Devicetree Specification v0.4
Devicetree Specification v0.4
Release v0.4
devicetree.org
28 June 2023
CONTENTS
1 Introduction 3
1.1 Purpose and Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Relationship to IEEE™ 1275 and ePAPR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 32-bit and 64-bit Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Definition of Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 The Devicetree 6
2.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Devicetree Structure and Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2.1 Node Names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2.2 Generic Names Recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2.3 Path Names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2.4 Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.3 Standard Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3.1 compatible . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3.2 model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3.3 phandle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3.4 status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3.5 #address-cells and #size-cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.3.6 reg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.7 virtual-reg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.8 ranges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.3.9 dma-ranges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.3.10 dma-coherent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.3.11 dma-noncoherent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.3.12 name (deprecated) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.3.13 device_type (deprecated) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4 Interrupts and Interrupt Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4.1 Properties for Interrupt Generating Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.4.2 Properties for Interrupt Controllers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.4.3 Interrupt Nexus Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.4.4 Interrupt Mapping Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.5 Nexus Nodes and Specifier Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.5.1 Nexus Node Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.5.2 Specifier Mapping Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
i
3.5.1 /reserved-memory parent node . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.5.2 /reserved-memory/ child nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.5.3 Device node references to reserved memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.5.4 /reserved-memory and UEFI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.5.5 /reserved-memory Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.6 /chosen Node . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.7 /cpus Node Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.8 /cpus/cpu* Node Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.8.1 General Properties of /cpus/cpu* nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.8.2 TLB Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.8.3 Internal (L1) Cache Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.8.4 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.9 Multi-level and Shared Cache Nodes (/cpus/cpu*/l?-cache) . . . . . . . . . . . . . . . . . . . . . . 41
3.9.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4 Device Bindings 43
4.1 Binding Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.1.1 General Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.1.2 Miscellaneous Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.2 Serial devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.2.1 Serial Class Binding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.2.2 National Semiconductor 16450/16550 Compatible UART Requirements . . . . . . . . . . . . . 45
4.3 Network devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.3.1 Network Class Binding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.3.2 Ethernet specific considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.4 Power ISA Open PIC Interrupt Controllers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.5 simple-bus Compatible Value . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Bibliography 60
Index 61
ii
Devicetree Specification, Release v0.4
Copyright
The Linaro and devicetree.org word marks and the Linaro and devicetree.org logos and related marks are trademarks and
service marks licensed by Linaro Ltd. Implementation of certain elements of this document may require licenses under
third party intellectual property rights, including without limitation, patent rights. Linaro and its Members are not, and
shall not be held, responsible in any manner for identifying or failing to identify any or all such third party intellectual
property rights.
The Power Architecture and Power.org word marks and the Power and Power.org logos and related marks are trademarks
and service marks licensed by Power.org. Implementation of certain elements of this document may require licenses under
third party intellectual property rights, including without limitation, patent rights. Power.org and its Members are not, and
shall not be held, responsible in any manner for identifying or failing to identify any or all such third party intellectual
property rights.
THIS SPECIFICATION PROVIDED “AS IS” AND WITHOUT ANY WARRANTY OF ANY KIND, INCLUD-
ING, WITHOUT LIMITATION, ANY EXPRESS OR IMPLIED WARRANTY OF NON-INFRINGEMENT, MER-
CHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT SHALL LINARO OR ANY MEM-
BER OF LINARO BE LIABLE FOR ANY DIRECT, INDIRECT, SPECIAL, EXEMPLARY, PUNITIVE, OR CON-
SEQUENTIAL DAMAGES, INCLUDING, WITHOUT LIMITATION, LOST PROFITS, EVEN IF ADVISED OF THE
POSSIBILITY OF SUCH DAMAGES.
Questions pertaining to this document, or the terms or conditions of its provision, should be addressed to:
Linaro, Ltd
Harston Mill,
Royston Road,
Harston CB22 7GG
Attn: Devicetree.org Board Secretary
License Information
Licensed under the Apache License, Version 2.0 (the “License”); you may not use this file except in compliance with the
License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an “AS
IS” BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License
for the specific language governing permissions and limitations under the License.
CONTENTS 1
Devicetree Specification, Release v0.4
Acknowledgements
The devicetree.org Technical Steering Committee would like to thank the many individuals and companies that contributed
to the development of this specification through writing, technical discussions and reviews.
