. If that same location reports multiple errors, the repeat error count is
incremented. If any other RAM locations report errors, the other error count is incremented. Double-bit
ECC errors set the fatal bit. When the register is written with zeros, the register clears all counts and
starts to monitor for a new first error again.

Related information
4.3.68 CPU Memory Error Syndrome Register, EL1 on page 4-209.

6.4.9 Load/store hardware prefetcher

The Load/store unit includes a hardware prefetcher that is responsible for generating prefetches targeting
both the L1D cache and L2 cache.

Prefetching on loads
The load side prefetcher uses a hybrid mechanism which is based on both physical-address (PA) and
virtual-address (VA) prefetching to either or both of the L1D cache and L2 cache, depending on the
memory access patterns.

Prefetching on stores
Prefetching on store accesses is managed by a PA based prefetcher and only prefetches to the L2 cache.
The Load/Store HW prefetcher can be controlled in the following manner using software programmable
bits:
1. Disable the Load/Store HW prefetcher: The load/store HW prefetcher can be disabled by setting the
CPUACTLR_EL1 bit [56].
2. Disable VA based prefetch: Prefetching using VA can be disabled by setting the CPUACTLR_EL1
bit [43]. When set, prefetch is restricted to within the page boundary of the demand request triggering
that triggers the prefetch.
3. Disable prefetch on store: Prefetching on stores can be disabled by setting the CPUACTLR_EL1 bit
[42].
4. Maximum load prefetch distance to L2: You can control the maximum prefetch distance to the L2, for
load side prefetching, by programming bits [33:32] of the CPUECTLR_EL1 register.

Related information
4.3.66 CPU Auxiliary Control Register, EL1 on page 4-194.
4.3.67 CPU Extended Control Register, EL1 on page 4-206.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 6-294
reserved.
Non-Confidential
 6 Level 1 Memory System
6.5 Program flow prediction

6.5 Program flow prediction

The Cortex-A72 processor contains program flow prediction hardware, also known as branch prediction.
With program flow prediction disabled, all taken branches incur a penalty associated with flushing the
pipeline. To avoid this penalty, the branch prediction hardware operates at the front of the instruction
pipeline. The branch prediction hardware consists of:
• A Branch Target Buffer (BTB) to identify branches and provide targets for direct branches.
• 2-level global history-based direction predictor.
• Indirect predictor to provide targets for indirect branches.
• Return stack.
• Static predictor.
The combination of global history-based direction predictor and BTB are called dynamic predictor.
This section contains the following subsections:
• 6.5.1 Predicted and non-predicted instructions on page 6-295.
• 6.5.2 Return stack predictions on page 6-295.
• 6.5.3 Indirect predictor on page 6-296.
• 6.5.4 Static predictor on page 6-296.
• 6.5.5 Enabling program flow prediction on page 6-296.
• 6.5.6 BTB invalidation and context switches on page 6-296.

6.5.1 Predicted and non-predicted instructions

This section describes the instructions that the processor predicts.
Unless otherwise specified, the list applies to A32, T32, and A64 instructions. As a general rule, the
branch prediction hardware predicts all branch instructions regardless of the addressing mode, including:
• Conditional branches.
• Unconditional branches.
• Indirect branches.
• Branches that switch between ARM and Thumb states.
• PC destination data processing operations.
• BXJ, because of the inclusion of the trivial Jazelle implementation, this degenerates to a BX
instruction. There is no BXJ instruction in A64.
However, the following branch instructions are not predicted:
• AArch32 instructions with the S suffix are not predicted because they are typically used when
returning from exceptions and have side effects that can change privilege mode and Security state.
• All mode or Exception level changing instructions.
In Thumb state, you can make a branch that is normally encoded as unconditional conditional by
including an If-Then (IT) block. It is then treated as a normal conditional branch.

6.5.2 Return stack predictions

The return stack stores the address and the ARM or Thumb state of the instruction after a function-call
type branch instruction. This address is the same as the Link Register value stored in r14 in AArch32
state or X30 in AArch64 state. The following instructions cause a return stack push if predicted:
• BL immediate.
• BLX(1) immediate in AArch32 state.
• BLX(2) register in AArch32 state.
• BLR register in AArch64 state.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 6-295
reserved.
Non-Confidential
 6 Level 1 Memory System
6.5 Program flow prediction

The following AArch32 instructions cause a return stack pop if predicted:

• BX r14
• MOV pc, r14.
• LDMIA sp!, {..pc}.
• LDR pc, [sp], #4.
The LDMIA and LDR instruction address modes are correspondent with popping the return address of a full
descending stack. In AArch64 state, the RET instruction causes a return stack pop. There is no
dependency on a specific return address target register, for example X30.
Because return-from-exception instructions can change the processor privilege mode and Security state,
they are not predicted. This includes the ERET, RFE, and LDM(3) instruction, and the MOVS pc, r14
instruction.

6.5.3 Indirect predictor

The indirect predictor can predict indirect branches that are not return-type instructions.
This predictor augments the branch address with an additional state that predicts the target address of an
indirect branch. The conditional branch predictor still predicts the direction of conditional indirect
branches. The indirect predictor only provides the address on a predicted taken conditional indirect
branch.

6.5.4 Static predictor

Branches must be resolved one time to be predicted by the dynamic predictor. To accelerate cold startup
of code, the processor includes a static predictor that detects branches in the code stream as follows:
• Direct unconditional branches, B immediate, are predicted taken.
• Direct unconditional call-type branches, BL immediate and BLX immediate, are predicted taken, and
the preferred return address value is pushed on the return stack.
• Unconditional return-type branches are predicted taken and the target is popped from the return stack.
To avoid potential illegal speculation, the static predictor is disabled when the MMU is disabled.

Related information
6.5.2 Return stack predictions on page 6-295.

6.5.5 Enabling program flow prediction

Program flow prediction is always enabled and no programming is required to take advantage of
program flow prediction.
When reset, the processor:
• Invalidates the BTB.
• Resets the GHB and indirect predictor to a known state.
No software intervention is required to prepare the prediction logic before enabling program flow
prediction.

6.5.6 BTB invalidation and context switches

The BTB is tagged by all memory space information required to uniquely identify a virtual memory
space, ASID, VMID, security, and Exception level. All predictions are checked at branch resolution time
to ensure that a legal branch is resolved. Therefore, flushing the BTB on a context switch is not required.
AArch64 state does not implement BTB flush instructions.
The processor automatically invalidates the BTB when either stage of the MMU is disabled.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 6-296
reserved.
Non-Confidential
 6 Level 1 Memory System
6.6 L1 RAM memories

6.6 L1 RAM memories

The L1 memory system contains several RAM memories that can be configured to use ECC or parity
error detection mechanisms.
Any RAM memory that uses ECC support can perform single bit error correction and double bit error
detection. Contents of the RAM memories with parity support can invalidate entries if a parity error is
detected because this data is associated with read-only structures.
The following table shows all RAM memories contained in the L1 memory system.

Table 6-2 L1 RAM memories

RAM memory ECC or Parity

L1 instruction Tag RAM Parity

L1 instruction Data RAM Parity

L1 instruction BTB RAM None

L1 instruction GHB RAM None

L1 instruction indirect predictor RAM None

L1 data Tag RAM ECC

L1 Data RAM ECC

L1 PF PHT RAM None

L2 TLB RAMdn Parity

dn The L2 TLB RAM is a unified TLB structure that supports L1 instruction and L1 data TLB misses.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 6-297
reserved.
Non-Confidential
Chapter 7
Level 2 Memory System

This chapter describes the Level 2 (L2) memory system.

It contains the following sections:

• 7.1 About the L2 memory system on page 7-299.
• 7.2 Cache organization on page 7-300.
• 7.3 L2 RAM memories on page 7-306.
• 7.4 L2 cache prefetcher on page 7-307.
• 7.5 Cache coherency on page 7-308.
• 7.6 Asynchronous errors on page 7-309.
• 7.7 External coherent interfaces on page 7-310.
• 7.8 ACP on page 7-317.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-298
reserved.
Non-Confidential
 7 Level 2 Memory System
7.1 About the L2 memory system

7.1 About the L2 memory system

The L2 memory system consists of a tightly-coupled L2 cache and an integrated Snoop Control Unit
(SCU), connecting up to four cores within a cluster and a configurable coherent external interface
supporting AMBA4 (ACE) or CHI architectures.
The L2 memory system also interfaces with an optional Accelerator Coherency Port (ACP) that is
implemented as an AXI slave interface.
The features of the L2 memory system include:
• Configurable L2 cache size of 512KB, 1MB, 2MB and 4MB.
• Fixed line length of 64 bytes.
• Physically indexed and tagged cache.
• 16-way set-associative cache structure.
• Banked pipeline structures.
• Inclusion property with L1 data caches.
• Software-programmable pseudo-least-recently-used or pseudo-random cache-replacement policy.
• Configurable 128-bit wide ACE or 128-bit wide CHI interface with support for multiple outstanding
requests.
• Optional 128-bit wide ACP with support for multiple incoming requests.
• Duplicate copies of the L1 data cache directories for coherency support.
• Configurable number of Fill/Eviction Queue (FEQ) entries to 20, 24, or 28.
• Optional Error Correction Code (ECC) support.
• Optional hardware prefetch support.
• Software-programmable variable latency RAMs.
• Register slice support for large L2 cache sizes to minimize impact on routing delays.
• MBIST support.

Note
• The Cortex-A72 processor does not support TLB or cache lockdown.

Related information
7.2.2 Strictly-enforced inclusion property with L1 data caches on page 7-300.
7.2.4 Error Correction Code on page 7-301.
7.2.5 Register slice support for large cache sizes on page 7-301.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-299
reserved.
Non-Confidential
 7 Level 2 Memory System
7.2 Cache organization

7.2 Cache organization

The L2 cache is 16-way set-associative of configurable size. The cache is physically-addressed. The
cache sizes are configurable with sizes of 512KB, 1MB, 2MB, and 4MB.
You can configure the L2 memory system pipeline to insert wait states to take into account the latencies
of the compiled memories for the implementation of the RAMs.
The L2 cache incorporates a single dirty bit per cache line. A write to a cache line results in the line
being written back to memory after the line is evicted from the L2 cache.
This section contains the following subsections:
• 7.2.1 L2 cache bank structure on page 7-300.
• 7.2.2 Strictly-enforced inclusion property with L1 data caches on page 7-300.
• 7.2.3 Enabling and disabling the L2 cache on page 7-301.
• 7.2.4 Error Correction Code on page 7-301.
• 7.2.5 Register slice support for large cache sizes on page 7-301.

7.2.1 L2 cache bank structure

The L2 cache is partitioned into multiple banks to enable parallel operations. The following levels of
banking exist:
• The Tag array is partitioned into multiple banks to enable up to two requests to access different tag
banks of the L2 cache simultaneously.
• Each tag bank is partitioned into multiple data banks to enable streaming accesses to the data banks.
Each tag bank consists of four data banks.
Figure 7-1 L2 cache bank structure on page 7-300 shows the logical representation of an L2 cache bank
structure with a configuration of all possible tag and data bank combinations.
Tag bank selected by PA [6]
Data bank selected by PA [5:4]

Data bank 0 Data bank 0

Data bank 1 Data bank 1

Tag bank 0 requests Tag bank 1 requests
Data bank 2 Data bank 2

Data bank 3 Data bank 3

Tag bank 0 Tag bank 1

Figure 7-1 L2 cache bank structure

7.2.2 Strictly-enforced inclusion property with L1 data caches

The L2 memory system requires support for inclusion between the L1 data caches and the L2 cache. A
line that resides in any of the L1 data caches must also reside in the L2 cache. However, the data can
differ between the two caches when the L1 cache line is in a dirty state. If another agent, a core in the
cluster or another cluster, accesses this line in the L2 then it knows the line is present in the L1 of a
processor and then it queries that core for the most recent data.
This strictly-enforced inclusion property has the following benefits:
• Any AXI or CHI ReadClean operation that results in a line being in shared state in the L1 data caches
can be returned from the L2 cache. This yields the highest performance for delivering data to a core.
• When powering down the processor, it reduces the time to clean and invalidate the entire L1 data
cache.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-300
reserved.
Non-Confidential
 7 Level 2 Memory System
7.2 Cache organization

7.2.3 Enabling and disabling the L2 cache

For processor requests, the L2 cache is enabled when the C bit of the SCTLR register is enabled. The
cache attributes are provided with each request, taking into account the page attributes that the MMU
page tables provided and overriding these attributes if the corresponding cache enable bit in the SCTLR
is disabled.
To enable the L2 cache to cache both instructions and data following the reset sequence, you must:
1. Complete the processor reset sequence.
2. Enable L2 ECC, if present and required, by programming bit[21] of the L2 Control Register.
3. Program the I bit and C bit of the SCTLR.
To disable the L2 cache, you must use the following sequence:
1. Disable the C bit.
2. Clean and invalidate the L1 and L2 caches.
For ACP requests, the L2 cache is enabled if the request uses Normal Write-Back memory attributes. The
processor searches the L2 cache to determine if the request is valid before allocating the line for Normal
Write-Back Read-Write-Allocate memory.

Related information
4.5.5 System Control Register on page 4-242.
4.3.58 L2 Control Register, EL1 on page 4-168.

7.2.4 Error Correction Code

The L2 cache supports optional ECC in most of its memories. For core instruction and data accesses
resulting in an L2 cache hit, where a single-bit error is detected on the Data array, the L2 memory system
supports in-line ECC correction. Uncorrected data is forwarded to the requesting unit, and in parallel, the
ECC circuitry checks for accuracy. If a single-bit error is detected, any uncorrected data returned within
two cycles before the error indicator must be discarded. The L2 memory system begins to stream
corrected data to the requestor.
When there is no data transfers, the L2 memory system shifts back to return uncorrected data until it
detects the next single-bit error. Forwarding uncorrected data can be disabled by programming bit[20] of
the L2 Control Register. This avoids the requirement to flush requests associated with single-bit ECC
errors on L2 cache hits, but adds an additional 2 cycles to the L2 hit latency.
For all other single-bit ECC errors detected, the request is flushed from the L2 pipeline and is forced to
reissue. The tag bank where the single-bit error occurred, performs a read-modify-write sequence to
correct the single-bit error in the array. The request is then reissued.

Related information
4.3.68 CPU Memory Error Syndrome Register, EL1 on page 4-209.

7.2.5 Register slice support for large cache sizes

As the L2 cache size is increased, the area of the implementation increases. This increase adds significant
route delays to and from the RAM memories. This increase can impact the maximum frequency of the
implementation.
To counter this, you can insert register slices before and after the RAM memories to offset the longer
route delays. This enables the frequency target of the implementation to remain high. Additional slices
can impact the overall L2 hit latency but they can enable requests to be streamed in a more efficient
manner. You can increase the programmed latency values of the RAMs to cover the additional route
delays without adding the slices. However, this method has an impact on performance because requests
cannot be streamed as efficiently.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-301
reserved.
Non-Confidential
 7 Level 2 Memory System
7.2 Cache organization

The L2 RAMs support one inserted register slice. Each register slice introduces a pair of registers, one
before the RAM and one after the RAM.
Bits[12] and [10] of the CP15 L2 Control Register, L2CTLR, indicate the presence of RAM register
slices in the design. In addition, the L2CTLR contains bits to program the setup and latency for the L2
Tag and Data RAMs.

Overall RAM latency calculation

The RAM latency is a function of the following:

• Programmed latency in the L2 Control Register.
• Additional strobe clock setup required value in the L2CTLR.
RAM latency = programmed value + strobe setup.
The RAM latency determines the rate at which back to back operations to the RAM can be scheduled.
The total effective latency = RAM latency + 2×N, where N is the number of register slices to insert.
The slices are considered pipeline registers and do not affect the throughput rate of RAM accesses.
The following table shows the total effective L2 Tag latency with the register slice and setup factored in.

Table 7-1 Total effective L2 Tag latency with slice and setup factored in

L2CTLR[8:6] register bits Total effective Tag latency

Tag slice =0 Tag slice =0 Tag slice =1 Tag slice =1

Tag setup =0 Tag setup =1 Tag setup =0 Tag setup =1

000do 2 3 4 5

001 2 3 4 5

010 3 4 5 5

011 4 5 5 5

100 5 5 5 5

1xx, ≥ 4 5 5 5 5

Note
• The total effective L2 Tag latency is set to a maximum of 5 cycles.
• Each tag slice adds 2 cycles and affects the L2 Tag, Snoop Tag, Dirty, and Inclusion PLRU RAMs.
• Setting tag setup to 1 adds 1 cycle.
• Slice and setup have priority over programmed latency in determining the total effective L2 Tag
latency.

The following example shows a Tag RAM access with 3 cycles total effective Tag latency.

do This is the reset value.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-302
reserved.
Non-Confidential
 7 Level 2 Memory System
7.2 Cache organization

Example 7-1 Examples

Tag RAM access with 3 cycles total latency

When tag slice = 0, L2CTLR[9] = 0, L2CTLR[8:6] = 0b010, the following applies:
• No slice cycle.
• No setup cycle.
• 3 cycles Tag RAM access.
• 3 cycles total effective Tag latency.

The following example shows a Tag RAM access with 4 cycles total effective Tag latency.

Example 7-2 Examples

Tag RAM access with 4 cycles total latency

When tag slice = 0, L2CTLR[9] = 1, L2CTLR[8:6] = 0b010, the following applies:
• No slice cycle.
• 1 setup cycle.
• 4 cycles Tag RAM access (Programmed + Setup).
• 4 cycles total effective Tag latency.

The following example shows a Tag RAM access with 5 cycles total effective Tag latency.

Example 7-3 Examples

Tag RAM access with 5 cycles total latency

When tag slice = 1, L2CTLR[9] = 1, L2CTLR[8:6] = 0b010, the following applies:
• 2 slice cycles.
• 1 setup cycle.
• 3 cycles Tag RAM access (Programmed + Setup, capped to 3 because of slices).
• 5 cycles total effective Tag latency.

The following table shows the total effective L2 Data latency with the register slice and setup factored in.

Table 7-2 Total effective L2 Data latency with slice and setup factored in

L2CTLR[2:0] register Total effective Data latency

bits
Data slice =0 Data slice =0 Data slice =1 Data slice =1 Data slice =2 Data slice =2
Data setup =0 Data setup =1 Data setup =0 Data setup =1 Data setup =0 Data setup =1

000dp 2 3 4 5 6 7

001 2 3 4 5 6 7

010 3 4 5 6 7 8

011 4 5 6 6 8 8

dp This is the reset value.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-303
reserved.
Non-Confidential
 7 Level 2 Memory System
7.2 Cache organization

Table 7-2 Total effective L2 Data latency with slice and setup factored in (continued)

L2CTLR[2:0] register Total effective Data latency

bits
Data slice =0 Data slice =0 Data slice =1 Data slice =1 Data slice =2 Data slice =2
Data setup =0 Data setup =1 Data setup =0 Data setup =1 Data setup =0 Data setup =1

100 5 6 6 6 8 8

101, 11x >=5 6 6 6 6 8 8

Note
• The total effective L2 Data latency is set to a maximum of 8 cycles for configurations supporting
Data slice=2, otherwise the maximum is set to 6 cycles.
• Each data slice adds 2 cycles and affects the L2 data and data ECC RAMs.
• Setting data setup to 1 adds 1 cycle.
• Slice and setup have priority over programmed latency in determining the total effective L2 Data
latency.

The following example shows a Data RAM access with 4 cycles total effective Data latency.

Example 7-4 Data RAM access with 4 cycles total latency

When data slice = 0, L2CTLR[5] = 0, L2CTLR[2:0] = 0b011, the following applies:

• No slice cycle.
• No setup cycle.
• 4 cycles Data RAM access.
• 4 cycles total effective Data latency.

The following example shows a Data RAM access with 5 cycles total effective Data latency.

Example 7-5 Data RAM access with 5 cycles total latency

When data slice = 0, L2CTLR[5] = 1, L2CTLR[2:0] = 0b011, the following applies:

• No slice cycle.
• 1 setup cycle.
• 5 cycles Data RAM access (Programmed + Setup).
• 5 cycles total effective Data latency.

The following example shows a Data RAM access with 6 cycles total effective Data latency.

Example 7-6 Data RAM access with 6 cycles total latency

When data slice = 1, L2CTLR[5] = 1, L2CTLR[2:0] = 0b011, the following applies:

• 2 slice cycles.
• 1 setup cycle.
• 4 cycles Data RAM access (Programmed + Setup, capped to 4 due to register slices).
• 6 cycles total effective Data latency.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-304
reserved.
Non-Confidential
 7 Level 2 Memory System
7.2 Cache organization

The following example shows a Data RAM access with 8 cycles total effective Data latency.

Example 7-7 Data RAM access with 8 cycles total latency

When data slice = 2, L2CTLR[5] = 1, L2CTLR[2:0] = 0b011, the following applies:

• 4 slice cycles.
• 1 setup cycle.
• 4 cycles Data RAM access (Programmed + Setup, capped to 4 due to register slices).
• 8 cycles total effective Data latency.

Related information
4.3.58 L2 Control Register, EL1 on page 4-168.

Related information
4.5.5 System Control Register on page 4-242.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-305
reserved.
Non-Confidential
 7 Level 2 Memory System
7.3 L2 RAM memories

7.3 L2 RAM memories

The L2 memory system contains several RAM memories that support optional ECC or parity error
detection mechanisms.
Any RAM memory that uses ECC support can perform single-bit error correction and double-bit error
detection. Contents of the RAM memories with parity support can invalidate entries if a parity error is
detected because this data is associated with read-only structures.
The following table shows all RAM memories contained in the L2 memory system.

Table 7-3 L2 RAM memories

RAM memory ECC or Parity

L2 Tag RAM ECC

L2 Snoop Tag RAM ECC

L2 Data RAM ECC

L2 Dirty RAM ECC

L2 Inclusion PLRU RAM ECC

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-306
reserved.
Non-Confidential
 7 Level 2 Memory System
7.4 L2 cache prefetcher

7.4 L2 cache prefetcher

The Cortex-A72 processor includes a hardware L2 prefetcher that handles prefetch generation for
instruction fetch and TBW descriptor accesses.

Note
The Load/store unit handles prefetch generation for Load/store accesses targeting both the L1D cache
and L2 cache.

Some of the key features are:

• Software-programmable prefetches on any instruction fetch L2 miss of 0, 1, 2, or 3 prefetches. All
prefetches are allocated into the L2 cache.
• Separate mechanisms to detect and prefetch:
— Instruction fetch streams, to fetch consecutive cache lines on an L2 instruction fetch access.
— Table walk descriptor, to fetch the consecutive cache line on an L2 table walk descriptor access.

Note
The prefetcher is limited to prefetch within the 4KB page of the current request, if the page has been
mapped at a 4KB granularity.

• Support for forwarding from prefetched requests. If a read request was sent over AXI because of a
prefetch request, and a demand access for the same line was received, the read data can be forwarded
from the internal data buffers to the demand request, before waiting for the line to be allocated to the
cache.
You can program the CPUECTLR register to indicate the maximum number of prefetches to be allocated
in the PRQ on the following:
• An instruction fetch miss in the L2 cache by programming CPUECTLR_EL1[36:35].
The programmed distance is also used as the skip distance for any instruction fetch read with a stride
match that hits in the L2 cache. In these cases, a single prefetch request is allocate in the PRQ as:
prefetch address = current address + (stride × programmed distance)
Note
The stride for an instruction fetch access is always one cache line.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-307
reserved.
Non-Confidential
 7 Level 2 Memory System
7.5 Cache coherency

7.5 Cache coherency

The SCU uses hybrid Modified Exclusive Shared Invalid (MESI) and Modified Owned Exclusive Shared
Invalid (MOESI) protocols to maintain coherency between the individual L1 data caches and the L2
cache.
The L1 data caches support the MESI protocol. The L2 memory system contains a Snoop Tag array that
is a duplicate copy of each of the L1 data cache directories. The Snoop Tag array reduces the amount of
snoop traffic between the L2 memory system and the L1 memory system. Any line that resides in the
Snoop Tag array in the Modified/Exclusive state belongs to the L1 memory system. Any access that hits
against a line in this state must be serviced by the L1 memory system and passed to the L2 memory
system. If the line is invalid or in the shared state in the Snoop Tag array, then the L2 cache can supply
the data.
The SCU contains buffers that can handle direct cache-to-cache transfers between processors without
reading or writing any data on the ACE or CHI interface. Lines can migrate back and forth without any
change to the MOESI state of the line in the L2 cache.
Shareable transactions on the optional ACP are also coherent, so the Snoop Tag arrays are queried as a
result of ACP transactions. For reads where the Shareable line resides in one of the L1 data caches in the
Modified/Exclusive state, the line is transferred from the L1 memory system to the L2 memory system
and passed back on the ACP.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-308
reserved.
Non-Confidential
 7 Level 2 Memory System
7.6 Asynchronous errors

7.6 Asynchronous errors

The L2 memory system has two outputs that indicate asynchronous error conditions.
An asynchronous external error condition exists when either:
• The nEXTERRIRQ output is LOW.
• The nINTERRIRQ output is LOW.
If an asynchronous error condition is detected, the corresponding bit in the L2 Extended Control Register
is asserted. The asynchronous error condition can be cleared by writing 0b0 to the corresponding bit of
the L2ECTLR. Software can only clear the L2ECTLR. Any attempt to assert the error by writing the
L2ECTLR is ignored.
Any external error associated with a load instruction is reported back to the requestor along with an error
response and this might trigger an abort. Any external error associated with a Device, Non-cacheable, or
non-allocating write that misses in the L2, or a cache maintenance operation is reported to the core that
issued the transaction through a processor-specific interrupt request to the GIC.

Related information
4.3.59 L2 Extended Control Register, EL1 on page 4-172.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-309
reserved.
Non-Confidential
 7 Level 2 Memory System
7.7 External coherent interfaces

7.7 External coherent interfaces

The Cortex-A72 processor provides configurable options for either AMBA4 AXI Coherency Extensions
(ACE) or CHI interconnect architectures.
Each interface option provides a 128-bit wide data interface to the system and supports 1:1 clock ratios
with respect to the processor clock and N:1, integer multiple clock ratios, of the core clock.
Note
ACE is supported with the following restriction:
• ARQOS and AWQOS signals are not present.

This section contains the following subsections:

• 7.7.1 L2 memory interface attributes on page 7-310.
• 7.7.2 Interface modes on page 7-311.
• 7.7.3 Snoop filter support on page 7-312.
• 7.7.4 Distributed virtual memory transactions on page 7-312.
• 7.7.5 External memory attributes on page 7-312.
• 7.7.6 ACE ARID and AWID assignment on page 7-313.
• 7.7.7 CHI LPID assignment on page 7-314.
• 7.7.8 ACE supported transfers on page 7-314.
• 7.7.9 CHI link layer flow control on page 7-315.
• 7.7.10 CHI DVM acceptance capability on page 7-315.
• 7.7.11 L2 Auxiliary Control Register settings on page 7-315.

7.7.1 L2 memory interface attributes

The following table shows the L2 memory interface attributes for the processor. The table lists the
maximum possible values for the read and write issuing capabilities.

Table 7-4 L2 memory interface attributes

Attribute Value Description

Write issuing 16 16 outstanding writes supported that can be evictions, single writes, or write bursts of any
capability memory type.

Read issuing 19, 23, or 27 If the core implements:

capability
20-entry FEQ 19 outstanding reads supported that can be line fills, single reads, or read
bursts of any memory type.
24-entry FEQ 23 outstanding reads supported that can be line fills, single reads, or read
bursts of any memory type.
28-entry FEQ 27 outstanding reads supported that can be line fills, single reads, or read
bursts of any memory type.

Snoop acceptance 20 Up to 20 outstanding snoop requests are accepted on the AC channel in response to those
capability requests on the CR channel.

DVM issuing 38, 46, or 54 If the processor implements:

capability
20-entry FEQ 38 DVM Message transactions supported (19 two-part messages)
24-entry FEQ 46 DVM Message transactions supported (23 two-part messages)
28-entry FEQ 54 DVM Message transactions supported (27 two-part messages)

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-310
reserved.
Non-Confidential
 7 Level 2 Memory System
7.7 External coherent interfaces

7.7.2 Interface modes

The ACE and CHI coherent interconnect interfaces can be configured through input signals to change the
interface behavior. The multiprocessor implements the following configuration signals:
• SYSBARDISABLE on page 7-311.
• BROADCASTINNER on page 7-311.
• BROADCASTOUTER on page 7-311.
• BROADCASTCACHEMAINT on page 7-312.

SYSBARDISABLE

SYSBARDISABLE controls issuing barrier transactions on the coherent interconnect.

When SYSBARDISABLE is deasserted, barriers are broadcast on the coherent interconnect as a
Memory Barrier or Synchronization Barrier for an ACE interface, or an EOBarrier or ECBarrier for a
CHI interface.
When SYSBARDISABLE is asserted, barriers are not broadcast on the coherent interconnect. Barriers
are enforced internally to the Cortex-A72 processor by observing completion of transactions through the
Read data channel and Write response channel for an ACE interface, or the RXDAT data channel and
RXRSP response channel for a CHI interface. Systems that use this mode must ensure that ACE write
responses or CHI RXRSP completion responses guarantee that the transaction has been globally
observed and that barrier broadcasts are not required for any other system functionality.
For ACE configurations that require AXI3 compatibility you must:
• Assert SYSBARDISABLE.
• Deassert BROADCASTINNER.
• Deassert BROADCASTOUTER.
• Deassert BROADCASTCACHEMAINT.

BROADCASTINNER

BROADCASTINNER controls issuing coherent transactions targeting the Inner Shareable domain on
the coherent interconnect. When BROADCASTINNER is asserted, the processor is considered to be
part of an Inner Shareable domain that extends beyond the processor and any transaction that requires
coherency with other masters in this domain is broadcast on the ACE or CHI interface.
When BROADCASTINNER is asserted, BROADCASTOUTER must also be asserted. In this
configuration, coherent masters can share memory in the Inner or Outer Shareable domains.
When BROADCASTINNER is deasserted, the processor does not issue DVM requests on the ACE AR
channel or CHI TXREQ channel.

BROADCASTOUTER

BROADCASTOUTER controls issuing coherent transactions targeting the outer shareability domain on
the coherent interconnect. When BROADCASTOUTER is asserted, the processor is considered to be
part of the Outer Shareable domain and any transaction that requires coherency with other masters in this
domain is broadcast on the ACE or CHI interface.
It is possible to assert BROADCASTOUTER without asserting BROADCASTINNER. This selects a
configuration that limits coherent masters to sharing memory only in the outer shareability domain.
However, cores within the cluster can still share memory in the Inner Shareable domain.
When BROADCASTOUTER is deasserted, BROADCASTINNER must also be deasserted.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-311
reserved.
Non-Confidential
 7 Level 2 Memory System
7.7 External coherent interfaces

When BROADCASTINNER and BROADCASTOUTER are both deasserted, the processor does not
issue coherent read or write requests on the ACE AR and AW channels, or the CHI TXREQ channel.

BROADCASTCACHEMAINT

BROADCASTCACHEMAINT controls issuing cache maintenance transactions, such as CleanShared,

CleanInvalid and MakeInvalid, on the coherent interconnect. When BROADCASTCACHEMAINT is
asserted, cache maintenance instructions might cause cache maintenance transactions on ACE or CHI
interconnect. The cache maintenance transactions are broadcast even if the memory location is Non-
shareable or when the BROADCASTINNER or BROADCASTOUTER signals normally prevent such
a broadcast. This configuration allows the management of an external L3 cache that might cache Non-
shareable data.
When BROADCASTCACHEMAINT is deasserted, only those cache maintenance transactions
required for coherency as determined by Inner and Outer shareability and the BROADCASTINNER
and BROADCASTOUTER signals are issued on the coherent interconnect.
Systems that utilize a L3 cache that supports caching of Non-shareable memory must assert
BROADCASTCACHEMAINT.

7.7.3 Snoop filter support

In general, the processor can issue a Write-Back, WriteEvict, or an Evict transaction for any cache line
that is removed from the L2 cache. You can use these messages to manage an external snoop filter.
However, the snoop filter logic must not depend on such a message for every clean line dropped from the
processor caches, because in some circumstances the processor might not signal an eviction. For
example, clean evictions are not guaranteed to occur in cases involving L1 or L2 tag ECC errors.

7.7.4 Distributed virtual memory transactions

In a system where the processor can receive a Distributed Virtual Memory (DVM) synchronization
message over the AXI master snoop address channel, BRESP for any write transaction must not be
asserted to the core until all AXI masters that might have initiated the DVM synchronization request
observe the transaction.
Note
The Cortex-A72 processor does not support a multi-part DVM hint message.

7.7.5 External memory attributes

The Cortex-A72 processor uses a combination of inner and outer memory attributes from the MMU to
determine how its memory system handles each combination. Table 6-1 Memory attribute combinations
on page 6-290 shows the Inner and Outer memory attributes used by the L1 and L2 caches to form the
internal memory types. Table 7-5 External memory attributes on page 7-312 shows how these attributes
are used to form the external memory type presented on ARCACHE, AWCACHE, or
TXREQFLIT[MemAttr].

Table 7-5 External memory attributes

Outer MemAttr Inner MemAttr External memory type ARCACHE AWCACHE TREQFLIT[MemAttr]

Device nGnRnE Strongly-ordered 0b0000 0b0000 0b0010

Device nGnRE Device 0b0001 0b0001 0b0011

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-312
reserved.
Non-Confidential
 7 Level 2 Memory System
7.7 External coherent interfaces

Table 7-5 External memory attributes (continued)

Outer MemAttr Inner MemAttr External memory type ARCACHE AWCACHE TREQFLIT[MemAttr]

Device nGRE Device 0b0001 0b0001 0b0011

Device GRE Device 0b0001 0b0001 0b0011

Non-cacheable Non-cacheable Non-cacheable Bufferable 0b0011 0b0011 0b0001

Non-cacheable Write-Through Non-cacheable Bufferable 0b0011 0b0011 0b0001

Non-cacheable Write-Back Non-cacheable Bufferable 0b0011 0b0011 0b0001

Write-Through Non-cacheable Non-cacheable Bufferable 0b0011 0b0011 0b0001

Write-Through Write-Through Non-cacheable Bufferable 0b0011 0b0011 0b0001

Write-Through Write-Back Non-cacheable Bufferable 0b0011 0b0011 0b0001

Write-Back Non-cacheable Non-cacheable Bufferable 0b0011 0b0011 0b0001

Write-Back Write-Through Non-cacheable Bufferable 0b0011 0b0011 0b0001

Write-Back No-Allocate Write-Back Write-Back No-Allocate 0b1011 0b0111 0b0101

Write-Back Read-Allocate Write-Back Write-Back Read-Allocate 0b1111 0b0111 0b1101

Write-Back Write-Allocate Write-Back Write-Back Write-Allocate 0b1011 0b1111 0b1101

Write-Back Read-Write- Write-Back Write-Back Read-Write- 0b1111 0b1111 0b1101

Allocate Allocate

In addition to ARCACHE, AWCACHE, and TXREQFLIT[MemAttr] the processor also presents the
raw outer memory attributes, inner memory type, and Inner and Outer Shareable on dedicated external
interface signals RDMEMATTR, WRMEMATTR, and REQMEMATTR corresponding to
transactions on ACE read channel, ACE write channel and CHI TXREQ channel, respectively.

Related information
A.9 CHI interface signals on page Appx-A-543.
A.10 ACE interface signals on page Appx-A-548.

7.7.6 ACE ARID and AWID assignment

When the system issues multiple requests on the AR channel with the same ARID, or on the AW channel
with the same AWID, it must follow the appropriate ordering rules as described in the ARM® AMBA® AXI
and ACE Protocol Specification.
For certain transactions, the system must be able to identify which core generated the request. This
applies to requests affecting the global exclusive monitor in addition to Strongly-ordered or Device
memory type accesses to peripherals.
ARCACHEM[3:0] and AWCACHEM[3:0] identify whether the memory types are Strongly-ordered or
Device. See the ARM® AMBA® AXI and ACE Protocol Specification. For these memory types, if
ARIDM[2] or AWIDM[2] is LOW, then the request is generated from one of the cores. ARIDM[1:0] or
AWIDM[1:0] indicate which core generated the request.
For an exclusive read transaction such as ARLOCK asserted, ARID[1:0] indicates which core generated
the request. Only cores can generate exclusive read requests, and not the ACP or any other source.
For an exclusive write transaction such as AWLOCK asserted, AWID[1:0] indicates which core
generated the request. Only cores can generate exclusive write requests, and not the ACP or any other
source.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-313
reserved.
Non-Confidential
 7 Level 2 Memory System
7.7 External coherent interfaces

The system must not rely on specific values of ARID or AWID that correspond with specific transaction
sources or transaction types other than the information described in this section.

7.7.7 CHI LPID assignment

CHI TXREQ transactions include the Logical Processor ID (LPID) field. This field uniquely identifies
the logical core that generated the request transaction.
The processor uses the following LPID values:
0b000 Core 0 request.
0b001 Core 1 request.
0b010 Core 2 request.
0b011 Core 3 request.
0b100 ACP request.
0b111 L2 hardware flush request.

Secondary transactions such as copybacks from the L2, because of cache fills caused by core or ACP
access L2 misses, use the LPID of the request that caused the copyback.

7.7.8 ACE supported transfers

For Normal Inner-Cacheable memory transfers initiated from one of the Cortex-A72 processors, the
following transfers are supported on the ACE:
• WRAP 4× 128-bit read transfers.
• WRAP 4× 128-bit write transfers.
For Normal Non-Cacheable memory transfers initated from one of the Cortex-A72 processors, the
following transfers are supported on the ACE:
• WRAP 4× 128-bit read transfers.
• WRAP 4× 128-bit write transfers.
• INCR 1× 128-bit read transfers.
• INCR 1× 128-bit write transfers.
For Strongly-ordered or Device transactions initiated from one of the Cortex-A72 processors, the
following transfers are supported on the ACE:
• INCR 1× 8-bit read transfers.
• INCR 1× 16-bit read transfers.
• INCR 1× 32-bit read transfers.
• INCR 1× 64-bit read transfers.
• INCR N (N:1 or 4) 128-bit read transfers.
• INCR 1× 8-bit write transfers.
• INCR 1× 16-bit write transfers.
• INCR 1× 32-bit write transfers.
• INCR 1× 64-bit write transfers.
• INCR N (N:1 or 4) 128-bit write transfers.
The following table describes the use of the burst types for Non-Cacheable and Cacheable but not
allocated memory attributes.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-314
reserved.
Non-Confidential
 7 Level 2 Memory System
7.7 External coherent interfaces

Table 7-6 Use of WRAP and INCR burst types for Non-Cacheable and Cacheable but not allocated transactions

Burst type Used by

WRAP • Non-Cacheable read transactions, excluding tablewalk or exclusive accesses.

• Cacheable but not allocated read transactions initiated from one of the Cortex-A72 processors issued on AR-
channel.

INCR • Non-cacheable tablewalk and exclusive read transactions.

• Non-Cacheable or Cacheable but not allocated write transactions initiated from one of the Cortex-A72 processors
issued on AW-channel.

If there are requests on the ACP interface, the following transfers can be generated on the ACE if
comparable requests are received on the ACP:
• WRAP N 4× 128-bit read transfers.
• WRAP 4× 128-bit write transfers.
• INCR 1× 128-bit read transfers.
• INCR 1× 128-bit write transfers.

7.7.9 CHI link layer flow control

CHI link layer flow control uses a counter on each link to track the number of outstanding link layer
credits. The Cortex-A72 processor can receive a maximum of 15 link-layer credits on the TXREQ,
TXRSP, and TXDAT links and issues a maximum of nine link layer credits on the RXSNP, RXRSP, and
RXDAT links.

7.7.10 CHI DVM acceptance capability

The Cortex-A72 processor can have a maximum of four outstanding DVM transactions on its snoop
interface. When this limit is reached, the system cannot send any more DVM transactions on the RXRSP
link until the processor has provided a response to an older DVM transaction on the TXRSP link

7.7.11 L2 Auxiliary Control Register settings

This section describes the recommended performance settings for the Cortex-A72 L2 Auxiliary Control
Register in various system configurations.

Evict transactions
Evict and WriteEvict transactions indicate that a shareable cache line has been evicted from the master's
local caches. The downstream snoop filter can use this information to update its directory to indicate that
the issuing master no longer contains a copy of the cache line.
WriteEvict carries data and can be used to allow allocation into a system or Level 3 cache. In general,
ARM recommends the following:
• A system that contains a snoop filter enables Evict transactions.
• A system that contains a L3 cache that wants to behave like a victim cache for cache lines in the
Unique state enables WriteEvict transactions.
The Cortex-A72 L2ACTLR_EL1 register contains bits that can enable or disable Evict and WriteEvict
transactions individually. See L2ACTLR_EL1[3] and L2ACTLR_EL1[14], respectively, in the L2
Auxiliary Control Register, EL1.
When the Cortex-A72 processor is used with the ARM CCI-400 in an ACE-based system, ARM
recommends that you set L2ACTLR_EL1[3] to 1 to disable Evict transactions. The reset value of

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-315
reserved.
Non-Confidential
 7 Level 2 Memory System
7.7 External coherent interfaces

L2ACTLR[3] is 0 in Cortex-A72 ACE configurations. WriteEvict transactions are disabled by default in

Cortex-A72 ACE configurations.
When the Cortex-A72 processor is used with the ARM CCN-504 in a CHI-based system, no change is
required from the default reset value of L2ACTLR_EL1. By default, Cortex-A72 CHI configurations
generate WriteEvict transactions for allocating into the CCN-504 L3 cache but do not generate Evict
transactions because the CCN-504 snoop filter does not require them.

WriteUnique and WriteLineUnique transactions

If the Cortex-A72 processor is implemented in an ACE-based system that does not contain a snoop filter,
enabling WriteUnique and WriteLineUnique transactions might provide a small increase in system
performance. To enable WriteUnique and WriteLineUnique transactions, clear L2ACTLR_EL1[4]. The
reset value of L2ACTLR[4] is 1.
If the Cortex-A72 processor is implemented in an ACE-based system that does contain a snoop filter or
in a CHI based system, no change is recommended from the default value.

Related information
L2 Auxiliary Control Register, EL1.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-316
reserved.
Non-Confidential
 7 Level 2 Memory System
7.8 ACP

7.8 ACP
The optional Accelerator Coherency Port (ACP) is implemented as an AXI4 slave interface with the
following restrictions:
• 128-bit read and write interfaces.
• ARCACHE and AWCACHE are restricted to Normal, Write-Back, Read-Write-Allocate, Read-
Allocate, Write-Allocate, and No-Allocate memory. ARCACHE and AWCACHE are limited to the
values 0b0111, 0b1011, and 0b1111. Other values cause a SLVERR response on RRESP or BRESP.
• Exclusive accesses are not supported.
• Barriers are not supported. BRESP indicates global observation of all writes.
• ARSIZE and AWSIZE signals are not present and assume a value of 0b100,16 bytes.
• ARBURST and AWBURST signals are not present and assume a value of INCR.
• ARLOCK and AWLOCK signals are not present.
• ARQOS and AWQOS signals are not present.
• ARLEN and AWLEN are limited to values 0 and 3.
This section contains the following subsections:
• 7.8.1 Transfer size support on page 7-317.
• 7.8.2 ACP ARUSER and AWUSER signals on page 7-317.

7.8.1 Transfer size support

ACP supports the following read-request transfer size and length combinations:
• 64-byte INCR request characterized by:
— ARLEN is 0x03, 4 beats.
— ARADDR aligned to 64-byte boundary, so ARADDR[5:0] is 0b00 0000.
— ARSIZE and ARBURST assume values of 0b100 and INCR respectively.
• 16-byte INCR request characterized by:
— ARLEN is 0x00, 1 beat.
— ARADDR aligned to 16-byte boundary, so ARADDR[3:0] is 0x0.
ACP supports the following write-request transfer size and length combinations:
• 64-byte INCR request characterized by:
— AWLEN is 0x03, 4 beats.
— AWADDR aligned to 64-byte boundary, so AWADDR[5:0] is 0b00 0000.
— AWSIZE and AWBURST assume values of 0b100 and INCR respectively.
— WSTRB for all beats must be the same and either all asserted or all deasserted.
• 16-byte INCR request characterized by:
— AWLEN is 0x00, 1 beat.
— AWADDR aligned to 16-byte boundary, so AWADDR[3:0] is 0x0.
— AWSIZE and AWBURST assume values of 0b100 and INCR respectively.
— WSTRB can take any value.
Requests not meeting these restrictions cause a SLVERR response on RRESP or BRESP.

7.8.2 ACP ARUSER and AWUSER signals

ACP transactions can cause coherent requests to the system. Therefore ACP requests must pass the Inner
and Outer Shareable attributes to the L2. To pass the Inner Shareable attribute, use ARUSER[0] and
AWUSER[0]. To pass the Outer Shareable attribute, use ARUSER[1] and AWUSER[1].
The setting of AxUSER[1:0] to 0b11 is not allowed and causes a SLVERR response.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 7-317
reserved.
Non-Confidential
Chapter 8
Generic Interrupt Controller CPU Interface

This section describes the Cortex-A72 processor implementation of the GIC CPU interface.

It contains the following sections:

• 8.1 About the GIC on page 8-319.
• 8.2 GIC functional description on page 8-320.
• 8.3 GIC programmers model on page 8-325.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-318
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.1 About the GIC

8.1 About the GIC

The GIC is a resource for supporting and managing interrupts in a cluster system.
It implements the GIC CPU interface and provides:
• Registers for managing:
— Interrupt sources.
— Interrupt behavior.
— Interrupt routing to one or more processors.
The GIC supports:
• Two Security states.
• Interrupt virtualization.
• Software-generated Interrupts (SGIs).
• System Error Interrupts (SEIs).
• Message-based interrupts.
• System register access.
• Memory-mapped register access.
• Interrupt masking and prioritization.
• Interrupt routing based on processor affinity, in multiprocessor environments.
• Interrupt routing based on specifying target processors.
• Wake-up events in power-management environments.
The GIC includes interrupt grouping functionality that supports:
• Configuring each interrupt to belong to an interrupt group.
• Signaling Group 1 interrupts to the target processor using either the IRQ or the FIQ exception
request.
• Signaling Group 0 interrupts to the target processor using the FIQ exception request only.
• A unified scheme for handling the priority of Group 0 and Group 1 interrupts.
The Cortex-A72 processor implements the GIC CPU interface as described in the Generic Interrupt
Controller (GICv3) architecture. It can interface with a GICv3 Distributor component in the system.
This chapter only describes features that are specific to the Cortex-A72 processor implementation.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-319
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.2 GIC functional description

8.2 GIC functional description

Provides a functional description of the GIC.
The GIC is a single functional unit in the Cortex-A72 processor. The GIC consists of a common block
and several CPU interface blocks. For each core in the system, there is:
• A CPU interface.
• A virtual interface control block.
• A virtual CPU interface.
This section contains the following subsections:
• 8.2.1 GIC memory map on page 8-320.
• 8.2.2 Interrupt sources on page 8-322.
• 8.2.3 Interrupt priority levels on page 8-322.
• 8.2.4 GIC bypass modes on page 8-323.
• 8.2.5 nIRQ and nVFIQ inputs on page 8-323.

8.2.1 GIC memory map

The GIC registers are memory-mapped, with a physical base address specified by
PERIPHBASE[43:18]. This input must be tied to a constant value. The PERIPHBASE value is
sampled during reset into the Configuration Base Address Register (CBAR) for each processor in the
MPCore device.
The GIC registers are grouped into three contiguous 64KB pages. These blocks include the CPU
interface, virtual interface control, and virtual CPU interface blocks.
Memory regions used for these registers must be marked as Device, nGnRnE, nGnRE, nGRE, or GRE in
the translation tables. Memory regions marked as Normal memory cannot access any of the GIC
registers, but can access caches or external memory as required.
Access to these registers must be with the single word load and store instructions. Load/store-multiple,
load/store-double, and load/store exclusive instructions result in a Data Abort exception to the requesting
processor.
The Accelerator Coherency Port (ACP) cannot access any of the GIC registers. The registers must be
accessed through one of the processors. Any access from ACP to the GIC registers goes to external
memory and no Data Abort exception is generated.
The following table shows the GIC memory map of a Cortex-A72 processor. An external standalone GIC
such as the ARM GIC-400 or other proprietary GIC might differ.
The following table shows the GIC memory map of a Cortex-A72 processor. An external standalone GIC
such as the ARM GIC-400 or other proprietary GIC might differ. It lists the address offsets for the GIC
blocks relative to the PERIPHBASE base address.

Table 8-1 Cortex-A72 processor GIC memory map

Offset range from PERIPHBASE[43:18] GIC block

0x00000-0x01FFF CPU interface

0x02000-0x0FFFF Reserved

0x10000-0x10FFF Virtual interface control

0x11000-0x1FFFF Reserved

0x20000-0x21FFF Virtual CPU interface (4KB page offset)

0x22000-0x2EFFF Reserved

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-320
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.2 GIC functional description

Table 8-1 Cortex-A72 processor GIC memory map (continued)

Offset range from PERIPHBASE[43:18] GIC block

0x2F000-0x30FFF Alias of the Virtual CPU interface (64KB page offset alias)

0x31FFF-0x3FFFF Reserved

Related information
4.3.70 Configuration Base Address Register, EL1 on page 4-212.
4.5.24 Configuration Base Address Register on page 4-274.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-321
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.2 GIC functional description

8.2.2 Interrupt sources

The GIC CPU interface receives interrupts on the following signals:

nSEI, nREI, These signals generate System Error Interrupts (SEIs).
nVSEI
nIRQ, nFIQ When the GIC CPU interface is in bypass mode, these signals provide legacy IRQ and
FIQ inputs to the processor.
nVIRQ, These signals enable an external source to generate virtual IRQ and FIQ interrupts.
nVFIQ
AMBA GICv3 interrupt messages that are sent by an external Distributor. Each interrupt message
AXI4 has a unique interrupt ID.
Stream
The interrupt message types are:
interface
Local An interrupt generated by a peripheral that is destined for one or more
Peripheral processors within a specific affinity hierarchy.
Interrupts (LPIs)
Private An interrupt generated by a peripheral that is specific to a single
Peripheral processor. All PPIs must connect directly to the external Distributor.
Interrupts (PPIs)
Shared An interrupt generated by a peripheral that is destined for one or more
Peripheral processors. All SPIs must connect directly to the external Distributor.
Interrupts
(SPIs).
Software SGIs are generated by:
Generated • Writing to the ICC_SGI0R, ICC_SGI1R, or ICC_ASGI1R
Interrupts registers in the CPU interface when in System-register mode.
(SGIs) • Writing to the Software Generated Interrupt Register,
GICD_SGIR, in the external Distributor when in memory-mapped
mode.
A maximum of 16 SGIs, ID0-ID15, can be generated for each
processor interface. An SGI has edge-triggered properties. The
software triggering of the interrupt is equivalent to the edge transition
of the interrupt signal on a peripheral input.
System Error System Errors can be generated internally using the nSEI, nVSEI, or
Interrupts (SEIs) nREI signals. See the ARM® Generic Interrupt Controller
Architecture Specification, GICv3 for more information.
Virtual Local A virtual interrupt generated by a write to the Distributor that is
Peripheral destined for one or more processors within a specific affinity
Interrupts hierarchy. See the ARM® Generic Interrupt Controller Architecture
(vLPIs) Specification, GICv3 for more information.

Related information
A.5 GIC CPU interface signals on page Appx-A-535.

8.2.3 Interrupt priority levels

The processor implements a 5-bit version of the interrupt priority field, so it can support 32 interrupt
priority levels in Secure state.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-322
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.2 GIC functional description

8.2.4 GIC bypass modes

This section describes the two GIC bypass modes. The bypass modes are:
• GICCDISABLE bypass mode on page 8-323.
• Software bypass mode on page 8-323.

GICCDISABLE bypass mode

When using an external standalone interrupt controller such as the ARM GIC-400 or a proprietary
interrupt controller, you must set the GICCDISABLE signal HIGH. This forces the GIC CPU interface
to operate in bypass mode as described in the ARM® Generic Interrupt Controller Architecture
Specification, GICv3.
When the GICCDISABLE signal is tied HIGH, the PERIPHBASE[43:18] value can be read in the
Configuration Base Address Register, to permit software to read the location of the GIC if it exists in the
system external to the Cortex-A72 processor.
When the GICCDISABLE signal is HIGH, you must tie these CPU interface input signals LOW:
• ICDTVALID.
• ICDTDATA.
• ICDTLAST.
• ICDTDEST.
• ICCTREADY.
When the GICCDISABLE signal is HIGH, you must leave these CPU interface output signals
unconnected:
• ICCTVALID.
• ICCTDATA.
• ICCTLAST.
• ICCTID.
• ICDTREADY.
• nVCPUMNTIRQ[N:0].
If GICCDISABLE is tied HIGH, the nVIRQ and nVFIQ inputs can be:
• Tied off to HIGH if they are not in use.
• Driven by an external GIC in the SoC.