We want to thank the power.org Platform Architecture Technical Subcommittee who developed and published ePAPR. The
text of ePAPR was used as the starting point for this document.
Significant aspects of the Devicetree Specification are based on work done by the Open Firmware Working Group which
developed bindings for IEEE-1275. We would like to acknowledge their contributions.
We would also like to acknowledge the contribution of the PowerPC and ARM Linux communities that developed and
implemented the flattened devicetree concept.
0.3 2020-FEB-13
• Added more recommended generic node names
• Document generic nexus binding
• codespell support and spelling fixes
CONTENTS 2
CHAPTER
ONE
INTRODUCTION
To initialize and boot a computer system, various software components interact. Firmware might perform low-level initial-
ization of the system hardware before passing control to software such as an operating system, bootloader, or hypervisor.
Bootloaders and hypervisors can, in turn, load and transfer control to operating systems. Standard, consistent interfaces
and conventions facilitate the interactions between these software components. In this document the term boot program is
used to generically refer to a software component that initializes the system state and executes another software component
referred to as a client program. Examples of a boot program include: firmware, bootloaders, and hypervisors. Examples of
a client program include: bootloaders, hypervisors, operating systems, and special purpose programs. A piece of software
may be both a client program and a boot program (e.g. a hypervisor).
This specification, the Devicetree Specification (DTSpec), provides a complete boot program to client program interface
definition, combined with minimum system requirements that facilitate the development of a wide variety of systems.
This specification is targeted towards the requirements of embedded systems. An embedded system typically consists of
system hardware, an operating system, and application software that are custom designed to perform a fixed, specific set of
tasks. This is unlike general purpose computers, which are designed to be customized by a user with a variety of software
and I/O devices. Other characteristics of embedded systems may include:
• a fixed set of I/O devices, possibly highly customized for the application
• a system board optimized for size and cost
• limited user interface
• resource constraints like limited memory and limited nonvolatile storage
• real-time constraints
• use of a wide variety of operating systems, including Linux, real-time operating systems, and custom or proprietary
operating systems
Organization of this Document
• Chapter 1 introduces the architecture being specified by DTSpec.
• Chapter 2 introduces the devicetree concept and describes its logical structure and standard properties.
• Chapter 3 specifies the definition of a base set of device nodes required by DTSpec-compliant devicetrees.
• Chapter 4 describes device bindings for certain classes of devices and specific device types.
• Chapter 5 specifies the DTB encoding of devicetrees.
• Chapter 6 specifies the DTS source language.
Conventions Used in this Document
The word shall is used to indicate mandatory requirements strictly to be followed in order to conform to the standard and
from which no deviation is permitted (shall equals is required to).
3
Devicetree Specification, Release v0.4
The word should is used to indicate that among several possibilities one is recommended as particularly suitable, without
mentioning or excluding others; or that a certain course of action is preferred but not necessarily required; or that (in the
negative form) a certain course of action is deprecated but not prohibited (should equals is recommended that).
The word may is used to indicate a course of action permissible within the limits of the standard (may equals is permitted).
Examples of devicetree constructs are frequently shown in Devicetree Syntax form. See Section 6 for an overview of this
syntax.
DTSpec is loosely related to the IEEE 1275 Open Firmware standard—IEEE Standard for Boot (Initialization Configura-
tion) Firmware: Core Requirements and Practices [IEEE1275].
The original IEEE 1275 specification and its derivatives such as CHRP [CHRP] and PAPR [PAPR] address problems of
general purpose computers, such as how a single version of an operating system can work on several different computers
within the same family and the problem of loading an operating system from user-installed I/O devices.
Because of the nature of embedded systems, some of these problems faced by open, general purpose computers do not
apply. Notable features of the IEEE 1275 specification that are omitted from the DTSpec include:
• Plug-in device drivers
• FCode
• The programmable Open Firmware user interface based on Forth
• FCode debugging
• Operating system debugging
What is retained from IEEE 1275 are concepts from the devicetree architecture by which a boot program can describe and
communicate system hardware information to a client program, thus eliminating the need for the client program to have
hard-coded descriptions of system hardware.
This specification partially supersedes the ePAPR [EPAPR] specification. ePAPR documents how devicetree is used by
the Power ISA, and covers both general concepts, as well as Power ISA specific bindings. The text of this document was
derived from ePAPR, but either removes architecture specific bindings, or moves them into an appendix.
The DTSpec supports CPUs with both 32-bit and 64-bit addressing capabilities. Where applicable, sections of the DTSpec
describe any requirements or considerations for 32-bit and 64-bit addressing.
AMP
Asymmetric Multiprocessing. Computer available CPUs are partitioned into groups, each running a distinct oper-
ating system image. The CPUs may or may not be identical.
boot CPU
The first CPU which a boot program directs to a client program’s entry point.
Book III-E
Embedded Environment. Section of the Power ISA defining supervisor instructions and related facilities used in
embedded Power processor implementations.
boot program
Used to generically refer to a software component that initializes the system state and executes another software
component referred to as a client program. Examples of a boot program include: firmware, bootloaders, and hyper-
visors.
client program
Program that typically contains application or operating system software. Examples of a client program include:
bootloaders, hypervisors, operating systems, and special purpose programs.
cell
A unit of information consisting of 32 bits.
DMA
Direct memory access
DTB
Devicetree blob. Compact binary representation of the devicetree.
DTC
Devicetree compiler. An open source tool used to create DTB files from DTS files.
DTS
Devicetree syntax. A textual representation of a devicetree consumed by the DTC. See Appendix A Devicetree
Source Format (version 1).
effective address
Memory address as computed by processor storage access or branch instruction.
physical address
Address used by the processor to access external device, typically a memory controller.
Power ISA
Power Instruction Set Architecture.
interrupt specifier
A property value that describes an interrupt. Typically information that specifies an interrupt number and sensitivity
and triggering mechanism is included.
secondary CPU
CPUs other than the boot CPU that belong to the client program are considered secondary CPUs.
SMP
Symmetric multiprocessing. A computer architecture where two or more identical CPUs can share memory and IO
and operate under a single operating system.
SoC
System on a chip. A single computer chip integrating one or more CPU core as well as number of other peripherals.
unit address
The part of a node name specifying the node’s address in the address space of the parent node.
quiescent CPU
A quiescent CPU is in a state where it cannot interfere with the normal operation of other CPUs, nor can its state be
affected by the normal operation of other running CPUs, except by an explicit method for enabling or re-enabling
the quiescent CPU.
TWO
THE DEVICETREE
2.1 Overview
DTSpec specifies a construct called a devicetree to describe system hardware. A boot program loads a devicetree into a
client program’s memory and passes a pointer to the devicetree to the client.
This chapter describes the logical structure of the devicetree and specifies a base set of properties for use in describing
device nodes. Chapter 3 specifies certain device nodes required by a DTSpec-compliant devicetree. Chapter 4 describes
the DTSpec-defined device bindings – the requirements for representing certain device types or classes of devices. Chapter
5 describes the in-memory encoding of the devicetree.
A devicetree is a tree data structure with nodes that describe the devices in a system. Each node has property/value pairs
that describe the characteristics of the device being represented. Each node has exactly one parent except for the root node,
which has no parent.
A device in this context may be an actual hardware device, such as a UART. It may be part of a hardware device, such as the
random-number generator in a TPM. It may also be a device provided through virtualisation, such as a protocol providing
access to an I2C device attached to a remote CPU. A device may include functions implemented by firmware running in
higher privilege levels or remote processors. There is no requirement that nodes in a device tree be a physical hardware
device, but generally they have some correlation to physical hardware devices. Nodes should not be designed for OS- or
project- specific purposes. They should describe something which can be implemented by any OS or project.
A devicetree is often used to describe devices which cannot necessarily be dynamically detected by a client program. For
example, the architecture of PCI enables a client to probe and detect attached devices and thus devicetree nodes describing
PCI devices might not be required. However, a device node is often used to describe a PCI host bridge device in the system.
This node is required if the bridge cannot be detected by probing, but is otherwise optional. Also, a bootloader may do
PCI probing and produce a device tree containing the results of its scan, for passing to the Operating System.
Example
Fig. 2.1 shows an example representation of a simple devicetree that is nearly complete enough to boot a simple operating
system, with the platform type, CPU, memory and a single UART described. Device nodes are shown with properties and
values inside each node.