Related information
4.5.24 Configuration Base Address Register on page 4-274.
8.2.5 nIRQ and nVFIQ inputs on page 8-323.

Software bypass mode

The GIC CPU interface supports software interrupt bypass mode through interrupt disable bypass bits for
both memory-mapped accesses and System register accesses. Unlike the GICCDISABLE bypass mode,
the software bypass mode does not fully disable the internal GIC CPU interface.
For more information, see the ARM® Generic Interrupt Controller Architecture Specification, GICv3.

8.2.5 nIRQ and nVFIQ inputs

The Cortex-A72 processor includes the virtual interrupt signals, nVIRQ and nVFIQ. There is one
nVIRQ and one nVFIQ for each core.
• If GICCDISABLE is tied HIGH, nVIRQ and nVFIQ can be:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-323
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.2 GIC functional description

— Tied off to HIGH if they are not in use.

— Driven by an external GIC in the SoC.
• If GICCDISABLE is tied LOW and the GIC virtual CPU interface is enabled and in use, ARM
recommends to tie nVIRQ and nVFIQ off to HIGH. This is because the internal GIC CPU interface
generates the virtual interrupt signals to the cores.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-324
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.3 GIC programmers model

8.3 GIC programmers model

Describes the GIC programmers model for the Cortex-A72 processor.
This section contains the following subsections:
• 8.3.1 CPU interface register summary on page 8-325.
• 8.3.2 CPU interface memory-mapped register descriptions on page 8-328.
• 8.3.3 CPU interface System register descriptions on page 8-330.
• 8.3.4 Virtual interface control register summary on page 8-331.
• 8.3.5 Virtual interface control register descriptions on page 8-333.
• 8.3.6 Virtual CPU interface register summary on page 8-335.
• 8.3.7 Virtual CPU interface register descriptions on page 8-336.

8.3.1 CPU interface register summary

Each GIC CPU interface block provides the interface for a Cortex-A72 processor that operates with the
GIC. Each CPU interface provides a programming interface for:
• Enabling the signaling of interrupt requests by the CPU interface.
• Acknowledging an interrupt.
• Indicating completion of the processing of an interrupt.
• Setting an interrupt priority mask for the core.
• Defining the preemption policy for the core.
• Determining the highest priority pending interrupt for the core.
For more information on CPU interfaces, see the ARM® Generic Interrupt Controller Architecture
Specification, GICv3.

AArch32 GIC CPU interface memory-mapped register summary

The following table shows the GIC CPU interface register address offsets of the Cortex-A72 processor.
For information about an external standalone GIC such as the ARM GIC-400 or other proprietary GIC,
see the documentation of that product.
The following table shows the register memory map for the CPU interface in AArch32. The offsets in
this table are relative to the CPU interface block base address as shown in Table 8-1 Cortex-A72
processor GIC memory map on page 8-320.
All the registers in the following table are word-accessible. Registers not described in this table are
Reserved.

Table 8-2 GIC CPU interface memory-mapped register summary

Offset Name Type Reset Description

0x0000 GICC_CTLR RW 0x00000000 CPU Interface Control Register dq
0x0004 GICC_PMR RW 0x00000000 Interrupt Priority Mask Register
0x0008 GICC_BPR RW 0x00000002 (S)dr Binary Point Register
0x00000003 (NS)ds

0x000C GICC_IAR RO 0x000003FF Interrupt Acknowledge Register

0x0010 GICC_EOIR WO - End Of Interrupt Register
0x0014 GICC_RPR RO 0x000000FF Running Priority Register

dq See the ARM® Generic Interrupt Controller Architecture Specification, GICv3 for more information.
dr S = Secure.
ds NS = Non-secure.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-325
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.3 GIC programmers model

Table 8-2 GIC CPU interface memory-mapped register summary (continued)

Offset Name Type Reset Description

0x0018 GICC_HPPIR RO 0x000003FF Highest Priority Pending Interrupt Register
0x001C GICC_ABPR RW 0x00000003 Aliased Binary Point Register
0x0020 GICC_AIAR RO 0x000003FF Aliased Interrupt Acknowledge Register
0x0024 GICC_AEOIR WO - Aliased End of Interrupt Register
0x0028 GICC_AHPPIR RO 0x000003FF Aliased Highest Priority Pending Interrupt Register
0x00D0 GICC_APR0 RW 0x00000000 Active Priority Register on page 8-328
0x00E0 GICC_NSAPR0 RW 0x00000000 Non-secure Active Priority Register on page 8-329
0x00FC GICC_IIDR RO 0x0084043B CPU Interface Identification Register on page 8-329
0x1000 GICC_DIR WO - Deactivate Interrupt Register

AArch32 GIC CPU interface System register summary

The following table shows the System register map for the CPU interface in AArch32. See the ARM®
Generic Interrupt Controller Architecture Specification, GICv3 for more information about the registers.

Table 8-3 AArch32 GIC CPU interface System register summary

Name CRn op1 CRm op2 Type Description

ICC_PMR c4 0 c6 0 RW Priority Mask Register

ICC_IAR0 c12 0 c8 0 RO Group0 Interrupt Acknowledge Register

ICC_EOIR0 1 WO Group0 End of Interrupt Register

ICC_HPPIR0 2 RO Group0 Highest Priority Pending Interrupt Register

ICC_BPR0 3 RW Group0 Binary Pointer Register

ICC_AP0R0 4 RW Active Priority Group0 Register on page 8-330

ICC_AP1R0 c9 0 RW Active Priority Group1 Register on page 8-331

ICC_DIR c11 1 WO Deactivate Register

ICC_RPR 3 RO Running Priority Register

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-326
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.3 GIC programmers model

Table 8-3 AArch32 GIC CPU interface System register summary (continued)

Name CRn op1 CRm op2 Type Description

ICC_IAR1 c12 0 RO Group1 Interrupt Acknowledge Register

ICC_EOIR1 1 WO Group1 End of Interrupt Register

ICC_HPPIR1 2 RO Group1 Highest Priority Pending Interrupt Register

ICC_BPR1 3 RW Bdt Group1 Binary Pointer Register

ICC_CTLR 4 RW B Control Register

ICC_SRE 5 RW B System Register Enable

ICC_IGRPEN0 6 RW Group0 Interrupt Group Enable

ICC_IGRPEN1 7 RW B Group1 Interrupt Group Enable

ICC_SGI1Rdu - WO Group1 Software Generated Interrupt Register

ICC_ASGI1R 0 c12 - WO Aliased Group1 Software Generated Interrupt Register

ICC_SGI0R 2 c12 - WO Group0 Software Generated Interrupt Register

ICC_MCTLR 6 c12 4 RW Monitor Control Register

ICC_MSRE 5 RW Monitor System Register Enable

ICC_MGRPEN1 7 RW Monitor Group1 Interrupt Group Enable

AArch64 GIC CPU interface System register summary

The following table shows the System register map for the GIC CPU interface in AArch64. See the
ARM® Generic Interrupt Controller Architecture Specification, GICv3 for more information about the
registers.

Table 8-4 AArch64 GIC CPU interface System register summary

Name Type Description

ICC_PMR_EL1 RW Priority Mask Register

ICC_IAR0_EL1 RO Group0 Interrupt Acknowledge Register

ICC_EOIR0_EL1 WO Group0 End of Interrupt Register

ICC_HPPIR0_EL1 RO Group0 Highest Priority Pending Interrupt Register

ICC_BPR0_EL1 RW Group0 Binary Pointer Register

ICC_AP0R0_EL1 RW Active Priority Group0 Register on page 8-330

ICC_AP1R0_EL1 RW Active Priority Group1 Register on page 8-331

ICC_DIR_EL1 WO Deactivate Register

ICC_RPR_EL1 RO Running Priority Register

ICC_SGI1R_EL1 WO Group1 Software Generated Interrupt Register

ICC_ASGI1R_EL1 WO Aliased Group1 Software Generated Interrupt Register

dt When operating in EL3, accesses to Banked EL1 registers access the copy designated by the current value of the SCR_EL3.NS. When EL3 is using AArch32, there
is no Secure EL1 interrupt regime and accesses in any Secure EL3 mode, except Monitor mode, access the Secure copy.
du Use MCRR instructions to access this register in AArch32 state.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-327
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.3 GIC programmers model

Table 8-4 AArch64 GIC CPU interface System register summary (continued)

Name Type Description

ICC_SGI0R_EL1 WO Group0 Software Generated Interrupt Register

ICC_IAR1_EL1 RO Group1 Interrupt Acknowledge Register

ICC_EOIR1_EL1 WO Group1 End of Interrupt Register

ICC_HPPIR1_EL1 RO Group1 Highest Priority Pending Interrupt Register

ICC_BPR1_EL1 RW Bdv Group1 Binary Pointer Register

ICC_CTLR_EL1 RW B Control Register

ICC_SRE_EL1 RW B System Register Enable

ICC_IGRPEN0_EL1 RW Group0 Interrupt Group Enable Register

ICC_IGRPEN1_EL1 RW B Group1 Interrupt Group Enable

ICC_CTLR_EL3 RW EL3 Control Register

ICC_SRE_EL3 RW EL3 System Register Enable

ICC_GRPEN1_EL3 RW EL3 Group1 Interrupt Group Enable

8.3.2 CPU interface memory-mapped register descriptions

This section only describes registers whose implementation is specific to the Cortex-A72 processor. All
other registers are described in the ARM® Generic Interrupt Controller Architecture Specification,
GICv3. Table 8-2 GIC CPU interface memory-mapped register summary on page 8-325 provides cross-
references to individual registers.

Active Priority Register

The GICC_APR0 characteristics are:

Purpose
Provides support for preserving and restoring state in power-management applications.
Usage constraints
This register is Banked to provide Secure and Non-secure copies. This ensures that Non-secure
accesses do not interfere with Secure operation.
Configurations
Available if the GIC is implemented and setup for memory-mapped accesses.
Attributes
See the register summary in Table 8-2 GIC CPU interface memory-mapped register summary
on page 8-325.
The processor implements the GICC_APR0 according to the recommendations described in the ARM®
Generic Interrupt Controller Architecture Specification, GICv3.
The following table shows the Cortex-A72 processor GICC_APR0 implementation.

dv When operating in EL3, accesses to Banked EL1 registers access the copy designated by the current value of the SCR_EL3.NS. When EL3 is using AArch32, there
is no Secure EL1 interrupt regime and accesses in any Secure EL3 mode, except Monitor mode, access the Secure copy.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-328
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.3 GIC programmers model

Table 8-5 Active Priority Register implementation

Number of Preemption Minimum legal Minimum legal Active Priority View of Active Priority
group levels value of Secure value of Non- Registers Registers for Non-secure
priority bits GICC_BPR secure implemented accesses
GICC_BPR

5 32 2 3 GICC_APR0[31:0] GICC_NSAPR0[31:16] appears

as GICC_APR0[15:0]

Non-secure Active Priority Register

The GICC_NSAPR0 characteristics are:

Purpose
Provides support for preserving and restoring state in power-management applications.
Usage constraints
This register is only accessible from a Secure access.
Configurations
Available if the GIC is implemented.
Attributes
See the register summary in Table 8-2 GIC CPU interface memory-mapped register summary
on page 8-325.
The processor implements the GICC_NSAPR0 according to the recommendations described in the ARM®
Generic Interrupt Controller Architecture Specification, GICv3. It is consistent with the GICC_APR0
Register.

CPU Interface Identification Register

The GICC_IIDR characteristics are:

Purpose
Provides information about the implementer and revision of the CPU interface.
Usage constraints.
There are no usage constraints.
Configurations
Available if the GIC is implemented.
Attributes
See the register summary in Table 8-2 GIC CPU interface memory-mapped register summary
on page 8-325.
The following figure shows the GICC_IIDR bit assignments.
31 20 19 16 15 12 11 0
Architecture
ProductID Revision Implementer
version

Figure 8-1 GICC_IIDR bit assignments

The following table shows the GICC_IIDR bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-329
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.3 GIC programmers model

Table 8-6 GICC_IIDR bit assignments

Bit Name Function

[31:20] ProductID Identifies the product:

0x008 Product ID.

[19:16] Architecture version Identifies the architecture version of the GIC:

0x4 Version 4.

[15:12] Revision Identifies the revision number for the CPU interface:
0x0 Revision 0.

[11:0] Implementer Contains the JEP106 code of the company that implemented the CPU interface. For an ARM
implementation, these values are:
Bits[11:8] = 0x4 The JEP106 continuation code of the implementer.
Bit[7] Always 0.
Bits[6:0] = 0x3B The JEP106 identity code of the implementer.

8.3.3 CPU interface System register descriptions

This section only describes registers whose implementation is specific to the Cortex-A72 processor. All
other registers are described in the ARM® Generic Interrupt Controller Architecture Specification,
GICv3. Table 8-4 AArch64 GIC CPU interface System register summary on page 8-327 provides cross-
references to individual registers.

Active Priority Group0 Register

The ICC_AP0R0_EL1 characteristics are:

Purpose
Provides support for preserving and restoring state in power-management applications.
Usage constraints
Accessibility and constraints on this register are described in the ARM® Generic Interrupt
Controller Architecture Specification, GICv3.
Configurations
Available if the GIC is implemented for System register mode.
Attributes
See the register summary in Table 8-4 AArch64 GIC CPU interface System register summary
on page 8-327.
The multiprocessor implements the ICC_AP0R0_EL1 according to the recommendations described in
the ARM® Generic Interrupt Controller Architecture Specification, GICv3.
The following table shows the Cortex-A72 MPCore processor ICC_AP0R0_EL1 implementation.

Table 8-7 Active Priority Group0 Register implementation

Number of group priority Preemption levels Minimum legal value of Active Priority Group0 Registers
bits BPR implemented

5 32 2 ICC_AP0R0_EL1[31:0]

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-330
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.3 GIC programmers model

Active Priority Group1 Register

The ICC_AP1R0_EL1 characteristics are:

Purpose
Provides support for preserving and restoring state in power-management applications.
Usage constraints
This register is Banked to provide Secure and Non-secure copies. This ensures that Non-secure
accesses do not interfere with Secure operation. Accessibility and constraints on this register are
described in the ARM® Generic Interrupt Controller Architecture Specification, GICv3.
Configurations
Available if the GIC is implemented for System register mode.
Attributes
See the register summary in Table 8-4 AArch64 GIC CPU interface System register summary
on page 8-327.
The multiprocessor implements the ICC_AP1R0_EL1 according to the recommendations described in
the ARM® Generic Interrupt Controller Architecture Specification, GICv3.
The following table shows the Cortex-A72 processor ICC_AP1R0_EL1 implementation.

Table 8-8 Active Priority Group1 Register implementation

Number of group Preemption Minimum legal Minimum legal value Active Priority Group1
priority bits levels value of Secure of Non-secure BPR Registers implemented
BPR

5 32 2 3 ICC_AP1R0_EL1[31:0]

8.3.4 Virtual interface control register summary

The virtual interface control registers are management registers. The processor configuration software
must ensure that these registers are accessible only by a hypervisor, or similar software.

Virtual interface control register summary

The following table shows the register map for the virtual interface control registers. The offsets in this
table are relative to the virtual interface control registers block base address as shown in Table 8-1
Cortex-A72 processor GIC memory map on page 8-320.
All the registers in the following table are word-accessible. Registers not described in this table are
Reserved.

Table 8-9 Virtual interface control register summary

Offset Name Type Reset Description

0x000 GICH_HCR RW 0x00000000 Hypervisor Control Register dw

0x004 GICH_VTR RO 0x90000003 VGIC Type Memory-Mapped Register on page 8-333

0x008 GICH_VMCR RW 0x004C0000 Virtual Machine Control Register dw

0x010 GICH_MISR RO 0x00000000 Maintenance Interrupt Status Register dw

0x020 GICH_EISR0 RO 0x00000000 End of Interrupt Status Register dw

dw See the ARM® Generic Interrupt Controller Architecture Specification GICv3 for more information.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-331
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.3 GIC programmers model

Table 8-9 Virtual interface control register summary (continued)

Offset Name Type Reset Description

0x030 GICH_ELRSR0 RO 0x0000000F Empty List Register Status Register dw

0x0F0 GICH_APR RW 0x00000000 Active Priorities Register dw

0x100 GICH_LR0 RW 0x00000000 List Register 0 dw

0x104 GICH_LR1 RW 0x00000000 List Register 1 dw

0x108 GICH_LR2 RW 0x00000000 List Register 2 dw

0x10C GICH_LR3 RW 0x00000000 List Register 3 dw

AArch32 virtual interface System register summary

The following table shows the register map for the AArch32 virtual interface System registers. The
offsets in this table are relative to the virtual interface control registers block base address as shown in
Table 8-1 Cortex-A72 processor GIC memory map on page 8-320.
All the registers in the following table are word-accessible. Registers not described in this table are
Reserved.

Table 8-10 AArch32 virtual interface System register summary

Name CRn op1 CRm op2 Type Description

ICH_APR0 c12 4 c8 0 RW Hypervisor Active Priority Register 0

ICH_APR1 c9 0 RW Hypervisor Active Priority Register 1

ICH_VSEIR 4 RW Virtual System Error Interrupt Register

ICH_SRE 5 RW Hypervisor System Register

ICH_HCR 4 c11 0 RW Hypervisor Control Register

ICH_VTR 1 RO VGIC Type Register

ICH_MISR 2 RO Maintenance Interrupt Status Register

ICH_EISR 3 RO End of Interrupt Status Register

ICH_ELRSR 5 RO Empty List Register Status Register

ICH_VMCR 7 RW Virtual Machine Control Register

ICH_LR0 c12 0 RW List Register 0 to 3

ICH_LR1 1 RW

ICH_LR2 2 RW

ICH_LR3 3 RW

ICH_LRC0 c14 0 RW List Register Extension 0 to 3

ICH_LRC1 1 RW

ICH_LRC2 2 RW

ICH_LRC3 3 RW

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-332
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.3 GIC programmers model

AArch64 virtual interface System register summary

The following table shows the register map for the AArch64 virtual interface System registers. The
offsets in this table are relative to the virtual interface control registers block base address as shown in
Table 8-1 Cortex-A72 processor GIC memory map on page 8-320.
All the registers in the following table are word-accessible. Registers not described in this table are
Reserved.

Table 8-11 AArch64 virtual interface System register summary

Name Type Description

ICH_APR0_EL2 RW Hypervisor Active Priority Register

ICH_VSEIR_EL2 RW Virtual System Error Interrupt Register

ICH_HCR_EL2 RW Hypervisor Control Register

ICH_VTR_EL2 RO VGIC Type Register

ICC_SRE_EL2 RW Hypervisor System Register Enable

ICH_MISR_EL2 RO Maintenance Interrupt Status Register

ICH_EISR_EL2 RO End of Interrupt Status Register

ICH_ELRSR_EL2 RO Empty List Register Status Register

ICH_VMCR_EL2 RW Virtual Machine Control Register

ICH_LR0_EL2 RW List Register 0

ICH_LR1_EL2 RW List Register 1

ICH_LR2_EL2 RW List Register 2

ICH_LR3_EL2 RW List Register 3

8.3.5 Virtual interface control register descriptions

This section only describes registers whose implementation is specific to the Cortex-A72 processor. All
other registers are described in the ARM® Generic Interrupt Controller Architecture Specification,
GICv3.

VGIC Type Memory-Mapped Register

The GICH_VTR characteristics are:

Purpose
Holds information on number of priority bits, number of preemption bits, and number of List
Registers implemented.
Usage constraints
There are no usage constraints.
Configurations
Available if the GIC is implemented and setup for memory-mapped accesses.
Attributes
See the register summary in Table 8-9 Virtual interface control register summary
on page 8-331.
The following figure shows the GICH_VTR bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-333
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.3 GIC programmers model

31 29 28 26 25 6 5 0

PRIbits PREbits Reserved ListRegs

Figure 8-2 GICH_VTR bit assignments

The following table shows the GICH_VTR bit assignments.

Table 8-12 GICH_VTR bit assignments

Bit Name Description

[31:29] PRIbits Indicates the number of priority bits implemented, minus one:
0b100 Five bits of priority and 32 priority levels.

[28:26] PREbits Indicates the number of preemption bits implemented, minus one:
0b100 Five bits of preemption and 32 preemption levels.

[25:6] - Reserved, RAZ.

[5:0] ListRegs Indicates the number of implemented List Registers, minus one:
0b00 0011 Four List Registers.

VGIC Type System Register

The ICH_VTR_EL2 characteristics are:

Purpose
Holds information on number of priority bits, number of preemption bits, and number of List
Registers implemented.
Usage constraints
There are no usage constraints.
Configurations
Available if the GIC is implemented and setup for System register accesses.
Attributes
See the register summary in Table 8-9 Virtual interface control register summary
on page 8-331..
The following figure shows the ICH_VTR_EL2 bit assignments.
31 29 28 26 25 23 22 21 20 5 4 0

PRIbits PREbits IDbits Reserved ListRegs

SEIS A3V

Figure 8-3 ICH_VTR_EL2 bit assignments

The following table shows the ICH_VTR_EL2 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-334
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.3 GIC programmers model

Table 8-13 ICH_VTR_EL2 bit assignments

Bit Name Description

[31:29] PRIbits Indicates the number of priority bits implemented, minus one:
0b100 Five bits of priority and 32 priority levels.

[28:26] PREbits Indicates the number of preemption bits implemented, minus one:
0b100 Five bits of preemption and 32 preemption levels.

[25:23] IDbits Indicates the number of virtual interrupt identifier bits supported:
0b000 16 bits of virtual interrupt identifier.

[22] SEIS Indicates if locally generated virtual System Errors are supported:
0b0 Locally generated virtual System Errors are not supported.

[21] A3V Indicates if affinity level 3 is supported in SGI generation from System registers:
0b0 SGI generation from System registers does not support affinity level 3.

[20:5] - Reserved, RAZ.

[4:0] ListRegs Indicates the number of implemented List Registers, minus one:
0b000011 Four List Registers.

8.3.6 Virtual CPU interface register summary

The virtual CPU interface forwards virtual interrupts to a connected Cortex-A72 processor, subject to the
normal GIC handling and prioritization rules. The virtual interface control registers control virtual CPU
interface operation, and in particular, the virtual CPU interface uses the contents of the List registers to
determine when to signal virtual interrupts. When a core accesses the virtual CPU interface, the List
registers are updated.
For more information on virtual CPU interface, see the ARM® Generic Interrupt Controller Architecture
Specification, GICv3.
The following table shows the register map for the virtual CPU interface. The offsets in this table are
relative to the virtual CPU interface block base address as shown in Table 8-1 Cortex-A72 processor
GIC memory map on page 8-320.
All the registers in the following table are word-accessible. Registers not described in this table are
Reserved.

Table 8-14 Virtual CPU interface register summary

Offset Name Type Reset Description

0x0000 GICV_CTLR RW 0x00000000 VM Control Register 

0x0004 GICV_PMR RW 0x00000000 VM Priority Mask Register dx

0x0008 GICV_BPR RW 0x00000002 VM Binary Point Register dx

0x000C GICV_IAR RO 0x000003FF VM Interrupt Acknowledge Register dx

0x0010 GICV_EOIR WO - VM End Of Interrupt Register dx

dx See the ARM® Generic Interrupt Controller Architecture Specification GICv3 for more information. The System register counterparts of these registers are described
in the ARM® Generic Interrupt Controller Architecture Specification GICv3. The virtual CPU interface System registers do not have a separate encoding from the
physical CPU interface System registers but access is controlled from the appropriate system controls that the ARM® Generic Interrupt Controller Architecture
Specification GICv3 describes.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-335
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.3 GIC programmers model

Table 8-14 Virtual CPU interface register summary (continued)

Offset Name Type Reset Description

0x0014 GICV_RPR RO 0x000000FF VM Running Priority Register dx

0x0018 GICV_HPPIR RO 0x000003FF VM Highest Priority Pending Interrupt Register dx

0x001C GICV_ABPR RW 0x00000003 VM Aliased Binary Point Register dx

0x0020 GICV_AIAR RO 0x000003FF VM Aliased Interrupt Acknowledge Register dx

0x0024 GICV_AEOIR WO - VM Aliased End of Interrupt Register dx

0x0028 GICV_AHPPIR RO 0x000003FF VM Aliased Highest Priority Pending Interrupt Register dx

0x002C GICV_STATUSR RW - VM Error Reporting Status Register dx

0x00D0 GICV_APR0 RW 0x00000000 VM Active Priority Register on page 8-336

0x00FC GICV_IIDR RO 0x0084043B VM CPU Interface Identification Register on page 8-336

0x1000 GICV_DIR WO - VM Deactivate Interrupt Register dx

8.3.7 Virtual CPU interface register descriptions

This section only describes registers whose implementation is specific to the Cortex-A72 processor.
All other registers are described in the ARM® Generic Interrupt Controller Architecture Specification,
GICv3. Table 8-14 Virtual CPU interface register summary on page 8-335 provides cross-references to
individual registers.

VM Active Priority Register

The GICV_APR0 characteristics are:

Purpose
For software compatibility, this register is present in the virtual CPU interface. However, in
virtualized system, it is not used in the preserving and restoring state.
Usage constraints
Reading the content of this register and then writing the same values must not change any state
because there is no requirement to preserve and restore state during a power down.
Configurations
Available if the GIC is implemented.
Attributes
See the register summary in Table 8-14 Virtual CPU interface register summary on page 8-335.
The multiprocessor implements the GICV_APR0 as an alias of GICH_APR.

VM CPU Interface Identification Register

The GICV_IIDR characteristics are:

Purpose
Provides information about the implementer and revision of the virtual CPU interface.
Usage constraints
There are no usage constraints.
Configurations
Available if the GIC is implemented.
Attributes
See the register summary in Table 8-14 Virtual CPU interface register summary on page 8-335.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-336
reserved.
Non-Confidential
 8 Generic Interrupt Controller CPU Interface
8.3 GIC programmers model

The bit assignments for the VM CPU Interface Identification Register are identical to the corresponding
register in the CPU interface.

Related information
CPU Interface Identification Register on page 8-329.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 8-337
reserved.
Non-Confidential
Chapter 9
Generic Timer

This chapter describes the Cortex-A72 processor implementation of the ARM Generic Timer.

It contains the following sections:

• 9.1 About the Generic Timer on page 9-339.
• 9.2 Generic Timer functional description on page 9-340.
• 9.3 Generic Timer register summary on page 9-341.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 9-338
reserved.
Non-Confidential
 9 Generic Timer
9.1 About the Generic Timer

9.1 About the Generic Timer

The Generic Timer in the Cortex-A72 processor can schedule events and trigger interrupts based on an
incrementing counter value.
It provides:
• Generation of timer events as interrupt outputs.
• Generation of event streams.
The Generic Timer is compliant with the ARM® Architecture Reference Manual ARMv8, for ARMv8-A
architecture profile.
This chapter only describes features that are specific to the Cortex-A72 processor implementation.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 9-339
reserved.
Non-Confidential
 9 Generic Timer
9.2 Generic Timer functional description

9.2 Generic Timer functional description

The Cortex-A72 processor provides a set of timers for each core in the processor.
The timers are:
• A Non-secure EL1 physical timer.
• A Secure EL1 physical timer.
• A Non-secure EL2 physical timer.
• A virtual timer.
The processor does not include the system counter that resides in the SoC. The system counter value is
distributed to the Cortex-A72 processor with a synchronous binary encoded 64-bit bus,
CNTVALUEB[63:0].
Because CNTVALUEB is generated from a system counter that typically operates at a slower frequency
than the main processor CLK, the CNTCLKEN input is provided as a clock enable for the
CNTVALUEB bus.
Each timer provides an active-LOW interrupt output that is an external pin to the SoC.
The following table shows the signals that are the external interrupt output pins.

Table 9-1 Generic Timer signals

Signaldy Description

nCNTPNSIRQ[n:0] Non-secure EL1 physical timer interrupt

nCNTPSIRQ[n:0] Secure EL1 physical timer interrupt

nCNTHPIRQ[n:0] Non-secure EL2 physical timer interrupt

nCNTVIRQ[n:0] Virtual timer interrupt

Related information
2.3.1 Clocks on page 2-32.
2.3.1 Clocks on page 2-32.

dy n is the number of processors present in the MPCore device, minus one.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 9-340
reserved.
Non-Confidential
 9 Generic Timer
9.3 Generic Timer register summary

9.3 Generic Timer register summary

Within each core, a set of Generic Timer registers are allocated to the CP15 coprocessor space.
The Generic Timer registers are either 32-bit or 64-bit wide and accessible in the AArch32 and AArch64
states.
This section contains the following subsections:
• 9.3.1 AArch64 Generic Timer register summary on page 9-341.
• 9.3.2 AArch32 Generic Timer register summary on page 9-341.

9.3.1 AArch64 Generic Timer register summary

The following table shows the AArch64 Generic Timer registers. See the ARM® Architecture Reference
Manual ARMv8 for information about these registers.

Table 9-2 AArch64 Generic Timer registers

Name Type Reset Width Description

CNTKCTL_EL1 RW -dz 32-bit Timer Control register (EL1)

CNTFRQ_EL0 RW ea UNK 32-bit Timer Counter Frequency register

CNTPCT_EL0 RO UNK 64-bit Physical Timer Count register

CNTVCT_EL0 RO UNK 64-bit Virtual Timer Count register

CNTP_TVAL_EL0 RW UNK 32-bit Physical Timer TimerValue (EL0)

CNTP_CTL_EL0 RW -eb 32-bit Physical Timer Control register (EL0)

CNTP_CVAL_EL0 RW UNK 64-bit Physical Timer CompareValue register (EL0)

CNTV_TVAL_EL0 RW UNK 32-bit Virtual Timer TimerValue register

CNTV_CTL_EL0 RW -eb 32-bit Virtual Timer Control register

CNTV_CVAL_EL0 RW UNK 64-bit Virtual Timer CompareValue register

CNTVOFF_EL2 RW UNK 64-bit Virtual Timer Offset register

CNTHCTL_EL2 RW -ec 32-bit Timer Control register (EL2)

CNTHP_TVAL_EL2 RW UNK 32-bit Physical Timer TimerValue register (EL2)

CNTHP_CTL_EL2 RW -eb 32-bit Physical Timer Control register (EL2)

CNTHP_CVAL_EL2 RW UNK 64-bit Physical Timer CompareValue register (EL2)

CNTPS_TVAL_EL1 RW UNK 32-bit Physical Timer TimerValue register (EL2)

CNTPS_CTL_EL1 RW -eb 32-bit Physical Secure Timer Control register (EL1)

CNTPS_CVAL_EL1 RW UNK 64-bit Physical Secure Timer CompareValue register (EL1)

9.3.2 AArch32 Generic Timer register summary

The following table shows the AArch32 Generic Timer registers.
See the ARM® Architecture Reference Manual ARMv8 for information about these registers.

dz The reset value for bits[9:8, 2:0] is 0b00000.

ea Only at EL3, otherwise this register is RO.
eb The reset value for bit[0] is 0.
ec The reset value for bit[2] is 0 and for bits[1:0] is 0b11.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 9-341
reserved.
Non-Confidential
 9 Generic Timer
9.3 Generic Timer register summary

Table 9-3 AArch32 Generic Timer registers

Name CRn op1 CRm op2 Type Reset Width Description

CNTFRQ c14 0 c0 0 RW ed UNK 32-bit Timer Counter Frequency register

CNTPCT - 0 c14 - RO UNK 64-bit Physical Timer Count register

CNTKCTL c14 0 c1 0 RW -ee 32-bit EL1 Timer Control register

CNTP_TVAL c2 0 RW UNK 32-bit EL1 Physical Timer TimerValue register

CNTP_CTL 1 RW -ef 32-bit EL1 Physical Timer Control register

CNTV_TVAL c3 0 RW UNK 32-bit Virtual Timer TimerValue register

CNTV_CTL 1 RW ef 32-bit Virtual Timer Control register

CNTVCT - 1 c14 - RO UNK 64-bit Virtual Timer Count register

CNTP_CVAL 2 RW UNK 64-bit EL1 Physical Timer CompareValue

register

CNTV_CVAL 3 RW UNK 64-bit Virtual Timer CompareValue register

CNTVOFF 4 RW UNK 64-bit Virtual Timer Offset register

CNTHCTL c14 4 c1 0 RW -eg 32-bit EL2 Timer Control register

CNTHP_TVAL c2 0 RW UNK 32-bit EL2 Physical Timer TimerValue register

CNTHP_CTL 1 RW ef 32-bit EL2 Physical Timer Control register

CNTHP_CVAL - 6 c14 - RW UNK 64-bit EL2 Physical Timer CompareValue

register

ed Only at EL3, otherwise this register is RO.

ee The reset value for bits[9:8, 2:0] is 0b00000.
ef The reset value for bit[0] is 0.
eg The reset value for bit[2] is 0 and for bits[1:0] is 0b11.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 9-342
reserved.
Non-Confidential
Chapter 10
Debug

This section describes the Cortex-A72 processor debug registers and shows examples of how to use
them.

It contains the following sections:

• 10.1 About debug on page 10-344.
• 10.2 Debug register interfaces on page 10-346.
• 10.3 AArch64 debug register summary on page 10-348.
• 10.4 AArch64 debug register descriptions on page 10-350.
• 10.5 AArch32 debug register summary on page 10-356.
• 10.6 AArch32 debug register descriptions on page 10-358.
• 10.7 Memory-mapped register summary on page 10-362.
• 10.8 Memory-mapped register descriptions on page 10-366.
• 10.9 Debug events on page 10-380.
• 10.10 External debug interface on page 10-381.
• 10.11 ROM table on page 10-384.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-343
reserved.
Non-Confidential
 10 Debug
10.1 About debug

10.1 About debug

Provides an overview of debug and describes the debug components.
The processor forms one component of a debug system. You can use the following invasive debug
methods:
Conventional JTAG debug (external debug)
The processor halts execution when breakpoints and watchpoints are triggered. A debug
connection enables you to examine and modify registers and memory, and provide single-step
execution.
Conventional monitor debug (self-hosted debug)
The processor runs a debug monitor that resides in memory.
The following figure shows a typical JTAG debug system.

Debug
Debug host computer running suitable debugger tool
host

Protocol
for example, DSTREAM or RealView ICE
converter

Debug
Development system containing ARM processor
target

Figure 10-1 Typical debug system

This section contains the following subsections:
• 10.1.1 Debug host on page 10-344.
• 10.1.2 Protocol converter on page 10-344.
• 10.1.3 Debug target on page 10-345.
• 10.1.4 The debug unit on page 10-345.
• 10.1.5 Self-hosted debug on page 10-345.

10.1.1 Debug host

The debug host is a computer, for example a personal computer, running a software debugger such as the
DS-5 Debugger. The debug host enables you to issue high-level commands such as setting breakpoint at
a certain location, or examining the contents of a memory address.

10.1.2 Protocol converter

The debug host sends messages to the debug target using an interface such as Ethernet. However, the
debug target typically implements a different interface protocol. A device such as DSTREAM is required
to convert between the two protocols.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-344
reserved.
Non-Confidential
 10 Debug
10.1 About debug

10.1.3 Debug target

The debug target is the lowest level of the system. An example of a debug target is a development system
with a test chip or a silicon part with a processor.
The debug target implements system support for the protocol converter to access the debug unit using the
AMBA Advanced Peripheral Bus (APB) slave interface.

10.1.4 The debug unit

The processor debug unit assists in debugging software running on the processor. You can use the
processor debug unit, in combination with a software debugger program, to debug:
• Application software.
• Operating systems.
• Hardware systems based on an ARM processor.
The debug unit enables you to:
• Stop program execution.
• Examine and alter process and coprocessor state.
• Examine and alter memory and input/output peripheral state.
• Restart the processor.

10.1.5 Self-hosted debug

For self-hosted debug, the debug target runs additional debug monitor software, and uses the on-chip bus
fabric to send messages to the APB slave interface on the debug unit.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-345
reserved.
Non-Confidential
 10 Debug
10.2 Debug register interfaces

10.2 Debug register interfaces

This section describes the debug architecture and debug events of the processor.
The processor implements the ARMv8 Debug architecture and debug events as described in the ARM®
Architecture Reference Manual ARMv8.
The Debug architecture defines a set of debug registers. The debug register interfaces provide access to
these registers from:
• Software running on the processor.
• An external debugger.
This section contains the following subsections:
• 10.2.1 Processor interfaces on page 10-346.
• 10.2.2 Breakpoints and watchpoints on page 10-346.
• 10.2.3 Effects of resets on debug registers on page 10-346.
• 10.2.4 External access permissions on page 10-347.

10.2.1 Processor interfaces

System register access allows the processor to directly access certain debug registers. The external debug
interface allows both external and self-hosted debug agents to access debug registers.
Access to the debug registers is partitioned as follows:
Debug registers
This interface is System register based and memory-mapped. You can access the debug register
map using the APB slave port.
Performance monitor
This interface is System register based and memory-mapped. You can access the performance
monitor registers using the APB slave port.
Trace registers
This interface is memory-mapped.

Related information
10.10 External debug interface on page 10-381.

10.2.2 Breakpoints and watchpoints

The processor supports six hardware breakpoints, four watchpoints, and a standard Debug
Communications Channel (DCC). Four of the breakpoints match only to Virtual Address and the other
two match against either Virtual Address or context ID, or Virtual Machine Identifier (VMID). All the
watchpoints can be linked to two breakpoints to enable a memory request to be trapped in a given
process context.

10.2.3 Effects of resets on debug registers

The processor has the following reset signals that affect the debug registers:
nCPUPORESET
This signal initializes the processor logic, including the debug, Embedded Trace Macrocell
(ETM), breakpoint, watchpoint logic, and performance monitors logic. This maps to a Cold
reset that covers reset of the processor logic and the integrated debug functionality.
nCORERESET
This signal resets some of the debug and performance monitor logic. This maps to a Warm reset
that covers reset of the processor logic.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-346
reserved.
Non-Confidential
 10 Debug
10.2 Debug register interfaces

nPRESETDBG
This signal initializes the shared debug APB, Cross Trigger Interface (CTI), and Cross Trigger
Matrix (CTM) logic. This maps to an external debug reset that covers the resetting of the
external debug interface and has no impact on the processor functionality.

10.2.4 External access permissions

External access permission to the debug registers is subject to the conditions at the time of the access.
The following table describe the processor response to accesses through the external debug interface.

Table 10-1 External register access conditions

Condition Condition trigger Description

Off EDPRSR.PU is 0 Core power domain is completely off, or in a low-power state where the
Core power domain registers cannot be accessed.
Note
If debug power is off then all external debug and memory-mapped register
accesses return an error.

DLK EDPRSR.DLK is 1 OS Double Lock is locked.

OSLK OSLSR_EL1.OSLK is 1 OS Lock is locked.

EDAD AllowExternalDebugAccess() External debug access disabled. When an error is returned because of the
==FALSE EDAD condition, and this is the highest priority error condition,
EDPRSR.SDAD is set to 1. Otherwise SDAD is unchanged.

SLK Memory-mapped interface only Software Lock is locked. For the external debug interface, ignore this
condition.

Default - None of the conditions apply, normal access.

The following table shows an example of external register access conditions for access to a Performance
Monitors register. To determine the access permission for the register, scan the columns from left to
right. Stop at the first column whose condition is true, the entry gives the access permission of the
register and scanning stops.

Table 10-2 External register access conditions example

Off DLK OSLK EDAD SLK Default

- - - - RO/WI RO

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-347
reserved.
Non-Confidential
 10 Debug
10.3 AArch64 debug register summary

10.3 AArch64 debug register summary

Provides a summary of the debug control registers in the AArch64 state.
The following table shows the debug control registers that are accessible in AArch64 state. These
registers are accessed by the MRS and MSR instructions.
The table also shows the offset address for the AArch64 registers that are accessible from the internal
memory-mapped interface or the external debug interface. See 10.7 Memory-mapped register summary
on page 10-362 for a complete list of registers accessible from the internal memory-mapped or the
external debug interface.

Table 10-3 AArch64 debug register summary

Offset Name Type Width Description

- DBGDTR_EL0 RW 64-bit Debug Data Transfer Register, half-duplex eh

- DBGVCR32_EL2 RW 32-bit Debug Vector Catch Register eh

- MDCCINT_EL1 RW 32-bit Monitor Debug Comms Channel Interrupt Enable Register eh

- MDCCSR_EL0 RO 32-bit Monitor Debug Comms Channel Status Register eh

- MDRAR_EL1 RO 64-bit Monitor Debug ROM Address Register eh

- MDSCR_EL1 RW 32-bit Monitor Debug System Control Register eh

- OSDTRRX_EL1 RW 32-bit OS Lock Data Transfer Register, Receive, External View 

- OSDTRTX_EL1 RW 32-bit OS Lock Data Transfer Register, Transmit, External View eh

- OSDLR_EL1 RW 32-bit OS Double Lock Register eh

- OSLSR_EL1 RO 32-bit OS Lock Status Register

0x080 DBGDTRRX_EL0 RO 32-bit Debug Data Transfer Register, Receive, Internal View eh
0x08C DBGDTRTX_EL0 WO 32-bit Debug Data Transfer Register, Transmit, Internal View eh
0x098 OSECCR_EL1 RW 32-bit OS Lock Exception Catch Control Register eh
0x300 OSLAR_EL1 WO 32-bit Debug OS Lock Access Register eh
0x310 DBGPRCR_EL1 RW 32-bit Debug Power/Reset Control Register eh
0x400 DBGBVR0_EL1 RW 64-bit Debug Breakpoint Value Register 0 eh
0x408 DBGBCR0_EL1 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1 on page 10-350
0x410 DBGBVR1_EL1 RW 64-bit Debug Breakpoint Value Register 1 eh
0x418 DBGBCR1_EL1 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1 on page 10-350
0x420 DBGBVR2_EL1 RW 64-bit Debug Breakpoint Value Register 2 eh
0x428 DBGBCR2_EL1 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1 on page 10-350
0x430 DBGBVR3_EL1 RW 64-bit Debug Breakpoint Value Register 3 eh
0x438 DBGBCR3_EL1 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1 on page 10-350
0x440 DBGBVR4_EL1 RW 64-bit Debug Breakpoint Value Register 4 eh
0x448 DBGBCR4_EL1 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1 on page 10-350
0x450 DBGBVR5_EL1 RW 64-bit Debug Breakpoint Value Register 5 eh

eh See the ARM® Architecture Reference Manual ARMv8 for more information.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-348
reserved.
Non-Confidential
 10 Debug
10.3 AArch64 debug register summary

Table 10-3 AArch64 debug register summary (continued)

Offset Name Type Width Description

0x458 DBGBCR5_EL1 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1 on page 10-350
0x800 DBGWVR0_EL1 RW 64-bit Debug Watchpoint Value Register 0 eh
0x808 DBGWCR0_EL1 RW 32-bit 10.4.2 Debug Watchpoint Control Registers, EL1 on page 10-352
0x810 DBGWVR1_EL1 RW 64-bit Debug Watchpoint Value Register 1 eh
0x818 DBGWCR1_EL1 RW 32-bit 10.4.2 Debug Watchpoint Control Registers, EL1 on page 10-352
0x820 DBGWVR2_EL1 RW 64-bit Debug Watchpoint Value Register 2 eh
0x828 DBGWCR2_EL1 RW 32-bit 10.4.2 Debug Watchpoint Control Registers, EL1 on page 10-352
0x830 DBGWVR3_EL1 RW 64-bit Debug Watchpoint Value Register 3 eh
0x838 DBGWCR3_EL1 RW 32-bit 10.4.2 Debug Watchpoint Control Registers, EL1 on page 10-352
0xFA0 DBGCLAIMSET_EL1 RW 32-bit Debug Claim Tag Set Register eh
0xFA4 DBGCLAIMCLR_EL1 RW 32-bit Debug Claim Tag Clear Register eh
0xFB8 DBGAUTHSTATUS_EL1 RO 32-bit Debug Authentication Status Register eh

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-349
reserved.
Non-Confidential
 10 Debug
10.4 AArch64 debug register descriptions

10.4 AArch64 debug register descriptions

This section describes the debug registers in AArch64 state.
10.3 AArch64 debug register summary on page 10-348 provides cross-references to the individual
registers.
This section contains the following subsections:
• 10.4.1 Debug Breakpoint Control Registers, EL1 on page 10-350.
• 10.4.2 Debug Watchpoint Control Registers, EL1 on page 10-352.

10.4.1 Debug Breakpoint Control Registers, EL1

The DBGBCRn_EL1characteristics are:

Purpose
Holds control information for a breakpoint. Each DBGBVR_EL1 is associated with a
DBGBCR_EL1 to form a Breakpoint Register Pair (BRP). DBGBVRn_EL1 is associated with
DBGBCRn_EL1to form BRPn.
Note
The range of n for DBGBCRn_EL1 is 0 to 5.

Usage constraints
The accessibility to the DBGBCRn_EL1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

The external debug accessibility to the DBGBCRn_EL1 by condition code is:

Off DLK OSLK EDAD SLK Default

Error Error Error Error RO RW

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The DBGBCRn_EL1 is Common to Secure and Non-secure states and architecturally mapped
to:
• The AArch32 DBGBCRn registers.
• The external DBGBCRn_EL1 registers.
Attributes
See the register summary in Table 10-3 AArch64 debug register summary on page 10-348.
The debug logic reset value of a DBGBCRn_EL1 is UNKNOWN.
The following figure shows the DBGBCRn_EL1bit assignments.

31 24 23 20 19 16 15 14 13 12 9 8 5 4 3 2 1 0

RES0 BT LBN SSC RES0 BAS RES0 PMC E

HMC

Figure 10-2 DBGBCR

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-350
reserved.
Non-Confidential
 10 Debug
10.4 AArch64 debug register descriptions

The following table shows the DBGBCRn_EL1 bit assignments.

Table 10-4 DBGBCRn_EL1 bit assignments

Bits Name Function

[31:24] - Reserved, RES0.

[23:20] BT Breakpoint Type. This field controls the behavior of Breakpoint debug event generation. This includes the meaning
of the value held in the associated DBGBVR, indicating whether it is an instruction address match or mismatch or a
Context match. It also controls whether the breakpoint is linked to another breakpoint. The possible values are:
0b0000 Unlinked instruction address match.
0b0001 Linked instruction address match.
0b0010 Unlinked ContextIDR match.
0b0011 Linked ContextIDR match.
0b0100 Unlinked instruction address mismatch.
0b0101 Linked instruction address mismatch.
0b1000 Unlinked VMID match.
0b1001 Linked VMID match.
0b1010 Unlinked VMID + CONTEXTIDR match.
0b1011 Linked VMID + CONTEXTIDR match.

All other values are reserved.

The field break down is:
BT[0] Enable linking.
BT[3,1] Base type. If the breakpoint is not context-aware, these bits are RES0. Otherwise, the possible values are:
0b00 Match address.
0b01 Match context ID.
0b10 Match VMID.
0b11 Match VMID and context ID.

BT[2] Mismatch. This bit is ignored in AArch64 state, and in EL0 if EL1 is using AArch64. If EL1 using
AArch32 is not implemented, this bit is RES0. The address in DBGBVRn_EL1 is the address of an
instruction to be stepped.

[19:16] LBN Linked breakpoint number. For Linked address matching breakpoints, this specifies the index of the Context-
matching breakpoint linked to.

[15:14] SSC Security State Control. Determines the Security states that a breakpoint debug event for breakpoint n is generated.
This field must be interpreted along with the AMC and PMC fields.
This field is used with the Higher Mode Control (HMC), and Privileged Mode Control (PMC), fields to determine
the mode and Security states that can be tested.
See the ARM® Architecture Reference Manual ARMv8 for possible values of the fields.

[13] HMC Hyp Mode Control bit. Determines the debug perspective for deciding when a breakpoint debug event for breakpoint
n is generated. This bit must be interpreted along with the SSC and PMC fields.
This bit is used with the SSC and PMC fields to determine the mode and Security states that can be tested.
See the ARM® Architecture Reference Manual ARMv8 for possible values of the fields.

[12:9] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-351
reserved.
Non-Confidential
 10 Debug
10.4 AArch64 debug register descriptions

Table 10-4 DBGBCRn_EL1 bit assignments (continued)

Bits Name Function

[8:5] BASei Byte Address Select. Defines which halfwords a regular breakpoint matches, regardless of the instruction set and
Execution state. A debugger must program this field as follows:
0x3 Match the T32 instruction at DBGBVRn.
0xC Match the T32 instruction at DBGBVRn+2.
0xF Match the A64 or A32 instruction at DBGBVRn, or context match.

All other values are reserved.

Note
ARMv8 does not support direct execution of Java bytecodes. BAS[3] and BAS[1] ignore writes and on reads return
the values of BAS[2] and BAS[0] respectively.

[4:3] - Reserved, RES0.

[2:1] PMC Privileged Mode Control. Determines the Exception level or levels that a breakpoint debug event for breakpoint n is
generated. This field must be interpreted along with the SSC and AMC fields.
This field is used with the SSC and HMC fields to determine the mode and Security states that can be tested.
See the ARM® Architecture Reference Manual ARMv8 for possible values of the fields.
Note
Bits[2:1] has no effect for accesses made in Hyp mode.

[0] E Enable breakpoint. This bit enables the BRP:

0 BRP disabled.
1 BRP enabled.

A BRP never generates a Breakpoint debug event when it is disabled.

Note
The value of DBGBCR.E is UNKNOWN on reset. A debugger must ensure that DBGBCR.E has a defined value before
it programs DBGDSCR.MDBGen and DBGDSCR.HDBGen to enable debug.

To access the DBGBCRn_EL1 in AArch64 state, read or write the register with:
MRS <Xt>, DBGBCRn_EL1; Read Debug Breakpoint Control Register n
MSR DBGBCRn_EL1, <Xt>; Write Debug Breakpoint Control Register n

To access the DBGBCRn in AArch32 state, read or write the CP14 register with:
MRC p14, 0, <Rt>, c0, cn, 4; Read Debug Breakpoint Control Register n
MCR p14, 0, <Rt>, c0, cn, 4; Write Debug Breakpoint Control Register n

The DBGBCRn_EL1 can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0x4n8.

10.4.2 Debug Watchpoint Control Registers, EL1

The DBGWCRn_EL1characteristics are:

ei See the ARM® Architecture Reference Manual ARMv8 for more information on how the BAS field is interpreted by hardware.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-352
reserved.
Non-Confidential
 10 Debug
10.4 AArch64 debug register descriptions

Purpose Holds control information for a watchpoint. Each DBGWCR_EL1 is associated with a
DBGWVR_EL1 to form a Watchpoint Register Pair (WRP). DBGWCRn_EL1 is
associated with DBGWVRn_EL1 to form WRPn.
Note
The range of n for DBGBCRn_EL1is 0 to 3.

Usage The accessibility to the DBGWCRn_EL1 by Exception level is:

constraints
EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RW RW RW RW RW

The external debug accessibility to the DBGWCRn_EL1 by condition code is:

Off DLK OSLK EDAD SLK Default

Error Error Error Error RO RW

Table 11-1 External register access conditions on page 11-398 describes the access
conditions.
Configurations The DBGWCRn_EL1 is Common to Secure and Non-secure states and architecturally
mapped to:
• The AArch32 DBGWCRn registers.
• The external DBGWCRn_EL1 registers.
Attributes See the register summary in Table 10-3 AArch64 debug register summary
on page 10-348.
The debug logic reset value of a DBGWCR_EL1 is UNKNOWN.

The following figure shows the DBGWCRn_EL1 bit assignments.

31 29 28 24 23 21 20 19 16 15 14 13 12 5 4 3 2 1 0

RES0 MASK RES0 LBN SSC BAS LSC PAC E

WT HMC

Figure 10-3 DBGWCR

The following table shows the DBGWCRn_EL1 bit assignments.

Table 10-5 DBGWCRn_EL1 bit assignments

Bits Name Function

[31:29] - Reserved, RES0.

[28:24] MASK Address range mask. The processor supports watchpoint address range masking. This field can set a watchpoint on a
range of addresses by masking lower order address bits out of the watchpoint comparison. The value of this field is
the number of low order bits of the address that are masked off, except that values of 1 and 2 are reserved.
See the ARM® Architecture Reference Manual ARMv8 for the meanings of watchpoint address range mask values.

[23:21] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-353
reserved.
Non-Confidential
 10 Debug
10.4 AArch64 debug register descriptions

Table 10-5 DBGWCRn_EL1 bit assignments (continued)

Bits Name Function

[20] WT Watchpoint Type. This bit is set to 1 to link the watchpoint to a breakpoint to create a linked watchpoint that requires
both data address matching and Context matching:
0 Unlinked data address match.
1 Linked data address match.

When this bit is set to 1 the linked BRP number field indicates the BRP that this WRP is linked. See the ARM®
Architecture Reference Manual ARMv8 for more information.