6
Devicetree Specification, Release v0.4
cpu@0
device_type="cpu"
reg=<0>
timebase-frequency=<825000000>
cpus clock-frequency=<825000000>
#address-cells=<1>
#size-cells=<0> cpu@1
device_type="cpu"
memory@0 reg=<1>
timebase-frequency=<825000000>
device_type="memory" clock-frequency=<825000000>
reg=<0 0x20000000>
/
uart@fe001000
model="fsl,mpc8572ds"
compatible="fsl,mpc8572ds" compatible="ns16550"
#address-cells=<1> reg=<0xfe001000 0x100>
#size-cells=<1>
chosen
bootargs="root=/dev/sda2"
aliases
serial0="/uart@fe001000"
The node-name shall start with a lower or uppercase character and should describe the general class of device.
The unit-address component of the name is specific to the bus type on which the node sits. It consists of one or more ASCII
characters from the set of characters in Table 2.1. The unit-address must match the first address specified in the reg property
of the node. If the node has no reg property, the @unit-address must be omitted and the node-name alone differentiates the
node from other nodes at the same level in the tree. The binding for a particular bus may specify additional, more specific
requirements for the format of reg and the unit-address.
In the case of node-name without an @unit-address the node-name shall be unique from any property names at the same
level in the tree.
The root node does not have a node-name or unit-address. It is identified by a forward slash (/).
cpu@0
cpus
cpu@1
memory@0
/ uart@fe001000
ethernet@fe002000
ethernet@fe003000
In Fig. 2.2:
• The nodes with the name cpu are distinguished by their unit-address values of 0 and 1.
• The nodes with the name ethernet are distinguished by their unit-address values of fe002000 and fe003000.
The name of a node should be somewhat generic, reflecting the function of the device and not its precise programming
model. If appropriate, the name should be one of the following choices:
A node in the devicetree can be uniquely identified by specifying the full path from the root node, through all descendant
nodes, to the desired node.
The convention for specifying a device path is:
/node-name-1/node-name-2/node-name-N
For example, in Fig. 2.2, the device path to cpu #1 would be:
/cpus/cpu@1
The path to the root node is /.
A unit address may be omitted if the full path to the node is unambiguous.
If a client program encounters an ambiguous path, its behavior is undefined.
2.2.4 Properties
Each node in the devicetree has properties that describe the characteristics of the node. Properties consist of a name and
a value.
Property Names
Property names are strings of 1 to 31 characters from the characters show in Table 2.2
Nonstandard property names should specify a unique string prefix, such as a stock ticker symbol, identifying the name of
the company or organization that defined the property. Examples:
fsl,channel-fifo-len
ibm,ppc-interrupt-server#s
linux,network-index
Property Values
A property value is an array of zero or more bytes that contain information associated with the property.
Properties might have an empty value if conveying true-false information. In this case, the presence or absence of the
property is sufficiently descriptive.
Table 2.3 describes the set of basic value types defined by the DTSpec.
DTSpec specifies a set of standard properties for device nodes. These properties are described in detail in this section.
Device nodes defined by DTSpec (see Chapter 3) may specify additional requirements or constraints regarding the use
of the standard properties. Chapter 4 describes the representation of specific devices and may also specify additional
requirements.
Note: All examples of devicetree nodes in this document use the DTS (Devicetree Source) format for specifying nodes
and properties.
2.3.1 compatible
2.3.2 model
2.3.3 phandle
pic@10000000 {
phandle = <1>;
interrupt-controller;
reg = <0x10000000 0x100>;
};
A phandle value of 1 is defined. Another device node could reference the pic node with a phandle value of 1:
another-device-node {
interrupt-parent = <1>;
};
Note: Older versions of devicetrees may be encountered that contain a deprecated form of this property called linux,
phandle. For compatibility, a client program might want to support linux,phandle if a phandle property is not present.
The meaning and use of the two properties is identical.
Note: Most devicetrees in DTS (see Appendix A) will not contain explicit phandle properties. The DTC tool automatically
inserts the phandle properties when the DTS is compiled into the binary DTB format.