[19:16] LBN Linked Breakpoint Number. If this watchpoint is programmed with the watchpoint type set to linked, then this field
must be programmed with the number of the breakpoint that defines the Context match to be combined with data
address comparison. Otherwise, this field must be programmed to 0b0000.
Reading this register returns an UNKNOWN value for this field, and the generation of Watchpoint debug events is
UNPREDICTABLE, if either:
• This watchpoint does not have linking enabled and this field is not programmed to 0x0.
• This watchpoint has linking enabled and the breakpoint indicated by this field does not support Context
matching, is not programmed for Context matching, or does not exist.
See the ARM® Architecture Reference Manual ARMv8 for more information.

[15:14] SSC Security State Control. This field enables the watchpoint to be conditional on the Security state of the processor.
This field is used with the Hyp Mode Control (HMC) and Privileged Access Control (PAC) fields.
See the ARM® Architecture Reference Manual ARMv8 for possible values of the fields, and the access modes and
Security states that can be tested.

[13] HMC Hyp Mode Control. This field is used with the Security State Control (SSC) and PAC fields. The value of
DBGWCR.PAC has no effect for accesses made in Hyp mode.
See the ARM® Architecture Reference Manual ARMv8 for possible values of the fields, and the access modes and
Security states that can be tested.

[12:5] BAS Byte Address Select. The processor implements an 8-bit Byte address select field, DBGWCR[12:5].
A DBGWVR is programmed with a word-aligned address. This field enables the watchpoint to hit only if certain
bytes of the addressed word are accessed. The watchpoint hits if an access hits any byte being watched, even if:
• The access size is larger than the size of the region being watched.
• The access is unaligned, and the base address of the access is not in the same word of memory as the address in
the DBGWVR.
• The access size is smaller than the size of region being watched.
See the ARM® Architecture Reference Manual ARMv8 for more information.

[4:3] LSC Load/store access control. This field enables watchpoint matching for the type of access. The possible values are:
0b00 Reserved.
0b01 Match on any load, Load-Exclusive, or swap.
0b10 Match on any store, Store-Exclusive, or swap.
0b11 Match on all type of access.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-354
reserved.
Non-Confidential
 10 Debug
10.4 AArch64 debug register descriptions

Table 10-5 DBGWCRn_EL1 bit assignments (continued)

Bits Name Function

[2:1] PAC Privileged Access Control. This field enables watchpoint matching conditional on the mode of the processor. This
field is used with the SSC and PAC fields.
See the ARM® Architecture Reference Manual ARMv8 for possible values of the fields, and the access modes and
Security states that can be tested.
Note
• For all cases the match refers to the privilege level of the access, not the mode of the processor. For example, if
the watchpoint is configured to match only accesses at PL1 or higher, and the processor executes an LDRT
instruction in a PL1 mode, the watchpoint does not match.
• Permitted values of this field are not identical to those for the DBGBCR. In the DBGBCR the value 0b00
permitted.

[0] E Watchpoint Enable. This bit enables the watchpoint:

0 Watchpoint disabled.
1 Watchpoint enabled.

A watchpoint never generates a Watchpoint debug event when it is disabled.

Note
The value of DBGWCR.E is UNKNOWN on reset. A debugger must ensure that DBGWCR.E has a defined value
before it programs DBGDSCR[15:14] to enable debug.

To access the DBGWCRn in AArch32 state, read or write the CP14 register with:
MRC p14, 0, <Rt>, c0, cn, 7; Read Debug Watchpoint Control Register n
MCR p14, 0, <Rt>, c0, cn, 7; Write Debug Watchpoint Control Register n

To access the DBGWCRn_EL1 in AArch64 state, read or write the register with:
MRS <Xt>, DBGWCRn_EL1; Read Debug Watchpoint Control Register n
MSR DBGWCRn_EL1, <Xt>; Write Debug Watchpoint Control Register n

The DBGWCRn_EL1 can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0x8n8. The range of n for DBGWCRn_EL1 is 0 to 3.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-355
reserved.
Non-Confidential
 10 Debug
10.5 AArch32 debug register summary

10.5 AArch32 debug register summary

The following table summarizes the 32-bit and 64-bit debug control registers that are accessible in
AArch32 state from the internal CP14 interface.
These registers are accessed by the MCR and MRC instructions in the order of CRn, op1, CRm, op2 or MCRR
and MRRC instructions in the order of CRm, op1.
The table also shows the offset address for the AArch32 registers that are accessible from the internal
memory-mapped interface and the external debug interface. See 10.7 Memory-mapped register summary
on page 10-362 for a complete list of registers accessible from the internal memory-mapped and the
external debug interface.

Table 10-6 AArch32 debug register summary

Offset CRn op1 CRm op2 Name Type Width Description

- c0 0 c0 0 DBGDIDR RO 32-bit 10.6.1 Debug ID Register on page 10-358

- 2 DBGDTRRXext RW 32-bit Debug Data Transfer Register, Receive, External View 

0x400 4 DBGBVR0 RW 32-bit Debug Breakpoint Value Register 0 ej
0x408 5 DBGBCR0 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1
on page 10-350
0x800 6 DBGWVR0 RW 32-bit Debug Watchpoint Value Register 0 ej
0x808 7 DBGWCR0 RW 32-bit 10.4.2 Debug Watchpoint Control Registers, EL1
on page 10-352

- c1 0 DBGDSCRint RO 32-bit Debug Status and Control Register, Internal View ej

0x410 4 DBGBVR1 RW 32-bit Debug Breakpoint Value Register 1 ej
0x418 5 DBGBCR1 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1
on page 10-350
0x810 6 DBGWVR1 RW 32-bit Debug Watchpoint Value Register 1 ej
0x818 7 DBGWCR1 RW 32-bit 10.4.2 Debug Watchpoint Control Registers, EL1
on page 10-352

- c2 0 DBGDCCINT RW 32-bit Debug Comms Channel Interrupt Enable Register ej

- 2 DBGDSCRext RW 32-bit Debug Status and Control Register, External View ej

0x420 4 DBGBVR2 RW 32-bit Debug Breakpoint Value Register 2 ej
0x428 5 DBGBCR2 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1
on page 10-350
0x820 6 DBGWVR2 RW 32-bit Debug Watchpoint Value Register 2 ej
0x828 7 DBGWCR2 RW 32-bit 10.4.2 Debug Watchpoint Control Registers, EL1
on page 10-352

ej See the ARM® Architecture Reference Manual ARMv8 for more information.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-356
reserved.
Non-Confidential
 10 Debug
10.5 AArch32 debug register summary

Table 10-6 AArch32 debug register summary (continued)

Offset CRn op1 CRm op2 Name Type Width Description

- c3 2 DBGDTRTXext RW 32-bit Debug Data Transfer Register, Transmit, External View ej

0x430 4 DBGBVR3 RW 32-bit Debug Breakpoint Value Register 3 ej
0x438 5 DBGBCR3 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1
on page 10-350
0x830 6 DBGWVR3 RW 32-bit Debug Watchpoint Value Register 3 ej
0x838 7 DBGWCR3 RW 32-bit 10.4.2 Debug Watchpoint Control Registers, EL1
on page 10-352
0x440 c4 4 DBGBVR4 RW 32-bit Debug Breakpoint Value Register 4 ej
0x448 5 DBGBCR4 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1
on page 10-350
0x08C c0 0 c5 0 DBGDTRTXint WO 32-bit Debug Data Transfer Register, Transmit, Internal View ej

DBGDTRRXint RO 32-bit Debug Data Transfer Register, Receive, Internal View ej

0x450 4 DBGBVR5 RW 32-bit Debug Breakpoint Value Register 5 ej
0x458 5 DBGBCR5 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1
on page 10-350

- c6 0 DBGWFAR RW 32-bit Debug Watchpoint Fault Address Registerek, RES0.

0x098 2 DBGOSECCR RW 32-bit Debug OS Lock Exception Catch Control Register ej

- c7 0 DBGVCR RW 32-bit Debug Vector Catch Register ej

- c1 0 c0 0 DBGDRAR[31:0] RO 32-bit Debug ROM Address Register ej

- - 0 c1 - DBGDRAR[63:0] 64-bit
0x300 c1 0 c0 4 DBGOSLAR WO 32-bit Debug OS Lock Access Register ej

- c1 4 DBGOSLSR RO 32-bit Debug OS Lock Status Register ej

- c3 4 DBGOSDLR RW 32-bit Debug OS Double Lock Register ej

0x444 c4 1 DBGBXVR4 RW 32-bit Debug Breakpoint Extended Value Register 4 ej
0x310 4 DBGPRCR RW 32-bit Debug Power/Reset Control Register ej
0x454 c5 1 DBGBXVR5 RW 32-bit Debug Breakpoint Extended Value Register 5 ej

- c2 2 c0 0 DBGDSAR[31:0] RO 32-bit Debug Self Address Registerel RES0

- - 0 c2 - DBGDSAR[63:0]el 64-bit

- c7 0 c0 7 DBGDEVID2 RO 32-bit Debug Device ID Register 2, RES0

- c1 7 DBGDEVID1 RO 32-bit 10.6.2 Debug Device ID Register 1 on page 10-359

- c2 7 DBGDEVID RO 32-bit 10.6.3 Debug Device ID Register on page 10-360

0xFA0 c8 6 DBGCLAIMSET RW 32-bit Debug Claim Tag Set Register ej
0xFA4 c9 6 DBGCLAIMCLR RW 32-bit Debug Claim Tag Clear Register ej
0xFB8 c14 6 DBGAUTHSTATUS RO 32-bit Debug Authentication Status Register ej

ek Previously returned information about the address of the instruction that accessed a watchpoint address. This register is now deprecated and is RES0.
el Previously defined the offset from the base address defined in DBGDRAR of the physical base address of the debug registers for the processor. This register is now
deprecated and RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-357
reserved.
Non-Confidential
 10 Debug
10.6 AArch32 debug register descriptions

10.6 AArch32 debug register descriptions

This section describes the debug registers in AArch32 state.
10.5 AArch32 debug register summary on page 10-356 provides cross-references to the individual
registers.
This section contains the following subsections:
• 10.6.1 Debug ID Register on page 10-358.
• 10.6.2 Debug Device ID Register 1 on page 10-359.
• 10.6.3 Debug Device ID Register on page 10-360.

10.6.1 Debug ID Register

The DBGDIDR characteristics are:

Purpose Specifies:
• The version of the Debug architecture that is implemented.
• Some features of the debug implementation.
Usage The accessibility to the DBGDIDR by Exception level is:
constraints
EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

RO RO RO RO RO RO

Configurations The DBGDIDR is Common to Secure and Non-secure states.

Attributes See the register summary in Table 10-6 AArch32 debug register summary
on page 10-356.

The following figure shows the DBGDIDR bit assignments.

31 28 27 24 23 20 19 16 15 14 13 12 11 0

WRPs BRPs CTX_CMPs Version RES0

RES1 SE_imp
nSUHD_imp RES0

Figure 10-4 DBGDIDR bit assignments

The following table shows the DBGDIDR bit assignments.

Table 10-7 DBGDIDR bit assignments

Bits Name Function

[31:28] WRPs The number of Watchpoint Register Pairs (WRPs) implemented. The number of implemented WRPs is one
more than the value of this field. The value is:
0x3 The processor implements 4 WRPs.

This field has the same value as ID_AA64DFR0_EL1.WRPs.

[27:24] BRPs The number of Breakpoint Register Pairs (BRPs) implemented. The number of implemented BRPs is one
more than the value of this field. The value is:
0x5 The processor implements 6 BRPs.

This field has the same value as ID_AA64DFR0_EL1.BRPs.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-358
reserved.
Non-Confidential
 10 Debug
10.6 AArch32 debug register descriptions

Table 10-7 DBGDIDR bit assignments (continued)

Bits Name Function

[23:20] CTX_CMPs The number of BRPs that can be used for Context matching. This is one more than the value of this field. The
value is:
0x1 The processor implements two Context matching breakpoints, breakpoints 4 and 5.

This field has the same value as ID_AA64DFR0_EL1.CTX_CMPs.

[19:16] Version The Debug architecture version.

0x6 The processor implements ARMv8 Debug architecture.

[15] - Reserved, RES1.

[14] nSUHD_imp Secure User Halting Debug not implemented bit. The value is:
1 The processor does not implement Secure User Halting Debug.

[13] - Reserved, RES0.

[12] SE_imp Security Extensions implemented bit. The value is:

1 The processor implements Security Extensions.

[11:0] - Reserved, RES0.

To access the DBGDIDR in AArch32 state, read or write the CP14 register with:
MRC p14, 0, <Rt>, c0, c0, 0; Read Debug ID Register

10.6.2 Debug Device ID Register 1

The DBGDEVID1 characteristics are:

Purpose
Adds to the information given by the DBGDIDR by describing other features of the debug
implementation.
Usage constraints
The accessibility to the DBGDEVID1 by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations
The DBGDEVID1 is Common to Secure and Non-secure states.
Attributes
See the register summary in Table 10-6 AArch32 debug register summary on page 10-356.
The following figure shows the DBGDEVID1 bit assignments.
31 4 3 0

RES0 PCSROffset

Figure 10-5 DBGDEVID1 bit assignments

The following table shows the DBGDEVID1 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-359
reserved.
Non-Confidential
 10 Debug
10.6 AArch32 debug register descriptions

Table 10-8 DBGDEVID1 bit assignments

Bits Name Function

[31:4] - Reserved, RES0.

[3:0] PCSROffset Indicates the offset applied to PC samples returned by reads of EDPCSR. The value is:
0x2 EDPCSR samples have no offset applied and do not sample the instruction set state in the AArch32 state.

To access the DBGDEVID1 in AArch32 state, read the CP14 register with:
MRC p14, 0, <Rt>, c7, c1, 47 Read Debug Device ID Register 1

10.6.3 Debug Device ID Register

The DBGDEVID characteristics are:

Purpose Specifies the version of the Debug architecture is implemented, and some features of
the debug implementation.
Usage The accessibility to the DBGDEVID by Exception level is:
constraints
EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

Configurations The DBGDEVID is Common to Secure and Non-secure states.

Attributes See the register summary in Table 10-6 AArch32 debug register summary
on page 10-356.

The following figure shows the DBGDEVID bit assignments.

31 28 27 24 23 20 19 16 15 12 11 8 7 4 3 0

CIDMask AuxRegs DoubleLock VirtExtns VectorCatch PCsample

BPAddrMask WPAddrMask

Figure 10-6 DBGDEVID bit assignments

The following table shows the DBGDEVID bit assignments.

Table 10-9 DBGDEVID bit assignments

Bits Name Function

[31:28] CIDMask Specifies the level of support for the Context ID matching breakpoint masking capability. This value is:
0x0 Context ID masking is not implemented.

[27:24] AuxRegs Specifies support for the Debug External Auxiliary Control Register. This value is:
0x1 The processor supports Debug External Auxiliary Control Register.

[23:20] DoubleLock Specifies support for the Debug OS Double Lock Register. This value is:
0x1 The processor supports Debug OS Double Lock Register.

[19:16] VirExtns Specifies whether EL2 is implemented. This value is:

0x1 The processor implements EL2.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-360
reserved.
Non-Confidential
 10 Debug
10.6 AArch32 debug register descriptions

Table 10-9 DBGDEVID bit assignments (continued)

Bits Name Function

[15:12] VectorCatch Defines the form of the vector catch event implemented. This value is:
0x0 The processor implements address matching form of vector catch.

[11:8] BPAddrMask Indicates the level of support for the Immediate Virtual Address (IVA) matching breakpoint masking
capability. This value is:
0xF Breakpoint address masking not implemented. DBGBCRn[28:24] are UNK/SBZP.

[7:4] WPAddrMask Indicates the level of support for the DVA matching watchpoint masking capability. This value is:
0x1 Watchpoint address mask implemented.

[3:0] PCSample Indicates the level of support for Program Counter sampling using debug registers 40 and 41. This value is:
0x3 EDPCSR, EDCIDSR and EDVIDSR are implemented as debug registers 40, 41, and 42.

To access the DBGDEVID in AArch32 state, read the CP14 register with:
MRC p14, 0, <Rt>, c7, c2, 7; Read Debug Device ID Register 0

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-361
reserved.
Non-Confidential
 10 Debug
10.7 Memory-mapped register summary

10.7 Memory-mapped register summary

The following table shows the offset address for the registers that are accessible from the internal
memory-mapped interface or the external debug interface.

Table 10-10 Memory-mapped debug register summary

Offset Name Type Width Description

0x000-0x01C - - - Reserved

0x020 EDESR RW 32-bit External Debug Event Status Register 

0x024 EDECR RW 32-bit External Debug Execution Control Register em

0x028-0x02C - - - Reserved

0x030 EDWARlo RO 32-bit External Debug Watchpoint Address Register, low word em
0x034 EDWARhi RO 32-bit External Debug Watchpoint Address Register, high word em

0x038-0x07C - - - Reserved

0x080 DBGDTRRX_EL0 RW 32-bit Debug Data Transfer Register, Receive em

0x084 EDITR WO 32-bit External Debug Instruction Transfer Register em
0x088 EDSCR RW 32-bit External Debug Status and Control Register em
0x08C DBGDTRTX_EL0 RW 32-bit Debug Data Transfer Register, Transmit em
0x090 EDRCR WO 32-bit External Debug Reserve Control Register em
0x094 EDACR RW 32-bit 10.8.2 External Debug Auxiliary Control Register on page 10-367.
0x098 EDECCR RW 32-bit External Debug Exception Catch Control Register em
0x09C - - - Reserved
0x0A0 EDPCSRlo RO 32-bit External Debug Program Counter Sample Register, low word em
0x0A4 EDCIDSR RO 32-bit External Debug Context ID Sample Register em
0x0A8 EDVIDSR RO 32-bit External Debug Virtual Context Sample Register em
0x0AC EDPCSRhi RO 32-bit External Debug Program Counter Sample Register, high word em

0x0B0-0x2FC - - - Reserved

0x300 OSLAR_EL1 WO 32-bit Debug OS Lock Access Register em

0x304-0x30C - - - Reserved

0x310 EDPRCR RW 32-bit External Debug Power/Reset Control Register em

0x314 EDPRSR RO 32-bit External Debug Processor Status Register em

0x318-0x3FC - - - Reserved

0x400 DBGBVR0_EL1[31:0] RW 32-bit Debug Breakpoint Value Register 0 em

0x404 DBGBVR0_EL1[63:32]
0x408 DBGBCR0_EL1 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1 on page 10-350
0x40C - - - Reserved

em See the ARM® Architecture Reference Manual ARMv8 for more information.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-362
reserved.
Non-Confidential
 10 Debug
10.7 Memory-mapped register summary

Table 10-10 Memory-mapped debug register summary (continued)

Offset Name Type Width Description

0x410 DBGBVR1_EL1[31:0] RW 32-bit Debug Breakpoint Value Register 1 em
0x414 DBGBVR1_EL1[63:32]
0x418 DBGBCR1_EL1 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1 on page 10-350
0x41C - - - Reserved
0x420 DBGBVR2_EL1[31:0] RW 32-bit Debug Breakpoint Value Register 2 em
0x424 DBGBVR2_EL1[63:32]
0x428 DBGBCR2_EL1 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1 on page 10-350
0x42C - - - Reserved
0x430 DBGBVR3_EL1[31:0] RW 32-bit Debug Breakpoint Value Register 3 em
0x434 DBGBVR3_EL1[63:32]
0x438 DBGBCR3_EL1 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1 on page 10-350
0x43C - - - Reserved
0x440 DBGBVR4_EL1[31:0] RW 32-bit Debug Breakpoint Value Register 4 em
0x444 DBGBVR4_EL1[63:32]
0x448 DBGBCR4_EL1 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1 on page 10-350
0x44C - - - Reserved
0x450 DBGBVR5_EL1[31:0] RW 32-bit Debug Breakpoint Value Register 5 em
0x454 DBGBVR5_EL1[63:32]
0x458 DBGBCR5_EL1 RW 32-bit 10.4.1 Debug Breakpoint Control Registers, EL1 on page 10-350

0x45C-0x7FC - - - Reserved

0x800 DBGWVR0_EL1[31:0] RW 32-bit Debug Watchpoint Value Register 0 em

0x804 DBGWVR0_EL1[63:32]
0x808 DBGWCR0_EL1 RW 32-bit 10.4.2 Debug Watchpoint Control Registers, EL1 on page 10-352
0x80C - - - Reserved
0x810 DBGWVR1_EL1[31:0] RW 32-bit Debug Watchpoint Value Register 1 em
0x814 DBGWVR1_EL1[63:32]
0x818 DBGWCR1_EL1 RW 32-bit 10.4.2 Debug Watchpoint Control Registers, EL1 on page 10-352
0x81C - - - Reserved
0x820 DBGWVR2_EL1[31:0] RW 32-bit Debug Watchpoint Value Register 2 em
0x824 DBGWVR2_EL1[63:32]
0x828 DBGWCR2_EL1 RW 32-bit 10.4.2 Debug Watchpoint Control Registers, EL1 on page 10-352
0x82C - - - Reserved
0x830 DBGWVR3_EL1[31:0] RW 32-bit Debug Watchpoint Value Register 3 em
0x834 DBGWVR3_EL1[63:32]
0x838 DBGWCR3_EL1 RW 32-bit 10.4.2 Debug Watchpoint Control Registers, EL1 on page 10-352

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-363
reserved.
Non-Confidential
 10 Debug
10.7 Memory-mapped register summary

Table 10-10 Memory-mapped debug register summary (continued)

Offset Name Type Width Description

0x83C-0xCFC - - - Reserved

0xD00 MIDR_EL1 RO 32-bit 4.3.1 Main ID Register, EL1 on page 4-89

0xD04-0xD1C - - - Reserved

0xD20 ID_AA64PFR0_EL1[31:0] RO 32-bit 4.3.18 AArch64 Processor Feature Register 0, EL1 on page 4-111
0xD24 ID_AA64PFR0_EL1[63:32] RO 32-bit
0xD28 ID_AA64DFR0_EL1[31:0] RO 32-bit 4.3.19 AArch64 Debug Feature Register 0, EL1 on page 4-113
0xD2C ID_AA64DFR0_EL1[63:32] RO 32-bit
0xD30 ID_AA64ISAR0_EL1[31:0] RO 32-bit 4.3.20 AArch64 Instruction Set Attribute Register 0, EL1
on page 4-114
0xD34 ID_AA64ISAR0_EL1[63:32] RO 32-bit
0xD38 ID_AA64MMFR0_EL1[31:0] RO 32-bit 4.3.21 AArch64 Memory Model Feature Register 0, EL1
on page 4-115
0xD3C ID_AA64MMFR0_EL1[63:32] RO 32-bit
0xD40 ID_AA64PFR1_EL1[31:0] RO 32-bit AArch64 Processor Feature Register 1 low word, RES0
0xD44 ID_AA64PFR1_EL1[63:32] RO 32-bit AArch64 Processor Feature Register 1 high word, RES0
0xD48 ID_AA64DFR1_EL1[31:0] RO 32-bit AArch64 Debug Feature Register 1 low word, RES0
0xD4C ID_AA64DFR1_EL1[63:32] RO 32-bit AArch64 Debug Feature Register 1 high word, RES0
0xD50 ID_AA64ISAR1_EL1[31:0] RO 32-bit AArch64 Instruction Set Attribute Register 1 low word, RES0
0xD54 ID_AA64ISAR1_EL1[63:32] RO 32-bit AArch64 Instruction Set Attribute Register 1 high word, RES0
0xD58 ID_AA64MMFR1_EL1[31:0] RO 32-bit AArch64 Memory Model Feature Register 1 low word, RES0
0xD5C ID_AA64MMFR1_EL1[63:32] RO 32-bit AArch64 Memory Model Feature Register 1 high word, RES0

0xD60-0xEF4 - - - Reserved

0xEF8 EDITOCTRL WO 32-bit 10.8.3 External Debug Integration Output Control Register
on page 10-368
0xEFC EDITISR RO 32-bit 10.8.4 External Debug Integration Input Status Register
on page 10-369
0xF00 EDITCTRL RW 32-bit 10.8.5 External Debug Integration Mode Control Register
on page 10-370

0xF04-0xF9C - - - Reserved

0xFA0 DBGCLAIMSET_EL1 RW 32-bit Debug Claim Tag Set Register em

0xFA4 DBGCLAIMCLR_EL1 RW 32-bit Debug Claim Tag Clear Register em
0xFA8 EDDEVAFF0 RO 32-bit External Debug Device Affinity Register 0. See
4.3.2 Multiprocessor Affinity Register, EL1 on page 4-90
0xFAC EDDEVAFF1 RO 32-bit External Debug Device Affinity Register 1, RES0
0xFB0 EDLAR WO 32-bit External Debug Lock Access Register em
0xFB4 EDLSR RO 32-bit External Debug Lock Status Register em
0xFB8 DBGAUTHSTATUS_EL1 RO 32-bit Debug Authentication Status Register em
0xFBC EDDEVARCH RO 32-bit External Debug Device Architecture Register em

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-364
reserved.
Non-Confidential
 10 Debug
10.7 Memory-mapped register summary

Table 10-10 Memory-mapped debug register summary (continued)

Offset Name Type Width Description

0xFC0 EDDEVID2 RO 32-bit External Debug Device ID Register 2, RES0
0xFC4 EDDEVID1 RO 32-bit 10.8.6 External Debug Device ID Register 1 on page 10-371
0xFC8 EDDEVID RO 32-bit 10.8.7 External Debug Device ID Register 0 on page 10-371
0xFCC EDDEVTYPE RO 32-bit External Debug Device Type Register em
0xFD0 EDPIDR4 RO 32-bit External Debug Peripheral Identification Register 4 on page 10-376

0xFD4-0xFDC EDPIDR5-7 RO 32-bit External Debug Peripheral Identification Register 5-7

on page 10-376
0xFE0 EDPIDR0 RO 32-bit External Debug Peripheral Identification Register 0 on page 10-373
0xFE4 EDPIDR1 RO 32-bit External Debug Peripheral Identification Register 1 on page 10-373
0xFE8 EDPIDR2 RO 32-bit External Debug Peripheral Identification Register 2 on page 10-374
0xFEC EDPIDR3 RO 32-bit External Debug Peripheral Identification Register 3 on page 10-375
0xFF0 EDCIDR0 RO 32-bit External Debug Component Identification Register 0
on page 10-377
0xFF4 EDCIDR1 RO 32-bit External Debug Component Identification Register 1
on page 10-377
0xFF8 EDCIDR2 RO 32-bit External Debug Component Identification Register 2
on page 10-378
0xFFC EDCIDR3 RO 32-bit External Debug Component Identification Register 3
on page 10-379

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-365
reserved.
Non-Confidential
 10 Debug
10.8 Memory-mapped register descriptions

10.8 Memory-mapped register descriptions

This section describes the Cortex-A72 processor debug registers.
10.7 Memory-mapped register summary on page 10-362 provides cross-references to the individual
registers.
This section contains the following subsections:
• 10.8.1 External Debug Reserve Control Register on page 10-366.
• 10.8.2 External Debug Auxiliary Control Register on page 10-367.
• 10.8.3 External Debug Integration Output Control Register on page 10-368.
• 10.8.4 External Debug Integration Input Status Register on page 10-369.
• 10.8.5 External Debug Integration Mode Control Register on page 10-370.
• 10.8.6 External Debug Device ID Register 1 on page 10-371.
• 10.8.7 External Debug Device ID Register 0 on page 10-371.
• 10.8.8 External Debug Peripheral Identification Registers on page 10-372.
• 10.8.9 External Debug Component Identification Registers on page 10-376.

10.8.1 External Debug Reserve Control Register

The EDRCR characteristics are:
Purpose Used to cancel bus requests and clear sticky bits in the EDSCR.
Usage Accessible through the internal memory-mapped interface and the external debug
constraints interface. The access conditions are:

Off DLK OSLK EDAD SLK Default

ERROR ERROR ERROR - WI WO

Table 10-1 External register access conditions on page 10-347 describes the access
conditions.
Configurations The EDRCR is in the Core power domain.
Attributes See the register summary in Table 10-10 Memory-mapped debug register summary
on page 10-362.

The following figure shows the EDRCR bit assignments.

31 5 4 3 2 1 0

RES0 RES0

Cancel Bus Requests Request

Clear Sticky Pipeline Advance
Clear Sticky Error

Figure 10-7 EDRCR bit assignments

The following table shows the EDRCR bit assignments.

Table 10-11 EDRCR bit assignments

Bits Name Function

[31:5] - Reserved, RES0.

[4] CBRRQ Reserved.RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-366
reserved.
Non-Confidential
 10 Debug
10.8 Memory-mapped register descriptions

Table 10-11 EDRCR bit assignments (continued)

Bits Name Function

[3] CSPA Clear Sticky Pipeline Advance. This bit is used to clear the EDSCR.PipeAdv bit to 0. The possible values are:
0 No action.
1 Clear the EDSCR.PipeAdv bit to 0.

[2] CSE Clear Sticky Error. Used to clear the EDSCR cumulative error bits to 0. The possible values:
0 No action.
1 Clear the EDSCR.{TXU, RXO, ERR} bits, and, if the processor is in Debug state, the EDSCR.ITO bit, to 0.

[1:0] - Reserved, RES0.

10.8.2 External Debug Auxiliary Control Register

The EDACR characteristics are:

Purpose
Provides IMPLEMENTATION DEFINED configuration and control options.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - RO RW

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The EDACR is in the Debug power domain.
Attributes
See 10.7 Memory-mapped register summary on page 10-362.
The following figure shows the EDACR bit assignments.

31 4 3 2 1 0

RES0

Override idle acknowledgement

Core debug reset status
Debug powerdown override
Debug clock stop control

Figure 10-8 EDACR bit assignments

The following table shows the EDACR bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-367
reserved.
Non-Confidential
 10 Debug
10.8 Memory-mapped register descriptions

Table 10-12 EDACR bit assignments

Bits Name Function

[31:4] - Reserved, RES0.

[3] Override idle Override idle Acknowledgment signal to processor. The possible values are:
Acknowledgment
0 Processor waits for the debug register access logic to go idle before it enters the idle state.
This is the reset value.
1 Processor does not wait for the debug register access logic to go idle before it enters the idle
state.

[2] Core debug reset status Read-only status bit that reflects the current reset state of the debug logic in the processor power
domain:
0 Debug logic in processor power domain is not in reset state.
1 Debug logic in processor power domain is currently in reset state.

[1] Debug powerdown Debug powerdown control bit. If debug is enabled and this bit is:
override
0 Error response is generated for APB accesses to the processor domain debug registers when
the processor is powered down or OS Double Lock is set. This is the reset value.
1 APB accesses to the processor domain debug registers proceed normally when the processor
is powered down or OS Double Lock is set.

[0] Debug clock stop control Debug clock control bit. If debug is enabled and this bit is:
0 Does not prevent the clock generator from stopping the processor clock. This is the reset
value.
1 Prevents the clock generator from stopping the processor clock.

10.8.3 External Debug Integration Output Control Register

The EDITOCTRL characteristics are:

Purpose
Controls signal outputs when EDITCTRL.IME is set.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

Error Error Error - WI WO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
EDITOCTRL is in the Core power domain.
Attributes
See 10.7 Memory-mapped register summary on page 10-362.
The following figure shows the EDITOCTRL bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-368
reserved.
Non-Confidential
 10 Debug
10.8 Memory-mapped register descriptions

31 3 2 1 0

Reserved

nPMUIRQ
CTI PMUIRQ
CTI DBGTRIGGER

Figure 10-9 EDITOCTRL bit assignments

The following table shows the EDITOCTRL bit assignments.

Table 10-13 EDITOCTRL bit assignments

Bits Name Function

[31:3] - Reserved, RES0.

[2] nPMUIRQ Controls the nPMUIRQ output. When this bit is set to 1, the corresponding nPMUIRQ signal goes
LOW. The reset value is 0.

[1] CTI PMUIRQ Controls the internal signal equivalent to PMUIRQ that goes from the PMU to the CTI. The reset value
is 0.

[0] CTI DBGTRIGGER Controls the internal signal equivalent to DBGTRIGGER that goes from the Debug unit to the CTI.
The reset value is 0.

Related information
10.8.5 External Debug Integration Mode Control Register on page 10-370.

10.8.4 External Debug Integration Input Status Register

The EDITISR characteristics are:

Purpose
Enables the values of signal inputs to be read when EDITCTRL.IME is set.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

Error Error Error - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
EDITISR is in the Core power domain.
Attributes
See 10.7 Memory-mapped register summary on page 10-362.
The following figure shows the EDITISR bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-369
reserved.
Non-Confidential
 10 Debug
10.8 Memory-mapped register descriptions

31 3 2 1 0

Reserved

CTI DBGRESTART
CTI EDBGRQ
EDBGRQ

Figure 10-10 EDITISR bit assignments

The following table shows the EDITISR bit assignments.

Table 10-14 EDITISR bit assignments

Bits Name Function

[31:3] - Reserved, RES0.

[2] CTI DBGRESTART CTI debug restart bit. This bit reads the state of the debug restart input coming from the CTI into the
debug unit.

[1] CTI EDBGRQ CTI debug request bit. This bit reads the state of the debug request input coming from the CTI into the
debug unit.

[0] EDBGRQ This bit reads the state of the EDBGRQ input.

Related information
10.8.5 External Debug Integration Mode Control Register on page 10-370.

10.8.5 External Debug Integration Mode Control Register

The EDITCTRL characteristics are:

Purpose
Enables the external debug to switch from its default mode into integration mode, where test
software can control directly the inputs and outputs of the processor, for integration testing or
topology detection.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

Error Error Error - RO RW

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
EDITCTRL is in the Core power domain.
Attributes
See 10.7 Memory-mapped register summary on page 10-362.
The following figure shows the EDITCTRL bit assignments.
31 1 0

Reserved

IME

Figure 10-11 EDITCTRL bit assignments

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-370
reserved.
Non-Confidential
 10 Debug
10.8 Memory-mapped register descriptions

The following table shows the EDITCTRL bit assignments.

Table 10-15 EDITCTRL bit assignments

Bits Name Function

[31:1] - Reserved, RES0.

[0] IME When IME is set to 1, the device reverts to an integration mode to enable integration testing or topology detection:
0 Normal operation.
1 Integration mode enabled.

10.8.6 External Debug Device ID Register 1

The EDDEVID1 characteristics are:

Purpose
Provides extra information for external debuggers about features of the debug implementation.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The EDDEVID1 is in the Debug power domain.
Attributes
See 10.7 Memory-mapped register summary on page 10-362.
The following figure shows the EDDEVID1 bit assignments.
31 4 3 0

Reserved PCSROffset

Figure 10-12 EDDEVID1 bit assignments

The following table shows the EDDEVID1 bit assignments.

Table 10-16 EDDEVID1 bit assignments

Bits Name Function

[31:4] - Reserved, RES0.

[3:0] PCSROffset Indicates the offset applied to PC samples returned by reads of EDPCSR. For ARMv8 the value is:
0x2 EDPCSR samples have no offset applied and do not sample the instruction set state in AArch32 state.

10.8.7 External Debug Device ID Register 0

The EDDEVID characteristics are:

Purpose
Provides extra information for external debuggers about features of the debug implementation.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-371
reserved.
Non-Confidential
 10 Debug
10.8 Memory-mapped register descriptions

Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The EDDEVID is in the Debug power domain.
Attributes
See 10.7 Memory-mapped register summary on page 10-362.
The following figure shows the EDDEVID bit assignments.
31 28 27 24 23 4 3 0

Reserved AuxRegs Reserved PC Sample

Figure 10-13 EDDEVID bit assignments

The following table shows the EDDEVID bit assignments.

Table 10-17 EDDEVID bit assignments

Bits Name Function

[31:28] - Reserved, RES0.

[27:24] AuxRegs Indicates support for auxiliary registers. The possible values are:
0x1 External Debug Auxiliary Control Register, EDACR, is implemented.

[23:4] - Reserved, RES0.

[3:0] PC Sample Indicates the level of sample-based profiling support using external debug registers 40 through 43. Valid values
of this field in v8-A are:
0x3 EDPCSR, EDCIDSR, and EDVIDSR are implemented.

10.8.8 External Debug Peripheral Identification Registers

The External Debug Peripheral Identification Registers provide standard information required for all
components that conform to the ARM® Debug Interface Architecture Specification ADIv5.0 to ADIv5.2.
They are a set of eight registers, listed in register number order in the following table.

Table 10-18 Summary of the External Debug Peripheral Identification Registers

Register Value Offset

EDPIDR4 0x04 0xFD0

EDPIDR5 0x00 0xFD4

EDPIDR6 0x00 0xFD8

EDPIDR7 0x00 0xFDC

EDPIDR0 0x08 0xFE0

EDPIDR1 0xBD 0xFE4

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-372
reserved.
Non-Confidential
 10 Debug
10.8 Memory-mapped register descriptions

Table 10-18 Summary of the External Debug Peripheral Identification Registers (continued)

Register Value Offset

EDPIDR2 0x0B 0xFE8

EDPIDR3 0x00 0xFEC

Only bits[7:0] of each External Debug Peripheral ID Register are used, with bits[31:8] reserved.
Together, the eight External Debug Peripheral ID Registers define a single 64-bit Peripheral ID.
The External Debug Peripheral ID registers are:
• External Debug Peripheral Identification Register 0 on page 10-373.
• External Debug Peripheral Identification Register 1 on page 10-373.
• External Debug Peripheral Identification Register 2 on page 10-374.
• External Debug Peripheral Identification Register 3 on page 10-375.
• External Debug Peripheral Identification Register 4 on page 10-376.
• External Debug Peripheral Identification Register 5-7 on page 10-376.

External Debug Peripheral Identification Register 0

The EDPIDR0 characteristics are:
Purpose
Provides information to identify an external debug component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The EDPIDR0 is in the Debug power domain.
Attributes
See 10.7 Memory-mapped register summary on page 10-362.
The following figure shows the EDPIDR0 bit assignments.
31 8 7 0

Reserved Part_0

Figure 10-14 EDPIDR0 bit assignments

The following table shows the EDPIDR0 bit assignments.

Table 10-19 EDPIDR0 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:0] Part_0 0x08 Least significant byte of the debug part number.

External Debug Peripheral Identification Register 1

The EDPIDR1 characteristics are:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-373
reserved.
Non-Confidential
 10 Debug
10.8 Memory-mapped register descriptions

Purpose
Provides information to identify an external debug component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The EDPIDR1 is in the Debug power domain.
Attributes
See 10.7 Memory-mapped register summary on page 10-362.
The following figure shows the EDPIDR1 bit assignments.
31 8 7 4 3 0

Reserved DES_0 Part_1

Figure 10-15 EDPIDR1 bit assignments

The following table shows the EDPIDR1 bit assignments.

Table 10-20 EDPIDR1 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] DES_0 0xB ARM Limited. This is the least significant nibble of JEP106 ID code.

[3:0] Part_1 0xD Most significant nibble of the debug part number.

External Debug Peripheral Identification Register 2

The EDPIDR2 characteristics are:
Purpose
Provides information to identify an external debug component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The EDPIDR2 is in the Debug power domain.
Attributes
See 10.7 Memory-mapped register summary on page 10-362.
The following figure shows the EDPIDR2 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-374
reserved.
Non-Confidential
 10 Debug
10.8 Memory-mapped register descriptions

31 8 7 4 3 2 0

Reserved Revision DES_1

JEDEC

Figure 10-16 EDPIDR2 bit assignments

The following table shows the EDPIDR2 bit assignments.

Table 10-21 EDPIDR2 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] Revision 0x0 Part major revision.

[3] JEDEC 0b1 RAO. Indicates a JEP106 identity code is used.

[2:0] DES_1 0b011 ARM Limited. This is the most significant nibble of JEP106 ID code.

External Debug Peripheral Identification Register 3

The EDPIDR3 characteristics are:
Purpose
Provides information to identify an external debug component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The EDPIDR3 is in the Debug power domain.
Attributes
See 10.7 Memory-mapped register summary on page 10-362.
The following figure shows the EDPIDR3 bit assignments.
31 8 7 4 3 0

Reserved REVAND CMOD

Figure 10-17 EDPIDR3 bit assignments

The following table shows the EDPIDR3 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-375
reserved.
Non-Confidential
 10 Debug
10.8 Memory-mapped register descriptions

Table 10-22 EDPIDR3 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] REVAND 0x0 Part minor revision.

[3:0] CMOD 0x0 Customer modified.

External Debug Peripheral Identification Register 4

The EDPIDR4 characteristics are:
Purpose
Provides information to identify an external debug component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The EDPIDR4 is in the Debug power domain.
Attributes
See 10.7 Memory-mapped register summary on page 10-362.
The following figure shows the EDPIDR4 bit assignments.
31 8 7 4 3 0

Reserved Size DES_2

Figure 10-18 EDPIDR4 bit assignments

The following table shows the EDPIDR4 bit assignments.

Table 10-23 EDPIDR4 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] Size 0x0 Size of the component. Log2 the number of 4KB pages from the start of the component to the end of the
component ID registers.

[3:0] DES_2 0x4 ARM Limited. This is the least significant nibble JEP106 continuation code.

External Debug Peripheral Identification Register 5-7

No information is held in the Peripheral ID5, Peripheral ID6, and Peripheral ID7 Registers. They are
reserved for future use and are RES0.

10.8.9 External Debug Component Identification Registers

There are four read-only External Debug Component Identification Registers, Debug Component ID0
through Debug Component ID3. The following table shows these registers.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-376
reserved.
Non-Confidential
 10 Debug
10.8 Memory-mapped register descriptions

Table 10-24 Summary of the External Debug Component Identification Registers

Register Value Offset

Component ID0 0x0D 0xFF0

Component ID1 0x90 0xFF4

Component ID2 0x05 0xFF8

Component ID3 0xB1 0xFFC

The External Debug Component Identification Registers identify Debug as an ARM Debug Interface v5
component. The External Debug Component ID registers are:
• External Debug Component Identification Register 0 on page 10-377.
• External Debug Component Identification Register 1 on page 10-377.
• External Debug Component Identification Register 2 on page 10-378.
• External Debug Component Identification Register 3 on page 10-379.

External Debug Component Identification Register 0

The EDCIDR0 characteristics are:
Purpose
Provides information to identify an external debug component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The EDCIDR0 is in the Debug power domain.
Attributes
See 10.7 Memory-mapped register summary on page 10-362.
The following figure shows the EDCIDR0 bit assignments.
31 8 7 0

Reserved PRMBL_0

Figure 10-19 EDCIDR0 bit assignments

The following table shows the EDCIDR0 bit assignments.

Table 10-25 EDCIDR0 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:0] PRMBL_0 0x0D Preamble byte 0.

External Debug Component Identification Register 1

The EDCIDR1 characteristics are:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-377
reserved.
Non-Confidential
 10 Debug
10.8 Memory-mapped register descriptions

Purpose
Provides information to identify an external debug component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The EDCIDR1 is in the Debug power domain.
Attributes
See 10.7 Memory-mapped register summary on page 10-362.
The following figure shows the EDCIDR1 bit assignments.
31 8 7 4 3 0

Reserved CLASS PRMBL_1

Figure 10-20 EDCIDR1 bit assignments

The following table shows the EDCIDR1 bit assignments.

Table 10-26 EDCIDR1 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] CLASS 0x9 Debug component.

[3:0] PRMBL_1 0x0 Preamble.

External Debug Component Identification Register 2

The EDCIDR2 characteristics are:
Purpose
Provides information to identify an external debug component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The EDCIDR2 is in the Debug power domain.
Attributes
See 10.7 Memory-mapped register summary on page 10-362.
The following figure shows the EDCIDR2 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-378
reserved.
Non-Confidential
 10 Debug
10.8 Memory-mapped register descriptions

31 8 7 0

Reserved PRMBL_2

Figure 10-21 EDCIDR2 bit assignments

The following table shows the EDCIDR2 bit assignments.

Table 10-27 EDCIDR2 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:0] PRMBL_2 0x05 Preamble byte 2.

External Debug Component Identification Register 3

The EDCIDR3 characteristics are:
Purpose
Provides information to identify an external debug component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The EDCIDR3 is in the Debug power domain.
Attributes
See 10.7 Memory-mapped register summary on page 10-362.
The following figure shows the EDCIDR3 bit assignments.
31 8 7 0

Reserved PRMBL_3

Figure 10-22 EDCIDR3 bit assignments

The following table shows the EDCIDR3 bit assignments.

Table 10-28 EDCIDR3 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:0] PRMBL_3 0xB1 Preamble byte 3.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-379
reserved.
Non-Confidential
 10 Debug
10.9 Debug events

10.9 Debug events

Describes debug events in the processor.
A debug event can be either:
• A software debug event.
• A halting debug event.
A processor responds to a debug event in one of the following ways:
• Ignores the debug event.
• Takes a debug exception.
• Enters Debug state.
See the ARM® Architecture Reference Manual ARMv8 for more information on debug events.
This section contains the following subsections:
• 10.9.1 Watchpoint debug events on page 10-380.
• 10.9.2 Debug OS Lock on page 10-380.

10.9.1 Watchpoint debug events

In the Cortex-A72 processor, watchpoint debug events are always synchronous. Memory hint
instructions and cache clean operations, except DC ZVA, DC IVAC, and DC IVAU do not generate
watchpoint debug events. Store exclusive instructions generate a watchpoint debug event even when the
check for the control of exclusive monitor fails.
For watchpoint debug events, the value reported in DFAR is guaranteed to be no lower than the address
of the watchpointed location rounded down to a multiple of 16 bytes.

10.9.2 Debug OS Lock

Debug OS Lock is set by the powerup reset, nCPUPORESET. For normal behavior of debug events and
debug register accesses, Debug OS Lock must be cleared. For more information, see the ARM®
Architecture Reference Manual ARMv8.

Related information
2.3.2 Resets on page 2-36.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-380
reserved.
Non-Confidential
 10 Debug
10.10 External debug interface

10.10 External debug interface

The system can access memory-mapped debug registers through the APB interface.
The APB interface is compliant with the AMBA 3 APB interface.
The following figure shows the debug interface implemented in the Cortex-A72 processor. For more
information on these signals, see the ARM® CoreSight™ Architecture Specification.

Processor

PCLKDBG
DBGEN
PCLKENDBG
Authentication SPIDEN
PSELDBG
interface NIDEN
PADDRDBG
SPNIDEN
PRDATADBG Debug slave
COMMTX PWDATADBG port, APBv3
DCC
COMMRX PENABLEDBG
handshake
nCOMMIRQ PREADYDBG
PSLVERRDBG
Debug state DBGACK PWRITEDBG
entry EDBGRQ

DBGPWRDUP
Power DBGROMADDR
controller DBGPWRUPREQ Configuration
interface DBGROMADDRV
DBGNOPWRDWN

Reset nPRESETDBG
controller
interface DBGRSTREQ

Figure 10-23 External debug interface, including APBv3 slave port

This section contains the following subsections:
• 10.10.1 Debug memory map on page 10-381.
• 10.10.2 DBGPWRDUP debug signal on page 10-382.
• 10.10.3 DBGL1RSTDISABLE debug signal on page 10-382.
• 10.10.4 Changing the authentication signals on page 10-383.

10.10.1 Debug memory map

The memory map supports up to four cores in a cluster. The following table shows the address mapping
for the debug trace components.

Table 10-29 Address mapping for debug trace components

Address range Component en

0x000000 - 0x00FFFF ROM table

0x010000 - 0x01FFFF Core 0 Debug

0x020000 - 0x02FFFF Core 0 CTI

0x030000 - 0x03FFFF Core 0 PMU

0x040000 - 0x04FFFF Core 0 Trace

0x050000 - 0x10FFFF Reserved

en Indicates the mapped component if present, otherwise reserved.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-381
reserved.
Non-Confidential
 10 Debug
10.10 External debug interface

Table 10-29 Address mapping for debug trace components (continued)

Address range Component en

0x110000 - 0x11FFFF Core 1 Debug

0x120000 - 0x12FFFF Core 1 CTI

0x130000 - 0x13FFFF Core 1 PMU

0x140000 - 0x14FFFF Core 1 Trace

0x150000 - 0x20FFFF Reserved

0x210000 - 0x21FFFF Core 2 Debug

0x220000 - 0x22FFFF Core 2 CTI

0x230000 - 0x23FFFF Core 2 PMU

0x240000 - 0x24FFFF Core 2 Trace

0x250000 - 0x30FFFF Reserved

0x310000 - 0x31FFFF Core 3 Debug

0x320000 - 0x32FFFF Core 3 CTI

0x330000 - 0x33FFFF Core 3 PMU

0x340000 - 0x34FFFF Core 3 Trace

0x350000 - 0x3FFFFF Reserved

10.10.2 DBGPWRDUP debug signal

This section describes the DBGPWRDUP debug input signal.

DBGPWRDUP

You must set the DBGPWRDUP signal LOW before removing power to the core domain. After power
is restored to the core domain, the DBGPWRDUP signal must be asserted HIGH. The EDPRSR.PU bit
reflects the value of this DBGPWRDUP signal.
Note
DBGPWRDUP must be tied HIGH if the particular implementation does not support separate core and
debug power domains.

10.10.3 DBGL1RSTDISABLE debug signal

When set HIGH, the DBGL1RSTDISABLE input signal disables the automatic hardware controlled
invalidation of the L1 data cache after the processor is reset using nCORERESET or nCPUPORESET.
It also disables the automatic hardware-controlled invalidation of the L2 snoop tag RAMs after the L2 is
reset using nL2RESET.
The DBGL1RSTDISABLE must be used only to assist debug of an external watchdog triggered reset
by allowing the contents of the L1 data cache before the reset to be observable after the reset. If reset is
asserted, while an L1 data cache eviction or L1 data cache fetch is performed, the accuracy of those

en Indicates the mapped component if present, otherwise reserved.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-382
reserved.
Non-Confidential
 10 Debug
10.10 External debug interface

cache entries is not guaranteed. Similarly, the contents of the L2 snoop tag RAMs might be observed
following reset of the L2 if DBGL1RSTDISABLE is asserted before resetting the L2.
You must not use the DBGL1RSTDISABLE signal to disable automatic hardware-controlled
invalidation of the L1 data cache or the L2 snoop tag RAMs in normal processor powerup sequences.
This is because synchronization of the L1 data cache invalidation sequence with the duplicate L1 tags in
the Level 2 Memory System is not guaranteed.
The DBGL1RSTDISABLE signal applies to all processors in the multiprocessor. Each processor
samples the signal when nCORERESET or nCPUPORESET is asserted. The L2 samples the signal
when nL2RESET is asserted.
If the functionality offered by the DBGL1RSTDISABLE input signal is not required, the input must be
tied to LOW.

10.10.4 Changing the authentication signals

The NIDEN, DBGEN, SPIDEN, and SPNIDEN input signals are either tied off to some fixed value or
controlled by some external device.
If software running on the processor has control over an external device that drives the authentication
signals, it must make the change using a safe sequence:
1. Execute an implementation-specific sequence of instructions to change the signal value. For example,
this might be a single STR instruction that writes certain value to a control register in a system
peripheral.
2. If the prior step involves any memory operation, issue a DSB instruction.
3. Poll the DBGAUTHSTATUS_EL1 register to check whether the processor has already detected the
changed value of these signals. This is required because the system might not issue the signal change
to the processor until several cycles after the DSB instruction completes.
4. Issue an ISB instruction or exception entry or exception return.
The software cannot perform debug or analysis operations that depend on the new value of the
authentication signals until this procedure is complete. The same rules apply when the debugger has
control of the processor through the Instruction Transfer Register, EDITR, while in Debug state. The
relevant combinations of the DBGEN, NIDEN, SPIDEN, and SPNIDEN values can be determined by
polling DBGAUTHSTATUS_EL1.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-383
reserved.
Non-Confidential
 10 Debug
10.11 ROM table

10.11 ROM table

This section provides details on the processor's ROM table.
The Cortex-A72 processor includes a ROM table that complies with the ARM® CoreSight™ Architecture
Specification. This table contains a list of components such as processor debug units, processor Cross
Trigger Interfaces (CTIs), processor Performance Monitoring Units (PMUs) and processor Embedded
Trace Macrocells (ETMs). Debuggers can use the ROM table to determine which components are
implemented inside the processor.
If a component is not included in your configuration of the processor, the corresponding debug APB
ROM table entry is still present but the component is marked as not present.
This section contains the following subsections:
• 10.11.1 ROM table register interface on page 10-384.
• 10.11.2 ROM table register summary on page 10-384.
• 10.11.3 ROM table register descriptions on page 10-385.
• 10.11.4 ROM table Debug Peripheral Identification Registers on page 10-387.
• 10.11.5 ROM table Debug Component Identification Registers on page 10-391.

10.11.1 ROM table register interface

The interface to the ROM table entries is the APB slave port.

Related information
10.10 External debug interface on page 10-381.

10.11.2 ROM table register summary

The following table shows the offsets from the physical base address of the ROM table.