2.3.4 status
soc {
#address-cells = <1>;
#size-cells = <1>;
serial@4600 {
compatible = "ns16550";
reg = <0x4600 0x100>;
clock-frequency = <0>;
interrupts = <0xA 0x8>;
interrupt-parent = <&ipic>;
};
};
In this example, the #address-cells and #size-cells properties of the soc node are both set to 1. This setting
specifies that one cell is required to represent an address and one cell is required to represent the size of nodes
that are children of this node.
The serial device reg property necessarily follows this specification set in the parent (soc) node—the address
is represented by a single cell (0x4600), and the size is represented by a single cell (0x100).
2.3.6 reg
2.3.7 virtual-reg
2.3.8 ranges
If the property is defined with an <empty> value, it specifies that the parent and child address space is identical,
and no address translation is required.
If the property is not present in a bus node, it is assumed that no mapping exists between children of the node
and the parent address space.
Address Translation Example:
soc {
compatible = "simple-bus";
#address-cells = <1>;
#size-cells = <1>;
ranges = <0x0 0xe0000000 0x00100000>;
serial@4600 {
device_type = "serial";
compatible = "ns16550";
reg = <0x4600 0x100>;
clock-frequency = <0>;
interrupts = <0xA 0x8>;
interrupt-parent = <&ipic>;
};
};
2.3.9 dma-ranges
2.3.10 dma-coherent
2.3.11 dma-noncoherent
DTSpec adopts the interrupt tree model of representing interrupts specified in Open Firmware Recommended Practice:
Interrupt Mapping, Version 0.9 [b7]. Within the devicetree a logical interrupt tree exists that represents the hierarchy and
routing of interrupts in the platform hardware. While generically referred to as an interrupt tree it is more technically a
directed acyclic graph.
The physical wiring of an interrupt source to an interrupt controller is represented in the devicetree with the interrupt-parent
property. Nodes that represent interrupt-generating devices contain an interrupt-parent property which has a phandle
value that points to the device to which the device’s interrupts are routed, typically an interrupt controller. If an interrupt-
generating device does not have an interrupt-parent property, its interrupt parent is assumed to be its devicetree parent.
Each interrupt generating device contains an interrupts property with a value describing one or more interrupt sources
for that device. Each source is represented with information called an interrupt specifier. The format and meaning of an
interrupt specifier is interrupt domain specific, i.e., it is dependent on properties on the node at the root of its interrupt
domain. The #interrupt-cells property is used by the root of an interrupt domain to define the number of <u32> values
needed to encode an interrupt specifier. For example, for an Open PIC interrupt controller, an interrupt-specifer takes two
32-bit values and consists of an interrupt number and level/sense information for the interrupt.
An interrupt domain is the context in which an interrupt specifier is interpreted. The root of the domain is either (1) an
interrupt controller or (2) an interrupt nexus.
1. An interrupt controller is a physical device and will need a driver to handle interrupts routed through it. It may also
cascade into another interrupt domain. An interrupt controller is specified by the presence of an interrupt-controller
property on that node in the devicetree.
2. An interrupt nexus defines a translation between one interrupt domain and another. The translation is based on both
domain-specific and bus-specific information. This translation between domains is performed with the interrupt-
map property. For example, a PCI controller device node could be an interrupt nexus that defines a translation from
the PCI interrupt namespace (INTA, INTB, etc.) to an interrupt controller with Interrupt Request (IRQ) numbers.
The root of the interrupt tree is determined when traversal of the interrupt tree reaches an interrupt controller node without
an interrupts property and thus no explicit interrupt parent.
See Fig. 2.3 for an example of a graphical representation of a devicetree with interrupt parent relationships shown. It shows
both the natural structure of the devicetree as well as where each node sits in the logical interrupt tree.
Devicetree
open-pic device3
interrupt-parent=<&gpioctrl>
simple-bus gpioctrl
interrupt-parent=<&open-pic>
pci-host
soc
interrupt-parent=<&open-pic> slot1
interrupt-parent=<&pci-host>
device2
interrupt-parent=<&gpioctrl> slot0
interrupt-parent=<&pci-host>
device1
interrupt-parent=<&open-pic>
Interrupt tree
device1 device3
pci-host slot1
Nexus Node
slot0
– GPIO Controller
• Three interrupt domains exist; one rooted at the open-pic node, one at the PCI host bridge node, and one at
the GPIO Controller node.