Table 10-30 ROM table registers

Offset Name Type Description

0x000 ROMENTRY0 RO Core 0 Debug, see ROM entry registers on page 10-385
0x004 ROMENTRY1 RO Core 0 CTI, see ROM entry registers on page 10-385
0x008 ROMENTRY2 RO Core 0 PMU, see ROM entry registers on page 10-385
0x00C ROMENTRY3 RO Core 0 ETM, see ROM entry registers on page 10-385
0x010 ROMENTRY4 RO Core 1 Debug, see ROM entry registers on page 10-385
0x014 ROMENTRY5 RO Core 1 CTI, see ROM entry registers on page 10-385
0x018 ROMENTRY6 RO Core 1 PMU, see ROM entry registers on page 10-385
0x01C ROMENTRY7 RO Core 1 ETM, see ROM entry registers on page 10-385
0x020 ROMENTRY8 RO Core 2 Debug, see ROM entry registers on page 10-385
0x024 ROMENTRY9 RO Core 2 CTI, see ROM entry registers on page 10-385
0x028 ROMENTRY10 RO Core 2 PMU, see ROM entry registers on page 10-385
0x02C ROMENTRY11 RO Core 2 ETM, see ROM entry registers on page 10-385
0x030 ROMENTRY12 RO Core 3 Debug, see ROM entry registers on page 10-385
0x034 ROMENTRY13 RO Core 3 CTI, see ROM entry registers on page 10-385

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-384
reserved.
Non-Confidential
 10 Debug
10.11 ROM table

Table 10-30 ROM table registers (continued)

Offset Name Type Description

0x038 ROMENTRY14 RO Core 3 PMU, see ROM entry registers on page 10-385
0x03C ROMENTRY15 RO Core 3 ETM, see ROM entry registers on page 10-385

0x040-0xFCC - RO Reserved, RES0

0xFD0 ROMPIDR4 RO ROM table Debug Peripheral Identification Register 4 on page 10-390
0xFD4 ROMPIDR5 RO ROM table Debug Peripheral Identification Register 5-7 on page 10-391
0xFD8 ROMPIDR6 RO
0xFDC ROMPIDR7 RO
0xFE0 ROMPIDR0 RO ROM table Debug Peripheral Identification Register 0 on page 10-387
0xFE4 ROMPIDR1 RO ROM table Debug Peripheral Identification Register 1 on page 10-388
0xFE8 ROMPIDR2 RO ROM table Debug Peripheral Identification Register 2 on page 10-389
0xFEC ROMPIDR3 RO ROM table Debug Peripheral Identification Register 3 on page 10-389
0xFF0 ROMCIDR0 RO ROM table Debug Component Identification Register 0 on page 10-391
0xFF4 ROMCIDR1 RO ROM table Debug Component Identification Register 1 on page 10-392
0xFF8 ROMCIDR2 RO ROM table Debug Component Identification Register 2 on page 10-393
0xFFC ROMCIDR3 RO ROM table Debug Component Identification Register 3 on page 10-393

10.11.3 ROM table register descriptions

This section describes the ROM table registers. 10.11.2 ROM table register summary on page 10-384
provides cross-references to individual registers.

ROM entry registers

The characteristics of the ROMENTRYn are:
Purpose Indicates to a debugger whether the debug component is present in the processor's
debug logic. There are 16 ROMENTRY registers in the Cortex-A72 processor.
Usage The accessibility to the ROMENTRYn by condition code is:
constraints
Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access
conditions.
Configurations The ROMENTRYn is Common to Secure and Non-secure states.
Attributes See the register summary in Table 10-30 ROM table registers on page 10-384.

The following figure shows the ROMENTRY bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-385
reserved.
Non-Confidential
 10 Debug
10.11 ROM table

31 12 11 2 1 0

Address offset RES0

Format
Component present

Figure 10-24 ROMENTRY bit assignments

The following table shows the ROMENTRY bit assignments.

Table 10-31 ROMENTRY bit assignments

Bits Name Function

[31:12] Address offset Address offset for the debug component.

Note
Negative values of address offsets are permitted using the two’s complement of the offset.

[11:2] - Reserved, RES0.

[1] Format Format of the ROM table entry. The value for all ROMENTRY registers is:
0 End marker.
1 32-bit format.

[0] Component presenteo Indicates whether the component is present:

0 Component is not present.
1 Component is present.

The Physical Address of a debug component is determined by shifting the address offset 12 places to the
left and adding the result to Physical Address of processor ROM table.
The following table shows the offset values for all ROMENTRY values. If a processor is not
implemented, the ROMENTRY registers for its debug, CTI, PMU, and ETM components are
0x00000000.

Table 10-32 ROMENTRY values

Name Debug component Offset value ROMENTRY value

ROMENTRY0 Core 0 Debug 0x00010 0x00010003

ROMENTRY1 Core 0 CTI 0x00020 0x00020003

ROMENTRY2 Core 0 PMU 0x00030 0x00030003

ROMENTRY3 Core 0 ETM 0x00040 0x00040003

ROMENTRY4 Core 1 Debug 0x00110 0x00110003

ROMENTRY5 Core 1 CTI 0x00120 0x00120003ep

ROMENTRY6 Core 1 PMU 0x00130 0x00130003ep

ROMENTRY7 Core 1 ETM 0x00140 0x00140003ep

ROMENTRY8 Core 2 Debug 0x00210 0x00210003ep

eo core 0 is always present. The component entries for core 1, 2, and 3 depend on your configuration.
ep If the component is present.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-386
reserved.
Non-Confidential
 10 Debug
10.11 ROM table

Table 10-32 ROMENTRY values (continued)

Name Debug component Offset value ROMENTRY value

ROMENTRY9 Core 2 CTI 0x00220 0x00220003ep

ROMENTRY10 Core 2 PMU 0x00230 0x00230003ep

ROMENTRY11 Core 2 ETM 0x00240 0x00240003ep

ROMENTRY12 Core 3 Debug 0x00310 0x00310003ep

ROMENTRY13 Core 3 CTI 0x00320 0x00320003ep

ROMENTRY14 Core 3 PMU 0x00330 0x00330003ep

ROMENTRY15 Core 3 ETM 0x00340 0x00340003ep

10.11.4 ROM table Debug Peripheral Identification Registers

The ROM table Debug Peripheral Identification Registers provide standard information required for all
components that conform to the ARM® Debug Interface Architecture Specification ADIv5.0 to ADIv5.2.
There is a set of eight registers, listed in register number order in the following table.

Table 10-33 Summary of the ROM table Debug Peripheral Identification Registers

Register Value Offset

ROMPID4 0x04 0xFD0

ROMPID5 0x00 0xFD4

ROMPID6 0x00 0xFD8

ROMPID7 0x00 0xFDC

ROMPID0 0xA4 0xFE0

ROMPID1 0xB4 0xFE4

ROMPID2 0x0B 0xFE8

ROMPID3 0x00 0xFEC

Only bits[7:0] of each ROM table Debug Peripheral ID Register are used, with bits[31:8] reserved.
Together, the eight ROM table Debug Peripheral ID Registers define a single 64-bit Peripheral ID.
The ROM table Debug Peripheral ID registers are:
• ROM table Debug Peripheral Identification Register 0 on page 10-387.
• ROM table Debug Peripheral Identification Register 1 on page 10-388.
• ROM table Debug Peripheral Identification Register 2 on page 10-389.
• ROM table Debug Peripheral Identification Register 3 on page 10-389.
• ROM table Debug Peripheral Identification Register 4 on page 10-390.
• ROM table Debug Peripheral Identification Register 5-7 on page 10-391.

ROM table Debug Peripheral Identification Register 0

The ROMPIDR0 characteristics are:

Purpose
Provides information to identify an external debug component.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-387
reserved.
Non-Confidential
 10 Debug
10.11 ROM table

Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The ROMPIDR0 is in the Debug power domain.
Attributes
See the register summary in Table 10-30 ROM table registers on page 10-384.
The following figure shows the ROMPIDR0 bit assignments.
31 8 7 0

Reserved Part_0

Figure 10-25 ROMPIDR0 bit assignments

The following table shows the ROMPIDR0 bit assignments.

Table 10-34 ROMPIDR0 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:0] Part_0 0xA4 Least significant byte of the ROM table part number.

ROM table Debug Peripheral Identification Register 1

The ROMPIDR1 characteristics are:

Purpose
Provides information to identify an external debug component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The ROMPIDR1 is in the Debug power domain.
Attributes
See the register summary in Table 10-30 ROM table registers on page 10-384.
The following figure shows the ROMPIDR1 bit assignments.
31 8 7 4 3 0

Reserved DES_0 Part_1

Figure 10-26 ROMPIDR1 bit assignments

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-388
reserved.
Non-Confidential
 10 Debug
10.11 ROM table

The following table shows the ROMPIDR1 bit assignments.

Table 10-35 ROMPIDR1 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] DES_0 0xB Least significant nibble of JEP106 ID code. For ARM Limited.

[3:0] Part_1 0x4 Most significant nibble of the ROM table part number.

ROM table Debug Peripheral Identification Register 2

The ROMPIDR2 characteristics are:
Purpose
Provides information to identify an external debug component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The ROMPIDR2 is in the Debug power domain.
Attributes
See the register summary in Table 10-30 ROM table registers on page 10-384.
The following figure shows the ROMPIDR2 bit assignments.
31 8 7 4 3 2 0

Reserved Revision DES_1

JEDEC

Figure 10-27 ROMPIDR2 bit assignments

The following table shows the ROMPIDR2 bit assignments.

Table 10-36 ROMPIDR2 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] Revision 0x0 Part major revision.

[3] JEDEC 0b1 RAO. Indicates a JEP106 identity code is used.

[2:0] DES_1 0b011 Designer, most significant bits of JEP106 ID code. For ARM Limited.

ROM table Debug Peripheral Identification Register 3

The ROMPIDR3 characteristics are:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-389
reserved.
Non-Confidential
 10 Debug
10.11 ROM table

Purpose
Provides information to identify an external debug component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The ROMPIDR3 is in the Debug power domain.
Attributes
See the register summary in Table 10-30 ROM table registers on page 10-384.
The following figure shows the ROMPIDR3 bit assignments.
31 8 7 4 3 0

Reserved REVAND CMOD

Figure 10-28 ROMPIDR3 bit assignments

The following table shows the ROMPIDR3 bit assignments.

Table 10-37 ROMPIDR3 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] REVAND 0x0 Part minor revision.

[3:0] CMOD 0x0 Customer modified.

ROM table Debug Peripheral Identification Register 4

The ROMPIDR4 characteristics are:
Purpose Provides information to identify an external debug component.
Usage constraints Accessible through the internal memory-mapped interface and the external debug
interface. The access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access
conditions.
Configurations The ROMPIDR4 is in the Debug power domain.
Attributes See the register summary in Table 10-30 ROM table registers on page 10-384.

The following figure shows the ROMPIDR4 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-390
reserved.
Non-Confidential
 10 Debug
10.11 ROM table

31 8 7 4 3 0

Reserved Size DES_2

Figure 10-29 ROMPIDR4 bit assignments

The following table shows the ROMPIDR4 bit assignments.

Table 10-38 ROMPIDR4 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] Size 0x0 Size of the component. Log2 the number of 4KB pages from the start of the component to the end of the
component ID registers.

[3:0] DES_2 0x4 Designer, JEP106 continuation code, least significant nibble. For ARM Limited.

ROM table Debug Peripheral Identification Register 5-7

No information is held in the Peripheral ID5, Peripheral ID6, and Peripheral ID7 Registers. They are
reserved for future use and are RES0.

10.11.5 ROM table Debug Component Identification Registers

There are four read-only Component Identification Registers, Component ID0 through Component ID3.
The following table shows these registers.

Table 10-39 Summary of the ROM table Debug Component Identification registers

Register Value Offset

ROMCIDR0 0x0D 0xFF0

ROMCIDR1 0x10 0xFF4

ROMCIDR2 0x05 0xFF8

ROMCIDR3 0xB1 0xFFC

The ROM table Debug Component Identification Registers identify Debug as an ARM Debug Interface
v5 component. The ROM table Component ID registers are:
• ROM table Debug Component Identification Register 0 on page 10-391.
• ROM table Debug Component Identification Register 1 on page 10-392.
• ROM table Debug Component Identification Register 2 on page 10-393.
• ROM table Debug Component Identification Register 3 on page 10-393.

ROM table Debug Component Identification Register 0

The ROMCIDR0 characteristics are:
Purpose
Provides information to identify an external debug component.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-391
reserved.
Non-Confidential
 10 Debug
10.11 ROM table

Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The ROMCIDR0 is in the Debug power domain.
Attributes
See the register summary in Table 10-30 ROM table registers on page 10-384.
The following figure shows the ROMCIDR0 bit assignments.
31 8 7 0

Reserved PRMBL_0

Figure 10-30 ROMCIDR0 bit assignments

The following table shows the ROMCIDR0 bit assignments.

Table 10-40 ROMCIDR0 bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:0] PRMBL_0 0x0D Preamble byte 0

ROM table Debug Component Identification Register 1

The ROMCIDR1 characteristics are:
Purpose Provides information to identify an external debug component.
Usage constraints Accessible through the internal memory-mapped interface and the external debug
interface. The access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access
conditions.
Configurations The ROMCIDR1 is in the Debug power domain.
Attributes See the register summary in Table 10-30 ROM table registers on page 10-384.

The following figure shows the ROMCIDR1 bit assignments.

31 8 7 4 3 0

Reserved CLASS PRMBL_1

Figure 10-31 ROMCIDR1 bit assignments

The following table shows the ROMCIDR1 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-392
reserved.
Non-Confidential
 10 Debug
10.11 ROM table

Table 10-41 ROMCIDR1 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] CLASS 0x1 Component Class. For a ROM table.

[3:0] PRMBL_1 0x0 Preamble.

ROM table Debug Component Identification Register 2

The ROMCIDR2 characteristics are:
Purpose Provides information to identify an external debug component.
Usage constraints Accessible through the internal memory-mapped interface and the external debug
interface. The access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access
conditions.
Configurations The ROMCIDR2 is in the Debug power domain.
Attributes See the register summary in Table 10-30 ROM table registers on page 10-384.

The following figure shows the ROMCIDR2 bit assignments.

31 8 7 0

Reserved PRMBL_2

Figure 10-32 ROMCIDR2 bit assignments

The following table shows the ROMCIDR2 bit assignments.

Table 10-42 ROMCIDR2 bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:0] PRMBL_2 0x05 Preamble byte 2

ROM table Debug Component Identification Register 3

The ROMCIDR3 characteristics are:
Purpose
Provides information to identify an external debug component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - - RO

Table 11-1 External register access conditions on page 11-398 describes the access conditions.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-393
reserved.
Non-Confidential
 10 Debug
10.11 ROM table

Configurations
The ROMCIDR3 is in the Debug power domain.
Attributes
See the register summary in Table 10-30 ROM table registers on page 10-384.
The following figure shows the ROMCIDR3 bit assignments.
31 8 7 0

Reserved PRMBL_3

Figure 10-33 ROMCIDR3 bit assignments

The following table shows the ROMCIDR3 bit assignments.

Table 10-43 ROMCIDR3 bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:0] PRMBL_3 0xB1 Preamble byte 3

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 10-394
reserved.
Non-Confidential
Chapter 11
Performance Monitor Unit

This section describes the Performance Monitor Unit (PMU) and the registers that it uses.

It contains the following sections:

• 11.1 About the PMU on page 11-396.
• 11.2 PMU functional description on page 11-397.
• 11.3 AArch64 PMU register summary on page 11-399.
• 11.4 AArch64 PMU register descriptions on page 11-401.
• 11.5 AArch32 PMU register summary on page 11-406.
• 11.6 Memory-mapped register summary on page 11-408.
• 11.7 Memory-mapped register descriptions on page 11-411.
• 11.8 Events on page 11-428.
• 11.9 Interrupts on page 11-432.
• 11.10 Exporting PMU events on page 11-433.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-395
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.1 About the PMU

11.1 About the PMU

The processor includes logic to gather various statistics on the operation of the processor and memory
system during runtime, based on PMUv3 architecture.
These events provide useful information about the behavior of the processor that you can use when
debugging or profiling code.
The processor PMU provides six counters. Each counter can count any of the events available in the
processor. The absolute counts recorded might vary because of pipeline effects. This has negligible effect
except in cases where the counters are enabled for a very short time.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-396
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.2 PMU functional description

11.2 PMU functional description

Describes the functionality of the PMU.
The following figure shows the PMU block diagram.

CLK
Cycle counter

Count Enable
System control, Set/Clear and
processor, and Event Performance
APB interface Selection counter
registers

Performance
counter

Interrupt and
Performance
overflow nPMUIRQ
counter
registers

Performance
counter

Performance
Events from counter
other units

Performance
counter

Figure 11-1 PMU block diagram

This section contains the following subsections:
• 11.2.1 Event interface on page 11-397.
• 11.2.2 System register and APB interface on page 11-397.
• 11.2.3 Counters on page 11-397.
• 11.2.4 PMU register interfaces on page 11-398.
• 11.2.5 External register access permissions on page 11-398.

11.2.1 Event interface

Events from all other units from across the design are provided to the PMU.

11.2.2 System register and APB interface

You can program the PMU registers using the System registers or the external APB interface.

11.2.3 Counters

The Cortex-A72 processor has six counters. Each counter can count any of the events available in the
processor.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-397
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.2 PMU functional description

11.2.4 PMU register interfaces

The Cortex-A72 processor supports access to the Performance Monitor registers from the System
registers and a memory-mapped interface. External access to the Performance Monitor registers is also
provided with the APB slave interface.

Related information
10.10 External debug interface on page 10-381.

11.2.5 External register access permissions

External access permission to the PMU registers is subject to the conditions at the time of the access. The
following table describes the processor response to accesses through the external debug and memory-
mapped interfaces.

Table 11-1 External register access conditions

Condition Condition trigger Description

Off EDPRSR.PU is 0 Core power domain is completely off, or in a low-power state where the
Core power domain registers cannot be accessed.
Note
If debug power is off then all external debug and memory-mapped register
accesses return an error.

DLK EDPRSR.DLK is 1 OS Double Lock is locked.

OSLK OSLSR_EL1.OSLK is 1 OS Lock is locked.

EPMAD AllowExternalPMUAccess() == External performance monitors access disabled. When an error is returned
FALSE because of the EPMAD condition, and this is the highest priority error
condition, EDPRSR.SPMAD is set to 1. Otherwise SPMAD is unchanged.

SLK Memory-mapped interface only Software Lock is locked. For the external debug interface, ignore this
condition.

Default - None of the conditions apply, normal access.

The following table shows an example of external register access conditions for access to a Performance
Monitors register. To determine the access permission for the register, scan the columns from left to
right. Stop at the first column whose condition is true, the entry gives the register’s access permission
and scanning stops.

Table 11-2 External register access conditions example

Off DLK OSLK EPMAD SLK Default

- - - - RO/WI RO

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-398
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.3 AArch64 PMU register summary

11.3 AArch64 PMU register summary

The PMU counters and their associated control registers are accessible in AArch64 state with MRS and
MSR instructions.

The following table shows the PMU registers in AArch64 state. It also shows the offset address for the
registers that are accessible from the internal memory-mapped interface or the external debug interface.

Table 11-3 PMU register summary in AArch64 state

Offset Name Type Width Description

0xE04 PMCR_EL0 RW 32-bit 11.4.1 Performance Monitors Control Register, EL0 on page 11-401
0xC00 PMCNTENSET_EL0 RW 32-bit Performance Monitors Count Enable Set Register 
0xC20 PMCNTENCLR_EL0 RW 32-bit Performance Monitors Count Enable Clear Register eq
0xC80 PMOVSCLR_EL0 RW 32-bit Performance Monitors Overflow Flag Status Register eq
0xCA0 PMSWINC_EL0 WO 32-bit Performance Monitors Software Increment Register eq

- PMSELR_EL0 RW 32-bit Performance Monitors Event Counter Selection Register eq

0xE20 PMCEID0_EL0 RO 32-bit 11.4.2 Performance Monitors Common Event Identification Register 0, EL0
on page 11-403
0xE24 PMCEID1_EL0 RO 32-bit Performance Monitors Common Event ID Register 1 eq

- PMCCNTR_EL0 RW 64-bit Performance Monitors Cycle Count Register eq

- PMXEVTYPER_EL0 RW 32-bit Performance Monitors Selected Event Type Register eq

0x47C PMCCFILTR_EL0 RW 32-bit Performance Monitors Cycle Count Filter Register eq er

- PMXEVCNTR0_EL0 RW 32-bit Performance Monitors Selected Event Count Register eq

- PMUSERENR_EL0 RW 32-bit Performance Monitors User Enable Register eq

0xC40 PMINTENSET_EL1 RW 32-bit Performance Monitors Interrupt Enable Set Register eq
0xC60 PMINTENCLR_EL1 RW 32-bit Performance Monitors Interrupt Enable Clear Register eq
0xCC0 PMOVSSET_EL0 RW 32-bit Performance Monitors Overflow Flag Status Set Register eq
0x000 PMEVCNTR0_EL0 RW 32-bit Performance Monitors Event Count Registers eq
0x008 PMEVCNTR1_EL0

0x010 PMEVCNTR2_EL0

0x018 PMEVCNTR3_EL0

0x020 PMEVCNTR4_EL0

0x028 PMEVCNTR5_EL0

eq See the ARM® Architecture Reference Manual ARMv8 for more information.
er The CP15 encoding provides access to PMCCFILTR_EL0 only when PMSELR_EL0.SEL==31.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-399
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.3 AArch64 PMU register summary

Table 11-3 PMU register summary in AArch64 state (continued)

Offset Name Type Width Description

0x400 PMEVTYPER0_EL0 RW 32-bit Performance Monitors Event Type Registers eq
0x404 PMEVTYPER1_EL0

0x408 PMEVTYPER2_EL0

0x40C PMEVTYPER3_EL0

0x410 PMEVTYPER4_EL0

0x414 PMEVTYPER5_EL0

0x47C PMCCFILTR_EL0 RW 32-bit Performance Monitors Cycle Count Filter Register eq

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-400
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.4 AArch64 PMU register descriptions

11.4 AArch64 PMU register descriptions

This section describes the Cortex-A72 processor PMU registers in AArch64 state.
Table 11-3 PMU register summary in AArch64 state on page 11-399 provides cross-references to
individual registers.
This section contains the following subsections:
• 11.4.1 Performance Monitors Control Register, EL0 on page 11-401.
• 11.4.2 Performance Monitors Common Event Identification Register 0, EL0 on page 11-403.

11.4.1 Performance Monitors Control Register, EL0

The PMCR_EL0 characteristics are:
Purpose Provides information on the Performance Monitors implementation, including the
number of counters implemented, and configures and controls the counters.
Usage The accessibility of the PMCR_EL0 by Exception level is:
constraints
EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

Config RW RW RW RW RW

The external accessibility to the PMCR_EL0 by condition code is:

Off DLK OSLK EPMAD SLK Default

Error Error Error Error RO/WI RW

Table 11-1 External register access conditions on page 11-398 describes the access
conditions.
Configurations The PMCR_EL0 is Common to Secure and Non-secure states and architecturally
mapped to:
• The AArch32 PMCR register.
• The external PMCR_EL0 register.
Attributes See the register summary in Table 11-3 PMU register summary in AArch64 state
on page 11-399.

The following figure shows the PMCR_EL0 bit assignments for a System register access.
31 24 23 16 15 11 10 7 6 5 4 3 2 1 0

IMP IDCODE N RES0 LC DP X D C P E

Figure 11-2 PMCR_EL0 bit assignments

The following table shows the PMCR_EL0 bit assignments for a System register access.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-401
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.4 AArch64 PMU register descriptions

Table 11-4 PMCR_EL0 bit assignments

Bits Name Function

[31:24] IMP Implementer code:

0x41 ARM.

This is a read-only field.

[23:16] IDCODE Identification code:

0x02 Cortex-A72 processor.

This is a read-only field.

[15:11] N Number of event counters.

In Non-secure modes other than Hyp mode, this field reads the value of HDCR.HPMN. See 4.5.12 Hyp Debug
Control Register on page 4-261.
In Secure state and Hyp mode, this field returns 0x6 that indicates the number of counters implemented.
This is a read-only field.

[10:7] - Reserved, RES0.

[6] LC Long cycle count enable. Selects which PMCCNTR_EL0 bit generates an overflow recorded in PMOVSR[31]:
0 Overflow on increment that changes PMCCNTR_EL0[31] from 1 to 0.
1 Overflow on increment that changes PMCCNTR_EL0[63] from 1 to 0.

[5] DP Disable cycle counter, PMCCNTR_EL0 when event counting is prohibited:

0 Cycle counter operates regardless of the non-invasive debug authentication settings.
1 Cycle counter is disabled if non-invasive debug is not permitted and enabled.

This bit is read/write.

[4] X Export enable. This bit permits events to be exported to another debug device, such as a trace macrocell, over an
event bus:
0 Export of events is disabled.
1 Export of events is enabled.

This bit is read/write and does not affect the generation of Performance Monitors interrupts, that can be
implemented as a signal exported from the processor to an interrupt controller.

[3] D Clock divider:

0 When enabled, PMCCNTR_EL0 counts every clock cycle.
1 When enabled, PMCCNTR_EL0 counts every 64 clock cycles.

This bit is read/write.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-402
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.4 AArch64 PMU register descriptions

Table 11-4 PMCR_EL0 bit assignments (continued)

Bits Name Function

[2] C Clock counter reset:

0 No action.
1 Reset PMCCNTR_EL0 to 0.

Note
Resetting PMCCNTR does not clear the PMCCNTR_EL0 overflow bit to 0. See the ARM® Architecture Reference
Manual ARMv8 for more information.

This bit is write-only, and always RAZ.

[1] P Event counter reset:

0 No action.
1 Reset all event counters, not including PMCCNTR_EL0, to 0.

In Non-secure modes other than Hyp mode, a write of 1 to this bit does not reset event counters that the
HDCR.HPMN field reserves for Hyp mode use. See 4.5.12 Hyp Debug Control Register on page 4-261.
In Secure state and Hyp mode, a write of 1 to this bit resets all the event counters.

[0] E Enable bit. This bit does not disable or enable, counting by event counters reserved for Hyp mode by
HDCR.HPMN. It also does not suppress the generation of performance monitor overflow interrupt requests by
those counters:
0 All counters, including PMCCNTR_EL0, are disabled. This is the reset value.
1 All counters are enabled.

This bit is read/write.

To access the PMCR_EL0 in AArch64 state, read or write the register with:
MRS <Xt>, PMCR_EL0; Read Performance Monitors Control Register
MSR PMCR_EL0, <Xt>; Write Performance Monitors Control Register

To access the PMCR in AArch32 state, read or write the CP15 registers with:
MRC p15, 0, <Rt>, c9, c12, 0; Read Performance Monitors Control Register
MCR p15, 0, <Rt>, c9, c12, 0; Write Performance Monitors Control Register

See 11.7.1 Performance Monitors Control Register, EL0 on page 11-411 for information about accessing
the PMCR_EL0 through the internal memory-mapped interface and the external debug interface.

11.4.2 Performance Monitors Common Event Identification Register 0, EL0

The PMCEID0_EL0 characteristics are:

Purpose
Defines which common architectural and common micro-architectural feature events are
implemented.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-403
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.4 AArch64 PMU register descriptions

Usage constraints
The accessibility to the PMCEID0_EL0 by Exception level is:

EL0 (NS) EL0 (S) EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

Config Config RO RO RO RO RO

The external accessibility to the PMCEID0_EL0 by condition code is:

Off DLK OSLK EPMAD SLK Default

Error Error Error Error RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
The PMCEID0_EL0 is Common to Secure and Non-secure states and architecturally mapped to:
• The AArch32 PMCEID0 register.
• The external PMCEID0_EL0 register.
Attributes
See the register summary in Table 11-3 PMU register summary in AArch64 state
on page 11-399.
The following figure shows the PMCEID0_EL0 bit assignments
31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

BC IS BA BP UL BR BI ER ET IA ST LD SI

RES0 TW ME BM CW IC1R
CH DC2W CC PW IT1R
DC2R MA DC1R
DC2A IC1A DC1A
DC1W DT1R

Figure 11-3 PMCEID0_EL0 bit assignments

The following table shows the PMCEID0_EL0 bit assignments with event implemented or not
implemented when the associated bit is set to 1 or 0.
PMCEID1_EL0[31:0] is reserved.

Table 11-5 Common Event Identification Register 0 bit assignments

Bit Name Event number Value Event implemented if bit set to 1 or not implemented if bit set to 0

[31] - 0x1F 0 Reserved, RES0.

[30] CH 0x1E 1 Chain.es An odd-numbered counter increments when an overflow occurs on the preceding
even-numbered counter. For even-numbered counters, does not count.

[29] BC 0x1D 1 Bus cycle.

[28] TW 0x1C 1 TTBR write, architecturally executed, condition check pass - write to translation table base.

[27] IS 0x1B 1 Instruction speculatively executed.

[26] ME 0x1A 1 Local memory error.

[25] BA 0x19 1 Bus access.

[24] DC2W 0x18 1 Level 2 data cache Write-Back.

es See the ARM® Architecture Reference Manual ARMv8 for more information about the chain event.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-404
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.4 AArch64 PMU register descriptions

Table 11-5 Common Event Identification Register 0 bit assignments (continued)

Bit Name Event number Value Event implemented if bit set to 1 or not implemented if bit set to 0

[23] DC2R 0x17 1 Level 2 data cache refill.

[22] DC2A 0x16 1 Level 2 data cache access.

[21] DC1W 0x15 1 Level 1 data cache Write-Back.

[20] IC1A 0x14 1 Level 1 instruction cache access.

[19] MA 0x13 1 Data memory access.

[18] BP 0x12 1 Predictable branch speculatively executed.

[17] CC 0x11 1 Cycle.

[16] BM 0x10 1 Mispredicted or not predicted branch speculatively executed.

[15] UL 0x0F 0 Instruction architecturally executed, condition check pass - unaligned load or store.

[14] BR 0x0E 0 Instruction architecturally executed, condition check pass - procedure return.

[13] BI 0x0D 0 Instruction architecturally executed - immediate branch.

[12] PW 0x0C 0 Instruction architecturally executed, condition check pass - software change of the PC.

[11] CW 0x0B 1 Instruction architecturally executed, condition check pass - write to CONTEXTIDR.

[10] ER 0x0A 1 Instruction architecturally executed, condition check pass - exception return.

[9] ET 0x09 1 Exception taken.

[8] IA 0x08 1 Instruction architecturally executed.

[7] ST 0x07 0 Instruction architecturally executed, condition check pass - store.

[6] LD 0x06 0 Instruction architecturally executed, condition check pass - load.

[5] DT1R 0x05 1 Level 1 data TLB refill.This event is implemented.

[4] DC1A 0x04 1 Level 1 data cache access.

[3] DC1R 0x03 1 Level 1 data cache refill.

[2] IT1R 0x02 1 Level 1 instruction TLB refill.

[1] IC1R 0x01 1 Level 1 instruction cache refill.

[0] SI 0x00 1 Instruction architecturally executed, condition check pass - software increment.

To access the PMCEID0_EL0 in AArch64 state, read or write the register with:
MRS <Xt>, PMCEID0_EL0; Read Performance Monitors Common Event Identification Register 0

To access the PMCEID0 in AArch32 state, read or write the CP15 register with:
MRC p15, 0, <Rt>, c9, c12, 6; Read Performance Monitors Common Event Identification Register
0

The PMCEID0_EL0 can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0xE20.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-405
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.5 AArch32 PMU register summary

11.5 AArch32 PMU register summary

The PMU counters and their associated control registers are accessible in AArch32 state from the System
registers with MCR and MRC instructions for 32-bit registers and MCRR and MRRC for 64-bit registers.
The following table gives a summary of the PMU registers in AArch32 state.
The table also shows the offset address for the AArch32 registers that are accessible from the internal
memory-mapped interface or the external debug interface.
See 11.7 Memory-mapped register descriptions on page 11-411 for a complete list of registers that are
accessible from the internal memory-mapped interface or the external debug interface.

Table 11-6 PMU register summary in AArch32 state

Offset CRn op1 CRm op2 Name Type Width Description

0xE04 c9 0 c12 0 PMCR RW 32-bit 11.4.1 Performance Monitors Control Register, EL0
on page 11-401
0xC00 1 PMCNTENSET RW 32-bit Performance Monitors Count Enable Set Register 
0xC20 2 PMCNTENCLR RW 32-bit Performance Monitors Count Enable Clear Register et
0xC80 3 PMOVSR RW 32-bit Performance Monitors Overflow Flag Status Register et
0xCA0 4 PMSWINC WO 32-bit Performance Monitors Software Increment Register et

- 5 PMSELR RW 32-bit Performance Monitors Event Counter Selection Register et

0xE20 6 PMCEID0 RO 32-bit 11.4.2 Performance Monitors Common Event Identification
Register 0, EL0 on page 11-403
0xE24 7 PMCEID1 RO 32-bit Performance Monitors Common Event Identification
Register 1 et
0x0F8 c9 0 c13 0 PMCCNTR[31:0] RW 32-bit Performance Monitors Cycle Count Register et
0x0FC - - - - PMCCNTR[63:32]

- - 0 c9 - PMCCNTR[63:0] 64-bit

- c9 0 c13 1 PMXEVTYPER RW 32-bit Performance Monitors Selected Event Type Register et

0x47C PMCCFILTR RW 32-bit Performance Monitors Cycle Count Filter Register et

- c9 0 c13 2 PMXEVCNTR RW 32-bit Performance Monitors Selected Event Count Register et

- c14 0 PMUSERENR RW 32-bit Performance Monitors User Enable Register et

0xC40 1 PMINTENSET RW 32-bit Performance Monitors Interrupt Enable Set Register et
0xC60 2 PMINTENCLR RW 32-bit Performance Monitors Interrupt Enable Clear Register et
0xCC0 3 PMOVSSET RW 32-bit Performance Monitors Overflow Flag Status Set Register et
0x000 c14 0 c8 0 PMEVCNTR0 RW 32-bit Performance Monitors Event Count Registers et
0x008 1 PMEVCNTR1
0x010 2 PMEVCNTR2
0x018 3 PMEVCNTR3
0x020 4 PMEVCNTR4
0x028 5 PMEVCNTR5

et See the ARM® Architecture Reference Manual ARMv8 for more information.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-406
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.5 AArch32 PMU register summary

Table 11-6 PMU register summary in AArch32 state (continued)

Offset CRn op1 CRm op2 Name Type Width Description

0x400 c12 0 PMEVTYPER0 RW 32-bit Performance Monitors Event Type Registers et
0x404 1 PMEVTYPER1
0x408 2 PMEVTYPER2
0x40C 3 PMEVTYPER3
0x410 4 PMEVTYPER4
0x414 5 PMEVTYPER5
0x47C c15 7 PMCCFILTR RW 32-bit Performance Monitors Cycle Count Filter Register et

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-407
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.6 Memory-mapped register summary

11.6 Memory-mapped register summary

The following table shows the PMU registers that are accessible through the internal memory-mapped
interface and the external debug interface.

Table 11-7 Memory-mapped PMU register summary

Offset Name Type Width Description

0x000 PMEVCNTR0_EL0 RW 32-bit Performance Monitors Event Count Register 0 
0x004 - - - Reserved
0x008 PMEVCNTR1_EL0 RW 32-bit Performance Monitors Event Count Register 1 eu
0x00C - - - Reserved
0x010 PMEVCNTR2_EL0 RW 32-bit Performance Monitors Event Count Register 2 eu
0x014 - - - Reserved
0x018 PMEVCNTR3_EL0 RW 32-bit Performance Monitors Event Count Register 3 eu
0x01C - - - Reserved
0x020 PMEVCNTR4_EL0 RW 32-bit Performance Monitors Event Count Register 4 eu
0x024 - - - Reserved
0x028 PMEVCNTR5_EL0 RW 32-bit Performance Monitors Event Count Register 5 eu

0x02C-0x0F4 - - - Reserved

0x0F8 PMCCNTR_EL0[31:0] RW 32-bit Performance Monitors Cycle Count Register eu

0x0FC PMCCNTR_EL0[63:32] RW 32-bit

0x100-0x3FC - - Reserved

0x400 PMEVTYPER0_EL0 RW 32-bit Performance Monitors Event Type Register eu

0x404 PMEVTYPER1_EL0
0x408 PMEVTYPER2_EL0
0x40C PMEVTYPER3_EL0
0x410 PMEVTYPER4_EL0
0x414 PMEVTYPER5_EL0

0x418-0x478 - - - Reserved

0x47C PMCCFILTR_EL0 RW 32-bit Performance Monitors Cycle Count Filter Register eu

0x480-0x5FC - - - Reserved

0x600 PMPCSR[31:0] RO 32-bit 11.7.2 Performance Monitors Program Counter Sample Register
on page 11-412
0x604 PMPCSR[63:32]
0x608 PMCIDSR RO 32-bit 11.7.3 Performance Monitors Context ID Sample Register on page 11-412
0x60C PMVIDSR RO 32-bit 11.7.4 Performance Monitors Virtual Context Sample Register
on page 11-413
0x610 PMSSR RO 32-bit 11.7.5 Performance Monitors Snapshot Status Register on page 11-413

eu See the ARM® Architecture Reference Manual ARMv8 for more information.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-408
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.6 Memory-mapped register summary

Table 11-7 Memory-mapped PMU register summary (continued)

Offset Name Type Width Description

0x614 PMOVSSR RO 32-bit 11.7.6 Performance Monitors Overflow Status Snapshot Register
on page 11-414
0x618 PMCCNTSR[31:0] RO 32-bit 11.7.7 Performance Monitors Cycle Counter Snapshot Register
on page 11-415
0x61C PMCCNTSR[63:32] RO 32-bit
0x620 PMEVCNTSR0 RO 32-bit 11.7.8 Performance Monitors Event Counters Snapshot Registers
on page 11-415
0x624 PMEVCNTSR1
0x628 PMEVCNTSR2
0x62C PMEVCNTSR3
0x630 PMEVCNTSR4
0x634 PMEVCNTSR5

0x638-0x6EC - - - Reserved

0x6F0 PMSCR WO 32-bit 11.7.9 Performance Monitors Snapshot Control Register on page 11-416
0x6F4 PMSRR RW 32-bit 11.7.10 Performance Monitors Snapshot Reset Register on page 11-417

0x6F8-0xBFC - - - Reserved

0xC00 PMCNTENSET_EL0 RW 32-bit Performance Monitors Count Enable Set Register eu

0xC04-0xC1C - - - Reserved

0xC20 PMCNTENCLR_EL0 RW 32-bit Performance Monitors Count Enable Clear Register eu

0xC24-0xC3C - - - Reserved

0xC40 PMINTENSET_EL1 RW 32-bit Performance Monitors Interrupt Enable Set Register eu

0xC44-0xC5C - - - Reserved

0xC60 PMINTENCLR_EL1 RW 32-bit Performance Monitors Interrupt Enable Clear Register eu

0xC64-0xC7C - - - Reserved

0xC80 PMOVSCLR_EL0 RW 32-bit Performance Monitors Overflow Flag Status Register eu

0xC84-0xC9C - - - Reserved

0xCA0 PMSWINC_EL0 WO 32-bit Performance Monitors Software Increment Register eu

0xCA4-0xCBC - - - Reserved

0xCC0 PMOVSSET_EL0 RW 32-bit Performance Monitors Overflow Flag Status Set Register eu

0xCC4-0xDFC - - - Reserved

0xE00 PMCFGR RO 32-bit 11.7.11 Performance Monitors Configuration Register on page 11-418
0xE04 PMCR_EL0 RW 32-bit 11.7.1 Performance Monitors Control Register, EL0 on page 11-411

0xE08-0xE1C - - - Reserved

0xE20 PMCEID0_EL0 RO 32-bit 11.4.2 Performance Monitors Common Event Identification Register 0, EL0
on page 11-403
0xE24 PMCEID1_EL0 RO 32-bit Performance Monitors Common Event Identification Register 1 eu

0xE28-0xFA4 - - - Reserved

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-409
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.6 Memory-mapped register summary

Table 11-7 Memory-mapped PMU register summary (continued)

Offset Name Type Width Description

0xFA8 PMDEVAFF0 RO 32-bit Performance Monitors Device Affinity Register 0, see 4.3.2 Multiprocessor
Affinity Register, EL1 on page 4-90
0xFAC PMDEVAFF1 RO 32-bit Performance Monitors Device Affinity Register 1, RES0.
0xFB0 PMLAR WO 32-bit Performance Monitors Lock Access Register eu
0xFB4 PMLSR RO 32-bit Performance Monitors Lock Status Register eu
0xFB8 PMAUTHSTATUS RO 32-bit Performance Monitors Authentication Status Register eu
0xFBC PMDEVARCH 32-bit Performance Monitors Device Architecture Register eu

0xFC0-0xFC8 - - - Reserved

0xFCC PMDEVTYPE RO 32-bit Performance Monitors Device Type Register eu

0xFD0 PMPIDR4 RO 32-bit Performance Monitors Peripheral Identification Register 4 on page 11-423
0xFD4 PMPIDR5 RO 32-bit Performance Monitors Peripheral Identification Register 5-7
on page 11-424
0xFD8 PMPIDR6
0xFDC PMPIDR7
0xFE0 PMPIDR0 RO 32-bit Performance Monitors Peripheral Identification Register 0 on page 11-420
0xFE4 PMPIDR1 RO 32-bit Performance Monitors Peripheral Identification Register 1 on page 11-420
0xFE8 PMPIDR2 RO 32-bit Performance Monitors Peripheral Identification Register 2 on page 11-421
0xFEC PMPIDR3 RO 32-bit Performance Monitors Peripheral Identification Register 3 on page 11-422
0xFF0 PMCIDR0 RO 32-bit Performance Monitors Component Identification Register 0
on page 11-424
0xFF4 PMCIDR1 RO 32-bit Performance Monitors Component Identification Register 1
on page 11-425
0xFF8 PMCIDR2 RO 32-bit Performance Monitors Component Identification Register 2
on page 11-426
0xFFC PMCIDR3 RO 32-bit Performance Monitors Component Identification Register 3
on page 11-426

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-410
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

11.7 Memory-mapped register descriptions

This section describes the Cortex-A72 processor PMU registers accessible through the memory-mapped
and debug interfaces.
11.6 Memory-mapped register summary on page 11-408 provides cross-references to individual registers.
This section contains the following subsections:
• 11.7.1 Performance Monitors Control Register, EL0 on page 11-411.
• 11.7.2 Performance Monitors Program Counter Sample Register on page 11-412.
• 11.7.3 Performance Monitors Context ID Sample Register on page 11-412.
• 11.7.4 Performance Monitors Virtual Context Sample Register on page 11-413.
• 11.7.5 Performance Monitors Snapshot Status Register on page 11-413.
• 11.7.6 Performance Monitors Overflow Status Snapshot Register on page 11-414.
• 11.7.7 Performance Monitors Cycle Counter Snapshot Register on page 11-415.
• 11.7.8 Performance Monitors Event Counters Snapshot Registers on page 11-415.
• 11.7.9 Performance Monitors Snapshot Control Register on page 11-416.
• 11.7.10 Performance Monitors Snapshot Reset Register on page 11-417.
• 11.7.11 Performance Monitors Configuration Register on page 11-418.
• 11.7.12 Performance Monitors Peripheral Identification Registers on page 11-419.
• 11.7.13 Performance Monitors Component Identification Registers on page 11-424.

11.7.1 Performance Monitors Control Register, EL0

The PMCR_EL0 characteristics are:

Purpose
Configures and controls the counters.
Usage constraints
The external accessibility to the PMCR_EL0 by condition code is:

Off DLK OSLK EPMAD SLK Default

Error Error Error Error RO/WI RW

Table 11-1 External register access conditions on page 11-398 describes the access conditions.
Configurations
The PMCR_EL0 is Common to Secure and Non-secure states and architecturally mapped to:
• The AArch32 PMCR register.
• The external PMCR_EL0 register.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
The following figure shows the PMCR_EL0 bit assignments for a memory-mapped access.
31 7 6 5 4 3 2 1 0

RES0 LC DP X D C P E

Figure 11-4 PMCR_EL0 bit assignments, memory-mapped view

The following table shows the PMCR_EL0 bit assignments for a memory-mapped access.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-411
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

Table 11-8 PMCR_EL0 bit assignments, memory-mapped view

Bits Name Function

[31:7] - Reserved, RES0.

[6] LC The function of these bits is the same as when a System register access occurs. See Table 11-4 PMCR_EL0 bit
assignments on page 11-402 for a description of these bits.
[5] DP

[4] X

[3] D

[2] C

[1] P

[0] E

The PMCR_EL0 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xE04.

11.7.2 Performance Monitors Program Counter Sample Register

The PMPCSR characteristics are:

Purpose
The PMPCSR registers are aliases of the EDPCSR debug registers. Reads of the PMPCSR
registers return a copy of the EDPCSR debug registers but does not:
• Cause a new EDPCSR capture.
• Change the EDCIDSR and EDVIDSR registers.
Usage constraints
The external accessibility to the PMPCSR by condition code is:

Off DLK OSLK EPMAD SLK Default

Error Error RO RO RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
PMPCSR[31:0] copies the EDPCSRlo debug register.
PMPCSR[63:32] copies the EDPCSRhi debug register.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
See the ARM® Architecture Reference Manual ARMv8 for more information about the EDPCSR debug
registers.
PMPCSR[31:0] can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x600.
PMPCSR[63:32] can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x604.

11.7.3 Performance Monitors Context ID Sample Register

The PMCIDSR characteristics are:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-412
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

Purpose
The PMCIDSR register is an alias of the EDCIDSR debug register. Reads of the PMCIDSR
return a copy of the EDCIDSR debug register.
Usage constraints
The external accessibility to the PMCIDSR by condition code is:

Off DLK OSLK EPMAD SLK Default

Error Error RO RO RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
There is no configuration information for PMCIDSR.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
See the ARM® Architecture Reference Manual ARMv8 for more information about the EDCIDSR debug
register.
PMCIDSR can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x608.

11.7.4 Performance Monitors Virtual Context Sample Register

The PMVIDSR characteristics are:

Purpose
The PMVIDSR register is an alias of the EDVIDSR debug register. Reads of the PMVIDSR
return a copy of the EDVIDSR debug register.
Usage constraints
The external accessibility to the PMVIDSR by condition code is:

Off DLK OSLK EPMAD SLK Default

Error Error RO RO RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
There is no configuration information for PMVIDSR.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
See the ARM® Architecture Reference Manual ARMv8 for more information about the EDVIDSR debug
register.
PMVIDSR can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x60C.

11.7.5 Performance Monitors Snapshot Status Register

The PMSSR characteristics are:

Purpose
Provides status information on whether the PMU counters have been captured.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-413
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

Usage constraints
The external accessibility to the PMSSR by condition code is:

Off DLK OSLK EPMAD SLK Default

Error Error RO RO RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
A security violation prevents the capture of the event counters.
The external monitor must keep track of whether the snapshot registers were captured by the
processor.
To prevent loss of data, software must save and restore the PMU state, including the PMSCR
and PMSRR registers, when capturing over a reset or power down.
Configurations
There is no configuration information for PMSSR.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
The following figure shows the PMSSR bit assignments.
31 1 0

Reserved

NC

Figure 11-5 PMSSR bit assignments

The following table shows the PMSSR bit assignments.

Table 11-9 PMSSR bit assignments

Bits Name Function

[31:1] - Reserved, RES0.

[0] NC No capture. The possible values are:

0 PMU counters captured.
1 PMU counters not captured.

The NC bit:
• Is reset to 1 by a Warm reset but overwritten at the first capture.
• Does not reflect the status of the captured Program Counter Sample registers.

PMSSR can be accessed through the internal memory-mapped interface and the external debug interface,
offset 0x610.

11.7.6 Performance Monitors Overflow Status Snapshot Register

The PMOVSSR characteristics are:

Purpose
Captures a copy of the PMOVSR register. After capture, writes to PMOVSSET_EL0 and
PMOVSCLR_EL0 do not affect the PMOVSSR value.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-414
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

Usage constraints
The external accessibility to the PMOVSSR by condition code is:

Off DLK OSLK EPMAD SLK Default

Error Error RO RO RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
There is no configuration information for PMOVSSR.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
See the ARM® Architecture Reference Manual ARMv8 for more information about the PMOVSR register.
PMOVSSR can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x614.

11.7.7 Performance Monitors Cycle Counter Snapshot Register

The PMCCNTSR characteristics are:

Purpose
Captures a copy of the PMCCNTR_EL0 register. After capture, writes to PMCCNTR_EL0 and
PMCR_EL0.C do not affect the PMCCNTSR value.
Usage constraints
The external accessibility to the PMCCNTSR by condition code is:

Off DLK OSLK EPMAD SLK Default

Error Error RO RO RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
There is no configuration information for PMCCNTSR.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
See the ARM® Architecture Reference Manual ARMv8 for more information about the PMCCNTR_EL0
register.
PMCCNTSR[31:0] can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0x618.
PMCCNTSR[63:32] can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0x61C.

11.7.8 Performance Monitors Event Counters Snapshot Registers

The PMEVCNTSRn characteristics are:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-415
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

Purpose
Captures a copies of the PMEVCNTRn_EL0 registers. After capture, writes to
PMEVCNTRn_EL0 and PMCR_EL0.P do not affect the PMEVCNTSRn value.
Note
The range of n for PMEVCNTSRn is 0 to 5.

Usage constraints
The external accessibility to the PMEVCNTSRn by condition code is:

Off DLK OSLK EPMAD SLK Default

Error Error RO RO RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
There is no configuration information for PMEVCNTSRn.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
See the ARM® Architecture Reference Manual ARMv8 for more information about the
PMEVCNTRn_EL0 registers.
The PMEVCNTRn_EL0 registers can be accessed through the internal memory-mapped interface and
the external debug interface, offsets 0x620-0x634.

11.7.9 Performance Monitors Snapshot Control Register

The PMSCR characteristics are:

Purpose
Initiates an immediate capture.
Usage constraints
The external accessibility to the PMSCR by condition code is:

Off DLK OSLK EPMAD SLK Default

Error Error WO WO WO WO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
There is no configuration information for PMSCR.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
The following figure shows the PMSCR bit assignments.
31 1 0

Reserved

SS

Figure 11-6 PMSCR bit assignments

The following table shows the PMSCR bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-416
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

Table 11-10 PMSCR bit assignments

Bits Name Function

[31:1] - Reserved, RES0.

[0] SS Capture now. The possible values are:

0 Capture ignored.
1 Initiate a capture immediately.

The PMSCR can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x6F0.

11.7.10 Performance Monitors Snapshot Reset Register

The PMSRR characteristics are:

Purpose Reset the cycle counter and the performance counters.
Usage constraints The external accessibility to the PMSRR by condition code is:

Off DLK OSLK EPMAD SLK Default

Error Error RW RW RW RW

Table 11-1 External register access conditions on page 11-398 describes the
condition codes.
Configurations There is no configuration information for PMSRR.
Attributes See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.

The following figure shows the PMSRR bit assignments.

31 30 6 5 4 3 2 1 0

Reserved

RC RP5
RP4
RP3
RP2
RP1
RP0

Figure 11-7 PMSRR bit assignments

The following table shows the PMSRR bit assignments.

Table 11-11 PMSRR bit assignments

Bits Name Function

[31] RC Reset cycle counter. Indicates whether the PMCCNTR_EL0 and PMOVSR[31] are reset after a capture:
0 PMCCNTR_EL0 and PMOVSR[31] are not reset on capture.
1 PMCCNTR_EL0 and PMOVSR[31] are reset on capture.

[30:6] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-417
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

Table 11-11 PMSRR bit assignments (continued)

Bits Name Function

[5] RP5 Reset performance counter 5. Indicates whether PMEVCNTR5_EL0 and PMOVSR[5] are reset after a capture:
0 PMEVCNTR5_EL0 and PMOVSR[5] are not reset on capture.
1 PMEVCNTR5_EL0 and PMOVSR[5] are reset on capture.

[4] RP4 Reset performance counter 4. Indicates whether PMEVCNTR4_EL0 and PMOVSR[4] are reset after a capture:
0 PMEVCNTR4_EL0 and PMOVSR[4] are not reset on capture.
1 PMEVCNTR4_EL0 and PMOVSR[4] are reset on capture.

[3] RP3 Reset performance counter 3. Indicates whether PMEVCNTR3_EL0 and PMOVSR[3] are reset after a capture:
0 PMEVCNTR3_EL0 and PMOVSR[3] are not reset on capture.
1 PMEVCNTR3_EL0 and PMOVSR[3] are reset on capture.

[2] RP2 Reset performance counter 2. Indicates whether PMEVCNTR2_EL0 and PMOVSR[2] are reset after a capture:
0 PMEVCNTR2_EL0 and PMOVSR[2] are not reset on capture.
1 PMEVCNTR2_EL0 and PMOVSR[2] are reset on capture.

[1] RP1 Reset performance counter 1. Indicates whether PMEVCNTR1_EL0 and PMOVSR[1] are reset after a capture:
0 PMEVCNTR1_EL0 and PMOVSR[1] are not reset on capture.
1 PMEVCNTR1_EL0 and PMOVSR[1] are reset on capture.

[0] RP0 Reset performance counter 0. Indicates whether PMEVCNTR0_EL0 and PMOVSR[0] are reset after a capture:
0 PMEVCNTR0_EL0 and PMOVSR[0] are not reset on capture.
1 PMEVCNTR0_EL0 and PMOVSR[0] are reset on capture.

The PMSRR can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x6F4.

11.7.11 Performance Monitors Configuration Register

The PMCFGR characteristics are:

Purpose
Contains PMU specific configuration data.
Usage constraints
The accessibility to the PMCFGR by condition code is:

Off DLK OSLK EPMAD SLK Default

Error Error Error Error RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
The PMCFGR is in the Core power domain.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
The following figure shows the PMCFGR bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-418
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

31 17 16 15 14 13 8 7 0

RES0 Size N

EX CC
CCD

Figure 11-8 PMCFGR bit assignments

The following table shows the PMCFGR bit assignments.

Table 11-12 PMCFGR bit assignments

Bits Name Function

[31:17] - Reserved, RES0.

[16] EX Export supported. The value is:

1 Export is supported. PMCR_EL0.EX is read/write.

[15] CCD Cycle counter has pre-scale. The value is:

1 PMCR_EL0.D is read/write.

[14] CC Dedicated cycle counter supported. The value is:

1 Dedicated cycle counter is supported.

[13:8] Size Counter size. The value is:

0b111111 64-bit counters.

[7:0] N Number of event counters. The value is:

0x06 Six counters.

The PMCFGR can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xE00.

11.7.12 Performance Monitors Peripheral Identification Registers

The Performance Monitors Peripheral Identification Registers provide standard information required for
all components that conform to the ARM PMUv3 architecture. There is a set of eight registers, listed in
register number order in the following table.

Table 11-13 Summary of the Performance Monitors Peripheral Identification Registers

Register Value Offset

PMPIDR4 0x04 0xFD0

PMPIDR5 0x00 0xFD4

PMPIDR6 0x00 0xFD8

PMPIDR7 0x00 0xFDC

PMPIDR0 0xD8 0xFE0

PMPIDR1 0xB9 0xFE4

PMPIDR2 0x0B 0xFE8

PMPIDR3 0x00 0xFEC

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-419
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

Only bits[7:0] of each PMU Peripheral ID Register are used, with bits[31:8] reserved. Together, the eight
PMU Peripheral ID Registers define a single 64-bit Peripheral ID.
The PMU Peripheral ID registers are:
• Performance Monitors Peripheral Identification Register 0 on page 11-420.
• Performance Monitors Peripheral Identification Register 1 on page 11-420.
• Performance Monitors Peripheral Identification Register 2 on page 11-421.
• Performance Monitors Peripheral Identification Register 3 on page 11-422.
• Performance Monitors Peripheral Identification Register 4 on page 11-423.
• Performance Monitors Peripheral Identification Register 5-7 on page 11-424.

Performance Monitors Peripheral Identification Register 0

The PMPIDR0 characteristics are:
Purpose
Provides information to identify a Performance Monitors component.
Usage constraints
The PMPIDR0 can be accessed through the internal memory-mapped interface and the external
debug interface.
The accessibility to the PMPIDR0 by condition code is:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
The PMPIDR0 is in the Debug power domain.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
The following figure shows the PMPIDR0 bit assignments.
31 8 7 0

Reserved Part_0

Figure 11-9 PMPIDR0 bit assignments

The following table shows the PMPIDR0 bit assignments.

Table 11-14 PMPIDR0 bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:0] Part_0 0xD8 Least significant byte of the performance monitor part number

The PMPIDR0 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFE0.

Performance Monitors Peripheral Identification Register 1

The PMPIDR1 characteristics are:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-420
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

Purpose
Provides information to identify a Performance Monitors component.
Usage constraints
The PMPIDR1 can be accessed through the internal memory-mapped interface and the external
debug interface.
The accessibility to the PMPIDR1 by condition code is:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
The PMPIDR1 is in the Debug power domain.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
The following figure shows the PMPIDR1 bit assignments.
31 8 7 4 3 0

Reserved DES_0 Part_1

Figure 11-10 PMPIDR1 bit assignments

The following table shows the PMPIDR1 bit assignments.

Table 11-15 PMPIDR1 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] DES_0 0xB ARM Limited. This is the least significant nibble of JEP106 ID code.

[3:0] Part_1 0x9 Most significant nibble of the performance monitor part number.

The PMPIDR1 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFE4.

Performance Monitors Peripheral Identification Register 2

The PMPIDR2 characteristics are:
Purpose
Provides information to identify a Performance Monitors component.
Usage constraints
The accessibility to the PMPIDR2 by condition code is:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
The PMPIDR2 can be accessed through the internal memory-mapped interface and the external
debug interface.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-421
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

Configurations
The PMPIDR2 is in the Debug power domain.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
The following figure shows the PMPIDR2 bit assignments.
31 8 7 4 3 2 0

Reserved Revision DES_1

JEDEC

Figure 11-11 PMPIDR2 bit assignments

The following table shows the PMPIDR2 bit assignments.

Table 11-16 PMPIDR2 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] Revision 0x0 Part major revision.

[3] JEDEC 0b1 RAO. Indicates a JEP106 identity code is used.

[2:0] DES_1 0b011 ARM Limited. This is the most significant nibble of JEP106 ID code.

The PMPIDR2 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFE8.

Performance Monitors Peripheral Identification Register 3

The PMPIDR3 characteristics are:

Purpose
Provides information to identify a Performance Monitors component.
Usage constraints
The PMPIDR3 can be accessed through the internal memory-mapped interface and the external
debug interface.
The accessibility to the PMPIDR3 by condition code is:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
The PMPIDR3 is in the Debug power domain.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
The following figure shows the PMPIDR3 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-422
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

31 8 7 4 3 0

Reserved REVAND CMOD

Figure 11-12 PMPIDR3 bit assignments

The following table shows the PMPIDR3 bit assignments.

Table 11-17 PMPIDR3 bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:4] REVAND 0x0 Part minor revision

[3:0] CMOD 0x0 Customer modified

The PMPIDR3 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFEC.

Performance Monitors Peripheral Identification Register 4

The PMPIDR4 characteristics are:
Purpose
Provides information to identify a Performance Monitors component.
Usage constraints
PMPIDR4 can be accessed through the internal memory-mapped interface and the external
debug interface.
The accessibility to the PMPIDR4 by condition code is:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
The PMPIDR4 is in the Debug power domain.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
The following figure shows the PMPIDR4 bit assignments.
31 8 7 4 3 0

Reserved Size DES_2

Figure 11-13 PMPIDR4 bit assignments

The following table shows the PMPIDR4 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-423
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

Table 11-18 PMPIDR4 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] Size 0x0 Size of the component. Log2 the number of 4KB pages from the start of the component to the end of the
component ID registers.

[3:0] DES_2 0x4 ARM Limited. This is the least significant nibble JEP106 continuation code.

The PMPIDR4 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFD0.

Performance Monitors Peripheral Identification Register 5-7

No information is held in the Performance Monitors Peripheral ID5, Performance Monitors Peripheral
ID6, and Performance Monitors Peripheral ID7 Registers. They are reserved for future use and are RES0.

11.7.13 Performance Monitors Component Identification Registers

There are four read-only Performance Monitors Component Identification Registers, Performance
Monitors Component ID0 through Performance Monitors Component ID3. The following table shows
these registers.

Table 11-19 Summary of the Performance Monitors Component Identification Registers

Register Value Offset

PMCIDR0 0x0D 0xFF0

PMCIDR1 0x90 0xFF4

PMCIDR2 0x05 0xFF8

PMCIDR3 0xB1 0xFFC

The Performance Monitors Component Identification Registers identify Performance Monitors as ARM
PMUv3 architecture. The Component ID registers are:
• Performance Monitors Component Identification Register 0 on page 11-424.
• Performance Monitors Component Identification Register 1 on page 11-425.
• Performance Monitors Component Identification Register 2 on page 11-426.
• Performance Monitors Component Identification Register 3 on page 11-426.

Performance Monitors Component Identification Register 0

The PMCIDR0 characteristics are:
Purpose
Provides information to identify a Performance Monitors component.
Usage constraints
The PMCIDR0 can be accessed through the internal memory-mapped interface and the external
debug interface.
The accessibility to the PMCIDR0 by condition code is:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-424
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

Configurations
The PMCIDR0 is in the Debug power domain.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
The following figure shows the PMCIDR0 bit assignments.
31 8 7 0

Reserved PRMBL_0

Figure 11-14 PMCIDR0 bit assignments

The following table shows the PMCIDR0 bit assignments.

Table 11-20 PMCIDR0 bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:0] PRMBL_0 0x0D Preamble byte 0

The PMCIDR0 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFF0.

Performance Monitors Component Identification Register 1

The PMCIDR1 characteristics are:
Purpose
Provides information to identify a Performance Monitors component.
Usage constraints
The PMCIDR1 can be accessed through the internal memory-mapped interface and the external
debug interface.
The accessibility to the PMCIDR1 by condition code is:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
The PMCIDR1 is in the Debug power domain.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
The following figure shows the PMCIDR1 bit assignments.
31 8 7 4 3 0

Reserved CLASS PRMBL_1

Figure 11-15 PMCIDR1 bit assignments

The following table shows the PMCIDR1 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-425
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

Table 11-21 PMCIDR1 bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:4] CLASS 0x9 Debug component

[3:0] PRMBL_1 0x0 Preamble

The PMCIDR1 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFF4.

Performance Monitors Component Identification Register 2

The PMCIDR2 characteristics are:
Purpose
Provides information to identify a Performance Monitors component.
Usage constraints
The PMCIDR2 can be accessed through the internal memory-mapped interface and the external
debug interface.
The accessibility to the PMCIDR2 by condition code is:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
The PMCIDR2 is in the Debug power domain.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
The following figure shows the PMCIDR2 bit assignments.
31 8 7 0

Reserved PRMBL_2

Figure 11-16 PMCIDR2 bit assignments

The following table shows the PMCIDR2 bit assignments.

Table 11-22 PMCIDR2 bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:0] PRMBL_2 0x05 Preamble byte 2

The PMCIDR2 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFF8.

Performance Monitors Component Identification Register 3

The PMCIDR3 characteristics are:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-426
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.7 Memory-mapped register descriptions

Purpose
Provides information to identify a Performance Monitors component.
Usage constraints
The PMCIDR3 can be accessed through the internal memory-mapped interface and the external
debug interface.
The accessibility to the PMCIDR3 by condition code is:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 11-1 External register access conditions on page 11-398 describes the condition codes.
Configurations
The PMCIDR3 is in the Debug power domain.
Attributes
See the register summary in Table 11-7 Memory-mapped PMU register summary
on page 11-408.
The following figure shows the PMCIDR3 bit assignments.
31 8 7 0

Reserved PRMBL_3

Figure 11-17 PMCIDR3 bit assignments

The following table shows the PMCIDR3 bit assignments.

Table 11-23 PMCIDR3 bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:0] PRMBL_3 0xB1 Preamble byte 3

The PMCIDR3 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFFC.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-427
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.8 Events

11.8 Events
The following table shows the events that are generated and the numbers that the PMU uses to reference
the events.
The table also shows the bit position of each event on the event bus. Event reference numbers that are not
listed are reserved.

Table 11-24 PMU events

Event Event mnemonic PMUEVENTx[24:0] PMU event Event name

number busev bus (to
trace)ev
0x00 SW_INCR - [0] Instruction architecturally executed
(condition check pass) - Software
increment
0x01 L1I_CACHE_REFILL [0] [1] Level 1 instruction cache refill
0x02 L1I_TLB_REFILL [1] [2] Level 1 instruction TLB refill
0x03 L1D_CACHE_REFILL [2] [3] Level 1 data cache refill
0x04 L1D_CACHE - [5:4] Level 1 data cache access
0x05 L1D_TLB_REFILL - [7:6] Level 1 data TLB refill
0x08 INST_RETIRED [6:3] [11:8] Instruction architecturally executed
0x09 EXC_TAKEN [7] [12] Exception taken
0x0A EXC_RETURN [8] [13] Instruction architecturally executed
(condition check pass) - Exception
return
0x0B CID_WRITE_RETIRED - [14] Instruction architecturally executed
(condition check pass) - Write to
CONTEXTIDR
0x10 BR_MIS_PRED [9] [15] Mispredicted or not predicted branch
speculatively executed
0x11 CPU_CYCLES - [16] Cycle
0x12 BR_PRED [10] [17] Predictable branch speculatively
executed
0x13 MEM_ACCESS - [19:18] Data memory access
0x14 L1I_CACHE [11] [20] Level 1 instruction cache access
0x15 L1D_CACHE_WB [12] [21] Level 1 data cache Write-Back
0x16 L2D_CACHE - [23:22] Level 2 data cache access
0x17 L2D_CACHE_REFILL [13] [24] Level 2 data cache refill
0x18 L2D_CACHE_WB [14] [25] Level 2 data cache Write-Back
0x19 BUS_ACCESS - [27:26] Bus access
0x1A MEMORY_ERROR - [28] Local memory error
0x1B INST_SPEC - [30:29] Operation speculatively executed

ev Event count is encoded as a plain binary number to accommodate count values of more than one in the same cycle.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-428
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.8 Events

Table 11-24 PMU events (continued)

Event Event mnemonic PMUEVENTx[24:0] PMU event Event name

number busev bus (to
trace)ev
0x1C TTBR_WRITE_RETIRED - [31] Instruction architecturally executed
(condition check pass) - Write to
translation table base
0x1D BUS_CYCLES - [32] Bus cycle
0x1E CHAIN - [33] Odd performance counter chain mode
0x40 L1D_CACHE_LD [15] [34] Level 1 data cache access - Read
0x41 L1D_CACHE_ST [16] [35] Level 1 data cache access - Write
0x42 L1D_CACHE_REFILL_LD - [36] Level 1 data cache refill - Read
0x43 L1D_CACHE_REFILL_ST - [37] Level 1 data cache refill - Write
0x46 L1D_CACHE_WB_VICTIM - [38] Level 1 data cache Write-back -
Victim
0x47 L1D_CACHE_WB_CLEAN - [39] Level 1 data cache Write-back -
Cleaning and coherency
0x48 L1D_CACHE_INVAL - [40] Level 1 data cache invalidate
0x4C L1D_TLB_REFILL_LD [17] [41] Level 1 data TLB refill - Read
0x4D L1D_TLB_REFILL_ST [18] [42] Level 1 data TLB refill - Write
0x50 L2D_CACHE_LD [19] [43] Level 2 data cache access - Read
0x51 L2D_CACHE_ST [20] [44] Level 2 data cache access - Write
0x52 L2D_CACHE_REFILL_LD - [45] Level 2 data cache refill - Read
0x53 L2D_CACHE_REFILL_ST - [46] Level 2 data cache refill - Write
0x56 L2D_CACHE_WB_VICTIM - [47] Level 2 data cache Write-back -
Victim
0x57 L2D_CACHE_WB_CLEAN - [48] Level 2 data cache Write-back -
Cleaning and coherency
0x58 L2D_CACHE_INVAL - [49] Level 2 data cache invalidate
0x60 BUS_ACCESS_LD - [50] Bus access - Read
0x61 BUS_ACCESS_ST - [51] Bus access - Write
0x62 BUS_ACCESS_SHARED - [53:52] Bus access - Shared
0x63 BUS_ACCESS_NOT_SHARED - [55:54] Bus access - Not shared
0x64 BUS_ACCESS_NORMAL - [57:56] Bus access - Normal
0x65 BUS_ACCESS_PERIPH - [59:58] Bus access - Peripheral
0x66 MEM_ACCESS_LD - [60] Data memory access - Read
0x67 MEM_ACCESS_ST - [61] Data memory access - Write
0x68 UNALIGNED_LD_SPEC - [62] Unaligned access - Read

ev Event count is encoded as a plain binary number to accommodate count values of more than one in the same cycle.
ew For this event, unaligned access means data access related memory operation that crosses line boundary.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-429
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.8 Events

Table 11-24 PMU events (continued)

Event Event mnemonic PMUEVENTx[24:0] PMU event Event name

number busev bus (to
trace)ev
0x69ew UNALIGNED_ST_SPEC - [63] Unaligned access - Write
0x6Aew UNALIGNED_LDST_SPEC - [65:64] Unaligned access
0x6C LDREX_SPEC [21] [66] Exclusive operation speculatively
executed - LDREX
0x6D STREX_PASS_SPEC [22] [67] Exclusive instruction speculatively
executed - STREX pass
0x6E STREX_FAIL_SPEC [23] [68] Exclusive operation speculatively
executed - STREX fail
0x70 LD_SPEC - [70:69] Operation speculatively executed -
Load
0x71 ST_SPEC - [72:71] Operation speculatively executed -
Store
0x72 LDST_SPEC - [74:73] Operation speculatively executed -
Load or store
0x73 DP_SPEC - [76:75] Operation speculatively executed -
Integer data processing
0x74 ASE_SPEC - [78:77] Operation speculatively executed -
Advanced SIMD
0x75 VFP_SPEC - [80:79] Operation speculatively executed -
VFP
0x76 PC_WRITE_SPEC - [82:81] Operation speculatively executed -
Software change of the PC
0x77 CRYPTO_SPEC - [84:83] Operation speculatively executed,
crypto data processing
0x78 BR_IMMED_SPEC - [85] Branch speculatively executed -
Immediate branch
0x79 BR_RETURN_SPEC - [86] Branch speculatively executed -
Procedure return
0x7A BR_INDIRECT_SPEC - [87] Branch speculatively executed -
Indirect branch
0x7C ISB_SPEC - [88] Barrier speculatively executed - ISB
0x7D DSB_SPEC [24] [89] Barrier speculatively executed - DSB
0x7E DMB_SPEC [24] [90] Barrier speculatively executed - DMB
0x81 EXC_UNDEF - [91] Exception taken, other synchronous
0x82 EXC_SVC - [92] Exception taken, Supervisor Call
0x83 EXC_PABORT - [93] Exception taken, Instruction Abort

ev Event count is encoded as a plain binary number to accommodate count values of more than one in the same cycle.
ev Event count is encoded as a plain binary number to accommodate count values of more than one in the same cycle.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-430
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.8 Events

Table 11-24 PMU events (continued)

Event Event mnemonic PMUEVENTx[24:0] PMU event Event name

number busev bus (to
trace)ev
0x84 EXC_DABORT - [94] Exception taken, Data Abort or
SError
0x86 EXC_IRQ - [95] Exception taken, IRQ
0x87 EXC_FIQ - [96] Exception taken, FIQ
0x88 EXC_SMC - [97] Exception taken, Secure Monitor Call
0x8A EXC_HVC - [98] Exception taken, Hypervisor Call
0x8B EXC_TRAP_PABORT - [99] Exception taken, Instruction Abort
not taken locally
0x8C EXC_TRAP_DABORT - [100] Exception taken, Data Abort, or
SError not taken locally
0x8D EXC_TRAP_OTHER - [101] Exception taken – Other traps not
taken locally
0x8E EXC_TRAP_IRQ - [102] Exception taken, IRQ not taken
locally
0x8F EXC_TRAP_FIQ - [103] Exception taken, FIQ not taken
locally
0x90 RC_LD_SPEC - [104] Release consistency instruction
speculatively executed – Load-
Acquire
0x91 RC_ST_SPEC - [105] Release consistency instruction
speculatively executed – Store-
Release

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-431
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.9 Interrupts

11.9 Interrupts
The Cortex-A72 processor asserts the nPMUIRQ signal when an interrupt is generated by the PMU.
You can route this signal to an external interrupt controller for prioritization and masking. This is the
only mechanism that signals this interrupt to the processor.
Interrupt is also driven as a trigger input to the CTI.

Related information
Chapter 12 Cross Trigger on page 12-434.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-432
reserved.
Non-Confidential
 11 Performance Monitor Unit
11.10 Exporting PMU events

11.10 Exporting PMU events

This section describes exporting PMU events.
This section contains the following subsections:
• 11.10.1 External hardware on page 11-433.
• 11.10.2 Debug trace hardware on page 11-433.

11.10.1 External hardware

In addition to the counters in the processor, some of the events that 11.8 Events on page 11-428 describes
are exported on the PMUEVENT bus and can be connected to external hardware.

11.10.2 Debug trace hardware

Some of the events that 11.8 Events on page 11-428 describes are exported to the ETM unit, other
external debug, or trace hardware, to enable the events to be monitored.

Related information
Chapter 12 Cross Trigger on page 12-434.
Chapter 13 Embedded Trace Macrocell on page 13-457.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 11-433
reserved.
Non-Confidential
Chapter 12
Cross Trigger

This chapter describes the cross trigger interfaces for the Cortex-A72 processor.

It contains the following sections:

• 12.1 About the cross trigger on page 12-435.
• 12.2 Trigger inputs and outputs on page 12-436.
• 12.3 CTI on page 12-437.
• 12.4 CTM on page 12-438.
• 12.5 Cross trigger register summary on page 12-439.
• 12.6 Cross trigger register descriptions on page 12-442.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-434
reserved.
Non-Confidential
 12 Cross Trigger
12.1 About the cross trigger

12.1 About the cross trigger

The Cortex-A72 processor has a single external cross trigger channel interface.
This external interface is connected to the CoreSight CTI interface corresponding to each processor
through a simplified Cross Trigger Matrix (CTM). A number of Embedded Cross Trigger (ECT), trigger
inputs and trigger outputs are connected between debug components in the processor and CoreSight CTI
blocks.
The CoreSight Cross Trigger Interface (CTI) enables the debug logic, ETM, and PMU, to interact with
each other and with other CoreSight components. This is called cross triggering. For example, you
configure the CTI to generate an interrupt when the ETM trigger event occurs.
The following figure shows the debug system components and the available trigger inputs and trigger
outputs.

Cortex-A72 processor

Optional core 4
Optional core 3
Optional core 2
core 1

PMUEVENT
PMU nPMUIRQ
Non-processor
Optional CSCTI 4
PMUEVENT[24:0] Optional CSCTI 3
Optional CSCTI 2
CSCTI 1

EXTOUT[3:0]
ETM
EXTIN[3:0]
CTIIRQ
CTIIRQACK

DBGRQ
DBGTRIGGER
DBGRESTART

CTICHOUT
CTICHIN

CTICHOUT[3:0]
Debug
CTICHOUTACK[3:0]
CTM
CTICHIN[3:0]
CTICHINACK[3:0]

nCOMMIRQ
COMMRX
COMMTX
EDBGRQ

Figure 12-1 Debug system components

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-435
reserved.
Non-Confidential
 12 Cross Trigger
12.2 Trigger inputs and outputs

12.2 Trigger inputs and outputs

This section describes the trigger inputs and outputs that are available to the CTI.
The following table shows the CTI inputs.

Table 12-1 Trigger inputs

CTI input Name Description

0 DBGTRIGGER, pulsed Pulsed on entry to Debug state

1 PMUIRQex PMU generated interrupt

2 - -

3 - -

4 EXTOUT[0] ETM external output

5 EXTOUT[1] ETM external output

6 EXTOUT[2] ETM external output

7 EXTOUT[3] ETM external output

The following table shows the CTI outputs.

Table 12-2 Trigger outputs

CTI output Name Description

0 EDBGRQ Causes the processor to enter Debug state

1 DBGRESTART Causes the processor to exit Debug state

2 CTIIRQ CTI interrupt

3 - -

4 EXTIN[0] ETM external input

5 EXTIN[1] ETM external input

6 EXTIN[2] ETM external input

7 EXTIN[3] ETM external input

ex This signal is the same as nPMUIRQ with inverted polarity.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-436
reserved.
Non-Confidential
 12 Cross Trigger
12.3 CTI

12.3 CTI
In the Cortex-A72 processor, the CTI operates in the PCLKDBG domain and it synchronizes the trigger
inputs and outputs to PCLKDBG.
Handshaking is required for all trigger outputs. Because the simplified CTM is implemented in the same
clock domain, synchronization and handshaking is not required for channel interface. In addition, APB
synchronization is not required. Trigger inputs are not masked by internal NIDEN. Trigger outputs are
not masked by internal DBGEN.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-437
reserved.
Non-Confidential
 12 Cross Trigger
12.4 CTM

12.4 CTM
The CoreSight CTI channel signals from all the cores are combined using a simplified Cross Trigger
Matrix (CTM) so that a single cross trigger channel interface is presented in the Cortex-A72 processor.
The CTM can combine up to four internal channel interfaces, corresponding to each core, and one
external channel interface.
In the simplified CTM:
• The external channel output is driven by the OR output of all internal channel outputs.
• Each internal channel input is driven by the OR output of the internal channel outputs of all other
CTIs, in addition to the external channel input.
The internal channel acknowledgment signals from the CTIs are not used because the CTIs and the CTM
are in the same PCLKDBG domain.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-438
reserved.
Non-Confidential
 12 Cross Trigger
12.5 Cross trigger register summary

12.5 Cross trigger register summary

This section describes the cross trigger registers in the Cortex-A72 processor.
These registers are accessed through the internal memory-mapped interface or the external debug
interface.
The following table shows the cross trigger registers in the Cortex-A72 processor.

Table 12-3 Cross trigger register summary

Offset Name Type Width Description

0x000 CTICONTROL RW 32-bit CTI Control register 

0x000-0x00C - - - Reserved

0x010 CTIINTACK WO 32-bit CTI Output Trigger Acknowledge register ey

0x014 CTIAPPSET RW 32-bit CTI Application Trigger Set register ey
0x018 CTIAPPCLEAR WO 32-bit CTI Application Trigger Clear register ey
0x01C CTIAPPPULSE WO 32-bit CTI Application Pulse register ey
0x020 CTIINEN0 RW 32-bit CTI Input Trigger to Output Channel Enable registers ey
0x024 CTIINEN1
0x028 CTIINEN2
0x02C CTIINEN3
0x030 CTIINEN4
0x034 CTIINEN5
0x038 CTIINEN6
0x03C CTIINEN7

0x040-0x09C - - - Reserved

0x0A0 CTIOUTEN0 RW 32-bit CTI Input Channel to Output Trigger Enable registers ey
0x0A4 CTIOUTEN1
0x0A8 CTIOUTEN2
0x0AC CTIOUTEN3
0x0B0 CTIOUTEN4
0x0B4 CTIOUTEN5
0x0B8 CTIOUTEN6
0x0BC CTIOUTEN7

0x0C0-0x12C - - - Reserved

0x130 CTITRIGINSTATUS RO 32-bit CTI Trigger In Status register ey

0x134 CTITRIGOUTSTATUS RO 32-bit CTI Trigger Out Status register ey
0x138 CTICHINSTATUS RO 32-bit CTI Channel In Status register ey
0x13C CTICHOUTSTATUS RO 32-bit CTI Channel Out Status register ey

ey See the ARM® Architecture Reference Manual ARMv8 for more information.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-439
reserved.
Non-Confidential
 12 Cross Trigger
12.5 Cross trigger register summary

Table 12-3 Cross trigger register summary (continued)

Offset Name Type Width Description

0x140 CTIGATE RW 32-bit CTI Channel Gate Enable register ey

0x144-0xED8 - - - Reserved

0xEDC CTIITCHINACK WO 32-bit 12.6.3 CTI Integration Test Channel In Acknowledge register
on page 12-444
0xEE0 CTIITTRIGINACK WO 32-bit 12.6.4 CTI Integration Test Trigger In Acknowledge register
on page 12-444
0xEE4 CTIITCHOUT WO 32-bit 12.6.5 CTI Integration Test Channel Out register on page 12-445
0xEE8 CTIITTRIGOUT WO 32-bit 12.6.6 CTI Integration Test Trigger Out register on page 12-446
0xEEC CTIITCHOUTACK RO 32-bit 12.6.7 CTI Integration Test Channel Out Acknowledge register
on page 12-446
0xEF0 CTIITTRIGOUTACK RO 32-bit 12.6.8 CTI Integration Test Trigger Out Acknowledge register
on page 12-447
0xEF4 CTIITCHIN RO 32-bit 12.6.9 CTI Integration Test Channel In register on page 12-448
0xEF8 CTIITTRIGIN RO 32-bit 12.6.10 CTI Integration Test Trigger In register on page 12-448

0xEFC-0xF7C - - - Reserved

0xF00 CTIICTRL RW 32-bit 12.6.2 CTI Integration Mode Control register on page 12-443

0xF04-0xFAC - - - Reserved

0xFB0 CTILAR WO 32-bit CTI Lock Access Register ey

0xFB4 CTILSR RO 32-bit CTI Lock Status Register ey
0xFB8 CTIAUTHSTATUS RO 32-bit CTI Authentication Status register ey

0xFBC-0xFC4 - - - Reserved

0xFC8 CTIDEVID RO 32-bit 12.6.1 CTI Device Identification register on page 12-442
0xFCC CTIDEVTYPE RO 32-bit CTI Device Type register ey
0xFD0 CTIPIDR4 RO 32-bit CTI Peripheral Identification Register 4 on page 12-452
0xFD4 CTIPIDR5 RO 32-bit CTI Peripheral Identification Register 5-7 on page 12-453
0xFD8 CTIPIDR6
0xFDC CTIPIDR7
0xFE0 CTIPIDR0 RO 32-bit CTI Peripheral Identification Register 0 on page 12-450
0xFE4 CTIPIDR1 RO 32-bit CTI Peripheral Identification Register 1 on page 12-450
0xFE8 CTIPIDR2 RO 32-bit CTI Peripheral Identification Register 2 on page 12-451
0xFEC CTIPIDR3 RO 32-bit CTI Peripheral Identification Register 3 on page 12-452
0xFF0 CTICIDR0 RO 32-bit CTI Component Identification Register 0 on page 12-454
0xFF4 CTICIDR1 RO 32-bit CTI Component Identification Register 1 on page 12-454
0xFF8 CTICIDR2 RO 32-bit CTI Component Identification Register 2 on page 12-455
0xFFC CTICIDR3 RO 32-bit CTI Component Identification Register 3 on page 12-456

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-440
reserved.
Non-Confidential
 12 Cross Trigger
12.5 Cross trigger register summary

This section contains the following subsections:

• 12.5.1 External register access permissions on page 12-441.

12.5.1 External register access permissions

External access permission to the cross trigger registers is subject to the conditions at the time of the
access. The following table describe the processor response to accesses through the external debug and
memory-mapped interfaces.

Table 12-4 External register access conditions

Condition Condition Description

code

Off EDPRSR.PU is 0 Core power domain is completely off, or in a low-power state where
the core power domain registers cannot be accessed.
Note
If debug is powered down, all external debug and memory-mapped
register accesses return an error.

DLK EDPRSR.DLK is 1 OS Double Lock is locked.

OSLK OSLSR_EL1.OSLK is 1 OS Lock is locked.

EDAD AllowExternalDebugAccess() External debug access disabled. When an error is returned because of
==FALSE the EDAD condition code, and this is the highest priority error
condition, EDPRSR.SDAD is set to 1. Otherwise EDPRSR.SDAD is
unchanged.

SLK Memory-mapped interface only Software Lock is locked. For the external debug interface, ignore this
code.

Default - None of the conditions apply, normal access.

The following table shows an example of external register access conditions for access to a cross trigger
register. To determine the access permission for the register, scan the columns from left to right. Stop at
the first column whose condition is true, the entry gives the access permission of the register and
scanning stops.

Table 12-5 External register access conditions example

Off DLK OSLK EDAD SLK Default

- - - - RO/WI RO

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-441
reserved.
Non-Confidential
 12 Cross Trigger
12.6 Cross trigger register descriptions

12.6 Cross trigger register descriptions

This section describes the Cortex-A72 processor cross trigger registers.
12.5 Cross trigger register summary on page 12-439 provides cross-references to the individual registers.
The Integration Test registers are provided to simplify the process of verifying the integration of the ECT
with other devices in a CoreSight system. These registers enable direct control of outputs and the ability
to read the value of inputs. You must only use these registers when the CTIITCTRL.IME bit is set to 1.
See the ARM® Architecture Reference Manual ARMv8 for more information.
This section contains the following subsections:
• 12.6.1 CTI Device Identification register on page 12-442.
• 12.6.2 CTI Integration Mode Control register on page 12-443.
• 12.6.3 CTI Integration Test Channel In Acknowledge register on page 12-444.
• 12.6.4 CTI Integration Test Trigger In Acknowledge register on page 12-444.
• 12.6.5 CTI Integration Test Channel Out register on page 12-445.
• 12.6.6 CTI Integration Test Trigger Out register on page 12-446.
• 12.6.7 CTI Integration Test Channel Out Acknowledge register on page 12-446.
• 12.6.8 CTI Integration Test Trigger Out Acknowledge register on page 12-447.
• 12.6.9 CTI Integration Test Channel In register on page 12-448.
• 12.6.10 CTI Integration Test Trigger In register on page 12-448.
• 12.6.11 CTI Peripheral Identification Registers on page 12-449.
• 12.6.12 CTI Component Identification Registers on page 12-453.

12.6.1 CTI Device Identification register

The CTIDEVID characteristics are:

Purpose
Describes the CTI component to the debugger.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - RO RO

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.
Configurations
CTIDEVID is in the Debug power domain.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
The following figure shows the CTIDEVID bit assignments.
31 26 25 24 23 22 21 16 15 14 13 8 7 5 4 0

RES0 RES0 NUMCHAN RES0 NUMTRIG RES0 EXTMAXNUM

INOUT

Figure 12-2 CTIDEVID bit assignments

The following table shows the CTIDEVID bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-442
reserved.
Non-Confidential
 12 Cross Trigger
12.6 Cross trigger register descriptions

Table 12-6 CTIDEVID bit assignments

Bits Name Function

[31:26] - Reserved, RES0.

[25:24] INOUT Input and output options. Indicates the presence of an input gate. The possible values are:
0b00 CTIGATE does not mask propagation of input events from external channels.
0b01 CTIGATE masks propagation of input events from external channels.

[23:22] - Reserved, RES0.

[21:16] NUMCHAN Number of channels implemented. The value is:

0b000100 Four channels implemented.

[15:14] - Reserved, RES0.

[13:8] NUMTRIG Number of triggers implemented. The value is:

0b001000 Eight triggers implemented.

[7:5] - Reserved, RES0.

[4:0] EXTMAXNUM Maximum number of external triggers implemented. The value is:
0b00000 No external triggers implemented.

12.6.2 CTI Integration Mode Control register

The CTIITCTRL characteristics are:

Purpose
Enables the CTI to switch from its default mode into integration mode, where test software can
control directly the inputs and outputs of the processor, for integration testing or topology
detection.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - RO/WI RW

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.
Configurations
CTIITCTRL is in the Debug power domain.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
shows the CTIITCTRL bit assignments.
31 1 0

RES0

IME

Figure 12-3 CTIITCTRL bit assignment

The following table shows the CTIITCTRL bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-443
reserved.
Non-Confidential
 12 Cross Trigger
12.6 Cross trigger register descriptions

Table 12-7 CTIITCTRL bit assignment

Bits Name Function

[31:1] - Reserved, RES0.

[0] IME Integration mode enable. The values are:

0 Normal operation.
1 Enables integration mode.

12.6.3 CTI Integration Test Channel In Acknowledge register

The CTIITCHINACK characteristics are:

Purpose
Provides direct control of the channel in acknowledge signals.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - RO WO

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.
Configurations
CTIITCHINACK is in the Debug power domain.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
The following figure shows the CTIITCHINACK bit assignments.
31 4 3 0

RES0

CTCHINACK

Figure 12-4 CTIITCHINACK bit assignments

The following table shows the CTIITCHINACK bit assignments.

Table 12-8 CTIITCHINACK bit assignments

Bits Name Function

[31:4] - Reserved, RES0.

[3:0] CTCHINACK Set the value of the CTCHINACK outputs.

12.6.4 CTI Integration Test Trigger In Acknowledge register

The CTIITTRIGINACK characteristics are:

Purpose
Provides direct control of the trigger in acknowledge signals.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-444
reserved.
Non-Confidential
 12 Cross Trigger
12.6 Cross trigger register descriptions

Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - WI WO

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.
Configurations
CTIITTRIGINACK is in the Debug power domain.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
The following figure shows the CTIITTRIGINACK bit assignments.
31 8 7 0

Reserved CTTRIGINACK

Figure 12-5 CTIITTRIGINACK bit assignments

The following table shows the CTIITTRIGINACK bit assignments.

Table 12-9 CTIITTRIGINACK bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:0] CTTRIGINACK Set the value of the CTTRIGINACK outputs.

12.6.5 CTI Integration Test Channel Out register

The CTIITCHOUT characteristics are:

Purpose
Provides direct control of the channel out signals.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - WI WO

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.
Configurations
CTIITCHOUT is in the Debug power domain.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
The following figure shows the CTIITCHOUT bit assignments.
31 4 3 0

Reserved CTCHOUT

Figure 12-6 CTIITCHOUT bit assignments

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-445
reserved.
Non-Confidential
 12 Cross Trigger
12.6 Cross trigger register descriptions

The following table shows the CTIITCHOUT bit assignments.

Table 12-10 CTIITCHOUT bit assignments

Bits Name Function

[31:4] - Reserved, RES0.

[3:0] CTCHOUT Set the value of the CTCHOUT outputs.

12.6.6 CTI Integration Test Trigger Out register

The CTIITTRIGOUT characteristics are:

Purpose
Provides direct observation of the trigger out signals.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - WI WO

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.
Configurations
CTIITTRIGOUT is in the Debug power domain.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
The following figure shows the CTIITTRIGOUT bit assignments.
31 8 7 0

Reserved CTTRIGOUT

Figure 12-7 CTIITTRIGOUT bit assignments

The following table shows the CTIITTRIGOUT bit assignments.

Table 12-11 CTIITTRIGOUT bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:0] CTTRIGOUT Set the value of the CTTRIGOUT outputs.

12.6.7 CTI Integration Test Channel Out Acknowledge register

The CTIITCHOUTACK characteristics are:

Purpose
Provides direct observation of the channel out acknowledge signals.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-446
reserved.
Non-Confidential
 12 Cross Trigger
12.6 Cross trigger register descriptions

Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - RO RO

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.
Configurations
CTIITCHOUTACK is in the Debug power domain.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
The following figure shows the CTIITCHOUTACK bit assignments.
31 4 3 0

RES0

CTCHOUTACK

Figure 12-8 CTIITCHOUTACK bit assignments

The following table shows the CTIITCHOUTACK bit assignments.

Table 12-12 CTIITCHOUTACK bit assignments

Bits Name Function

[31:4] - Reserved, RES0.

[3:0] CTCHOUTACK Read values of the CTCHOUTACK signals.

12.6.8 CTI Integration Test Trigger Out Acknowledge register

The CTIITTRIGOUTACK characteristics are:

Purpose
Provides direct observation of the trigger out acknowledge signals.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - RO RO

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.
Configurations
CTIITTRIGOUTACK is in the Debug power domain.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
The following figure shows the CTIITTRIGOUTACK bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-447
reserved.
Non-Confidential
 12 Cross Trigger
12.6 Cross trigger register descriptions

31 8 7 0

Reserved CTTRIGOUTACK

Figure 12-9 CTIITTRIGOUTACK bit assignments

The following table shows the CTIITTRIGOUTACK bit assignments.

Table 12-13 CTIITTRIGOUTACK bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:0] CTTRIGOUTACK Read values of the CTTRIGOUTACK signals

12.6.9 CTI Integration Test Channel In register

The CTIITCHIN characteristics are:

Purpose
Provides direct observation of the channel in signals.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - RO RO

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.
Configurations
CTIITCHIN is in the Debug power domain.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
The following figure shows the CTIITCHIN bit assignments.
31 4 3 0

Reserved CTCHIN

Figure 12-10 CTIITCHIN bit assignments

The following table shows the CTIITCHIN bit assignments.

Table 12-14 CTIITCHIN bit assignments

Bits Name Function

[31:4] - Reserved, RES0

[3:0] CTCHIN Read values of the CTCHIN signals

12.6.10 CTI Integration Test Trigger In register

The CTIITTRIGIN characteristics are:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-448
reserved.
Non-Confidential
 12 Cross Trigger
12.6 Cross trigger register descriptions

Purpose
Provides direct observation of the trigger in signals.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EDAD SLK Default

- - - - RO RO

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.
Configurations
CTIITTRIGIN is in the Debug power domain.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
The following figure shows the CTIITTRIGIN bit assignments.
31 8 7 0

Reserved CTTRIGIN

Figure 12-11 CTIITTRIGIN bit assignments

The following table shows the CTIITTRIGIN bit assignments.

Table 12-15 CTIITTRIGIN bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:0] CTTRIGIN Read values of the CTTRIGIN signals

12.6.11 CTI Peripheral Identification Registers

The Peripheral Identification Registers provide standard information required for all components that
conform to the ARM CoreSight architecture. There is a set of eight registers, listed in register number
order in the following table.

Table 12-16 Summary of the CTI Peripheral Identification Registers

Register Value Offset

CTIPIDR4 0x04 0xFD0

CTIPIDR5 0x00 0xFD4

CTIPIDR6 0x00 0xFD8

CTIPIDR7 0x00 0xFDC

CTIPIDR0 0x06 0xFE0

CTIPIDR1 0xB9 0xFE4

CTIPIDR2 0x4B 0xFE8

CTIPIDR3 0x00 0xFEC

Only bits[7:0] of each CTI Peripheral ID Register are used, with bits[31:8] reserved. Together, the eight
CTI Peripheral ID Registers define a single 64-bit Peripheral ID.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-449
reserved.
Non-Confidential
 12 Cross Trigger
12.6 Cross trigger register descriptions

The CTI Peripheral ID registers are:

• CTI Peripheral Identification Register 0 on page 12-450.
• CTI Peripheral Identification Register 1 on page 12-450.
• CTI Peripheral Identification Register 2 on page 12-451.
• CTI Peripheral Identification Register 3 on page 12-452.
• CTI Peripheral Identification Register 4 on page 12-452.
• CTI Peripheral Identification Register 5-7 on page 12-453.

CTI Peripheral Identification Register 0

The CTIPIDR0 characteristics are:
Purpose Provides information to identify a CTI component.
Usage constraints Accessible through the internal memory-mapped interface and the external debug
interface. The access conditions are:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 12-3 Cross trigger register summary on page 12-439 describes the access
conditions.
Configurations CTIPIDR0 is in the Debug power domain.
CTIPIDR0 is optional to implement in the external register interface.
Attributes See 12.5 Cross trigger register summary on page 12-439.

The following figure shows the CTIPIDR0 bit assignments.

31 8 7 0

RES0 Part_0

Figure 12-12 CTIPIDR0 bit assignments

The following table shows the CTIPIDR0 bit assignments.

Table 12-17 CTIPIDR0 bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:0] Part_0 0x06 Least significant byte of the cross trigger part number

CTI Peripheral Identification Register 1

The CTIPIDR1 characteristics are:
Purpose
Provides information to identify a CTI component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-450
reserved.
Non-Confidential
 12 Cross Trigger
12.6 Cross trigger register descriptions

Configurations
CTIPIDR1 is in the Debug power domain.
CTIPIDR1 is optional to implement in the external register interface.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
The following figure shows the CTIPIDR1 bit assignments.
31 8 7 4 3 0

RES0 DES_0 Part_1

Figure 12-13 CTIPIDR1 bit assignments

The following table shows the CTIPIDR1 bit assignments.

Table 12-18 CTIPIDR1 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] DES_0 0xB ARM Limited. This is the least significant nibble of JEP106 ID code.

[3:0] Part_1 0x9 Most significant nibble of the cross trigger interface part number.

CTI Peripheral Identification Register 2

The CTIPIDR2 characteristics are:
Purpose Provides information to identify a CTI component.
Usage constraints Accessible through the internal memory-mapped interface and the external debug
interface. The access conditions are:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 12-3 Cross trigger register summary on page 12-439 describes the access
conditions.
Configurations CTIPIDR2 is in the Debug power domain.
CTIPIDR2 is optional to implement in the external register interface.
Attributes See 12.5 Cross trigger register summary on page 12-439.

The following figure shows the CTIPIDR2 bit assignments.

31 8 7 4 3 2 0

RES0 Revision DES_1

JEDEC

Figure 12-14 CTIPIDR2 bit assignments

The following table shows the CTIPIDR2 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-451
reserved.
Non-Confidential
 12 Cross Trigger
12.6 Cross trigger register descriptions

Table 12-19 CTI PIDR2 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] Revision 0x4 Part major revision.

[3] JEDEC 0b1 RES1. Indicates a JEP106 identity code is used.

[2:0] DES_1 0b011 ARM Limited. This is the most significant nibble of JEP106 ID code.

CTI Peripheral Identification Register 3

The CTIPIDR3 characteristics are:
Purpose
Provides information to identify a CTI component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.
Configurations
CTIPIDR3 is in the Debug power domain.
CTIPIDR3 is optional to implement in the external register interface.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
The following figure shows the CTIPIDR3 bit assignments.
31 8 7 4 3 0

RES0 REVAND CMOD

Figure 12-15 CTIPIDR3 bit assignments

The following table shows the CTIPIDR3 bit assignments.

Table 12-20 CTIPIDR3 bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:4] REVAND 0x0 Part minor revision

[3:0] CMOD 0x0 Customer modified

CTI Peripheral Identification Register 4

The CTIPIDR4 characteristics are:
Purpose
Provides information to identify a CTI component.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-452
reserved.
Non-Confidential
 12 Cross Trigger
12.6 Cross trigger register descriptions

Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.
Configurations
CTIPIDR4 is in the Debug power domain.
CTIPIDR4 is optional to implement in the external register interface.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
The following figure shows the CTIPIDR4 bit assignments.
31 8 7 4 3 0

RES0 Size DES_2

Figure 12-16 CTIPIDR4 bit assignments

The following table shows the CTIPIDR4 bit assignments.

Table 12-21 CTIPIDR4 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] Size 0x0 Size of the component. Log2 the number of 4KB pages from the start of the component to the end of the
component ID registers.

[3:0] DES_2 0x4 ARM Limited. This is the least significant nibble JEP106 continuation code.

CTI Peripheral Identification Register 5-7

No information is held in the Peripheral ID5, Peripheral ID6, and Peripheral ID7 Registers. They are
reserved for future use and are RES0.

12.6.12 CTI Component Identification Registers

There are four read-only CTI Component Identification Registers, Component ID0 through Component
ID3. The following table shows these registers.

Table 12-22 Summary of the CTI Component Identification Registers

Register Value Offset

CTICIDR0 0x0D 0xFF0

CTICIDR1 0x90 0xFF4

CTICIDR2 0x05 0xFF8

CTICIDR3 0xB1 0xFFC

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-453
reserved.
Non-Confidential
 12 Cross Trigger
12.6 Cross trigger register descriptions

The CTI Component ID registers are:

• CTI Component Identification Register 0 on page 12-454.
• CTI Component Identification Register 1 on page 12-454.
• CTI Component Identification Register 2 on page 12-455.
• CTI Component Identification Register 3 on page 12-456.

CTI Component Identification Register 0

The CTICIDR0 characteristics are:
Purpose
Provides information to identify a CTI component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.
Configurations
CTICIDR0 is in the Debug power domain.
CTICIDR0 is optional to implement in the external register interface.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
The following figure shows the CTICIDR0 bit assignments.
31 8 7 0

Reserved PRMBL_0

Figure 12-17 CTICIDR0 bit assignments

The following table shows the CTICIDR0 bit assignments.

Table 12-23 CTICIDR0 bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:0] PRMBL_0 0x0D Preamble byte 0

CTI Component Identification Register 1

The CTICIDR1 characteristics are:
Purpose
Provides information to identify a CTI component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-454
reserved.
Non-Confidential
 12 Cross Trigger
12.6 Cross trigger register descriptions

Configurations
CTICIDR1 is in the Debug power domain.
CTICIDR1 is optional to implement in the external register interface.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
The following figure shows the CTICIDR1 bit assignments.
31 8 7 4 3 0

Reserved CLASS PRMBL_1

Figure 12-18 CTICIDR1 bit assignments

The following table shows the CTICIDR1 bit assignments.

Table 12-24 CTICIDR1 bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:4] CLASS 0x9 Debug component

[3:0] PRMBL_1 0x0 Preamble

CTI Component Identification Register 2

The CTICIDR2 characteristics are:
Purpose
Provides information to identify a CTI component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.
Configurations
CTICIDR2 is in the Debug power domain.
CTICIDR2 is optional to implement in the external register interface.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
The following figure shows the CTICIDR2 bit assignments.
31 8 7 0

Reserved PRMBL_2

Figure 12-19 CTICIDR2 bit assignments

The following table shows the CTICIDR2 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-455
reserved.
Non-Confidential
 12 Cross Trigger
12.6 Cross trigger register descriptions

Table 12-25 CTICIDR2 bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:0] PRMBL_2 0x05 Preamble byte 2

CTI Component Identification Register 3

The CTICIDR3 characteristics are:
Purpose
Provides information to identify a CTI component.
Usage constraints
Accessible through the internal memory-mapped interface and the external debug interface. The
access conditions are:

Off DLK OSLK EPMAD SLK Default

- - - - RO RO

Table 12-3 Cross trigger register summary on page 12-439 describes the access conditions.
Configurations
CTICIDR3 is in the Debug power domain.
CTICIDR3 is optional to implement in the external register interface.
Attributes
See 12.5 Cross trigger register summary on page 12-439.
The following figure shows the CTICIDR3 bit assignments.
31 8 7 0

Reserved PRMBL_3

Figure 12-20 CTICIDR3 bit assignments

The following table shows the CTICIDR3 bit assignments.

Table 12-26 CTICIDR3 bit assignments

Bits Name Function

[31:8] - Reserved, RES0

[7:0] PRMBL_3 0xB1 Preamble byte 3

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 12-456
reserved.
Non-Confidential
Chapter 13
Embedded Trace Macrocell

This section describes the Embedded Trace Macrocell (ETM) for the Cortex-A72 processor.

It contains the following sections:

• 13.1 About ETM on page 13-458.
• 13.2 ETM trace generation options and resources on page 13-459.
• 13.3 ETM functional description on page 13-461.
• 13.4 Reset on page 13-462.
• 13.5 ETM register interfaces on page 13-463.
• 13.6 Register summary on page 13-464.
• 13.7 Register descriptions on page 13-468.
• 13.8 Interaction with debug and the Performance Monitor Unit on page 13-510.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-457
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.1 About ETM

13.1 About ETM

Describes the Embedded Trace Macrocell used in the Cortex-A72 processor.
The ETM is a module that performs real-time instruction flow tracing based on the ARM® Embedded
Trace Macrocell Architecture Specification, ETMv4. The ETM is a CoreSight component, and is an
integral part of the ARM Real-time Debug solution, RealView. See the CoreSight SoC-400
documentation for more information.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-458
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.2 ETM trace generation options and resources

13.2 ETM trace generation options and resources

The following table shows the trace generation options that the Cortex-A72 processor implements.

Table 13-1 ETM trace generation options implemented

Description Configuration

Instruction address size in bytes 8

Data address size in bytes 0

Data value size in bytes 0

Virtual Machine ID size in bytes 1

Context ID size in bytes 4

Support for conditional instruction tracing Not implemented

Support for tracing of data Not implemented

Support for tracing of load and store instructions as P0 elements Not implemented

Support for cycle counting in the instruction trace Implemented

Support for branch broadcast tracing Implemented

Exception Levels implemented in Non-secure state 0b0111

Exception Levels implemented in Secure state 0b1011

Number of events supported in the trace 4

Return stack support Implemented

Tracing of SError exception support Implemented

Instruction trace cycle counting minimum threshold 4

Size of Trace ID 7-bit

Synchronization period support Read-write

Global timestamp size 64-bit

Number of cores available for tracing 1

ATB trigger support Implemented

Low power behavior override Not implemented

Stall control support Not implemented

Support for no overflows in the trace Not implemented

The following table shows the ETM resources that the Cortex-A72 processor implements.