• There are two nexus nodes; one at the PCI host bridge and one at the GPIO controller.
interrupts
Property: interrupts
Value type: <prop-encoded-array> encoded as arbitrary number of interrupt specifiers
Description:
The interrupts property of a device node defines the interrupt or interrupts that are generated by the device.
The value of the interrupts property consists of an arbitrary number of interrupt specifiers. The format of an
interrupt specifier is defined by the binding of the interrupt domain root.
interrupts is overridden by the interrupts-extended property and normally only one or the other should be
used.
Example:
A common definition of an interrupt specifier in an open PIC–compatible interrupt domain consists of two
cells; an interrupt number and level/sense information. See the following example, which defines a single
interrupt specifier, with an interrupt number of 0xA and level/sense encoding of 8.
interrupts = <0xA 8>;
interrupt-parent
Property: interrupt-parent
Value type: <phandle>
Description:
Because the hierarchy of the nodes in the interrupt tree might not match the devicetree, the interrupt-parent
property is available to make the definition of an interrupt parent explicit. The value is the phandle to the
interrupt parent. If this property is missing from a device, its interrupt parent is assumed to be its devicetree
parent.
interrupts-extended
Property: interrupts-extended
Value type: <phandle> <prop-encoded-array>
Description:
The interrupts-extended property lists the interrupt(s) generated by a device. interrupts-extended should be
used instead of interrupts when a device is connected to multiple interrupt controllers as it encodes a parent
phandle with each interrupt specifier.
Example:
This example shows how a device with two interrupt outputs connected to two separate interrupt controllers
would describe the connection using an interrupts-extended property. pic is an interrupt controller with an
#interrupt-cells specifier of 2, while gic is an interrupt controller with an #interrupts-cells specifier of 1.
interrupts-extended = <&pic 0xA 8>, <&gic 0xda>;
The interrupts and interrupts-extended properties are mutually exclusive. A device node should use one or the other, but
not both. Using both is only permissible when required for compatibility with software that does not understand interrupts-
extended. If both interrupts-extended and interrupts are present then interrupts-extended takes precedence.
#interrupt-cells
Property: #interrupt-cells
Value type: <u32>
Description:
The #interrupt-cells property defines the number of cells required to encode an interrupt specifier for an
interrupt domain.
interrupt-controller
Property: interrupt-controller
Value type: <empty>
Description:
The presence of an interrupt-controller property defines a node as an interrupt controller node.
interrupt-map
Property: interrupt-map
Value type: <prop-encoded-array> encoded as an arbitrary number of interrupt mapping entries.
Description:
An interrupt-map is a property on a nexus node that bridges one interrupt domain with a set of parent inter-
rupt domains and specifies how interrupt specifiers in the child domain are mapped to their respective parent
domains.
The interrupt map is a table where each row is a mapping entry consisting of five components: child unit
address, child interrupt specifier, interrupt-parent, parent unit address, parent interrupt specifier.
child unit address
The unit address of the child node being mapped. The number of 32-bit cells required to specify this is
described by the #address-cells property of the bus node on which the child is located.
child interrupt specifier
The interrupt specifier of the child node being mapped. The number of 32-bit cells required to specify
this component is described by the #interrupt-cells property of this node—the nexus node containing
the interrupt-map property.
interrupt-parent
A single <phandle> value that points to the interrupt parent to which the child domain is being mapped.
parent unit address
The unit address in the domain of the interrupt parent. The number of 32-bit cells required to specify
this address is described by the #address-cells property of the node pointed to by the interrupt-parent
field.
Note: Both the child node and the interrupt parent node are required to have #address-cells and #interrupt-
cells properties defined. If a unit address component is not required, #address-cells shall be explicitly defined
to be zero.
interrupt-map-mask
Property: interrupt-map-mask
Value type: <prop-encoded-array> encoded as a bit mask
Description:
An interrupt-map-mask property is specified for a nexus node in the interrupt tree. This property specifies a
mask that is ANDed with the incoming unit interrupt specifier being looked up in the table specified in the
interrupt-map property.
#interrupt-cells
Property: #interrupt-cells
Value type: <u32>
Description:
The #interrupt-cells property defines the number of cells required to encode an interrupt specifier for an
interrupt domain.