Table 13-2 ETM resources implemented

Description Configuration

Number of resource selection pairs implemented 8

Number of external input selectors implemented 4

Number of external inputs implemented 110, 4 external + 106 PMU

Number of counters implemented 2

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-459
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.2 ETM trace generation options and resources

Table 13-2 ETM resources implemented (continued)

Description Configuration

Reduced function counter implemented Not implemented

Number of sequencer states implemented 4

Number of Virtual Machine ID comparators implemented 1

Number of Context ID comparators implemented 1

Number of address comparator pairs implemented 4

Number of single-shot comparator controls 1

Number of processor comparator inputs implemented 0

Data address comparisons implemented Not implemented

Number of data value comparators implemented 0

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-460
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.3 ETM functional description

13.3 ETM functional description

Describes the functional blocks of the ETM.
The following figure shows the main functional blocks of the ETM.
ETM
CLK

ATCLKEN Trace out SyncBridge

Trace interface
SyncBridge
(ATB)
Debug interface Filtering and triggering
FIFO
(APB) resources

Processor Processor
Trace generation
interface interface block

Figure 13-1 ETM functional blocks

The ETM blocks are:

Processor interface
This block monitors the behavior of the processor and generates P0 elements that are essentially
executed instructions and exceptions traced in program order.
Trace generation
The trace generation block generates various trace packets based on P0 elements.
Filtering and triggering resources
You can filter the ETM trace such as configuring it to trace only in certain address ranges. More
complicated logic analyzer style filtering options are also available.
The ETM can also generate a trigger that is a signal to the trace capture device to stop capturing
trace.
FIFO
The trace generated by the ETM is in a highly-compressed form. The FIFO enables trace bursts
to be flattened out. When the FIFO becomes full, the FIFO signals an overflow. The trace
generation logic does not generate any new trace until the FIFO is emptied. This causes a gap in
the trace when viewed in the debugger.
Trace out
Trace from FIFO is output on the synchronous AMBA Advanced Trace Bus (ATB) interface.
Syncbridge
The ATB interface from the trace out block goes through two slices of the CoreSight SoC ATB
Syncbridge IP.
See the ARM® AMBA® 3 ATB Protocol Specification for information about the ATB protocol.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-461
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.4 Reset

13.4 Reset
The reset for ETM is the same as Cold reset for processor.
The ETM is not reset when a Warm reset is applied to processor, so that tracing through the reset is
possible.
If the ETM is reset, tracing stops until the ETM is reprogrammed and re-enabled. However, if the
processor is reset using a Warm reset, the last few instructions provided by the processor before the reset
might not be traced.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-462
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.5 ETM register interfaces

13.5 ETM register interfaces

The reset for ETM is the same as Cold reset for processor.

This section contains the following subsections:

• 13.5.1 Access permissions on page 13-463.

13.5.1 Access permissions

See the ARM® Embedded Trace Macrocell Architecture Specification, ETMv4 for information on the
behaviors on register accesses for different trace unit states and the different access mechanisms.

Related information
10.10 External debug interface on page 10-381.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-463
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.6 Register summary

13.6 Register summary

This section summarizes the ETM registers.
For full descriptions of the ETM registers, see:
• 10.10 External debug interface on page 10-381, for the IMPLEMENTATION DEFINED registers and the
ARM® Embedded Trace Macrocell Architecture Specification, ETMv4, for the other registers.

Note
• In the following table, access type is described as follows:
RW Read and write.
RO Read only.
WO Write only.

All ETM registers are 32 bits wide. The following table lists all of the registers and their offsets from a
base address. The base address is defined by the system integrator when placing the ETM in the Debug-
APB memory map.

Table 13-3 ETM register summary

Offset Name Type Description

0x000 - - Reserved
0x004 TRCPRGCTLR RW Trace Programming Control Register
0x008 - - Reserved
0x00C TRCSTATR RO Trace Status Register
0x010 TRCCONFIGR RW 13.7.1 Trace Configuration Register on page 13-468
0x014 - - Reserved
0x018 TRCAUXCTLR RW 13.7.2 Trace Auxiliary Control Register on page 13-470
0x01C - - Reserved
0x020 TRCEVENTCTL0R RW 13.7.3 Trace Event Control 0 Register on page 13-472
0x024 TRCEVENTCTL1R RW 13.7.4 Trace Event Control 1 Register on page 13-473

0x028-0x2C - - Reserved

0x030 TRCTSCTLR RW Global Timestamp Control Register

0x034 TRCSYNCPR RW  13.7.5 Trace Synchronization Period Register on page 13-474
0x038 TRCCCCTLR RW 13.7.6 Trace Cycle Count Control Register on page 13-475
0x03C TRCBBCTLR RW Branch Broadcast Control Register
0x040 TRCTRACEIDR RW 13.7.7 Trace ID Register on page 13-476

0x044 -0x07C - - Reserved

0x080 TRCVICTLR RW 13.7.8 ViewInst Main Control Register on page 13-476

0x084 TRCVIIECTLR RW ViewInst Include-Exclude Control Register
0x088 TRCVISSCTLR RW ViewInst Start-Stop Control Register

0x08C -0x0FC - - Reserved

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-464
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.6 Register summary

Table 13-3 ETM register summary (continued)

Offset Name Type Description

0x100   TRCSEQEVR0 RW Sequencer State Transition Control Register 0

0x104 TRCSEQEVR1 RW Sequencer State Transition Control Register 1

0x108 TRCSEQEVR2 RW Sequencer State Transition Control Register 2

0x10C -0x114 - - Reserved

0x118 TRCSEQRSTEVR RW Sequencer Reset Control Register

0x11C TRCSEQSTR RW Sequencer State Register
0x120 TRCEXTINSELR RW 13.7.9 External Input Select Register on page 13-478

0x124 -0x13C - - Reserved

0x140  TRCCNTRLDVR0 RW Counter Reload Value Register 0

0x144 TRCCNTRLDVR1 RW Counter Reload Value Register 1

0x148-0x14C - - Reserved

0x150  TRCCNTCTLR0 RW Counter Control Register 0

0x154 TRCCNTCTLR1 RW Counter Control Register 1

0x158-0x15C - - Reserved

0x160  TRCCNTVR0 RW Counter Value Register 0

0x164 TRCCNTVR1 RW Counter Value Register 1

0x168-0x16C - - Reserved

0x170 -0x17C - - Reserved

0x180 TRCIDR8 RO 13.7.10 ID Register 8 on page 13-479

0x184 TRCIDR9 RO 13.7.11 ID Register 9 on page 13-479
0x188 TRCIDR10 RO 13.7.12 ID Register 10 on page 13-480
0x18C TRCIDR11 RO 13.7.13 ID Register 11 on page 13-480>
0x190 TRCIDR12 RO 13.7.14 ID Register 12 on page 13-481
0x194 TRCIDR13 RO 13.7.15 ID Register 13 on page 13-482

0x198-0x1BC  - - Reserved

0x1C0 TRCIMSPEC0 RW 13.7.16 Implementation Defined Register 0 on page 13-482

0x1C4-0x1DC - - Reserved

0x1E0 TRCIDR0 RO 13.7.17 Trace ID Register 0 on page 13-483

0x1E4 TRCIDR1 RO 13.7.18 Trace ID Register 1 on page 13-484
0x1E8 TRCIDR2 RO 13.7.19 Trace ID Register 2 on page 13-485
0x1EC TRCIDR3 RO 13.7.20 Trace ID Register 3 on page 13-486
0x1F0 TRCIDR4 RO 13.7.21 Trace ID Register 4 on page 13-488
0x1F4 TRCIDR5 RO 13.7.22 Trace ID Register 5 on page 13-489

0x1F8 -0x204 - - Reserved

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-465
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.6 Register summary

Table 13-3 ETM register summary (continued)

Offset Name Type Description

0x208-0x23C TRCRSCTLRn RW 13.7.23 Resource Selection Control Registers on page 13-491, n is 2, 15

0x240-0x27C - - Reserved

0x280 TRCSSCCR0 RW Single-shot Comparator Control Register 0

0x284-0x29C - - Reserved

0x2A0 TRCSSCSR0 RW, RO Single-shot Comparator Status Register 0

0x2A4-0x2FC - - Reserved

0x300 TRCOSLAR WO OS Lock Access Register

0x304 TRCOSLSR RO OS Lock Status Register

0x308 -0x30C - - Reserved

0x310 TRCPDCR RW PowerDown Control Register

0x314 TRCPDSR RO PowerDown Status Register

0x318 -0x3FC - - Reserved

0x400-0x438  TRCACVRn RW Address Comparator Value Register n, n = 0 to 7

0x440-0x47C - - Reserved

0x480-0x4B8  TRCACATRn RW 13.7.24 Address Comparator Access Type Registers on page 13-492, n is 0 to 7

0x4C0-0x5FC - - Reserved

0x600   TRCCIDCVR0 RW 13.7.25 Context ID Comparator Value Register 0 on page 13-494

0x608-0x63F - - Reserved

0x640   TRCVMIDCVR0 RW 13.7.26 VMID Comparator Value Register 0 on page 13-495

0x648-0x67F - - Reserved

0x680 TRCCIDCCTLR0 RW 13.7.27 Context ID Comparator Control Register 0 on page 13-496

0x684-0xED8 - - Reserved

0xEDC TRCITMISCOUT WO 13.7.28 Trace Integration Miscellaneous Outputs Register on page 13-496
0xEE0 TRCITMISCIN RO 13.7.29 Trace Integration Miscellaneous Input Register on page 13-497

0xEE4-0xEE8 - - Reserved

0xEEC TRCITATBDATA0 WO 13.7.30 Trace Integration Test ATB Data Register 0 on page 13-498
0xEF0 TRCITATBCTR2 RO 13.7.31 Trace Integration Test ATB Control Register 2 on page 13-498
0xEF4 TRCITATBCTR1 WO 13.7.32 Trace Integration Test ATB Control Register 1 on page 13-499
0xEF8 TRCITATBCTR0 WO 13.7.33 Trace Integration Test ATB Control Register 0 on page 13-500
0xEFC - - Reserved
0xF00 TRCITCTRL RW 13.7.34 Trace Integration Mode Control register on page 13-501

0xF04 -0xF9C - - Reserved

0xFA0 TRCCLAIMSET RW Trace Claim Tag Set register

0xFA4 TRCCLAIMCLR RW Trace Claim Tag Clear register

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-466
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.6 Register summary

Table 13-3 ETM register summary (continued)

Offset Name Type Description

0xFA8 TRCDEVAFF0 RO 13.7.35 Trace Device Affinity register 0 on page 13-501
0xFAC TRCDEVAFF1 RO 13.7.36 Trace Device Affinity register 1 on page 13-502
0xFB0 TRCLAR WO Trace Software Lock Access Register
0xFB4 TRCLSR RO Trace Software Lock Status Register
0xFB8 TRCAUTHSTATUS RO Trace Authentication Status register
0xFBC TRCDEVARCH RO Trace Device Architecture register

0xFC0 -0xFC4 - - Reserved

0xFC8 TRCDEVID RO Trace Device ID register

0xFCC TRCDEVTYPE RO Trace Device Type register
0xFD0 TRCPIDR4 RO Trace Peripheral Identification Register 4 on page 13-505

0xFD4  TRCPIDR5 RO Trace Peripheral Identification Register 5-7 on page 13-506

0xFD8 TRCPIDR6 RO
0xFDC TRCPIDR7 RO
0xFE0 TRCPIDR0 RO Trace Peripheral Identification Register 0 on page 13-503
0xFE4 TRCPIDR1 RO Trace Peripheral Identification Register 1 on page 13-503
0xFE8 TRCPIDR2 RO Trace Peripheral Identification Register 2 on page 13-504
0xFEC TRCPIDR3 RO Trace Peripheral Identification Register 3 on page 13-505
0xFF0 TRCCIDR0 RO Trace Component Identification Register 0 on page 13-507
0xFF4 TRCCIDR1 RO Trace Component Identification Register 1 on page 13-507
0xFF8 TRCCIDR2 RO Trace Component Identification Register 2 on page 13-508
0xFFC TRCCIDR3 RO Trace Component Identification Register 3 on page 13-508

Related information
10.11 ROM table on page 10-384.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-467
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

13.7 Register descriptions

This section describes the implementation-specific ETM registers in the Cortex-A72 processor.
13.6 Register summary on page 13-464 provides cross-references to individual registers.
The ARM® Embedded Trace Macrocell Architecture Specification, ETMv4 describes the other ETM
registers.
This section contains the following subsections:
• 13.7.1 Trace Configuration Register on page 13-468.
• 13.7.2 Trace Auxiliary Control Register on page 13-470.
• 13.7.3 Trace Event Control 0 Register on page 13-472.
• 13.7.4 Trace Event Control 1 Register on page 13-473.
• 13.7.5 Trace Synchronization Period Register on page 13-474.
• 13.7.6 Trace Cycle Count Control Register on page 13-475.
• 13.7.7 Trace ID Register on page 13-476.
• 13.7.8 ViewInst Main Control Register on page 13-476.
• 13.7.9 External Input Select Register on page 13-478.
• 13.7.10 ID Register 8 on page 13-479.
• 13.7.11 ID Register 9 on page 13-479.
• 13.7.12 ID Register 10 on page 13-480.
• 13.7.13 ID Register 11 on page 13-480.
• 13.7.14 ID Register 12 on page 13-481.
• 13.7.15 ID Register 13 on page 13-482.
• 13.7.16 Implementation Defined Register 0 on page 13-482.
• 13.7.17 Trace ID Register 0 on page 13-483.
• 13.7.18 Trace ID Register 1 on page 13-484.
• 13.7.19 Trace ID Register 2 on page 13-485.
• 13.7.20 Trace ID Register 3 on page 13-486.
• 13.7.21 Trace ID Register 4 on page 13-488.
• 13.7.22 Trace ID Register 5 on page 13-489.
• 13.7.23 Resource Selection Control Registers on page 13-491.
• 13.7.24 Address Comparator Access Type Registers on page 13-492.
• 13.7.25 Context ID Comparator Value Register 0 on page 13-494.
• 13.7.26 VMID Comparator Value Register 0 on page 13-495.
• 13.7.27 Context ID Comparator Control Register 0 on page 13-496.
• 13.7.28 Trace Integration Miscellaneous Outputs Register on page 13-496.
• 13.7.29 Trace Integration Miscellaneous Input Register on page 13-497.
• 13.7.30 Trace Integration Test ATB Data Register 0 on page 13-498.
• 13.7.31 Trace Integration Test ATB Control Register 2 on page 13-498.
• 13.7.32 Trace Integration Test ATB Control Register 1 on page 13-499.
• 13.7.33 Trace Integration Test ATB Control Register 0 on page 13-500.
• 13.7.34 Trace Integration Mode Control register on page 13-501.
• 13.7.35 Trace Device Affinity register 0 on page 13-501.
• 13.7.36 Trace Device Affinity register 1 on page 13-502.
• 13.7.37 Trace Peripheral Identification Registers on page 13-502.
• 13.7.38 Trace Component Identification Registers on page 13-506.

13.7.1 Trace Configuration Register

The TRCCONFIGR characteristics are:

Purpose
Controls the tracing options.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-468
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Usage constraints
Only accepts writes when the trace unit is disabled.
Configurations
Available in all configurations.
Attributes
A 32-bit RW trace register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCCONFIGR bit assignments.
31 13 12 11 10 8 7 6 5 4 3 2 1 0

RES0 RES0 RES0

RS VMID RES1
TS CID
RES0
CCI
BB

Figure 13-2 TRCCONFIGR bit assignments

The following table shows the TRCCONFIGR bit assignments.

Table 13-4 TRCCONFIGR bit assignments

Bits Name Function

[31:13] - Reserved, RES0.

[12] RS Enables the return stack. The possible values are:

0 Disables the return stack.
1 Enables the return stack.

[11] TS Enables global timestamp tracing. The possible values are:

0 Disables global timestamp tracing.
1 Enables global timestamp tracing.

[10:8] - Reserved, RES0.

[7] VMID Enables VMID tracing. The possible values are:

0 Disables VMID tracing.
1 Enables VMID tracing.

[6] CID Enables context ID tracing. The possible values are:

0 Disables context ID tracing.
1 Enables context ID tracing.

[5] - Reserved, RES0.

[4] CCI Enables cycle counting instruction trace. The possible values are:
0 Disables cycle counting instruction trace
1 Enables cycle counting instruction trace

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-469
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Table 13-4 TRCCONFIGR bit assignments (continued)

Bits Name Function

[3] BB Enables branch broadcast mode. The possible values are:

0 Disables branch broadcast mode.
1 Enables branch broadcast mode.

[2:1] - Reserved, RES0.

[0] - Reserved, RES1.

The TRCCONFIGR can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0x010.

13.7.2 Trace Auxiliary Control Register

The TRCAUXCTLR characteristics are:

Purpose
The function of this register is to provide IMPLEMENTATION DEFINED configuration and control
options.
Usage constraints
There are no usage constraints.
Configurations
Available in all configurations.
Attributes
A 32-bit RW trace register. This register is set to zero on a trace unit reset. Resetting this register
to zero ensures that none of the features are enabled by default, and that the trace unit resets to a
known state.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCAUXCTLR bit assignments.

31 10 9 8 7 6 5 4 7 2 1 0

Reserved

SBRCGFRCENABLE AFREADYOVERRIDE
DBGFLUSHOVERRIDE IDLEACKOVERRIDE
CIFOVERRIDE FRCSYNCOVERFLOW
CLKENOVERRIDE SYNCIOVERRIDE
FLUSHOVERRIDE TSIOVERRIDE

Figure 13-3 TRCAUXCTLR bit assignments

The following table shows the TRCAUXCTLR bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-470
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Table 13-5 TRCAUXCTLR bit assignments

Bits Name Function

[31:10] - Reserved, RES0.

[9] SBRCGFRCENABLE Force ETM trace sync bridge RCG enable active. The possible values are:
0
Enables ETM trace sync bridge RCG for additional clock gating and potentially reduce
dynamic power dissipation. This is the reset value.
1
Forces ETM trace sync bridge RCG enable high.

[8] DBGFLUSHOVERRIDE Override ETM flush behavior on Debug state entry. The possible values are:
0
ETM FIFO is flushed when the processor enters Debug state.
1
ETM FIFO is not flushed when the processor enters Debug state. This trace unit
behavior deviates from the architecturally-specified behavior.

[7] CIFOVERRIDE Override core interface register repeater clock enable. The possible values are:
0
Core interface is clock gated when DBGEN or NIDEN is LOW.
1
Core interface is not clock gated when DBGEN or NIDEN is LOW.

[6] CLKENOVERRIDE Override ETM clock enable. The possible values are:
0
ETM clock gating is enabled.
1
ETM clock gating is disabled.

[5] FLUSHOVERRIDE Override ETM flush behavior. The possible values are:
0
ETM FIFO is flushed and ETM enters idle state when DBGEN or NIDEN is LOW.
1
ETM FIFO is not flushed and ETM does not enter idle state when DBGEN or NIDEN
is LOW.
When this bit is set to 1, the trace unit behavior deviates from architecturally-specified behavior.

[4] TSIOVERRIDE Override TS packet insertion behavior. The possible values are:
0
Timestamp packets are inserted into FIFO when trace activity is LOW.
1
Timestamp packets are inserted into FIFO irrespective of trace activity.

[3] SYNCIOVERRIDE Override SYNC packet insertion behavior. The possible values are:
0
SYNC packets are inserted into FIFO when trace activity is LOW.
1
SYNC packets are inserted into FIFO irrespective of trace activity.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-471
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Table 13-5 TRCAUXCTLR bit assignments (continued)

Bits Name Function

[2] FRCSYNCOVERFLOW Force overflows to output synchronization packets. The possible values are:
0
No FIFO overflow when SYNC packets are delayed.
1
Forces FIFO overflow when SYNC packets are delayed.
When this bit is set to 1, the trace unit behavior deviates from architecturally-specified behavior.

[1] IDLEACKOVERRIDE Force ETM idle acknowledge. The possible values are:
0
ETM idle acknowledge is asserted only when ETM is in idle state.
1
ETM idle acknowledge is asserted irrespective of ETM idle state
When this bit is set to 1, trace unit behavior deviates from architecturally-specified behavior.

[0] AFREADYOVERRIDE Force assertion of AFREADYM output. The possible values are:
0
ETM AFREADYM output is asserted only when ETM is in idle state or when all the
trace bytes in FIFO before a flush request are output.
1
ETM AFREADYM output is always asserted HIGH. When this bit is set to 1, trace unit
behavior deviates from architecturally-specified behavior.

The TRCAUXCTLR can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0x018.

13.7.3 Trace Event Control 0 Register

The TRCEVENTCTL0R characteristics are:

Purpose
Controls the tracing of arbitrary events. Each of the event fields in this register is an event
selector.
If any of the selected events occur and the corresponding bit in TRCEVENTCTL1R.INSTEN is
1, then an event element is generated in the instruction trace stream.
If any of the selected events occur and the corresponding bit in TRCEVENTCTL1R.DATAEN
is 1, then an event element is generated in the data trace stream.
Usage constraints
Only accepts writes when the trace unit is disabled.
Configurations
Available in all configurations.
Attributes
A 32-bit RW trace register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCEVENTCTL0R bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-472
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

31 24 23 16 15 8 7 0

Event3 Event2 Event1 Event0

Figure 13-4 TRCEVENTCL0R bit assignments

The following table shows the TRCEVENTCTL0R bit assignments.

Table 13-6 TRCEVENTCL0R bit assignments

Bits Name Function

[31:24] Event3 Identifies the fourth event to trace

[23:16] Event2 Identifies the third event to trace

[15:8] Event1 Identifies the second event to trace

[7:0] Event0 Identifies the first event to trace

The TRCEVENTCTL0R can be accessed through the internal memory-mapped interface and the
external debug interface, offset 0x020.

13.7.4 Trace Event Control 1 Register

The TRCEVENTCTL1R characteristics are:

Purpose
Controls the behavior of the events that TRCEVENTCTL0R selects.
Usage constraints
Only accepts writes when the trace unit is disabled.
Configurations
Available in all configurations.
Attributes
A 32-bit RW trace register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCEVENTCTL1R bit assignments.
31 12 11 10 8 7 5 4 3 0

Reserved Reserved INSTEN

ATB

Figure 13-5 TRCEVENTCL1R bit assignments

The following table shows the TRCEVENTCTL1R bit assignments.

Table 13-7 TRCEVENTCL1R bit assignments

Bits Name Function

[31:12] - Reserved, RES0.

[11] ATB ATB trigger enable. This value is:

0 ATB trigger is disabled.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-473
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Table 13-7 TRCEVENTCL1R bit assignments (continued)

Bits Name Function

[10:4] - Reserved, RES0.

[3:0] INSTEN Instruction event enable field. Each bit represents an event, n=0-3. If event n occurs when INSTEN[n] is:
0 The trace unit does not generate an event element.
1 The trace unit generates an event element for event n, in the instruction trace stream.

The TRCEVENTCTL1R can be accessed through the internal memory-mapped interface and the
external debug interface, offset 0x024.

13.7.5 Trace Synchronization Period Register

The TRCSYNCPR characteristics are:

Purpose
Controls how often periodic trace synchronization requests occur.
Usage constraints
Only accepts writes when the trace unit is disabled.
This register must be programmed.
Configurations
Available in all configurations.
Attributes
A 32-bit RW trace register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCSYNCPR bit assignments.
31 5 4 0

Reserved Period

Figure 13-6 TRCSYNCPR bit assignments

The following table shows the TRCSYNCPR bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-474
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Table 13-8 TRCSYNCPR bit assignments

Bits Name Function

[31:5] - Reserved, RES0.

[4:0] Period Controls how many bytes of trace, the sum of instruction and data, that a trace unit can generate before a periodic
trace synchronization request occurs.
When 0b00000, periodic trace synchronization requests are disabled. This setting does not disable other types of
trace synchronization request.
The number of bytes is always a power of two and the permitted values are:
0b01000 Periodic trace synchronization request occurs after 28, or 256 bytes of trace.
0b01001 Periodic trace synchronization request occurs after 29, or 512 bytes of trace.
0b01010 Periodic trace synchronization request occurs after 210, or 1024 bytes of trace.
. .
. .
. .
0b10100 Periodic trace synchronization request occurs after 220, or 1048576 bytes of trace.

The TRCSYNCPR can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0x034.

13.7.6 Trace Cycle Count Control Register

The TRCCCCTLR characteristics are:

Purpose
Sets the threshold value for cycle counting.
Usage constraints
Only accepts writes when the trace unit is disabled.
This register must be programmed if TRCCONFIGR.CCI is set to 1.
Configurations
Available in all configurations.
Attributes
A 32-bit RW trace register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCCCCTLR bit assignments.
31 12 11 0

Reserved Threshold

Figure 13-7 TRCCCCTLR bit assignments

The following table shows the TRCCCCTLR bit assignments.

Table 13-9 TRCCCCTLR bit assignments

Bits Name Function

[31:12] - Reserved, RES0.

[11:0] Threshold Sets the threshold value for instruction trace cycle counting.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-475
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

The TRCCCCTLR can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0x038.

Related information
13.7.1 Trace Configuration Register on page 13-468.

13.7.7 Trace ID Register

The TRCTRACEIDR characteristics are:

Purpose
Sets the trace ID for instruction trace.
Usage constraints
Only accepts writes when the trace unit is disabled.
Configurations
Available in all configurations.
The TRACEID field width is set by TRCIDR5.TRACEIDSIZE.
Attributes
A 32-bit RW trace register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCTRACEIDR bit assignments.
31 7 6 0

Reserved TRACEID

Figure 13-8 TRCTRACEIDR bit assignments

The following table shows the TRCTRACEIDR bit assignments.

Table 13-10 TRCTRACEIDR bit assignments

Bits Name Function

[31:7] - Reserved, RES0.

[6:0] TRACEID Trace ID field. Sets the trace ID value for instruction trace. The width of this field is 7 bits.

The TRCTRACEIDR can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0x040.

Related information
13.7.22 Trace ID Register 5 on page 13-489.

13.7.8 ViewInst Main Control Register

The TRCVICTLR characteristics are:

Purpose
Controls instruction trace filtering.
Usage constraints
• Only accepts writes when the trace unit is disabled.
• Only returns stable data when TRCSTATR.PMSTABLE is set to 1.
• Must be programmed to set the value of the SSSTATUS bit, that sets the state of the start and
stop logic.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-476
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Configurations
Available in all configurations.
Attributes
A 32-bit RW trace register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCVICTLR bit assignments.
31 24 23 20 19 16 15 12 11 10 9 8 7 0

Reserved Reserved Event

EXLEVEL_S Reserved
EXLEVEL_NS SSSTATUS
TRCRESET
TRCERR

Figure 13-9 TRCVICTLR bit assignments

The following table shows the TRCVICTLR bit assignments.

Table 13-11 TRCVICTLR bit assignments

Bits Name Function

[31:24] - Reserved, RES0.

[23:20] EXLEVEL_NS Each bit controls whether instruction tracing in Non-secure state is enabled for the corresponding Exception
level. The bit to Exception level mapping is:
Bit[20] Exception level 0.
Bit[21] Exception level 1.
Bit[22] Exception level 2.
Bit[23] RES0.

For example, the value 0b0111 enables instruction tracing in Non-secure state for EL0, EL1, and EL2.

[19:16] EXLEVEL_S Each bit controls whether instruction tracing in Secure state is enabled for the corresponding Exception
level. The bit to Exception level mapping is:
Bit[16] Exception level 0.
Bit[17] Exception level 1.
Bit[18] RES0.

Bit[19] Exception level 3.

For example, the value 0b1011 enables instruction tracing in Secure state for EL0, EL1, and EL3.

[15:12] - Reserved, RES0.

[11] TRCERR Controls whether a trace unit must trace a System Error exception:
0 The trace unit does not trace a System Error exception unless it traces the exception or instruction
immediately prior to the System Error exception.
1 The trace unit always traces a System Error exception.

[10] TRCRESET Controls whether a trace unit must trace a reset exception:
0 The trace unit does not trace a reset exception unless it traces the exception or instruction immediately
prior to the reset exception.
1 The trace unit always traces a reset exception.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-477
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Table 13-11 TRCVICTLR bit assignments (continued)

Bits Name Function

[9] SSSTATUS Returns the status of the start and stop logic. The possible values are:
0 The start and stop logic is in the stopped state.
1 The start and stop logic is in the started state.

The bit only returns stable data when TRCSTATR.PMSTABLE is set to 1.

Before software enables the trace unit and TRCPRGCTLR.EN is set to 1, it must write to this bit to set the
initial state of the start and stop logic. If the start and stop logic is not used then set this bit to 1. ARM
recommends that the value of this bit is set before each trace run begins.

[8] - Reserved, RES0.

[7:0] Event Event selector.

The TRCVICTLR can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x080.

13.7.9 External Input Select Register

The TRCEXTINSELR characteristics are:

Purpose
Use this to set, or read, which external inputs are resources to the trace unit.
Usage constraints
Only accepts writes when the trace unit is disabled.
Configurations
The TRCIDR5.NUMEXTINSEL field controls how many input select resources are supported.
The TRCIDR5.NUMEXTIN field controls how many inputs, from a maximum of 256, are
supported.
Attributes
A 32-bit RW trace register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCEXTINSELR bit assignments.
31 24 23 16 15 8 7 0

SEL3 SEL2 SEL1 SEL0

Figure 13-10 TRCEXTINSELR bit assignments

The following table shows the TRCEXTINSELR bit assignments.

Table 13-12 TRCEXTINSELR bit assignments

Bits Name Function

[31:24] SEL3 A binary value that selects which external input is a resource for the trace unit. Bit[31] is reserved, RES0.

[23:16] SEL2 A binary value that selects which external input is a resource for the trace unit. Bit[23] is reserved, RES0.

[15:8] SEL1 A binary value that selects which external input is a resource for the trace unit. Bit[15] is reserved, RES0.

[7:0] SEL0 A binary value that selects which external input is a resource for the trace unit. Bit[7] is reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-478
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

The TRCEXTINSELR can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0x120.

13.7.10 ID Register 8

The TRCIDR8 characteristics are:

Purpose
Returns the maximum speculation depth of the instruction trace stream.
Usage constraints
There are no usage constraints.
Configurations
Available in all configurations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCIDR8 bit assignments.
31 0

MAXSPEC

Figure 13-11 TRCIDR8 bit assignments

The following table shows the TRCIDR8 bit assignments.

Table 13-13 TRCIDR8 bit assignments

Bits Name Function

[31:0] MAXSPEC The maximum number of P0 elements in the trace stream that can be speculative at any time:
0 Maximum speculation depth of the instruction trace stream.

The TRCIDR8 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x180.

13.7.11 ID Register 9

The TRCIDR9 characteristics are:

Purpose
Returns the number of P0 right-hand keys that the trace unit can use.
Usage constraints
There are no usage constraints.
Configurations
Available in all configurations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCIDR9 bit assignments.
31 0

NUMP0KEY

Figure 13-12 TRCID9 bit assignments

The following table shows the TRCIDR9 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-479
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Table 13-14 TRCID9 bit assignments

Bits Name Function

[31:0] NUMP0KEY The number of P0 right-hand keys that the trace unit can use:
0 The trace unit uses no P0 right-hand keys.

The TRCIDR9 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x184.

13.7.12 ID Register 10

The TRCIDR10 characteristics are:

Purpose
Returns the number of P1 right-hand keys that the trace unit can use.
Usage constraints
There are no usage constraints.
Configurations
Available in all configurations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCIDR10 bit assignments.
31 0

NUMP1KEY

Figure 13-13 TRCIDR10 bit assignments

The following table shows the TRCIDR10 bit assignments.

Table 13-15 TRCID10 bit assignments

Bits Name Function

[31:0] NUMP1KEY The number of P1 right-hand keys that the trace unit can use.
0 The trace unit uses no P1 right-hand keys.

The TRCIDR10 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x188.

13.7.13 ID Register 11

The TRCIDR11 characteristics are:

Purpose
Returns the number of special P1 right-hand keys that the trace unit can use.
Usage constraints
There are no usage constraints.
Configurations
Available in all configurations.
Attributes
See 13.6 Register summary on page 13-464.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-480
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

The following figure shows the TRCIDR11 bit assignments.

31 0

NUMP1SPC

Figure 13-14 TRCIDR11 bit assignments

The following table shows the TRCIDR11 bit assignments.

Table 13-16 TRCID11 bit assignments

Bits Name Function

[31:0] NUMP1SPC The number of special P1 right-hand keys that the trace unit can use:
0 The trace unit uses no special P1 right-hand keys.

The TRCIDR11 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x18C.

13.7.14 ID Register 12

The TRCIDR12 characteristics are:

Purpose
Returns the number of conditional instruction right-hand keys that the trace unit can use.
Usage constraints
There are no usage constraints.
Configurations
Available in all configurations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCIDR12 bit assignments.
31 0

NUMCONDKEY

Figure 13-15 TRCIDR12 bit assignments

The following table shows the TRCIDR12 bit assignments.

Table 13-17 TRCID12 bit assignments

Bits Name Function

[31:0] NUMCONDKEY The number of conditional instruction right-hand keys that the trace unit can use, including normal and
special keys:
0 The trace unit uses no conditional instruction right-hand keys.

The TRCIDR12 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x190.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-481
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

13.7.15 ID Register 13

The TRCIDR13 characteristics are:

Purpose
Returns the number of special conditional instruction right-hand keys that the trace unit can use.
Usage constraints
There are no usage constraints.
Configurations
Available in all configurations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCIDR13 bit assignments.
31 0

NUMCONDSPC

Figure 13-16 TRCIDR13 bit assignments

The following table shows the TRCIDR13 bit assignments.

Table 13-18 TRCID13 bit assignments

Bits Name Function

[31:0] NUMCONDSPC The number of special conditional instruction right-hand keys that the trace unit can use, including normal
and special keys:
0 The trace unit uses no special conditional instruction right-hand keys.

The TRCIDR13 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x194.

13.7.16 Implementation Defined Register 0

The TRCIMSPEC0 characteristics are:

Purpose
TRCIMSPEC0 is partially implemented for the future implementation of up to eight
IMPLEMENTATION DEFINED registers so that a debugger can implement a general mechanism for
detecting the IMPLEMENTATION DEFINED registers. This register must be implemented.
Usage constraints
There are no usage constraints.
Configurations
Available in all configurations.
Attributes
A 32-bit RW trace register. This register is reset by a trace unit reset.
See 13.6 Register summary on page 13-464>.
The following figure shows the TRCIMSPEC0 bit assignments.
31 8 7 4 3 0

Reserved EN SUPPORT

Figure 13-17 TRCIMSPEC0 bit assignments

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-482
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

The following table shows the TRCIMSPEC0 bit assignments.

Table 13-19 TRCIMSPEC0 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] EN EN is RES0 when the SUPPORT field is 0b0000.

[3:0] SUPPORT Indicates whether the implementation supports IMPLEMENTATION DEFINED features. This value is:
0b0000 No IMPLEMENTATION DEFINED features are supported.

The TRCIMSPEC0 can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0x1C0.

13.7.17 Trace ID Register 0

The TRCIDR0 characteristics are:

Purpose
Returns the tracing capabilities of the trace unit.
Usage constraints
There are no usage constraints.
Configurations
Available in all configurations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCIDR0 bit assignments.
31 30 29 28 24 23 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

TSSIZE Reserved

COMMOPT QSUPP Reserved

Reserved QFILT INSTP0
CONDTYPE TRCDATA
NUMEVENT TRCBB
RETSTACK TRCCOND
Reserved TRCCCI

Figure 13-18 TRCIDR0 bit assignments

The following table shows the TRCIDR0 bit assignments.

Table 13-20 TRCIDR0 bit assignments

Bits Name Function

[31:30] - Reserved, RES0.

[29] COMMOPT Commit mode field. This value is:

1 Commit mode 1.

[28:24] TSSIZE Global timestamp size field. This value is:

0b01000 Implementation supports a maximum global timestamp of 64 bits.

[23:17] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-483
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Table 13-20 TRCIDR0 bit assignments (continued)

Bits Name Function

[16:15] QSUPP Q element support field. This value is:

0b00 Q element support is not implemented. TRCCONFIGR is RES0.

[14] QFILT QFILT is RES0 when QSUPP is 0b00.

[13:12] CONDTYPE CONDTYPE is RES0 when TRCCOND is 0b0.

[11:10] NUMEVENT Number of events field. Indicates how many events the trace unit supports. This value is:
0b11 The trace unit supports 4 events.

[9] RETSTACK Return stack bit. Indicates whether the implementation supports a return stack. This value is:
1 Return stack is implemented. TRCCONFIGR.RS is supported.

[8] - Reserved, RES0.

[7] TRCCCI Cycle counting instruction bit. Indicates whether the trace unit supports cycle counting for instructions. This
value is:
1 Cycle counting in the instruction trace is implemented, therefore:
• TRCCONFIGR.CCI is supported.
• TRCCCTLR is supported.

[6] TRCCOND Conditional instruction tracing support bit. Indicates whether the trace unit supports conditional instruction
tracing. This value is:
0 Conditional instruction tracing is not supported.

[5] TRCBB Branch broadcast tracing support bit. Indicates whether the trace unit supports branch broadcast tracing. This
value is:
1 Branch broadcast tracing is supported, therefore:
• TRCCONFIGR.CCI is supported.
• TRCBBCTLR is supported.

[4:3] TRCDATA Conditional tracing field. This value is:

0b00 Data tracing is not supported.

[2:1] INSTP0 P0 tracing support field. This value is:

0b00 Tracing of load and store instructions as P0 elements is not supported.

[0] - Reserved, RES1.

The TRCIDR0 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x1E0.

13.7.18 Trace ID Register 1

The TRCIDR1 characteristics are:

Purpose
Returns the base architecture of the trace unit.
Usage constraints
There are no usage constraints.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-484
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Configurations
Available in all configurations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCIDR1 bit assignments.
31 24 23 16 15 12 11 8 7 4 3 0

Designer Reserved Reserved Revision

TRCARCHMAJ
TRCARCHMIN

Figure 13-19 TRCIDR1 bit assignments

The following table shows the TRCIDR1 bit assignments.

Table 13-21 TRCIDR1 bit assignments

Bits Name Function

[31:24] Designer Indicates which company designed the trace unit. The value is:
0x41 ARM

[23:16] - Reserved, RES0.

[15:12] - Reserved, RES1.

[11:8] TRCARCHMAJ Indicates the major version number of the trace unit architecture. The value is:
0x4 Indicates ETMv4.

All other values are reserved.

[7:4] TRCARCHMIN Identifies the minor version number of the trace unit architecture. The value is:
0x0 Identifies the minor version number of the trace unit architecture.

[3:0] Revision Identifies the revision of:

• The trace registers.
• The OS Lock registers.
0x1 Revision value.

The TRCIDR1 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x1E4.

13.7.19 Trace ID Register 2

The TRCIDR2 characteristics are:

Purpose
Returns the maximum size of the following parameters in the trace unit:
• Data value.
• Data address.
• VMID.
• Context ID.
• Instruction address.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-485
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Usage constraints
There are no usage constraints.
Configurations
Available in all configurations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCIDR2 bit assignments.
31 29 28 25 24 20 19 15 14 10 9 5 4 0

RES0 CCSIZE DVSIZE DASIZE VMIDSIZE CIDSIZE IASIZE

Figure 13-20 TRCIDR2 bit assignments

The following table shows the TRCIDR2 bit assignments.

Table 13-22 TRCIDR2 bit assignments

Bits Name Function

[31:29] - Reserved, RES0.

[28:25] CCSIZE Indicates the size of the cycle counter in bits minus 12. This value is:
0x0 The cycle counter is 12 bits in length.

[24:20] DVSIZE Indicates the data value size in bytes. This value is:
0x0 Data value tracing is not supported. TRCIDR0.TRCDATA must be 0b00.

[19:15] DASIZE Indicates the data address size in bytes. This value is:
0x0 Data address tracing is not supported. TRCIDR0.TRCDATA must be 0b00.

[14:10] VMIDSIZE Indicates the VMID size. This value is:

0x1 Maximum of 8-bit VMID size, therefore TRCCONFIGR.VMID is supported.

[9:5] CIDSIZE Indicates the Context ID size. This value is:

0x4 Maximum of 32-bit CID size, therefore TRCCONFIGR.CID is supported.

[4:0] IASIZE Indicates the instruction address size in bytes. This value is:
0x8 Maximum of 64-bit address size.

The TRCIDR2 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x1E8.

13.7.20 Trace ID Register 3

The TRCIDR3 characteristics are:
Purpose Indicates:
• Whether TRCVICTLR is supported.
• The number of cores available for tracing.
• If an Exception level supports instruction tracing.
• The minimum threshold value for instruction trace cycle counting.
• Whether the synchronization period is fixed.
• Whether TRCSTALLCTLR is supported and if so whether it supports trace
overflow prevention and supports stall control of the core.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-486
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Usage constraints There are no usage constraints.

Configurations Available in all configurations.
Attributes See 13.6 Register summary on page 13-464.

The following figure shows the TRCIDR3 bit assignments.

31 30 28 27 26 25 24 23 20 19 16 15 12 11 5 4 0

Reserved CCITMIN

EXLEVEL_S
EXLEVEL_NS
TRCERR
SYNCPR
STALLCTL
SYSTALL
NUMPROC
NOOVERFLOW

Figure 13-21 TRCIDR3 bit assignments

The following table shows the TRCIDR3 bit assignments.

Table 13-23 TRCIDR3 bit assignments

Bits Name Function

[31] NOOVERFLOW Indicates whether TRCSTALLCTLR.NOOVERFLOW is supported. This value is:

0 TRCSTALLCTLR.NOOVERFLOW is not supported. STALLCTL is 0.

[30:28] NUMPROC Indicates the number of cores available for tracing. This value is:
0b000 The trace unit can trace one core.

[27] SYSTALL Indicates whether stall control is supported. This value is:
0 The system does not support stall control of the core.

[26] STALLCTL Indicates whether TRCSTALLCTLR is supported. This value is:

0 TRCSTALLCTLR is not supported.

[25] SYNCPR Indicates whether there is a fixed synchronization period. This value is:
0 TRCSYNCPR is read-write so software can change the synchronization period.

[24] TRCERR Indicates whether TRCVICTLR.TRCERR is supported. This value is:

1 TRCVICTLR.TRCERR is supported.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-487
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Table 13-23 TRCIDR3 bit assignments (continued)

Bits Name Function

[23:20] EXLEVEL_NS Each bit controls whether instruction tracing in Non-secure state is supported for the corresponding
Exception level. The value is:
0b0111 Instruction tracing in Non-secure state is supported for EL0, EL1, and EL2.

Note
The bit to Exception level mapping is:
Bit[20] Exception level 0.
Bit[21] Exception level 1.
Bit[22] Exception level 2.
Bit[23] Always RES0.

[19:16] EXLEVEL_S Each bit controls whether instruction tracing in Secure state is supported for the corresponding Exception
level. The value is:
0b1011 Instruction tracing in Secure state is supported for EL0, EL1, and EL3.

Note
The bit to Exception level mapping is:
Bit[16] Exception level 0.
Bit[15] Exception level 1.
Bit[14] Always RES0.
Bit[13] Exception level 3.

[15:12] - Reserved, RES0.

[11:0] CCITMIN The minimum value that can be programmed in TRCCCCTLR.THRESHOLD. This value is:
0x004 Minimum value for cycle counting in the instruction trace.

The TRCIDR3 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x1EC.

13.7.21 Trace ID Register 4

The TRCIDR4 characteristics are:

Purpose Returns how many resources the trace unit supports.
Usage constraints There are no usage constraints.
Configurations Available in all configurations.
Attributes See 13.6 Register summary on page 13-464.

The following figure shows the TRCIDR4 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-488
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

31 28 27 24 23 20 19 16 15 12 11 9 8 7 4 3 0

NUMVMIDC NUMCIDC NUMSSCC NUMPC NUMDVC

NUMRSPAIR Reserved NUMASPAIRS

SUPPDAC

Figure 13-22 TRCIDR4 bit assignments

The following table shows the TRCIDR4 bit assignments.

Table 13-24 TRCIDR4 bit assignments

Bits Name Function

[31:28] NUMVMIDC Indicates the number of VMID comparators available for tracing. This value is:
0x1 One VMID comparator is available.

[27:24] NUMCIDC Indicates the number of CID comparators available for tracing. This value is:
0x1 One Context ID comparator is available.

[23:20] NUMSSCC Indicates the number of single-shot comparator controls available for tracing. This value is:
0x1 One single-shot comparator control is available.

[19:16] NUMRSPAIR Indicates the number of resource selection pairs available for tracing. This value is:
0x7 Eight resource selection pairs are available.

[15:12] NUMPC Indicates the number of processor comparator inputs available for tracing. This value is:
0x0 No processor comparator inputs are available.

[11:9] - Reserved, RES0.

[8] SUPPDAC Indicates whether the implementation supports data address comparisons: This value is:
0 Data address comparisons are not supported.

[7:4] NUMDVC Indicates the number of data value comparators available for tracing. This value is:
0x0 No data value comparators are available.

[3:0] NUMRSPAIRS Indicates the number of address comparator pairs available for tracing. This value is:
0x4 Four address comparator pairs are available.

The TRCIDR4 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x1F0.

13.7.22 Trace ID Register 5

The TRCIDR5 characteristics are:

Purpose
Returns how many resources the trace unit supports.
Usage constraints
There are no usage constraints.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-489
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Configurations
Available in all configurations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCIDR5 bit assignments.
31 30 28 27 25 24 23 22 21 16 15 12 11 9 8 0

TRACEIDSIZE Reserved NUMEXTIN

ATBTRIG NUMEXTINSEL
LPOVERRIDE
Reserved
NUMSEQSTATE
NUMCNTR
REDFUNCNTR

Figure 13-23 TRCIDR5 bit assignments

The following table shows the TRCIDR5 bit assignments.

Table 13-25 TRCIDR5 bit assignments

Bits Name Function

[31] REDFUNCNTR Indicates whether the reduced function counter is implemented. This value is:
0 Reduced function counter is not supported.

[30:28] NUMCNTR Indicates the number of counters available for tracing. This value is:
0b010 Two counters are available.

[27:25] NUMSEQSTATE Indicates the number of sequencer states implemented. This value is:
0b100 Four sequencer states are implemented.

[24] - Reserved, RES0.

[23] LPOVERRIDE Indicates whether low power state override is supported. This value is:
0 Low power state override is not supported.

[22] ATBTRIG Indicates whether ATB triggers are supported. This value is:
1 ATB triggers are supported and the TRCEVENTCTL1R.ATBTRIG field is implemented.

[21:16] TRACEIDSIZE Trace ID width. This value is:

0x07 A 7-bit trace ID width is supported. This defines the width of the TRCTRACEIDR.TRACEID
field.

Note
The CoreSight ATB requires a 7-bit trace ID width.

[15:12] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-490
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Table 13-25 TRCIDR5 bit assignments (continued)

Bits Name Function

[11:9] NUMEXTINSEL Indicates the number of external input select resources are implemented. If NUMEXTINSEL is 0,
NUMEXTIN must also be 0. This value is:
0b100 Four external input select resources are implemented.

[8:0] NUMEXTIN Indicates the number of external inputs are implemented. This value is:
0b001101110 110 external inputs are implemented.

The TRCIDR5 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0x1F4.

13.7.23 Resource Selection Control Registers

The TRCRSCTLRn characteristics are:

Purpose
Controls the selection of the resources in the trace unit.
Note
The range of n for TRCRSCTLRn is 2 to 15.

Usage constraints
• Only accepts writes when the trace unit is disabled.
• If software selects an non-implemented resource then constrained UNPREDICTABLE behavior of
the resource selector occurs. The resource selector might activate unexpectedly or might not
activate. Reads of the TRCRSCTLRn might return UNKNOWN.
Configurations
Resource selectors are implemented in pairs and there are eight pairs of TRCRSCTLR registers
implemented, set by TRCIDR4.NUMRSPAIR. Each odd numbered resource selector is part of a
pair with the even numbered resource selector that is numbered as one less than it. For example,
resource selectors 2 and 3 form a pair.
Resource selector pair 0 is always implemented and is reserved. Resource selector zero always
returns FALSE, and resource selector one always returns TRUE.
Attributes
A 32-bit RW trace register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCRSCTLRn bit assignments.
31 22 21 20 19 18 16 15 8 7 0

Reserved Group Reserved Select

Reserved
INV
PAIRINV

Figure 13-24 TRCSCTLR

The following table shows the TRCRSCTLRn bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-491
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Table 13-26 TRCSCTLRn bit assignments

Bits Name Function

[31:22] - Reserved, RES0.

[21] PAIRINV Controls whether the combined result from a resource pair is inverted when n is 2, 4, 6, 8, 10, 12, or 14. The
possible values are:
0 The combined result is not inverted.
1 The combined result is inverted.

PAIRINV is RES0 when n is 3, 5, 7, 9, 11, 13, or 15.

[20] INV Controls whether the resource, that GROUP and SELECT selects, is inverted. The possible values are:
0 The selected resource is not inverted.
1 The selected resource is inverted.

[19] - Reserved, RES0.

[18:16] Group Selects a group of resources. See the ARM® Embedded Trace Macrocell Architecture Specification, ETMv4 for
more information.

[15:8] - Reserved, RES0.

[7:0] Select Selects one or more resources from the group that the GROUP field selects. Each bit represents a resource from
the selected group. See the ARM® Embedded Trace Macrocell Architecture Specification, ETMv4 for more
information.

The TRCRSCTLRn can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0x208-023C.

13.7.24 Address Comparator Access Type Registers

The TRCACATRn characteristics are:

Purpose
Defines the type of access for the corresponding TRCACVRn Register. This register configures
the parameters of the address comparator for:
• Context type.
• Exception levels.
• Alignment.
• Masking.
• Behavior when it is one half of an address range comparator.

Note
The range of n for TRCACATRn is 0 to 7.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-492
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Usage constraints
• Only accepts writes when the trace unit is disabled.
• Constrained UNPREDICTABLE behavior of a comparator resource occurs if:
— TYPE is 0 and DATAMATCH is 0b01, 0b10, or 0b11.
— DATAMATCH is 0b01, 0b10, or 0b11 and software programs an address comparator to
control ViewData.
In these scenarios, the comparator might match unexpectedly or might not match.
• If software uses two single address comparators as an address range comparator then it must
program the corresponding TRCACATRs with identical values in the following fields:
— TYPE.
— CONTEXTTYPE.
— CONTEXT.
— EXLEVEL_S.
— EXLEVEL_NS.
Configurations
The number TRCACATRs is eight and is set by twice the size of TRCIDR4.NUMACPAIRS.
Attributes
A 64-bit RW trace register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCACATRn bit assignments.
63 16 15 12 11 8 7 4 3 2 1 0

Reserved Type

EXLEVEL_NS Context type

EXLEVEL_S Reserved

Figure 13-25 TRCACATR

The following table shows the TRCACATRn bit assignments.

Table 13-27 TRCACATRn bit assignments

Bits Name Function

[63:16] - Reserved, RES0.

[15:12] EXLEVEL_NS Each bit controls whether a comparison can occur in Non-secure state for the corresponding Exception
level. The possible values are:
0 The trace unit can perform a comparison, in Non-secure state, for Exception level n.
1 The trace unit does not perform a comparison, in Non-secure state, for Exception level n.

Note
The bit to Exception level mapping is:
Bit[12] Exception level 0.
Bit[13] Exception level 1.
Bit[14] Exception level 2.
Bit[15] Always RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-493
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Table 13-27 TRCACATRn bit assignments (continued)

Bits Name Function

[11:8] EXLEVEL_S Each bit controls whether a comparison can occur in Secure state for the corresponding Exception level.
The possible values are:
0 The trace unit can perform a comparison, in Secure state, for Exception level n.
1 The trace unit does not perform a comparison, in Secure state, for Exception level n.

Note
The bit to Exception level mapping is:
Bit[8] Exception level 0.
Bit[9] Exception level 1.
Bit[10] Always RES0.
Bit[11] Exception level 3.

[7:4] - Reserved, RES0.

[3:2] Context type Controls whether the trace unit performs a Context ID comparison, a VMID comparison, or both
comparisons:
0b00 The trace unit does not perform a Context ID comparison.
0b01 The trace unit performs a Context ID comparison using the Context ID comparator that the
CONTEXT field specifies, and signals a match if both the Context ID comparator matches and the
address comparator match.
0b10 The trace unit performs a VMID comparison using the VMID comparator that the CONTEXT field
specifies, and signals a match if both the VMID comparator and the address comparator match.
0b11 The trace unit performs a Context ID comparison and a VMID comparison using the comparators
that the CONTEXT field specifies, and signals a match if the Context ID comparator matches, the
VMID comparator matches, and the address comparator matches.

[1:0] Type Type is RES0 when TRCIDR4.SUPPDAC is 0b0.

The TRCACATRn can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0x480-0x4B8.

13.7.25 Context ID Comparator Value Register 0

The TRCCIDCVR0 characteristics are:

Purpose
Contains a Context ID value.
Usage constraints
Only accepts writes when the trace unit is disabled.
Configurations
There is one TRCCIDCVR register, set by TRCIDR4.NUMCIDC.
Attributes
A 64-bit RW trace register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCCIDCVR0 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-494
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

63 32 31 0

Reserved Context ID Value

Figure 13-26 TRCCIDCVR0 bit assignments

The following table shows the TRCCIDCVR0 bit assignments.

Table 13-28 TRCCIDCVR0 bit assignments

Bits Name Function

[63:32] - Reserved, RES0.

[31:0] Context ID Value The context ID value is 32 bits, set by TRCIDR2.CIDSIZE.

The TRCCIDCVR0 can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0x600.

13.7.26 VMID Comparator Value Register 0

The TRCVMIDCVR0 characteristics are:

Purpose
Contains a VMID value.
Usage constraints
Only accepts writes when the trace unit is disabled.
Configurations
There is one TRCVMIDCVR register, set by TRCIDR4.NUMVMIDC.
Attributes
A 64-bit RW trace register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCVMIDCVR0 bit assignments.
63 8 7 0

Reserved VMID Value

Figure 13-27 TRCVMIDCVR0 bit assignments

The following table shows the TRCVMIDCVR0 bit assignments.

Table 13-29 TRCVMIDCVR0 bit assignments

Bits Name Function

[63:8] - Reserved, RES0

[7:0] VMID Value The VMID value is 64 bits, set by TRCIDR2.VMIDSIZE

The TRCVMIDCVR0 can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0x640.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-495
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

13.7.27 Context ID Comparator Control Register 0

The TRCCIDCCTLR0 characteristics are:

Purpose
Contains Context ID mask values for the TRCCIDCVR0 register.
Usage constraints
• Only accepts writes when the trace unit is disabled.
• If software sets a mask bit to 1 then it must program the relevant byte in TRCCIDCVRn to
0x00.

Configurations
There is one Context ID comparator, set by TRCIDR4.NUMCIDC.
Attributes
A 32-bit RW trace register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCCIDCCTLR0 bit assignments.
31 4 3 0

Reserved COMP0

Figure 13-28 TRCCIDCCTLR0 bit assignments

The following table shows the TRCCIDCCTLR0 bit assignments.

Table 13-30 TRCCIDCCTLR0 bit assignments

Bits Name Function

[31:4] - Reserved, RES0.

[3:0] COMP0 Controls the mask value that the trace unit applies to TRCCIDCVR0. Each bit in this field corresponds to a byte in
TRCCIDCVR0. When a bit is:
0 The trace unit includes the relevant byte in TRCCIDCVR0 when it performs the Context ID comparison.
1 The trace unit ignores the relevant byte in TRCCIDCVR0 when it performs the Context ID comparison.

The TRCCIDCCTLR0 can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0x680.

13.7.28 Trace Integration Miscellaneous Outputs Register

The TRCITMISCOUT characteristics are:

Purpose
Controls signal outputs when TRCITCTRL.IME is set.
Usage constraints
There are no usage constraints.
Configurations
Available in all configurations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCITMISCOUT bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-496
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

31 12 11 8 7 0

Reserved Reserved

ETMEXTOUT[3:0]

Figure 13-29 TRCITMISCOUT bit assignments

The following table shows the TRCITMISCOUT bit assignments.

Table 13-31 TRCITMISCOUT bit assignments

Bits Name Function

[31:12] - Reserved, RES0.

[11:8] ETMEXTOUT[3:0] Drives the EXTOUT[3:0] outputs.

[7:0] - Reserved, RES0.

The TRCITMISCOUT can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0xEDC.

Related information
13.7.34 Trace Integration Mode Control register on page 13-501.

13.7.29 Trace Integration Miscellaneous Input Register

The TRCITMISCIN characteristics are:

Purpose
Enables the values of signal inputs to be read when TRCITCTRL.IME is set.
Usage constraints
There are no usage constraints.
Configurations
Available in all configurations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCITMISCIN bit assignments.
31 4 3 0

Reserved

ETMEXTIN[3:0]

Figure 13-30 TRCITMISCIN bit assignments

The following table shows the TRCITMISCIN bit assignments.

Table 13-32 TRCITMISCIN bit assignments

Bits Name Function

[31:4] - Reserved, RES0.

[3:0] ETMEXTIN[3:0] Returns the value of the ETMEXTIN[3:0] inputs.

The TRCITMISCIN can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0xEE0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-497
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Related information
13.7.34 Trace Integration Mode Control register on page 13-501.

13.7.30 Trace Integration Test ATB Data Register 0

The TRCITATBDATA0 characteristics are:

Purpose
Controls signal outputs when TRCITCTRL.IME is set.
Usage constraints
There are no usage constraints.
Configurations
Available in all configurations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCITATBDATA0 bit assignments.
31 5 4 3 2 1 0

RES0

ATDATAM[31]
ATDATAM[23]
ATDATAM[15]
ATDATAM[7]
ATDATAM[0]

Figure 13-31 TRCITATBDATA0 bit assignments

The following table shows the TRCITATBDATA0 bit assignments.

Table 13-33 TRCITATBDATA0 bit assignments

Bits Name Function

[31:5] - Reserved, RES0.

[4] ATDATAM[31] Drives the ATDATAM[31] output

[3] ATDATAM[23] Drives the ATDATAM[23] output

[2] ATDATAM[15] Drives the ATDATAM[15] output

[1] ATDATAM[7] Drives the ATDATAM[7] output

[0] ATDATAM[0] Drives the ATDATAM[0] output

The TRCITATBDATA0 can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0xEEC.

Related information
13.7.34 Trace Integration Mode Control register on page 13-501.

13.7.31 Trace Integration Test ATB Control Register 2

The TRCITATBCTR2 characteristics are:

Purpose
Enables the values of signal inputs to be read when bit[0] of the Integration Mode Control
Register is set.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-498
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Usage constraints
There are no usage constraints.
Configurations
Available in all configurations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCITATBCTR2 bit assignments.
31 2 1 0

Reserved

AFVALIDM
ATREADYM

Figure 13-32 TRCITATBCTR2 bit assignments

The following table shows the TRCITATBCTR2 bit assignments.

Table 13-34 TRCITATBCTR2 bit assignments

Bits Name Function

[31:2] - Reserved, RES0.

[1] AFVALIDM Returns the value of AFVALIDM input

[0] ATREADYM Returns the value of ATREADYM inputez

The TRCITATBCTR2 can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0xEF0.

Related information
13.7.34 Trace Integration Mode Control register on page 13-501.

13.7.32 Trace Integration Test ATB Control Register 1

The TRCITATBCTR1 characteristics are:

Purpose
Controls the ATIDM[6:0] signals when TRCITCTRL.IME is set.
Usage constraints
There are no usage constraints.
Configurations
Available in all configurations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCITATBCTR1 bit assignments.
31 7 6 0

Reserved ATIDM[6:0]

Figure 13-33 TRCITATBCTR1 bit assignments

The following table shows the TRCITATBCTR1 bit assignments.

ez To sample ATREADYM correctly from the processor signals, ATVALIDM must be asserted.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-499
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Table 13-35 TRCITATBCTR1 bit assignments

Bits Name Function

[31:7] - Reserved, RES0.

[6:0] ATIDM[6:0] Drives the ATIDM[6:0] outputs

The TRCITATBCTR2 can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0xEF4.

Related information
13.7.34 Trace Integration Mode Control register on page 13-501.

13.7.33 Trace Integration Test ATB Control Register 0

The TRCITATBCTR0 characteristics are:

Purpose
Controls signal outputs when TRCITCTRL.IME is set.
Usage constraints
There are no usage constraints.
Configurations
Available in all configurations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCITATBCTR0 bit assignments.
31 10 9 8 7 2 1 0

RES0 RES0

ATBYTESM[1:0] AFREADYM
ATVALIDM

Figure 13-34 TRCITATBCTR0 bit assignments

The following table shows the TRCITATBCTR0 bit assignments.

Table 13-36 TRCITATBCTR0 bit assignments

Bits Name Function

[31:10] - Reserved, RES0

[9:8] ATBYTESM[1:0] Drives the ATBYTESM outputs

[7:2] - Reserved, RES0

[1] AFREADYM Drives the AFREADYM output

[0] ATVALIDM Drives the ATVALIDM output

The TRCITATBCTR0 can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0xEF8.

Related information
13.7.34 Trace Integration Mode Control register on page 13-501.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-500
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

13.7.34 Trace Integration Mode Control register

The TRCITCTRL characteristics are:

Purpose
Controls whether the trace unit is in integration mode.
Usage constraints
• Accessible only from the memory-mapped interface or from an external agent such as a
debugger.
• If the IME bit changes from one to zero then ARM recommends that the trace unit is reset.
Otherwise the trace unit might generate incorrect or corrupt trace and the trace unit resources
might behave unexpectedly.
Configurations
Available in all configurations.
Attributes
A 32-bit RW management register. The register is reset to zero.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCITCTRL bit assignments.
31 1 0

Reserved

IME

Figure 13-35 TRCITCTRL bit assignments

The following table shows the TRCITCTRL bit assignments.

Table 13-37 TRCITCTRL bit assignments

Bits Name Function

[31:1] - Reserved, RES0

[0] IME Integration mode enable bit. The possible values are:
0 The trace unit is not in integration mode.
1 The trace unit is in integration mode. This mode enables:
• A debug agent to perform topology detection.
• SoC test software to perform integration testing.

The TRCITCTRL can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xF00.

13.7.35 Trace Device Affinity register 0

The TRCDEVAFF0 characteristics are:

Purpose
The value is a read-only copy of MPIDR_EL1[31:0] as seen from EL3, unaffected by
VMPIDR_EL2.
Usage constraints
Accessible only from the memory-mapped interface or from an external agent such as a
debugger.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-501
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Configurations
Available in all configurations.
Attributes
A 32-bit RO management register.
For the Cortex-A72 processor, MPIDR_EL1[31:0] is architecturally mapped to the AArch32
register MPIDR.
See 13.6 Register summary on page 13-464.
The TRCDEVAFF0 can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0xFA8.

Related information
4.5.3 Multiprocessor Affinity Register on page 4-240.

13.7.36 Trace Device Affinity register 1

The TRCDEVAFF1 characteristics are:

Purpose
The value is a read-only copy of MPIDR_EL1[63:32] as seen from EL3, unaffected by
VMPIDR_EL2.
Usage constraints
Accessible only from the memory-mapped interface or from an external agent such as a
debugger.
Configurations
Available in all configurations.
Attributes
A 32-bit RO management register.
For the Cortex-A72 processor, MPIDR_EL1[63:32] is RES0.
See 13.6 Register summary on page 13-464.
The TRCDEVAFF1 can be accessed through the internal memory-mapped interface and the external
debug interface, offset 0xFAC.

13.7.37 Trace Peripheral Identification Registers

The Trace Peripheral Identification Registers provide standard information required for all CoreSight
components. There is a set of eight registers, listed in register number order in the following table.

Table 13-38 Summary of the Trace Peripheral ID Registers

Register Value Offset

TRCPIDR4 0x04 0xFD0

TRCPIDR5 0x00 0xFD4

TRCPIDR6 0x00 0xFD8

TRCPIDR7 0x00 0xFDC

TRCPIDR0 0x5A 0xFE0

TRCPIDR1 0xB9 0xFE4

TRCPIDR2 0x0B 0xFE8

TRCPIDR3 0x00 0xFEC

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-502
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Only bits[7:0] of each Trace Peripheral ID Register are used, with bits[31:8] reserved. Together, the eight
Trace Peripheral ID Registers define a single 64-bit Peripheral ID.
The Peripheral ID registers are:
• Trace Peripheral Identification Register 0 on page 13-503.
• Trace Peripheral Identification Register 1 on page 13-503.
• Trace Peripheral Identification Register 2 on page 13-504.
• Trace Peripheral Identification Register 3 on page 13-505.
• Trace Peripheral Identification Register 4 on page 13-505.
• Trace Peripheral Identification Register 5-7 on page 13-506.

Trace Peripheral Identification Register 0

The TRCPIDR0 characteristics are:
Purpose
Provides information to identify a trace component.
Usage constraints
• Only bits[7:0] are valid.
• Accessible only from the memory-mapped interface or the external debugger interface.
Configurations
Available in all implementations.
Attributes
A 32-bit RO management register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCPIDR0 bit assignments.
31 8 7 0

Reserved Part_0

Figure 13-36 TRCPIDR0 bit assignments

The following table shows the TRCPIDR0 bit assignments.

Table 13-39 TRCPIDR0 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:0] Part_0 0x5A Least significant byte of the ETM part number.

TRCPIDR0 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFE0.

Trace Peripheral Identification Register 1

The TRCPIDR1 characteristics are:
Purpose
Provides information to identify a trace component.
Usage constraints
• Only bits[7:0] are valid.
• Accessible only from the memory-mapped interface or the external debugger interface.
Configurations
Available in all implementations.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-503
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Attributes
A 32-bit RO management register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCPIDR1 bit assignments.
31 8 7 4 3 0

Reserved DES_0 Part_1

Figure 13-37 TRCPIDR1 bit assignments

The following table shows the TRCPIDR1 bit assignments.

Table 13-40 TRCPIDR1 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] DES_0 0xB ARM Limited. This is the least significant nibble of JEP106 ID code.

[3:0] Part_1 0x9 Most significant nibble of the ETM part number.

TRCPIDR1 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFE4.

Trace Peripheral Identification Register 2

The TRCPIDR2 characteristics are:
Purpose
Provides information to identify a trace component.
Usage constraints
• Only bits[7:0] are valid.
• Accessible only from the memory-mapped interface or the external debugger interface.
Configurations
Available in all implementations.
Attributes
A 32-bit RO management register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCPIDR2 bit assignments.
31 8 7 4 3 2 0

Reserved Revision DES_1

JEDEC

Figure 13-38 TRCPIDR2 bit assignments

The following table shows the TRCPIDR2 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-504
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Table 13-41 TRCPIDR2 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] Revision 0x1 Part major revision.

[3] JEDEC 0b1 RES1. Indicates a JEP106 identity code is used.

[2:0] DES_1 0b011 ARM Limited. This is the most significant nibble of JEP106 ID code.

TRCPIDR2 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFE8.

Trace Peripheral Identification Register 3

The TRCPIDR3 characteristics are:
Purpose
Provides information to identify a trace component.
Usage constraints
• Only bits[7:0] are valid.
• Accessible only from the memory-mapped interface or the external debugger interface.
Configurations
Available in all implementations.
Attributes
A 32-bit RO management register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCPIDR3 bit assignments.
31 8 7 4 3 0

Reserved REVAND CMOD

Figure 13-39 TRCPIDR3 bit assignments

The following table shows the TRCPIDR3 bit assignments.

Table 13-42 TRCPIDR3 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] REVAND 0x0 Part minor revision.

[3:0] CMOD 0x0 Customer modified.

TRCPIDR3 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFEC.

Trace Peripheral Identification Register 4

The TRCPIDR4 characteristics are:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-505
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Purpose
Provides information to identify a trace component.
Usage constraints
• Only bits[7:0] are valid.
• Accessible only from the memory-mapped interface or the external debugger interface.
Configurations
Available in all implementations.
Attributes
A 32-bit RO management register.
See 13.6 Register summary on page 13-464.
The following figure shows the TRCPIDR4 bit assignments.
31 8 7 4 3 0

Reserved Size DES_2

Figure 13-40 TRCPIDR4 bit assignments

The following table shows the TRCPIDR4 bit assignments.

Table 13-43 TRCPIDR4 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] Size 0x0 Size of the component. Log2 the number of 4KB pages from the start of the component to the end of the ETM
component ID registers.

[3:0] DES_2 0x4 ARM Limited. This is the least significant nibble of the JEP106 continuation code.

TRCPIDR4 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFD0.

Trace Peripheral Identification Register 5-7

No information is held in the Peripheral ID5, Peripheral ID6 and Peripheral ID7 Registers.
They are reserved for future use and are RES0.

13.7.38 Trace Component Identification Registers

There are four read-only Trace Component Identification Registers, Trace Component ID0 to Trace
Component ID3. The following table shows these registers.

Table 13-44 Summary of the ETM Component Identification Registers

Register Value Offset

TRCCIDR0 0x0D 0xFF0

TRCCIDR1 0x90 0xFF4

TRCCIDR2 0x05 0xFF8

TRCCIDR3 0xB1 0xFFC

The Trace Component Identification Registers identify ETM as a CoreSight component.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-506
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

The Trace Component ID registers are:

• Trace Component Identification Register 0 on page 13-507.
• Trace Component Identification Register 1 on page 13-507.
• Trace Component Identification Register 2 on page 13-508.
• Trace Component Identification Register 3 on page 13-508.

Trace Component Identification Register 0

The TRCCIDR0 characteristics are:

Purpose
Provides information to identify a trace component.
Usage constraints
• Only bits[7:0] are valid.
• Accessible only from the memory-mapped interface or the external debugger interface.
Configurations
Available in all implementations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCCIDR0 bit assignments.
31 8 7 0

Reserved PRMBL_0

Figure 13-41 TRCCIDR0 bit assignments

The following table shows the TRCCIDR0 bit assignments.

Table 13-45 TRCCIDR0 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:0] PRMBL_0 0x0D Preamble byte 0.

TRCCIDR0 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFF0.

Trace Component Identification Register 1

The TRCCIDR1 characteristics are:
Purpose
Provides information to identify a trace component.
Usage constraints
• Only bits[7:0] are valid.
• Accessible only from the memory-mapped interface or the external debugger interface.
Configurations
Available in all implementations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCCIDR1 bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-507
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

31 8 7 4 3 0

Reserved CLASS PRMBL_1

Figure 13-42 TRCCIDR1 bit assignments

The following table shows the TRCCIDR1 bit assignments.

Table 13-46 TRCCIDR1 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] CLASS 0x9 Debug component.

[3:0] PRMBL_1 0x0 Preamble.

TRCCIDR1 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFF4.

Trace Component Identification Register 2

The TRCCIDR2 characteristics are:
Purpose
Provides information to identify a CTI component.
Usage constraints
• Only bits[7:0] are valid.
• Accessible only from the memory-mapped interface or the external debugger interface.
Configurations
Available in all implementations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCCIDR2 bit assignments.
31 8 7 0

Reserved PRMBL_2

Figure 13-43 TRCCIDR2 bit assignments

The following table shows the TRCCIDR2 bit assignments.

Table 13-47 TRCCIDR2 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:0] PRMBL_2 0x05 Preamble byte 2.

TRCCIDR2 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFF8.

Trace Component Identification Register 3

The TRCCIDR3 characteristics are:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-508
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.7 Register descriptions

Purpose
Provides information to identify a trace component.
Usage constraints
• Only bits[7:0] are valid.
• Accessible only from the memory-mapped interface or the external debugger interface.
Configurations
Available in all implementations.
Attributes
See 13.6 Register summary on page 13-464.
The following figure shows the TRCCIDR3 bit assignments.
31 8 7 0

Reserved PRMBL_3

Figure 13-44 TRCCIDR3 bit assignments

The following table shows the TRCCIDR3 bit assignments.

Table 13-48 TRCCIDR3 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:0] PRMBL_3 0xB1 Preamble byte 3.

TRCCIDR3 can be accessed through the internal memory-mapped interface and the external debug
interface, offset 0xFFC.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-509
reserved.
Non-Confidential
 13 Embedded Trace Macrocell
13.8 Interaction with debug and the Performance Monitor Unit

13.8 Interaction with debug and the Performance Monitor Unit

Describes the interaction with the Performance Monitor Unit and debug.
This section contains the following subsections:
• 13.8.1 Interaction with the Performance Monitor Unit on page 13-510.
• 13.8.2 Effect of debug double lock on trace register access on page 13-510.

13.8.1 Interaction with the Performance Monitor Unit

The Cortex-A72 processor includes a Performance Monitor Unit (PMU) that enables events, such as
cache misses and instructions executed, to be counted over a period of time. This section describes how
the PMU and ETM function together.

Use of PMU events by the ETM

All PMU architectural events are available to the ETM through the extended input facility. See the ARM®
Architectural Reference Manual ARMv8 for more information about PMU events.
The ETM uses four extended external input selectors to access the PMU events. Each selector can
independently select one of the PMU events, that are then active for the cycles where the relevant events
occur. These selected events can then be accessed by any of the event registers within the ETM.

Related information
Chapter 11 Performance Monitor Unit on page 11-395.

13.8.2 Effect of debug double lock on trace register access

All trace register accesses through the memory-mapped and external debug interfaces behave as if the
processor power domain is powered down when debug double lock is set. For more information on
debug double lock, see the ARM® Architecture Reference Manual ARMv8.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 13-510
reserved.
Non-Confidential
Chapter 14
Advanced SIMD and Floating-point

This chapter describes the Advanced SIMD and Floating-point features and registers in the Cortex-A72
processor.

It contains the following sections:

• 14.1 About Advanced SIMD and Floating-point on page 14-512.
• 14.2 Programmers model for Advanced SIMD and Floating-point on page 14-513.
• 14.3 AArch64 register summary on page 14-514.
• 14.4 AArch64 register descriptions on page 14-515.
• 14.5 AArch32 register summary on page 14-524.
• 14.6 AArch32 register descriptions on page 14-525.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-511
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.1 About Advanced SIMD and Floating-point

14.1 About Advanced SIMD and Floating-point

The Cortex-A72 processor supports the Advanced SIMD and Floating-point instructions in the A64,
A32, and T32 instruction sets.
The ARMv8-A architecture eliminates the concept of version numbers for Advanced SIMD and
Floating-point in AArch64 state because the instructions are always implicitly present.
This section contains the following subsections:
• 14.1.1 Advanced SIMD support on page 14-512.
• 14.1.2 Floating-point support on page 14-512.

14.1.1 Advanced SIMD support

The Cortex-A72 processor supports all addressing modes, data types, and operations of the Advanced
SIMD instructions.

14.1.2 Floating-point support

The Cortex-A72 processor supports all addressing modes, data types, and operations of the Floating-
point instructions. It does not support floating-point exception trapping.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-512
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.2 Programmers model for Advanced SIMD and Floating-point

14.2 Programmers model for Advanced SIMD and Floating-point

Software can identify the Cortex-A72 processor Advanced SIMD and Floating-point features by using
the feature identification registers in the AArch64 and AArch32 states.
You can access the feature identification registers in AArch64 state using the MRS instructions, for
example:
MRS <Xt>, MVFR0_EL1 ; Read MVFR0_EL1 into Xt
MRS <Xt>, MVFR1_EL1 ; Read MVFR1_EL1 into Xt
MRS <Xt>, MVFR2_EL1 ; Read MVFR2_EL1 into Xt

You can access the feature identification registers in AArch32 state using the VMRS instruction, for
example:
VMRS <Rt>, FPSID ; Read FPSID into Rt
VMRS <Rt>, MVFR0 ; Read MVFR0 into Rt
VMRS <Rt>, MVFR1 ; Read MFFR1 into Rt
VMRS <Rt>, MVFR2 ; Read MVFR2 into Rt

The following table lists the feature identification registers for the Advanced SIMD and Floating-point.

Table 14-1 Advanced SIMD and Floating-point feature identification registers

AArch64 name AArch32 name Description

- FPSID See 14.6.1 Floating-point System ID Register on page 14-525

MVFR0_EL1 MVFR0 See 14.4.3 Media and VFP Feature Register 0, EL1 on page 14-518

MVFR1_EL1 MVFR1 See 14.4.4 Media and VFP Feature Register 1, EL1 on page 14-519

MVFR2_EL1 MVFR2 See 14.4.5 Media and VFP Feature Register 2, EL1 on page 14-520

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-513
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.3 AArch64 register summary

14.3 AArch64 register summary

The following table gives a summary of the Cortex-A72 processor Advanced SIMD and Floating-point
System registers in AArch64 state. All AArch64 registers are 32-bit wide.

Table 14-2 AArch64 Advanced SIMD and Floating-point System registers

Name Type Reset Description

FPCR RW 0x00000000 See 14.4.1 Floating-point Control Register on page 14-515

FPSR RW 0x00000000 See 14.4.2 Floating-point Status Register on page 14-516

MVFR0_EL1 RO 0x10110222 See 14.4.3 Media and VFP Feature Register 0, EL1 on page 14-518

MVFR1_EL1 RO 0x12111111 See 14.4.4 Media and VFP Feature Register 1, EL1 on page 14-519

MVFR2_EL1 RW 0x00000043 See 14.4.5 Media and VFP Feature Register 2, EL1 on page 14-520

FPEXC32_EL2 RW 0x00000700 See 14.4.6 Floating-point Exception Control Register 32, EL2 on page 14-521

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-514
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.4 AArch64 register descriptions

14.4 AArch64 register descriptions

This section describes the AArch64 Advanced SIMD and Floating-point System registers in the Cortex-
A72 processor.
Table 14-2 AArch64 Advanced SIMD and Floating-point System registers on page 14-514 provides
cross-references to individual registers.
This section contains the following subsections:
• 14.4.1 Floating-point Control Register on page 14-515.
• 14.4.2 Floating-point Status Register on page 14-516.
• 14.4.3 Media and VFP Feature Register 0, EL1 on page 14-518.
• 14.4.4 Media and VFP Feature Register 1, EL1 on page 14-519.
• 14.4.5 Media and VFP Feature Register 2, EL1 on page 14-520.
• 14.4.6 Floating-point Exception Control Register 32, EL2 on page 14-521.

14.4.1 Floating-point Control Register

The FPCR characteristics are:

Purpose
Controls floating-point extension behavior.
Usage constraints
The accessibility to the FPCR by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

RW RW RW RW RW RW

Configurations
The FPCR is part of the Floating-point functional group.
The named fields in this register map to the equivalent fields in the AArch32 FPSCR.
Attributes
See the register summary in Table 14-2 AArch64 Advanced SIMD and Floating-point System
registers on page 14-514.
The following figure shows the FPCR bit assignments.
31 27 26 25 24 23 22 21 0

RES0 RES0

AHP RMode
DN FZ

Figure 14-1 FPCR bit assignments

The following table shows the FPCR bit assignments.

Table 14-3 FPCR bit assignments

Bits Name Function

[31:27] - Reserved, RES0.

[26] AHP Alternative half-precision control bit:

0 IEEE half-precision format selected.
1 Alternative half-precision format selected.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-515
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.4 AArch64 register descriptions

Table 14-3 FPCR bit assignments (continued)

Bits Name Function

[25] DN Default NaN mode control bit:

0 NaN operands propagate through to the output of a floating-point operation.
1 Any operation involving one or more NaNs returns the Default NaN.

[24] FZ Flush-to-zero mode control bit:

0 Flush-to-zero mode disabled. Behavior of the floating-point system is fully compliant with the IEEE 754
standard.
1 Flush-to-zero mode enabled.

[23:22] RMode Rounding Mode control field:

0b00 Round to Nearest (RN) mode.
0b01 Round towards Plus Infinity (RP) mode.
0b10 Round towards Minus Infinity (RM) mode.
0b11 Round towards Zero (RZ) mode.

[21:0] - Reserved, RES0.

To access FPCR in AArch64 state, read or write the register with:

MRS <Xt>, FPCR; Read Floating-point Control Register
MSR FPCR, <Xt>; Write Floating-point Control Register

14.4.2 Floating-point Status Register

The FPSR characteristics are:

Purpose
Provides floating-point system status information.
Usage constraints
The accessibility to the FPSR by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

RW RW RW RW RW RW

Configurations
The FPSR is part of the Floating-point functional group.
The named fields in this register map to the equivalent fields in the AArch32 FPSCR.
Attributes
See the register summary in Table 14-2 AArch64 Advanced SIMD and Floating-point System
registers on page 14-514.
The following figure shows the FPSR bit assignments.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-516
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.4 AArch64 register descriptions

31 30 29 28 27 26 8 7 6 5 4 3 2 1 0

N Z C V RES0 RES0

QC IDC IXC IOC

UFC DZC
OFC

Figure 14-2 FPSR bit assignments

The following table shows the FPSR bit assignments.

Table 14-4 FPSR bit assignments

Bits Name Function

[31] N Negative condition flag for floating-point comparison operations:

AArch32 Negative condition flag.
AArch64 Sets the N bit in the main processor state (PSTATE) condition code flag.

[30] Z Zero condition flag for floating-point comparison operations:

AArch32 Zero condition flag.
AArch64 Sets the PSTATE.Z condition code flag.

[29] C Carry condition flag for floating-point comparison operations:

AArch32 Carry condition flag.
AArch64 Sets the PSTATE.C condition code flag.

[28] V Overflow condition flag for floating-point comparison operations:

AArch32 Overflow condition flag.
AArch64 Sets the PSTATE.V condition code flag.

[27] QC Cumulative saturation bit, Advanced SIMD only. This bit is set to 1 to indicate that an Advanced SIMD integer
operation has saturated since 0 was last written to this bit.

[26:8] - Reserved, RES0.

[7] IDC Input Denormal cumulative exception bit. This bit is set to 1 to indicate that the Input Denormal exception has
occurred since 0 was last written to this bit.

[6:5] - Reserved, RES0.

[4] IXC Inexact cumulative exception bit. This bit is set to 1 to indicate that the Inexact exception has occurred since 0 was
last written to this bit.

[3] UFC Underflow cumulative exception bit. This bit is set to 1 to indicate that the Underflow exception has occurred since 0
was last written to this bit.

[2] OFC Overflow cumulative exception bit. This bit is set to 1 to indicate that the Overflow exception has occurred since 0
was last written to this bit.

[1] DZC Division by Zero cumulative exception bit. This bit is set to 1 to indicate that the Division by Zero exception has
occurred since 0 was last written to this bit.

[0] IOC Invalid Operation cumulative exception bit. This bit is set to 1 to indicate that the Invalid Operation exception has
occurred since 0 was last written to this bit.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-517
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.4 AArch64 register descriptions

To access FPSR in AArch64 state, read or write the register with:

MRS <Xt>, FPSR; Read Floating-point Status Register
MSR FPSR, <Xt>; Write Floating-point Status Register

14.4.3 Media and VFP Feature Register 0, EL1

The MVFR0_EL1 characteristics are:

Purpose
The MVFR0_EL1 must be interpreted with the MVFR1_EL1 and the MVFR2_EL1 to describe
the features provided by the Advanced SIMD and FP functions.
Usage constraints
The accessibility to the MVFR0_EL1 in AArch64 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

The accessibility to the MVFR0 in AArch32 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- Config RO Config Config RO

Configurations
MVFR0_EL1 is:
• Common to Secure and Non-secure states
• Architecturally mapped to AArch32 MVFR0 register.
Attributes
See the register summary in Table 14-2 AArch64 Advanced SIMD and Floating-point System
registers on page 14-514.
The following figure shows the MVFR0_EL1 bit assignments.
31 28 27 24 23 20 19 16 15 12 11 8 7 4 3 0

FPRound FPShVec FPSqrt FPDivide FPTrap FPDP FPSP SIMDReg

Figure 14-3 MVFR0_EL1 bit assignments

The following table shows the MVFR0_EL1 bit assignments.

Table 14-5 MVFR0_EL1 bit assignments

Bits Name Function

[31:28] FPRound Indicates the rounding modes supported by the FP floating-point hardware:
0x1 All rounding modes supported.

[27:24] FPShVec Indicates the hardware support for FP short vectors:

0x0 Not supported.

[23:20] FPSqrt Indicates the hardware support for FP square root operations:
0x1 Supported.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-518
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.4 AArch64 register descriptions

Table 14-5 MVFR0_EL1 bit assignments (continued)

Bits Name Function

[19:16] FPDivide Indicates the hardware support for FP divide operations:

0x1 Supported.

[15:12] FPTrap Indicates whether the FP hardware implementation supports exception trapping:
0x0 Not supported.

[11:8] FPDP Indicates the hardware support for FP double-precision operations:

0x2 Supported, VFPv3 or greater.

See the ARM® Architecture Reference Manual ARMv8 for more information.

[7:4] FPSP Indicates the hardware support for FP single-precision operations:

0x2 Supported, VFPv3 or greater.

See the ARM® Architecture Reference Manual ARMv8 for more information.

[3:0] SIMDReg Indicates support for the Advanced SIMD register bank:
0x2 32× 64-bit registers supported.

See the ARM® Architecture Reference Manual ARMv8 for more information.

To access the MVFR0_EL1 register, see 14.2 Programmers model for Advanced SIMD and Floating-
point on page 14-513.

14.4.4 Media and VFP Feature Register 1, EL1

The MVFR1_EL1 characteristics are:

Purpose
The MVFR1_EL1 must be interpreted with the MVFR0_EL1 and the MVFR2_EL1 to describe
the features provided by the Advanced SIMD and FP functions.
Usage constraints
The accessibility to the MVFR1_EL1 in AArch64 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

The accessibility to the MVFR1 in AArch32 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- Config RO Config Config RO

Configurations
The MVFR1_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to AArch32 MVFR1 register.
Attributes
See the register summary in Table 14-2 AArch64 Advanced SIMD and Floating-point System
registers on page 14-514.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-519
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.4 AArch64 register descriptions

The following figure shows the MVFR1_EL1 bit assignments.

31 28 27 24 23 20 19 16 15 12 11 8 7 4 3 0

SIMDFMAC FPHP SIMDHP SIMDSP SIMDInt SIMDLS FPDNaN FPFtZ

Figure 14-4 MVFR1_EL1 bit assignments

The following table shows the MVFR1_EL1 bit assignments.

Table 14-6 MVFR1_EL1 bit assignments

Bits Name Function

[31:28] SIMDFMAC Indicates whether the Advanced SIMD or FP supports fused multiply accumulate operations:
0x1 Supported.

[27:24] FPHP Indicates whether the FP supports half-precision floating-point conversion operations:
0x2 Supported.

[23:20] SIMDHP Indicates whether the Advanced SIMD supports half-precision floating-point conversion operations:
0x1 Supported.

[19:16] SIMDSP Indicates whether the Advanced SIMD supports single-precision floating-point operations:
0x1 Supported.

[15:12] SIMDInt Indicates whether the Advanced SIMD supports integer operations:
0x1 Supported.

[11:8] SIMDLS Indicates whether the Advanced SIMD supports load/store instructions:
0x1 Supported.

[7:4] FPDNaN Indicates whether the FP hardware implementation supports only the Default NaN mode:
0x1 Hardware supports propagation of NaN values.

[3:0] FPFtZ Indicates whether the FP hardware implementation supports only the Flush-to-zero mode of operation:
0x1 Hardware supports full denormalized number arithmetic.

To access the MVFR1_EL1 register, see 14.2 Programmers model for Advanced SIMD and Floating-
point on page 14-513.

14.4.5 Media and VFP Feature Register 2, EL1

The MVFR2_EL1 characteristics are:

Purpose
The MVFR2_EL1 must be interpreted with the MVFR0_EL1 and the MVFR1_EL1 to describe
the features provided by the Advanced SIMD and FP functions.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-520
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.4 AArch64 register descriptions

Usage constraints
The accessibility to the MVFR2_EL1 in AArch64 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- RO RO RO RO RO

The accessibility to the MVFR2 in AArch32 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- Config RO Config Config RO

Configurations
The MVFR2_EL1 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to AArch32 MVFR2 register.
Attributes
See the register summary in Table 14-2 AArch64 Advanced SIMD and Floating-point System
registers on page 14-514.
The following figure shows the MVFR2_EL1 bit assignments.
31 8 7 4 3 0

Reserved FPMisc SIMDMisc

Figure 14-5 MVFR2_EL1 bit assignments

The following table shows the MVFR2_EL1 bit assignments.

Table 14-7 MVFR2_EL1 bit assignments

Bits Name Function

[31:8] - Reserved, RES0.

[7:4] FPMisc Floating-point miscellaneous features supported. This value is:

0b0100 Includes support for the following features:
• Floating-point selection.
• Floating-point Conversion to Integer with Directed Rounding modes.
• Floating-point Round to Integral floating-point.
• Floating-point MaxNum and MinNum.

[3:0] SIMDMisc Advanced SIMD miscellaneous features supported. This value is:
0b011 Includes support for the following features:
• Floating-point Conversion to Integer with Directed Rounding modes.
• Floating-point Round to Integral floating-point.
• Floating-point MaxNum and MinNum.

To access the MVFR2_EL1 register, see 14.2 Programmers model for Advanced SIMD and Floating-
point on page 14-513.

14.4.6 Floating-point Exception Control Register 32, EL2

The FPEXC32_EL2 characteristics are:

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-521
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.4 AArch64 register descriptions

Purpose
Provides access to the AArch32 register FPEXC from AArch64 state only. Its value has no
effect on execution in AArch64 state.
Usage constraints
The accessibility to the FPEXC32_EL2 in AArch64 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- - - RW RW RW

The accessibility to the FPEXC in AArch32 state by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- Config RW Config Config RW

Configurations
The FPEXC32_EL2 is:
• Common to Secure and Non-secure states.
• Architecturally mapped to AArch32 FPEXC register.
Attributes
See the register summary in Table 14-2 AArch64 Advanced SIMD and Floating-point System
registers on page 14-514.
The following figure shows the FPEXC32_EL2 bit assignments.
31 30 29 11 10 8 7 0

EX EN RES0 RES1 RES0

Figure 14-6 FPEXC32_EL2 bit assignments

The following table shows the FPEXC32_EL2 bit assignments.

Table 14-8 FPEXC32_EL2 bit assignments

Bits Name Function

[31] EX Exception bit. The Cortex-A72 processor implementation does not generate asynchronous
FP exceptions, so this bit is RES0.

[30] EN bit, VFPdescriptionVFP Enable bit. A global enable for the Advanced SIMD and FP functions:
enable bitdescriptionEN
0 The Advanced SIMD and FP functions are disabled.
1 The Advanced SIMD and FP functions are enabled and operate normally.

The EN bit is cleared at reset. See the ARM® Architecture Reference Manual ARMv8 for
more information.

[29:11] - Reserved, RES0.

[10:8] - Reserved, RES1.

[7:0] - Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-522
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.4 AArch64 register descriptions

To access the FPEXC_EL2 register, see 14.2 Programmers model for Advanced SIMD and Floating-
point on page 14-513.
Note
The Cortex-A72 processor implementation does not support deprecated FP short vector feature. You can
use software to emulate the short vector feature, if required.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-523
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.5 AArch32 register summary

14.5 AArch32 register summary

The following table gives a summary of the Advanced SIMD and Floating-point System registers in the
Cortex-A72 processor when in AArch32 state.

Table 14-9 AArch32 Advanced SIMD and Floating-point System registers

Name Type Reset Description

FPSID RO 0x41034080 See 14.6.1 Floating-point System ID Register on page 14-525

FPSCR RW 0x00000000 See 14.6.2 Floating-point Status and Control Register on page 14-526

MVFR0 RO 0x10110222 See 14.4.3 Media and VFP Feature Register 0, EL1 on page 14-518

MVFR1 RO 0x12111111 See 14.4.4 Media and VFP Feature Register 1, EL1 on page 14-519

MVFR2 RW 0x00000043 See 14.4.5 Media and VFP Feature Register 2, EL1 on page 14-520

FPEXC RW 0x00000700 See 14.4.6 Floating-point Exception Control Register 32, EL2 on page 14-521

Note
The Floating-point Instruction Registers, FPINST and FPINST2 are not implemented, and any attempt to
access them is UNPREDICTABLE.

See the ARM® Architecture Reference Manual ARMv8 for information about permitted accesses to the
Advanced SIMD and Floating-point System registers.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-524
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.6 AArch32 register descriptions

14.6 AArch32 register descriptions

This section describes the AArch32 Advanced SIMD and Floating-point System registers in the Cortex-
A72 processor.
Table 14-9 AArch32 Advanced SIMD and Floating-point System registers on page 14-524 provides
cross-references to individual registers.
This section contains the following subsections:
• 14.6.1 Floating-point System ID Register on page 14-525.
• 14.6.2 Floating-point Status and Control Register on page 14-526.

14.6.1 Floating-point System ID Register

The FPSID characteristics are:

Purpose
Provides top-level information about the floating-point implementation.
Usage constraints
The accessibility to the FPSID by Exception level is:

EL0 EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

- Config RO Config Config RO

Configurations
The FPSID is Common to Secure and Non-secure states.
Attributes
See the register summary in Table 14-9 AArch32 Advanced SIMD and Floating-point System
registers on page 14-524.
The following figure shows the FPSID bit assignments.
31 24 23 22 16 15 8 7 4 3 0

Implementer Subarchitecture PartNum Variant Revision

SW

Figure 14-7 FPSID bit assignments

The following table shows the FPSID bit assignments.

Table 14-10 FPSID bit assignments

Bits Name Function

[31:24] Implementer Indicates the implementer:

0x41 ARM Limited.

[23] SW Software bit. This bit indicates whether a system provides only software emulation of the floating-point
instructions:
0x0 The system includes hardware support for floating-point operations.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-525
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.6 AArch32 register descriptions

Table 14-10 FPSID bit assignments (continued)

Bits Name Function

[22:16] Subarchitecture Subarchitecture version number:

0x03 VFPv3 architecture, or later, with no subarchitecture. The entire floating-point implementation is in
hardware, and no software support code is required.
The VFP architecture version is indicated by the MVFR0, MVFR1, and MVFR2 registers.

[15:8] PartNum Indicates the part number for the floating-point implementation:
0x40 VFP.

[7:4] Variant Indicates the variant number:

0x8 Cortex-A72 processor.

[3:0] Revision Indicates the revision number for the floating-point implementation:
0x0 Revision.

To access the FPSID register, see 14.2 Programmers model for Advanced SIMD and Floating-point
on page 14-513.

14.6.2 Floating-point Status and Control Register

The FPSCR characteristics are:

Purpose
Provides floating-point system status information and control.
Usage constraints
The accessibility to the FPSCR by Exception level is:

EL0 (NS) EL0 (S) EL1(NS) EL1(S) EL2 EL3(SCR.NS = 1) EL3(SCR.NS = 0)

Config RW Config RW Config Config RW

Configurations
The FPSCR is Common to Secure and Non-secure states.
The named fields in this register map to the equivalent fields in the AArch64 FPCR and FPSR.
Attributes
See the register summary in Table 14-9 AArch32 Advanced SIMD and Floating-point System
registers on page 14-524.
The following figure shows the FPSCR bit assignments.
31 30 29 28 27 26 25 24 23 22 21 20 19 18 16 15 8 7 6 5 4 3 2 1 0

N Z C V Len RES0 RES0

QC RES0 IDC IOC

AHP DZC
DN OFC
FZ UFC
RMode IXC
Stride

Figure 14-8 FPSCR bit assignments

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-526
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.6 AArch32 register descriptions

The following table shows the FPSCR bit assignments.

Table 14-11 FPSCR bit assignments

Bits Field Function

[31] N FP Negative condition code flag.

Set to 1 if a FP comparison operation produces a less than result.

[30] Z FP Zero condition code flag.

Set to 1 if a FP comparison operation produces an equal result.

[29] C FP Carry condition code flag.

Set to 1 if a FP comparison operation produces an equal, greater than, or unordered result.

[28] V FP Overflow condition code flag.

Set to 1 if a FP comparison operation produces an unordered result.

[27] QC Cumulative saturation bit.

This bit is set to 1 to indicate that an Advanced SIMD integer operation has saturated after 0 was last written to this
bit.

[26] AHP Alternative Half-Precision control bit:

0 IEEE half-precision format selected.
1 Alternative half-precision format selected.

[25] DN Default NaN mode control bit:

0 NaN operands propagate through to the output of a floating-point operation.
1 Any operation involving one or more NaNs returns the Default NaN.

The value of this bit only controls FP arithmetic. In AArch32 state, Advanced SIMD arithmetic always uses the
Default NaN setting, regardless of the value of the DN bit.

[24] FZ Flush-to-zero mode control bit:

0 Flush-to-zero mode disabled. Behavior of the floating-point system is fully compliant with the IEEE 754
standard.
1 Flush-to-zero mode enable.

The value of this bit only controls FP arithmetic. In AArch32 state, Advanced SIMD arithmetic always uses the
Flush-to-zero setting, regardless of the value of the FZ bit.

[23:22] RMode Rounding Mode control field:

0b00 Round to Nearest (RN) mode.
0b01 Round towards Plus Infinity (RP) mode.
0b10 Round towards Minus Infinity (RM) mode.
0b11 Round towards Zero (RZ) mode.

The specified rounding mode is used by almost all FP floating-point instructions. In AArch32 state, Advanced
SIMD arithmetic always uses the Round to Nearest setting, regardless of the value of the RMode bits.

[21:20] Stride Reserved, RES0.

[19] - Reserved, RES0.

[18:16] Len Reserved, RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-527
reserved.
Non-Confidential
 14 Advanced SIMD and Floating-point
14.6 AArch32 register descriptions

Table 14-11 FPSCR bit assignments (continued)

Bits Field Function

[15:8] - Reserved, RES0.

[7] IDC Input Denormal cumulative exception bit. This bit is set to 1 to indicate that the Input Denormal exception has
occurred since 0 was last written to this bit.

[6:5] - Reserved, RES0.

[4] IXC Inexact cumulative exception bit. This bit is set to 1 to indicate that the Inexact exception has occurred since 0 was
last written to this bit.

[3] UFC Underflow cumulative exception bit. This bit is set to 1 to indicate that the Underflow exception has occurred since
0 was last written to this bit.

[2] OFC Overflow cumulative exception bit. This bit is set to 1 to indicate that the Overflow exception has occurred since 0
was last written to this bit.

[1] DZC Division by Zero cumulative exception bit. This bit is set to 1 to indicate that the Division by Zero exception has
occurred since 0 was last written to this bit.

[0] IOC Invalid Operation cumulative exception bit. This bit is set to 1 to indicate that the Invalid Operation exception has
occurred since 0 was last written to this bit.

To access the FPSCR register, see 14.2 Programmers model for Advanced SIMD and Floating-point
on page 14-513.

Related information
14.4.1 Floating-point Control Register on page 14-515.
14.4.2 Floating-point Status Register on page 14-516.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights 14-528
reserved.
Non-Confidential
Appendix A
Signal Descriptions

This section describes the Cortex-A72 processor signals.

It contains the following sections:

• A.1 About the signal descriptions on page Appx-A-530.
• A.2 Clock signals on page Appx-A-531.
• A.3 Reset signals on page Appx-A-532.
• A.4 Configuration signals on page Appx-A-533.
• A.5 GIC CPU interface signals on page Appx-A-535.
• A.6 Generic Timer signals on page Appx-A-537.
• A.7 Power control signals on page Appx-A-538.
• A.8 ACE and CHI interface signals on page Appx-A-540.
• A.9 CHI interface signals on page Appx-A-543.
• A.10 ACE interface signals on page Appx-A-548.
• A.11 ACP interface signals on page Appx-A-553.
• A.12 Debug interface signals on page Appx-A-556.
• A.13 ETM interface on page Appx-A-559.
• A.14 Cross trigger channel interface on page Appx-A-561.
• A.15 PMU signals on page Appx-A-562.
• A.16 DFT and MBIST signals on page Appx-A-563.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-529
reserved.
Non-Confidential
 A Signal Descriptions
A.1 About the signal descriptions

A.1 About the signal descriptions

The tables in this appendix list the Cortex-A72 processor signals, along with their direction, input or
output, and a high-level description.
Some of the buses include a configurable width field, <signal>[N:0], where N = 0, 1, 2, or 3, to encode
up to four cores. For example:
• nIRQ[0] represents a core 0 interrupt request.
• nIRQ[2] represents a core 2 interrupt request.
Some signals are specified in the form <signal>x, where x = 0, 1, 2 or 3 references core 0, core 1, core 2,
or core 3, respectively. If a core is not present, the corresponding pin is removed. For example:
• PMUEVENT0[24:0] represents the core 0 PMU event bus.
• PMUEVENT3[24:0] represents the core 3 PMU event bus.
The number of signals changes depending on the configuration. For example, the CHI interface signals
are not present when the processor is configured to have an ACE interface.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-530
reserved.
Non-Confidential
 A Signal Descriptions
A.2 Clock signals

A.2 Clock signals

The following table shows the clock and clock enable signals.

Table A-1 Clock and clock enable signals

Signal Type Description

CLK Input Global clock.

CLKEN Input Global clock enable. This signal can only be deasserted when all the cores in the processor device and the L2 are in
WFI low-power state, and the ACE/CHI and ACP interfaces are idle.

Related information
2.3 Clocking and resets on page 2-32.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-531
reserved.
Non-Confidential
 A Signal Descriptions
A.3 Reset signals

A.3 Reset signals

The following table shows the reset and reset control signals.

Table A-2 Reset signals

Signal Type Description

nCPUPORESET[N:0] Input Individual processor powerup resets:

0 Apply reset to the processor including Debug, ETM, breakpoint and watchpoint logic.
1 Do not apply reset to the processor.

nCORERESET[N:0] Input Individual processor reset excluding Debug and ETM:

0 Apply reset to the processor excluding Debug, ETM, breakpoint and watchpoint logic.
1 Do not apply reset to the processor.

WARMRSTREQ[N:0] Output Individual processor Warm reset request:

0 Do not apply Warm reset to processor.
1 Apply Warm reset to processor.

This output is controlled by Reset request bit in the Reset Management Register (RMR or
RMR_EL3).

nL2RESET Input L2 reset:

0 Apply reset to shared L2 memory system controller.
1 Do not apply reset to shared L2 memory system controller.

L2RSTDISABLE Input Disable automatic L2 cache invalidate at reset:

0 L2 cache is reset by hardware.
1 L2 cache is not reset by hardware.

Related information
2.3.2 Resets on page 2-36.
2.3 Clocking and resets on page 2-32.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-532
reserved.
Non-Confidential
 A Signal Descriptions
A.4 Configuration signals

A.4 Configuration signals

The following table shows the configuration signals.

Table A-3 Configuration inputs

Signal Type Description

CFGEND[N:0] Input Individual core control of the endianness configuration at reset. It sets the initial value of the EE
bit in the System Control Register (SCTLR or SCTLR_EL3):
0 EE bit is 0.
1 EE bit is 1.

This signal is only sampled during powerup reset of the processor.

VINITHI[N:0] Input Individual core control of the location of the exception vectors at reset. It sets the initial value of
the V bit in the CP15 System Control Register (SCTLR when the highest Exception level is
AArch32):
0 Exception vectors start at address 0x00000000.
1 Exception vectors start at address 0xFFFF0000.

This signal is only sampled during powerup reset of the processor.

CFGTE[N:0] Input Individual core control of the default exception handling state. It sets the initial value of the TE
bit in the CP15 System Control Register (SCTLR when the highest Exception level is AArch32):
0 TE bit is 0.
1 TE bit is 1.

This signal is only sampled during powerup reset of the processor.

CP15SDISABLE[N:0] Input Disable write access to some Secure CP15 registers. See 4.1.1 Registers affected by
CP15SDISABLE on page 4-74.

CLUSTERIDAFF1[7:0] Input Value read in the Cluster ID Affinity Level-1 field, bits[15:8], of the Multiprocessor Affinity
Register (MPIDR).
This signal is only sampled during powerup reset of the processor.

CLUSTERIDAFF2[7:0] Input Value read in the Cluster ID Affinity Level-2 field, bits[23:16], of the Multiprocessor Affinity
Register (MPIDR).
This signal is only sampled during powerup reset of the processor.

AA64nAA32[N:0] Input Individual core register width state. The register width states are:
0 AArch32.
1 AArch64.

This signal is only sampled during powerup reset of the processor.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-533
reserved.
Non-Confidential
 A Signal Descriptions
A.4 Configuration signals

Table A-3 Configuration inputs (continued)

Signal Type Description

RVBARADDRx[43:2]fa Input Reset Vector Base Address for executing in AArch64 state. This signal is only sampled during
reset of the processor.

CRYPTODISABLE[N:0]fb Input Individual core Cryptography engine disable:

0 Enable the Cryptography engine.
1 Disable the Cryptography engine.

This signal is only sampled during powerup reset of the processor. This signal only exists if the
processor implements the Cryptography Extension.

fa x is 0, 1, 2, or 3 to reference a specific core.

fb The optional Cryptography engine is not included in the base product of the processor. ARM requires licensees to have contractual rights to obtain the Cortex-A72
processor Cryptography engine.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-534
reserved.
Non-Confidential
 A Signal Descriptions
A.5 GIC CPU interface signals

A.5 GIC CPU interface signals

The following table shows the Generic Interrupt Controller (GIC) CPU interface signals.

Table A-4 GIC CPU interface signals

Signal Type Description

nIRQ[N:0] Input Individual core IRQ request input. Active-LOW, interrupt request:
0 Activate IRQ request.
1 Do not activate IRQ request.

The processor treats nIRQ as level-sensitive. nIRQ must remain asserted until the processor
acknowledges the interrupt.
This signal is only used when IRQ is in bypass mode, and used as legacy IRQ.

nFIQ[N:0] Input Individual processor FIQ request input. Active-LOW, FIQ request:
0 Activate FIQ request.
1 Do not activate FIQ request.

The processor treats nFIQ as level-sensitive. nFIQ must remain asserted until the processor
acknowledges the interrupt.
This signal is only used when FIQ is in bypass mode, and used as legacy FIQ.

nVIRQ[N:0] Input Individual processor virtual IRQ request input. Active-LOW, virtual IRQ request:
0 Activate virtual IRQ request.
1 Do not activate virtual IRQ request.

The processor treats nVIRQ as level-sensitive. nVIRQ must remain asserted until the processor
acknowledges the interrupt.

nVFIQ[N:0] Input Individual processor virtual FIQ request input. Active-LOW, virtual FIQ request:
0 Activate virtual FIQ request.
1 Do not activate virtual FIQ request.

The processor treats nVFIQ as level-sensitive. nVFIQ must remain asserted until the processor
acknowledges the interrupt.

nSEI[N:0] Input Individual processor System Error Interrupt request. Active-LOW, SEI request:
0 Activate SEI request.
1 Do not activate SEI request.

The processor treats nSEI as edge-sensitive. The nSEI signal must be sent as a pulse to the
processor.

nREI[N:0] Input Individual core RAM Error Interrupt request. Active-LOW, REI request.
0 Activate REI request. Reports an asynchronous RAM error in the system.
1 Do not activate REI request.

The processor treats nREI as edge-sensitive. nREI must be sent as a pulse to the processor.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-535
reserved.
Non-Confidential
 A Signal Descriptions
A.5 GIC CPU interface signals

Table A-4 GIC CPU interface signals (continued)

Signal Type Description

nVSEI[N:0] Input Individual core virtual System Error Interrupt request. Active-LOW, virtual SEI request:
0 Activate virtual SEI request.
1 Do not activate virtual SEI request.

The processor treats nVSEI as edge-sensitive. nVSEI must be sent as a pulse to the processor.

nVCPUMNTIRQ[N:0] Output Individual core virtual CPU interface maintenance interrupt request. Processor N sets this signal
LOW to issue a maintenance interrupt request to the external Distributor.

PERIPHBASE[43:18] Input Specifies the base address for the GIC registers. This value is sampled into the Configuration Base
Address Register (CBAR) at reset.

GICCDISABLE Input Disables the GIC CPU interface logic and routes the legacy nIRQ, nFIQ, nVIRQ, and nVFIQ
signals directly to the processor:
0 Enable the GIC CPU interface logic.
1 Disable the GIC CPU interface logic.

This signal is only sampled during powerup reset of the processor.

AXI4 Stream protocol signals: fc

ICDTVALID Input When HIGH it indicates that the Distributor is driving a valid transfer.

ICDTREADY Output When HIGH it indicates that the processor can accept a transfer in the current cycle.

ICDTDATA[15:0] Input The primary payload that passes data from the Distributor to the processor.

ICDTLAST Input When HIGH it indicates the boundary of a packet.

ICDTDEST[1:0] Input Provides routing information for the data stream from the Distributor.

ICCTVALID Output When HIGH it indicates that the processor is driving a valid transfer.

ICCTREADY Input When HIGH it indicates that the Distributor can accept a transfer in the current cycle.

ICCTDATA[15:0] Output The primary payload that passes data from the processor to the Distributor.

ICCTLAST Output When HIGH it indicates the boundary of a packet.

ICCTID[1:0] Output The data stream identifier that indicates different streams of data.

Related information
4.3.70 Configuration Base Address Register, EL1 on page 4-212.
4.5.24 Configuration Base Address Register on page 4-274.
GICCDISABLE bypass mode on page 8-323.

fc See the ARM® AMBA® AXI4-Stream Protocol Specification for more information.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-536
reserved.
Non-Confidential
 A Signal Descriptions
A.6 Generic Timer signals

A.6 Generic Timer signals

The following table shows the Generic Timer signals.

Table A-5 Generic Timer signals

Signal Type Description

CNTVALUEB[63:0] Input Global system counter value in binary format

CNTCLKEN Input Counter clock enable

nCNTPNSIRQ[N:0] Output Non-secure physical timer interrupt

nCNTPSIRQ[N:0] Output Secure physical timer interrupt

nCNTHPIRQ[N:0] Output Hypervisor physical timer interrupt

nCNTVIRQ[N:0] Output Virtual timer interrupt

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-537
reserved.
Non-Confidential
 A Signal Descriptions
A.7 Power control signals

A.7 Power control signals

The following table shows the power control signals.

Table A-6 Power control signals

Signal Type Description

EVENTI Input Event input for processor wake-up from WFE low-power state. When this signal is asserted, it acts
as a WFE wake-up event to all the processors in the MPCore device. This signal must be asserted
for at least one CLK cycle.

EVENTO Output Event output. This signal is asserted HIGH for three CLK cycles when any of the processors in the
MPCore device executes an SEV instruction.

CLREXMONREQ Input Clearing of the external global exclusive monitor request. When this signal is asserted, it acts as a
WFE wake-up event to all the processors in the MPCore device.

CLREXMONACK Output Clearing of the external global exclusive monitor acknowledge.

STANDBYWFE[N:0] Output Indicates whether a processor is in WFE low-power state:

0 Processor not in WFE low-power state.
1 Processor in WFE low-power state.

STANDBYWFI[N:0] Output Indicates whether a processor is in WFI low-power state:

0 Processor not in WFI low-power state.
1 Processor in WFI low-power state.

STANDBYWFIL2 Output Indicates whether the L2 is in WFI low-power state. This signal is active when the following are
true:
• All processors are in WFI low-power state.
• ACINACTM or SINACT and AINACTS are asserted HIGH.
• L2 memory system is idle.

L2FLUSHREQ Input L2 hardware flush request. This signal indicates:

0 L2 hardware flush request is not asserted.
1 L2 hardware flush request is asserted.

L2FLUSHDONE Output L2 hardware flush done.

0 L2 hardware flush is not finished.
1 L2 hardware flush is finished.

SMPEN[N:0] Output CPUECTLR.SMPEN output. This signal indicates:

0 The CPUECTLR.SMPEN bit is not set.
1 The CPUECTLR.SMPEN bit is set.

CPUQACTIVE[N:0] Output When HIGH, it indicates that processor N is active.

CPUQREQn[N:0] Input The power controller sets this signal LOW, to request that processor N enters retention state.

CPUQACCEPTn[N:0] Output This signal goes LOW, if processor N accepts the power controller retention request.

CPUQDENY[N:0] Output When HIGH, it indicates that processor N denies the power controller retention request.

L2QACTIVE Output When HIGH, it indicates that the L2 Data and Tag RAMs are active.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-538
reserved.
Non-Confidential
 A Signal Descriptions
A.7 Power control signals

Table A-6 Power control signals (continued)

Signal Type Description

L2QREQn Input The power controller sets this signal LOW, to request that the L2 Data and Tag RAMs enter
retention state.

L2QACCEPTn Output This signal goes LOW, if the L2 Data and Tag RAMs accept the power controller retention request.

L2QDENY Output When HIGH, it indicates that the L2 Data and Tag RAMs deny the power controller retention
request.

Related information
Event communication using WFE and SEV instructions on page 2-43.
CLREXMON request and acknowledge signaling on page 2-43.
4.3.67 CPU Extended Control Register, EL1 on page 4-206.
Processor dynamic retention on page 2-46.
L2 RAMs dynamic retention on page 2-48.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-539
reserved.
Non-Confidential
 A Signal Descriptions
A.8 ACE and CHI interface signals

A.8 ACE and CHI interface signals

This section describes the ACE and CHI interface signals.
This section contains the following subsections:
• A.8.1 Configuration signals on page Appx-A-540.
• A.8.2 Asynchronous error signals on page Appx-A-542.

A.8.1 Configuration signals

The following table shows the configuration signals that are common to the ACE and CHI interfaces.

Table A-7 ACE or CHI configuration inputs

Signal Type Description

BROADCASTINNER Input Enable broadcasting of Inner Shareable

transactions:
0 Inner Shareable transactions are not
broadcasted externally.
1 Inner Shareable transactions are
broadcasted externally.

If BROADCASTINNER is tied HIGH,

BROADCASTOUTER must also be tied
HIGH.
This signal is only sampled during
powerup reset of the processor.

BROADCASTOUTER Input Enable broadcasting of Outer Shareable

transactions:
0 Outer Shareable transactions are not
broadcasted externally.
1 Outer Shareable transactions are
broadcasted externally.

This signal is only sampled during

powerup reset of the processor.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-540
reserved.
Non-Confidential
 A Signal Descriptions
A.8 ACE and CHI interface signals

Table A-7 ACE or CHI configuration inputs (continued)

Signal Type Description

BROADCASTCACHEMAINT Input Enable broadcasting of cache maintenance

operations to downstream caches:
0 Cache maintenance operations are not
broadcasted to downstream caches.
1 Cache maintenance operations are
broadcasted to downstream caches.

This signal is only sampled during

powerup reset of the processor.

SYSBARDISABLE Input Disable broadcasting of barriers on the

system bus:
0 Barriers are broadcast on the system
bus.
1 Barriers are not broadcast on the
system bus.

For AXI3 compatibility in ACE interface

configurations, SYSBARDISABLE must
be tied HIGH and the following signals
LOW:
• BROADCASTCACHEMAINT.
• BROADCASTINNER.
• BROADCASTOUTER.
This signal is only sampled during
powerup reset of the processor.

Related information
7.7.2 Interface modes on page 7-311.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-541
reserved.
Non-Confidential
 A Signal Descriptions
A.8 ACE and CHI interface signals

A.8.2 Asynchronous error signals

The following table shows the asynchronous error signals.

Table A-8 Asynchronous error signals

Signal Type Description

nEXTERRIRQ Output Error indicator for an AXI or CHI write transaction with a write response error condition.
Writing 0 to bit[29] of the L2ECTLR clears the error indicator.

nINTERRIRQ Output Error indicator for an L2 RAM double-bit ECC error.

Writing 0 to bit[30] of the L2ECTLR clears the error indicator.

Related information
4.3.59 L2 Extended Control Register, EL1 on page 4-172.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-542
reserved.
Non-Confidential
 A Signal Descriptions
A.9 CHI interface signals

A.9 CHI interface signals

This section shows the CHI interface signals.

Note
This interface only exists if the processor implements the CHI interface.

This section contains the following subsections:

• A.9.1 CHI clock and configuration signals on page Appx-A-543.
• A.9.2 Transmit request virtual channel signals on page Appx-A-543.
• A.9.3 Transmit response virtual channel signals on page Appx-A-544.
• A.9.4 Transmit data virtual channel signals on page Appx-A-544.
• A.9.5 Receive snoop virtual channel signals on page Appx-A-544.
• A.9.6 Receive response virtual channel signals on page Appx-A-545.
• A.9.7 Receive data virtual channel signals on page Appx-A-545.
• A.9.8 System address map signals on page Appx-A-545.

A.9.1 CHI clock and configuration signals

The following table shows the clock and configuration signals for the CHI interface.

Table A-9 CHI clock and configuration signals

Signal Type Description

SCLKEN Input CHI interface clock enable

SINACT Input CHI snoop active in

NODEID[6:0] Input CHI node identifier. This signal is only sampled during powerup reset of the processor.

RXSACTIVE Input Receive pending activity indicator

TXSACTIVE Output Transmit pending activity indicator

RXLINKACTIVEREQ Input Receive link active request

RXLINKACTIVEACK Output Receive link active acknowledge

TXLINKACTIVEREQ Output Transmit link active request

TXLINKACTIVEACK Input Transmit link active acknowledge

A.9.2 Transmit request virtual channel signals

The following table shows the transmit request virtual channel signals for the CHI interface.

Table A-10 Transmit request virtual channel signals

Signal Type Description

TXREQFLITPEND Output Transmit request flit pending.

TXREQFLITV Output Transmit request flit valid.

TXREQFLIT[99:0] Output Transmit request flit payload.fd

fd TXREQFLIT[MemAttr] allocation hints based on outer memory attributes from MMU.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-543
reserved.
Non-Confidential
 A Signal Descriptions
A.9 CHI interface signals

Table A-10 Transmit request virtual channel signals (continued)

Signal Type Description

TXREQLCRDV Input Transmit request link-layer credit valid.

REQMEMATTR[7:0] Output Transmit request raw memory attributes:

[7] Outer Shareable.
[6:3] Outer memory attribute in MAIR format.
[2] Inner Shareable.
[1:0] 0b00 Device.
0b01 Normal Non-cacheable.
0b10 Normal Write-Through.
0b11 Normal Write-Back.

A.9.3 Transmit response virtual channel signals

The following table shows the transmit response virtual channel signals for the CHI interface.

Table A-11 Transmit response virtual channel signals

Signal Type Description

TXRSPFLITPEND Output Transmit response flit pending

TXRSPFLITV Output Transmit response flit valid

TXRSPFLIT[44:0] Output Transmit response flit payload

TXRSPLCRDV Input Transmit response link-layer credit valid

A.9.4 Transmit data virtual channel signals

The following table shows the transmit data virtual channel signals for the CHI interface.

Table A-12 Transmit data virtual channel signals

Signal Type Description

TXDATFLITPEND Output Transmit data flit pending

TXDATFLITV Output Transmit data flit valid

TXDATFLIT[193:0] Output Transmit data flit payload

TXDATLCRDV Input Transmit data link-layer credit valid

A.9.5 Receive snoop virtual channel signals

The following table shows the receive snoop virtual channel signals for the CHI interface.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-544
reserved.
Non-Confidential
 A Signal Descriptions
A.9 CHI interface signals

Table A-13 Receive snoop virtual channel signals

Signal Type Description

RXSNPFLITPEND Input Receive snoop flit pending

RXSNPFLITV Input Receive snoop flit valid

RXSNPFLIT[64:0] Input Receive snoop flit payload

RXSNPLCRDV Output Receive snoop link-layer credit valid

A.9.6 Receive response virtual channel signals

The following table shows the receive response virtual channel signals for the CHI interface.

Table A-14 Receive response virtual channel signals

Signal Type Description

RXRSPFLITPEND Input Receive response flit pending

RXRSPFLITV Input Receive response flit valid

RXRSPFLIT[44:0] Input Receive response flit payload

RXRSPLCRDV Output Receive response link-layer credit valid

A.9.7 Receive data virtual channel signals

The following table shows the receive data virtual channel signals for the CHI interface.

Table A-15 Receive data virtual channel signals

Signal Type Description

RXDATFLITPEND Input Receive data flit pending

RXDATFLITV Input Receive data flit valid

RXDATFLIT[193:0] Input Receive data flit payload

RXDATLCRDV Output Receive data link-layer credit valid

A.9.8 System address map signals

The following table shows the system address map signals for the CHI interface. The SAM* signals are
only sampled during powerup reset of the processor.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-545
reserved.
Non-Confidential
 A Signal Descriptions
A.9 CHI interface signals

Table A-16 System address map signals

Signal Type Description

SAMMNBASE[43:24] Input MN base address.

SAMADDRMAP0[1:0] Input 0 to 512MB region mapping.

Encoding for all SAMADDRMAPx[1:0] signals:
0b00 HN-F.
0b01 HN-I.
0b10 Reserved.
0b11

SAMADDRMAP1[1:0] Input 512MB to 1GB region mapping.

SAMADDRMAP2[1:0] Input 1GB to 1.5GB region mapping.

SAMADDRMAP3[1:0] Input 1.5GB to 2GB region mapping.

SAMADDRMAP4[1:0] Input 2GB to 2.5GB region mapping.

SAMADDRMAP5[1:0] Input 2.5GB to 3GB region mapping.

SAMADDRMAP6[1:0] Input 3GB to 3.5GB region mapping.

SAMADDRMAP7[1:0] Input 3.5GB to 4GB region mapping.

SAMADDRMAP8[1:0] Input 4GB to 8GB region mapping.

SAMADDRMAP9[1:0] Input 8GB to 16GB region mapping.

SAMADDRMAP10[1:0] Input 16GB to 32GB region mapping.

SAMADDRMAP11[1:0] Input 32GB to 64GB region mapping.

SAMADDRMAP12[1:0] Input 64GB to 128GB region mapping.

SAMADDRMAP13[1:0] Input 128GB to 256GB region mapping.

SAMADDRMAP14[1:0] Input 256GB to 512GB region mapping.

SAMADDRMAP15[1:0] Input 512GB to 1TB region mapping.

SAMADDRMAP16[1:0] Input 1TB to 2TB region mapping.

SAMADDRMAP17[1:0] Input 2TB to 4TB region mapping.

SAMADDRMAP18[1:0] Input 4TB to 8TB region mapping.

SAMADDRMAP19[1:0] Input 8TB to 16TB region mapping.

SAMMNNODEID[6:0] Input MN node ID.

SAMHNI0NODEID[6:0] Input HN-I 0 node ID.

SAMHNI1NODEID[6:0] Input HN-I 1 node ID.

SAMHNF0NODEID[6:0] Input HN-F 0 node ID.

SAMHNF1NODEID[6:0] Input HN-F 1 node ID.

SAMHNF2NODEID[6:0] Input HN-F 2 node ID.

SAMHNF3NODEID[6:0] Input HN-F 3 node ID.

SAMHNF4NODEID[6:0] Input HN-F 4 node ID.

SAMHNF5NODEID[6:0] Input HN-F 5 node ID.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-546
reserved.
Non-Confidential
 A Signal Descriptions
A.9 CHI interface signals

Table A-16 System address map signals (continued)

Signal Type Description

SAMHNF6NODEID[6:0] Input HN-F 6 node ID.

SAMHNF7NODEID[6:0] Input HN-F 7 node ID.

SAMHNFMODE[2:0] Input HN-F interleaving mode:

0b000 1 HN-F.
0b001 2 HN-Fs.
0b010 4 HN-Fs.
0b100 8 HN-Fs.

All other values are reserved.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-547
reserved.
Non-Confidential
 A Signal Descriptions
A.10 ACE interface signals

A.10 ACE interface signals

This section shows the ACE interface signals.

Note
This interface only exists if the processor implements the ACE interface.

This section contains the following subsections:

• A.10.1 Clock and configuration signals on page Appx-A-548.
• A.10.2 Write address channel signals on page Appx-A-548.
• A.10.3 Write data channel signals on page Appx-A-549.
• A.10.4 Write response channel signals on page Appx-A-549.
• A.10.5 Read address channel signals on page Appx-A-550.
• A.10.6 Read data channel signals on page Appx-A-550.
• A.10.7 Snoop address channel signals on page Appx-A-551.
• A.10.8 Snoop response channel signals on page Appx-A-551.
• A.10.9 Snoop data channel handshake signals on page Appx-A-551.
• A.10.10 Read/Write acknowledge signals on page Appx-A-552.

A.10.1 Clock and configuration signals

The following table shows the clock and configuration signals for the ACE interface.

Table A-17 Clock and configuration signals

Signal Type Description

ACLKENM Input AXI master bus clock enable.

ACINACTM Input Snoop interface is inactive.

When this signal is HIGH, the snoop address channel stops accepting requests by deasserting ACREADYM.
Snoop requests that were accepted before deasserting ACREADYM are serviced.

Related information
2.3 Clocking and resets on page 2-32.

A.10.2 Write address channel signals

The following table shows the write address channel signals for the ACE master interface.

Table A-18 Write address channel signals

Signal Type Description

AWREADYM Input Write address ready.

AWVALIDM Output Write address valid.

AWIDM[6:0] Output Write request ID.

AWADDRM[43:0] Output Write address.

AWLENM[7:0] Output Write burst length. AWLENM[7:2] is always 0b000000.

AWSIZEM[2:0] Output Write burst size.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-548
reserved.
Non-Confidential
 A Signal Descriptions
A.10 ACE interface signals

Table A-18 Write address channel signals (continued)

Signal Type Description

AWBURSTM[1:0] Output Write burst type.

AWBARM[1:0] Output Write barrier type.

AWDOMAINM[1:0] Output Write shareability domain type.

AWLOCKM Output Write lock type.

AWCACHEM[3:0] Output Write cache type.fe

AWPROTM[2:0] Output Write protection type.

AWSNOOPM[2:0] Output Write snoop request type.

AWUNIQUEM Output Indicates the write operation for a WriteBack, WriteClean, or WriteEvict transaction is:
0 Shared.
1 Unique.

WRMEMATTR[7:0] Output Write request raw memory attributes:

[7] Outer Shareable.
[6:3] Outer memory attribute in MAIR format.
[2] Inner Shareable.
[1:0] 0b00 Device.
0b01 Normal Non-cacheable.
0b10 Normal Write-Through.
0b11 Normal Write-Back.

A.10.3 Write data channel signals

The following table shows the write data signals for the AXI master interface.

Table A-19 Write data channel signals

Signal Type Description

WREADYM Input Write data ready

WVALIDM Output Write data valid

WIDM[6:0] Output Write data ID

WDATAM[127:0] Output Write data

WSTRBM[15:0] Output Write byte-lane strobes

WLASTM Output Write data last transfer indication

A.10.4 Write response channel signals

The following table shows the write response channel signals for the ACE interface.

fe Allocation hints based on outer memory attributes from the MMU.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-549
reserved.
Non-Confidential
 A Signal Descriptions
A.10 ACE interface signals

Table A-20 Write response channel signals

Signal Type Description

BREADYM Output Write response ready

BVALIDM Input Write response valid

BIDM[6:0] Input Write response ID

BRESPM[1:0] Input Write response

A.10.5 Read address channel signals

The following table shows the read address channel signals for the ACE interface.

Table A-21 Read address channel signals

Signal Type Description

ARREADYM Input Read address ready.

ARVALIDM Output Read address valid.

ARIDM[6:0] Output Read request ID.

ARADDRM[43:0] Output Read address.

ARLENM[7:0] Output Read burst length. ARLENM[7:2] is always 0b000000.

ARSIZEM[2:0] Output Read burst size.

ARBURSTM[1:0] Output Burst type.

ARBARM[1:0] Output Read barrier type.

ARDOMAINM[1:0] Output Read shareability domain type.

ARLOCKM Output Read lock type.

ARCACHEM[3:0] Output Read cache type.ff

ARPROTM[2:0] Output Read protection type.

ARSNOOPM[3:0] Output Read snoop request type.

RDMEMATTR[7:0] Output Read request raw memory attributes:

[7] Outer Shareable.
[6:3] Outer memory attribute in MAIR format.
[2] Inner Shareable.
[1:0] 0b00 Device.
0b01 Normal Non-cacheable.
0b10 Normal Write-Through.
0b11 Normal Write-Back.

A.10.6 Read data channel signals

The following table shows the read data channel signals for the ACE interface.

ff Allocation hints based on outer memory attributes from the MMU.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-550
reserved.
Non-Confidential
 A Signal Descriptions
A.10 ACE interface signals

Table A-22 Read data channel signals

Signal Type Description

RREADYM Output Read data ready

RVALIDM Input Read data valid

RIDM[6:0] Input Read data ID

RDATAM[127:0] Input Read data

RRESPM[3:0] Input Read data response

RLASTM Input Read data last transfer indication

A.10.7 Snoop address channel signals

The following table shows the snoop address channel signals for the ACE interface.

Table A-23 Snoop address channel signals

Signal Type Description

ACREADYM Output Master ready to receive snoop address

ACVALIDM Input Snoop address valid

ACADDRM[43:0] Input Snoop address

ACPROTM[2:0] Input Snoop protection type

ACSNOOPM[3:0] Input Snoop request type

A.10.8 Snoop response channel signals

The following table shows the snoop response channel signals for the AXI master interface.

Table A-24 Snoop response channel signals

Signal Type Description

CRREADYM Input Slave ready to accept snoop response

CRVALIDM Output Snoop response valid

CRRESPM[4:0] Output Snoop response

A.10.9 Snoop data channel handshake signals

The following table shows the snoop data channel handshake signals for the ACE interface.

Table A-25 Snoop data channel handshake signals

Signal Type Description

CDREADYM Input Slave ready to accept snoop data

CDVALIDM Output Snoop data valid

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-551
reserved.
Non-Confidential
 A Signal Descriptions
A.10 ACE interface signals

Table A-25 Snoop data channel handshake signals (continued)

Signal Type Description

CDDATAM[127:0] Output Snoop data

CDLASTM Output Snoop data last transfer indication

A.10.10 Read/Write acknowledge signals

The following table shows the read/write acknowledge signals for the AXI master interface.

Table A-26 Read/write acknowledge signals

Signal Type Description

RACKM Output Read acknowledge

WACKM Output Write acknowledge

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-552
reserved.
Non-Confidential
 A Signal Descriptions
A.11 ACP interface signals

A.11 ACP interface signals

Describes the ACP interface signals.

Note
These signals only exist if the processor is configured to have an ACP interface.

This section contains the following subsections:

• A.11.1 Clock and configuration signals on page Appx-A-553.
• A.11.2 Write address channel signals on page Appx-A-553.
• A.11.3 Write data channel signals on page Appx-A-554.
• A.11.4 Write response channel signals on page Appx-A-554.
• A.11.5 Read address channel signals on page Appx-A-555.
• A.11.6 Read data channel signals on page Appx-A-555.

A.11.1 Clock and configuration signals

The following table shows the clock and configuration signals for the ACP interface.

Table A-27 Clock and configuration signals

Signal Type Description

ACLKENS Input ACP clock enable.

AINACTS Input ACP inactive control. When this signal is HIGH, the ACP stops accepting requests by deasserting ARREADYS
and AWREADYS. When AINACTS is asserted, the SoC must not assert ARVALIDS, AWVALIDS, or
WVALIDS.

Related information
2.3.1 Clocks on page 2-32.
2.4.1 Dynamic power management on page 2-42.

A.11.2 Write address channel signals

The following table shows the write address channel signals for the ACP interface.

Table A-28 Write address channel signals

Signal Type Description

AWREADYS Output Write address ready.

AWVALIDS Input Write address valid.

AWIDS[4:0] Input Write request ID.

AWADDRS[43:0] Input Write address.

AWLENS[7:0] Input Write burst length.

AWCACHES[3:0] Input Write cache type.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-553
reserved.
Non-Confidential
 A Signal Descriptions
A.11 ACP interface signals

Table A-28 Write address channel signals (continued)

Signal Type Description

AWUSERS[1:0] Input Write attributes:

[1] Outer Shareable.
[0] Inner Shareable.

AWPROTS[2:0] Input Write protection type.

Note
The ACP interface uses the AXI4 defined default values for the following input signals:
0b100 AWSIZES[2:0].
0b01 AWBURSTS[1:0].
0b0 AWLOCKS.

Related information
7.8.2 ACP ARUSER and AWUSER signals on page 7-317.

A.11.3 Write data channel signals

The following table shows the write data channel signals for the ACP interface.

Table A-29 Write data channel signals

Signal Type Description

WREADYS Output Write data ready

WVALIDS Input Write data valid

WDATAS[127:0] Input Write data

WSTRBS[15:0] Input Write byte-lane strobes

WLASTS Input Write data last transfer indication

A.11.4 Write response channel signals

The following table shows the write response channel signals for the ACP interface.

Table A-30 Write response channel signals

Signal Type Description

BREADYS Input Write response ready

BVALIDS Output Write response valid

BIDS[4:0] Output Write response ID

BRESPS[1:0] Output Write response

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-554
reserved.
Non-Confidential
 A Signal Descriptions
A.11 ACP interface signals

A.11.5 Read address channel signals

The following table shows the read address channel signals for the ACP interface.

Table A-31 Read address channel signals

Signal Type Description

ARREADYS Output Read address ready.

ARVALIDS Input Read address valid.

ARIDS[4:0] Input Read request ID.

ARADDRS[43:0] Input Read address.

ARLENS[7:0] Input Read burst length.

ARCACHES[3:0] Input Read cache type.

ARUSERS[1:0] Input Read attributes:

[1] Outer Shareable.
[0] Inner Shareable.

ARPROTS[2:0] Input Read protection type.

Note
The ACP interface uses the AXI4 defined default values for the following input signals:
0b100 ARSIZES[2:0].
0b01 ARBURSTS[1:0].
0b0 ARLOCKS.

Related information
7.8.2 ACP ARUSER and AWUSER signals on page 7-317.

A.11.6 Read data channel signals

The following table shows the read data channel signals for the ACP interface.

Table A-32 Read data channel signals

Signal Type Description

RREADYS Input Read data ready

RVALIDS Output Read data valid

RIDS[4:0] Output Read data ID

RDATAS[127:0] Output Read data

RRESPS[1:0] Output Read data response

RLASTS Output Read data last transfer indication

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-555
reserved.
Non-Confidential
 A Signal Descriptions
A.12 Debug interface signals

A.12 Debug interface signals

Describes the external debug interface signals.
This section contains the following subsections:
• A.12.1 APB interface signals on page Appx-A-556.
• A.12.2 Authentication interface signals on page Appx-A-556.
• A.12.3 Miscellaneous debug signals on page Appx-A-557.

A.12.1 APB interface signals

The following table shows the APB interface signals.

Table A-33 APB interface signals

Signal Type Description

PCLKDBG Input APB clock.

PCLKENDBG Input APB clock enable.

nPRESETDBG Input Active-LOW APB reset:

0 Reset APB.
1 Do not reset APB.

PSELDBG Input Debug registers select:

0 Debug registers not selected.
1 Debug registers selected.

PADDRDBG[21:2] Input APB address bus bits[21:2].

PADDRDBG31 Input APB address bus bit[31]:

0 Not an external debugger access.
1 External debugger access.

PENABLEDBG Input Indicates the second and subsequent cycles of an APB transfer.

PWRITEDBG Input APB read or write signal:

0 Reads from APB.
1 Writes to APB.

PWDATADBG[31:0] Input APB write data bus.

PRDATADBG[31:0] Output APB read data bus.

PREADYDBG Output APB slave ready. An APB slave can assert PREADYDBG to extend a transfer by inserting wait
states.

PSLVERRDBG Output APB slave transfer error:

0 No transfer error.
1 Transfer error.

A.12.2 Authentication interface signals

The following table shows the authentication interface signals.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-556
reserved.
Non-Confidential
 A Signal Descriptions
A.12 Debug interface signals

Table A-34 Authentication interface signals

Signal Type Description

DBGEN[N:0] Input Invasive debug enable:

0 Not enabled.
1 Enabled.

NIDEN[N:0] Input Non-invasive debug enable:

0 Not enabled.
1 Enabled.

SPIDEN[N:0] Input Secure privileged invasive debug enable:

0 Not enabled.
1 Enabled.

SPNIDEN[N:0] Input Secure privileged non-invasive debug enable:

0 Not enabled.
1 Enabled.

A.12.3 Miscellaneous debug signals

The following table shows the miscellaneous debug signals.

Table A-35 Miscellaneous debug signals

Signal Type Description

DBGROMADDR[43:12] Input Specifies bits[43:12] of the top-level ROM table Physical Address.
If the address cannot be determined, tie this signal LOW.
This signal is only sampled during powerup reset of the processor.

DBGROMADDRV Input Valid signal for DBGROMADDR.

If the address cannot be determined, tie this signal LOW.
This signal is only sampled during powerup reset of the processor.

DBGACK[N:0] Output Debug acknowledge:

0 Debug not acknowledged.
1 Debug acknowledged.

nCOMMIRQ[N:0] Output Communications channel receive or transmit interrupt request, active LOW:
0 Receive section data transfer register is full or transmit section data transfer register is empty.
1 Either or both:
• The receive section data transfer register is empty.
• The transmit section data transfer register is empty.

COMMRX[N:0] Output Communications channel receive. Receive portion of Data Transfer Register full flag:
0 Empty.
1 Full.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-557
reserved.
Non-Confidential
 A Signal Descriptions
A.12 Debug interface signals

Table A-35 Miscellaneous debug signals (continued)

Signal Type Description

COMMTX[N:0] Output Communication channel transmit. Transmit portion of Data Transfer Register empty flag:
0 Full.
1 Empty.

EDBGRQ[N:0] Input External debug request:

0 No external debug request.
1 External debug request.

The processor treats the EDBGRQ input as level-sensitive. The EDBGRQ input must be
asserted until the processor asserts DBGACK.

DBGRSTREQ[N:0] Output Warm reset request:

0 Warm reset is not requested.
1 Request Warm reset.

This output is controlled by Warm reset request bit in External Debug Power/Reset Control
Register, EDPRCR.

DBGNOPWRDWN[N:0] Output No powerdown request. On a powerdown request:

0 The SoC power controller powers down the processor.
1 The SoC power controller does not power down the processor.

DBGPWRDUP[N:0] Input Processor power status:

0 Processor is not powered up.
1 Processor is powered up.

DBGPWRUPREQ[N:0] Output Processor powerup request:

0 No request for processor power up.
1 Request for processor power up.

DBGL1RSTDISABLE Input Disable L1 data cache and L2 snoop tag RAM automatic invalidate on reset functionality.
0 Enable automatic invalidation of L1 data cache and L2 snoop tag RAMs on reset.
1 Disable automatic invalidation of L1 data cache and L2 snoop tag RAMs on reset

This signal is sampled only during reset of the processor.

Related information
WARMRSTREQ and DBGRSTREQ on page 2-40.
External debug over powerdown on page 2-57.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-558
reserved.
Non-Confidential
 A Signal Descriptions
A.13 ETM interface

A.13 ETM interface

Describes the ETM interface.
This section contains the following subsections:
• A.13.1 ATB interface on page Appx-A-559.
• A.13.2 Miscellaneous ETM signal on page Appx-A-559.

A.13.1 ATB interface

The following table shows the signals of the ATB interface.

Table A-36 ATB interface signals

Signal Type Description

ATCLKEN Input ATB clock enable

ATREADYMx Input ATB device ready:

0 Not ready.
1 Ready.

AFVALIDMx Input FIFO flush request.

ATDATAMx[31:0] Output ATB data bus.

ATVALIDMx Output ATB valid data:

0 No valid data.
1 Valid data.

ATBYTESMx[1:0] Output CoreSight ATB device data size:

0b00 1 byte.
0b01 2 byte.
0b10 3 byte.
0b11 4 byte.

AFREADYMx Output FIFO flush acknowledge:

0 FIFO flush not complete.
1 FIFO flush complete.

ATIDMx[6:0] Output ATB trace source identification.

SYNCREQMx Input Synchronization request.

The input must be driven HIGH for one ATCLK cycle.

A.13.2 Miscellaneous ETM signal

The following table shows the miscellaneous ETM interface signal.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-559
reserved.
Non-Confidential
 A Signal Descriptions
A.13 ETM interface

Table A-37 Miscellaneous ETM interface signal

Signal Type Description

TSVALUEB[63:0] Input Global system timestamp value in binary format

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-560
reserved.
Non-Confidential
 A Signal Descriptions
A.14 Cross trigger channel interface

A.14 Cross trigger channel interface

The following table shows the cross trigger channel interface signals.

Table A-38 Cross trigger channel interface signals

Signal Type Description

CIHSBYPASS[3:0] Input Cross trigger channel interface handshake bypass.

CISBYPASS Input Cross trigger channel interface sync bypass.

CTICHIN[3:0] Input Cross trigger channel input. Each bit represents a valid channel input:
0 Channel input inactive.
1 Channel input active.

CTICHINACK[3:0] Output Cross trigger channel input acknowledge.

CTICHOUT[3:0] Output Cross trigger channel output. Each bit represents a valid channel output:
0 Channel output inactive.
1 Channel output active.

CTICHOUTACK[3:0] Input Cross trigger channel output acknowledge.

CTIIRQ[N:0] Output Active-HIGH cross trigger interrupt output:

0 Interrupt not active.
1 Interrupt active.

CTIIRQACK[N:0] Input Cross trigger interrupt acknowledge.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-561
reserved.
Non-Confidential
 A Signal Descriptions
A.15 PMU signals

A.15 PMU signals

The following table shows the performance monitoring signals.

Table A-39 Performance monitoring signals

Signal Type Description

nPMUIRQ[N:0] Output PMU interrupt signal

PMUEVENTx[24:0] Output PMU event bus

PMUSNAPSHOTREQ[N:0] Input PMU snapshot trigger request

PMUSNAPSHOTACK[N:0] Output PMU snapshot trigger acknowledge

Related information
11.8 Events on page 11-428.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-562
reserved.
Non-Confidential
 A Signal Descriptions
A.16 DFT and MBIST signals

A.16 DFT and MBIST signals

Describes the DFT and MBIST signals.
This section contains the following subsections:
• A.16.1 DFT signals on page Appx-A-563.
• A.16.2 MBIST interface on page Appx-A-563.

A.16.1 DFT signals

The following table shows the DFT interface signals.

Table A-40 DFT interface signals

Signal Type Description

DFTCLKBYPASS Input Bypasses the strobe clock register to the L2 RAMs, forcing the L2 RAMs to be tested using
CLK as the source clock

DFTCRCLKDISABLE[N:0] Input Disables processor clock grid

DFTL2CLKDISABLE Input Disables L2 clock grid

DFTMCPHOLD Input Disables multi-cycle paths on RAM interfaces

DFTRAMHOLD Input Disables the RAM chip selects during scan shift

DFTRSTDISABLE Input Disables internal synchronized reset during scan shift

DFTSE Input Scan shift enable, forces on the clock grids during scan shift

A.16.2 MBIST interface

The following table shows the Memory Built-In Self Test (MBIST) interface signals.

Table A-41 MBIST interface signals

Signal Type Description

nMBISTRESET Input MBIST reset

MBISTREQ Input MBIST test request

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-A-563
reserved.
Non-Confidential
Appendix B
AArch32 Unpredictable Behaviors

This appendix describes specific Cortex-A72 processor UNPREDICTABLE behaviors that are of particular
interest.

It contains the following sections:

• B.1 Unpredictable behaviors on page Appx-B-565.
• B.2 Debug UNPREDICTABLE behaviors on page Appx-B-567.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-B-564
reserved.
Non-Confidential
 B AArch32 Unpredictable Behaviors
B.1 Unpredictable behaviors

B.1 Unpredictable behaviors

The following sections describe how the Cortex-A72 processor implementation differs from the
preferred AArch32 UNPREDICTABLE behaviors.
This section contains the following subsections:
• B.1.1 Use of R15 by instruction on page Appx-B-565.
• B.1.2 Load or store accesses that span a page boundary on page Appx-B-566.

B.1.1 Use of R15 by instruction

The Cortex-A72 processor does not implement a Read 0 policy on UNPREDICTABLE use of R15 by
instruction. Instead, the processor reads the PC with the standard offset that applies for the current
instruction set with alignment to a word boundary.
Word-alignment of the PC is imposed for all T32 instructions that are either:
• Defined as loads in the definition of PMU event 0x70.
• Defined as stores in the definition of PMU event 0x71.
With the notable exceptions to this alignment policy that:
• The PC value for TBB and TBH instructions is explicitly not forced to a word-aligned value. TBB and
TBH are technically PMU loads but for the processor to comply with the architecture, it cannot force
the PC to a word-aligned value for these instructions.
• The PC value for ADR instructions is explicitly forced to a word-aligned value. ADR is not a PMU load
or a PMU store, but the architecture specifies word-aligned PC for ADR instructions.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-B-565
reserved.
Non-Confidential
 B AArch32 Unpredictable Behaviors
B.1 Unpredictable behaviors

B.1.2 Load or store accesses that span a page boundary

This section describes load or store accesses that cross page boundaries.
The behavior of the Cortex-A72 processor is as follows:
• Store crosses a page boundary:
— The processor performs two stores, one to each page. The stores behave according to the
attributes of the page that each store hits.
• Load crosses a page boundary:
Device to Device, Normal to Normal
The processor performs two loads, one from each page. The loads behave according to the
attributes of the page that each load hits.
Device to Normal, Normal to Device
The processor generates an Alignment fault.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-B-566
reserved.
Non-Confidential
 B AArch32 Unpredictable Behaviors
B.2 Debug UNPREDICTABLE behaviors

B.2 Debug UNPREDICTABLE behaviors

This section describes the behavior that the Cortex-A72 processor implements when:
• A topic has multiple options.
• The behavior differs from either or both of the Options and Preferences behaviors.

Note
This section does not describe the behavior when a topic only has a single option and the processor
implements the preferred behavior.

This section contains the following subsections:

• B.2.1 A32 BKPT instruction with condition code not AL on page Appx-B-568.
• B.2.2 Address match breakpoint match only on second halfword of an instruction
on page Appx-B-568.
• B.2.3 Address matching breakpoint on A32 instruction with DBGBCRn.BAS=1100
on page Appx-B-568.
• B.2.4 Address match breakpoint match on T32 instruction at DBGBCRn+2 with
DBGBCRn.BAS=1111 on page Appx-B-568.
• B.2.5 Address mismatch breakpoint match only on second halfword of an instruction
on page Appx-B-568.
• B.2.6 Address mismatch breakpoint match on T32 instruction at DBGBCRn+2 with
DBGBCRn.BAS=1111 on page Appx-B-568.
• B.2.7 Other mismatch breakpoint matches any address in current mode and state
on page Appx-B-569.
• B.2.8 Mismatch breakpoint on branch to self on page Appx-B-569.
• B.2.9 Link to nonexistent breakpoint or breakpoint that is not context-aware on page Appx-B-569.
• B.2.10 DBGWCRn_EL1.MASK!=00000 and DBGWCRn_EL1.BAS!=11111111
on page Appx-B-569.
• B.2.11 Address-matching Vector catch on 32-bit T32 instruction at vector–2 on page Appx-B-569.
• B.2.12 Address-matching Vector catch on 32-bit T32 instruction at vector+2 on page Appx-B-569.
• B.2.13 Address-matching Vector catch and Breakpoint on same instruction on page Appx-B-569.
• B.2.14 Address match breakpoint with DBGBCRn_EL1.BAS=0000 on page Appx-B-569.
• B.2.15 DBGWCRn_EL1.BAS specifies a non-contiguous set of bytes within a doubleword
on page Appx-B-569.
• B.2.16 A32 HLT instruction with condition code not AL on page Appx-B-570.
• B.2.17 Execute instruction at a given EL when the corresponding EDECCR bit is 1 and Halting is
allowed on page Appx-B-570.
• B.2.18 Unlinked Context matching and Address mismatch breakpoints taken to Abort mode
on page Appx-B-570.
• B.2.19 Vector catch on Data or Prefetch Abort, and taken to Abort mode on page Appx-B-570.
• B.2.20 H > N or H = 0 at Non-secure EL1 and EL0, including value read from PMCR_EL0.N
on page Appx-B-571.
• B.2.21 H > N or H = 0: value read back in MDCR_EL2.HPMN on page Appx-B-571.
• B.2.22 P ≥ M and P ≠ 31: reads and writes of PMXEVTYPER_EL0 and PMXEVCNTR_EL0
on page Appx-B-571.
• B.2.23 P ≥ M and P ≠ 31: value read in PMSELR_EL0.SEL on page Appx-B-571.
• B.2.24 P = 31: reads and writes of PMXEVCNTR_EL0 on page Appx-B-571.
• B.2.25 n ≥ M: Direct access to PMEVCNTRn_EL0 and PMEVTYPERn_EL0 on page Appx-B-571.
• B.2.26 Exiting Debug state while instruction issued through EDITR is in flight
on page Appx-B-572.
• B.2.27 Using memory-access mode with a non-word-aligned address on page Appx-B-572.
• B.2.28 Access to memory-mapped registers mapped to Normal memory on page Appx-B-572.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-B-567
reserved.
Non-Confidential
 B AArch32 Unpredictable Behaviors
B.2 Debug UNPREDICTABLE behaviors

• B.2.29 Not word-sized accesses or (AArch64 only) doubleword-sized accesses

on page Appx-B-572.
• B.2.30 External debug write to register that is being reset on page Appx-B-572.
• B.2.31 Accessing reserved debug registers on page Appx-B-572.
• B.2.32 Clearing the clear-after-read EDPRSR bits when Core power domain is on, and
DoubleLockStatus() is TRUE on page Appx-B-573.

B.2.1 A32 BKPT instruction with condition code not AL

The processor implements the preferred option, that is:

• Executed unconditionally.

B.2.2 Address match breakpoint match only on second halfword of an instruction

The processor generates a breakpoint on the instruction, unless it is a breakpoint on the second half of the
first 32-bit instruction in an aligned 128-bit region or following a taken branch. In this case the
breakpoint is taken on the following instruction.

B.2.3 Address matching breakpoint on A32 instruction with DBGBCRn.BAS=1100

An address match occurs, unless the instruction is the first instruction within an instruction fetch, that is
the first instruction in a 128-bit aligned region for a sequential fetch, or first instruction following a taken
branch. In this case the breakpoint is taken on the following instruction.

B.2.4 Address match breakpoint match on T32 instruction at DBGBCRn+2 with

DBGBCRn.BAS=1111

The processor implements:

• Does match.

B.2.5 Address mismatch breakpoint match only on second halfword of an instruction

The processor implements:

• Does match.

B.2.6 Address mismatch breakpoint match on T32 instruction at DBGBCRn+2 with

DBGBCRn.BAS=1111

The processor behaves as follows:

• If BVRn+2 is directly jumped to, then the breakpoint is taken on the instruction following BVRn+2.
The instruction is stepped.
• If BVRn precedes a 16-bit instruction, then the breakpoint is taken on the instruction at DBGBVRn
+2.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-B-568
reserved.
Non-Confidential
 B AArch32 Unpredictable Behaviors
B.2 Debug UNPREDICTABLE behaviors

B.2.7 Other mismatch breakpoint matches any address in current mode and state

The processor implements:

• Immediate Breakpoint debug event.

B.2.8 Mismatch breakpoint on branch to self

The processor implements:

• Instruction is stepped an UNKNOWN number of times, while it continues to branch to itself.

B.2.9 Link to nonexistent breakpoint or breakpoint that is not context-aware

The processor implements:

• No Breakpoint or Watchpoint debug event is generated, and the LBN field of the linker reads
UNKNOWN.

B.2.10 DBGWCRn_EL1.MASK!=00000 and DBGWCRn_EL1.BAS!=11111111

The processor implements the preferred behavior:

• DBGWCRn_EL1.BAS is ignored and treated as if 0b11111111.

B.2.11 Address-matching Vector catch on 32-bit T32 instruction at vector–2

The processor implements:

• Does match, unless it is the first instruction following a discontinuity, a branch, in which case it
matches on the following instruction.

B.2.12 Address-matching Vector catch on 32-bit T32 instruction at vector+2

The processor implements:

• Does match, unless it is the first instruction following a discontinuity, a branch, in which case it
matches on the following instruction.

B.2.13 Address-matching Vector catch and Breakpoint on same instruction

The processor implements the preferred option, that is:

• Report Breakpoint.

B.2.14 Address match breakpoint with DBGBCRn_EL1.BAS=0000

The processor implements the preferred option, that is:

• As if disabled.

B.2.15 DBGWCRn_EL1.BAS specifies a non-contiguous set of bytes within a doubleword

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-B-569
reserved.
Non-Confidential
 B AArch32 Unpredictable Behaviors
B.2 Debug UNPREDICTABLE behaviors

The processor behaves as follows:

• A Watchpoint debug event is generated for each byte.

B.2.16 A32 HLT instruction with condition code not AL

The processor implements the preferred option, that is:

• Executed unconditionally.

B.2.17 Execute instruction at a given EL when the corresponding EDECCR bit is 1 and Halting is
allowed

The processor behaves as follows:

• Generates debug event and Halt no later than the instruction following the next Context
Synchronization operation (CSO) excluding ISB instruction.

B.2.18 Unlinked Context matching and Address mismatch breakpoints taken to Abort mode

The processor implements:

• A Prefetch Abort debug exception is generated. Because the breakpoint is configured to generate a
breakpoint at PL1, the instruction at the Prefetch Abort vector generates a Vector catch debug event.

Note
The debug event is subject to the same CONSTRAINED UNPREDICTABLE behavior, so the Breakpoint debug
event repeats for an UNKNOWN number of times.

B.2.19 Vector catch on Data or Prefetch Abort, and taken to Abort mode

The processor implements:

• A Prefetch Abort debug exception is generated. If Vector catch is enabled on the Prefetch Abort
vector, this generates a Vector catch debug event.

Note
The debug event is subject to the same CONSTRAINED UNPREDICTABLE behavior, so the Vector catch debug
event repeats for an UNKNOWN number of times.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-B-570
reserved.
Non-Confidential
 B AArch32 Unpredictable Behaviors
B.2 Debug UNPREDICTABLE behaviors

B.2.20 H > N or H = 0 at Non-secure EL1 and EL0, including value read from PMCR_EL0.N

The processor implements:

• HPMN[4:0], and in Non-secure EL1 and EL0:
— If H > N then M = N.
— If H = 0 then M = 0.

B.2.21 H > N or H = 0: value read back in MDCR_EL2.HPMN

The processor implements:

• HPMN[4:0], and reads return H.

B.2.22 P ≥ M and P ≠ 31: reads and writes of PMXEVTYPER_EL0 and PMXEVCNTR_EL0

The processor implements:

• SEL[4:0], and if P ≥ M and P ≠ 31 then the register is RES0.

B.2.23 P ≥ M and P ≠ 31: value read in PMSELR_EL0.SEL

The processor implements:

• SEL[4:0], and if P ≥ M and P ≠ 31 then the register is RES0.

B.2.24 P = 31: reads and writes of PMXEVCNTR_EL0

The processor implements:

• The register is RES0.

B.2.25 n ≥ M: Direct access to PMEVCNTRn_EL0 and PMEVTYPERn_EL0

The processor implements:

• If n ≥ N then the instruction is UNALLOCATED.
• Otherwise if n ≥ M then the register is RES0.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-B-571
reserved.
Non-Confidential
 B AArch32 Unpredictable Behaviors
B.2 Debug UNPREDICTABLE behaviors

B.2.26 Exiting Debug state while instruction issued through EDITR is in flight

The processor implements:

• The instruction completes in Debug state before executing the restart.

B.2.27 Using memory-access mode with a non-word-aligned address

The processor implements the preferred behavior, that is:

• Does unaligned accesses, faulting if these are not permitted for the memory type.

B.2.28 Access to memory-mapped registers mapped to Normal memory

The processor implements the preferred behavior, that is:

• The access is generated, and accesses might be repeated, gathered, split or resized, in accordance with
the rules for Normal memory, meaning the effect is UNPREDICTABLE.

B.2.29 Not word-sized accesses or (AArch64 only) doubleword-sized accesses

The processor implements the preferred behavior, that is:

• Reads occur and return UNKNOWN data.
• Writes set the accessed registers to UNKNOWN.

B.2.30 External debug write to register that is being reset

The processor implements the preferred behavior, that is:

• Takes reset value.

B.2.31 Accessing reserved debug registers

The processor deviates from the preferred behavior because the hardware cost to decode some of the
addresses in the debug power domain is significant.
The processor behavior is:
1. For reserved debug registers 0x000 - 0xCFC and reserved Performance Monitors registers 0x000 - 
0xF00, the response is CONSTRAINED UNPREDICTABLE Error, when any of the following apply:
Off Core power domain is either completely off, or in a low-power state where the Core power
domain registers are not accessible.
DLK DoubleLockStatus() is TRUE, OS double-lock is locked, that is, EDPRSR.DLK is 1.
OSLK OSLSR_EL1.OSLK is 1, OS Lock is locked.
2. For reserved debug registers in the address ranges 0x400 - 0x4FC and 0x800 - 0x8FC, the response is
CONSTRAINED UNPREDICTABLE Error when the conditions in 1 do not apply and:

EDAD AllowExternalDebugAccess() is FALSE, external debug access is disabled.

3. For reserved Performance Monitor registers in the address ranges 0x000 - 0x0FC and 0x400 - 0x47C,
the response is CONSTRAINED UNPREDICTABLE Error when the conditions in 1 and 2 do not apply but the
following condition applies:
EPMAD AllowExternalPMUAccess() is FALSE (external Performance Monitors access is
disabled).

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-B-572
reserved.
Non-Confidential
 B AArch32 Unpredictable Behaviors
B.2 Debug UNPREDICTABLE behaviors

B.2.32 Clearing the clear-after-read EDPRSR bits when Core power domain is on, and
DoubleLockStatus() is TRUE

The processor implements the preferred behavior, that is:

• Bits are not cleared to zero.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-B-573
reserved.
Non-Confidential
Appendix C
Revisions

This appendix describes the technical changes between released issues of this book.

It contains the following sections:

• C.1 Revisions on page Appx-C-575.

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-C-574
reserved.
Non-Confidential
 C Revisions
C.1 Revisions

C.1 Revisions

Table C-1 Issue 01

Change Location Affects

First release - -

Table C-2 Issue 02

Change Location Affects

Updated for support of 4MB L2 cache size Throughout document r0p1

Added configuration options for up to two L2 Data RAM slices Throughout document r0p1

L1 hardware prefetcher terminology clarified to load-store hardware prefetcher Throughout document r0p1

Updated Main ID Register Value (MIDR) Throughout document r0p1

Table C-3 Issue 03

Change Location Affects

Updated bit [23] of L2 Auxiliary Control Register, EL1 4.3.65 L2 Auxiliary Control Register, EL1 on page 4-188 r0p2

Updated bit [41] of CPU Auxiliary Control Register, EL1 4.3.66 CPU Auxiliary Control Register, EL1 on page 4-194 r0p2

Table C-4 Issue 04

Change Location Affects

Updated bits to write to disable L2 prefetch • Individual core powerdown on page 2-52 r0p2
• L2 hardware cache flush on page 2-45

Updated Main ID Register Value (MIDR) Throughout document r0p2

Table C-5 Issue 05

Change Location Affects

Updated Main ID Register Value (MIDR) Throughout document r0p3

Updated Trace ID Revision Value Throughout document r0p3

Table C-6 Issue 06

Change Location Affects

Removed statement about identifying requests coming from the master 7.7.6 ACE ARID and AWID assignment r0p3
connected to the optional slave ACP port. on page 7-313

ARM 100095_0003_06_en Copyright © 2014-2016 ARM Limited or its affiliates. All rights Appx-C-575
reserved.
Non-Confidential