The following shows the representation of a fragment of a devicetree with a PCI bus controller and a sample interrupt map
for describing the interrupt routing for two PCI slots (IDSEL 0x11,0x12). The INTA, INTB, INTC, and INTD pins for
slots 1 and 2 are wired to the Open PIC interrupt controller.
soc {
compatible = "simple-bus";
#address-cells = <1>;
#size-cells = <1>;
open-pic {
clock-frequency = <0>;
interrupt-controller;
#address-cells = <0>;
#interrupt-cells = <2>;
};
pci {
#interrupt-cells = <1>;
#size-cells = <2>;
(continues on next page)
One Open PIC interrupt controller is represented and is identified as an interrupt controller with an interrupt-controller
property.
Each row in the interrupt-map table consists of five parts: a child unit address and interrupt specifier, which is mapped to
an interrupt-parent node with a specified parent unit address and interrupt specifier.
• For example, the first row of the interrupt-map table specifies the mapping for INTA of slot 1. The components of
that row are shown here
– The child unit address is <0x8800 0 0>. This value is encoded with three 32-bit cells, which is determined
by the value of the #address-cells property (value of 3) of the PCI controller. The three cells represent the PCI
address as described by the binding for the PCI bus.
∗ The encoding includes the bus number (0x0 << 16), device number (0x11 << 11), and function number
(0x0 << 8).
– The child interrupt specifier is <1>, which specifies INTA as described by the PCI binding. This takes one
32-bit cell as specified by the #interrupt-cells property (value of 1) of the PCI controller, which is the child
interrupt domain.
– The interrupt parent is specified by a phandle which points to the interrupt parent of the slot, the Open PIC
interrupt controller.
– The parent has no unit address because the parent interrupt domain (the open-pic node) has an #address-cells
value of <0>.
– The parent interrupt specifier is <2 1>. The number of cells to represent the interrupt specifier (two cells) is
determined by the #interrupt-cells property on the interrupt parent, the open-pic node.
∗ The value <2 1> is a value specified by the device binding for the Open PIC interrupt controller (see
Section 4.5). The value <2> specifies the physical interrupt source number on the interrupt controller to
which INTA is wired. The value <1> specifies the level/sense encoding.
In this example, the interrupt-map-mask property has a value of <0xf800 0 0 7>. This mask is applied to a child unit
interrupt specifier before performing a lookup in the interrupt-map table.
To perform a lookup of the open-pic interrupt source number for INTB for IDSEL 0x12 (slot 2), function 0x3, the following
steps would be performed:
• The child unit address and interrupt specifier form the value <0x9300 0 0 2>.
– The encoding of the address includes the bus number (0x0 << 16), device number (0x12 << 11), and function
number (0x3 << 8).
– The interrupt specifier is 2, which is the encoding for INTB as per the PCI binding.
• The interrupt-map-mask value <0xf800 0 0 7> is applied, giving a result of <0x9000 0 0 2>.
• That result is looked up in the interrupt-map table, which maps to the parent interrupt specifier <4 1>.
A nexus node shall have a #<specifier>-cells property, where <specifier> is some specifier space such as ‘gpio’, ‘clock’,
‘reset’, etc.
<specifier>-map
Property: <specifier>-map
Value type: <prop-encoded-array> encoded as an arbitrary number of specifier mapping entries.
Description:
A <specifier>-map is a property in a nexus node that bridges one specifier domain with a set of parent specifier
domains and describes how specifiers in the child domain are mapped to their respective parent domains.
The map is a table where each row is a mapping entry consisting of three components: child specifier, specifier
parent, and parent specifier.
child specifier
The specifier of the child node being mapped. The number of 32-bit cells required to specify this com-
ponent is described by the #<specifier>-cells property of this node—the nexus node containing the
<specifier>-map property.
specifier parent
A single <phandle> value that points to the specifier parent to which the child domain is being mapped.
parent specifier
The specifier in the parent domain. The number of 32-bit cells required to specify this component is
described by the #<specifier>-cells property of the specifier parent node.
Lookups are performed on the mapping table by matching a specifier against the child specifier in the map.
Because some fields in the specifier may not be relevant or need to be modified, a mask is applied before the
lookup is done. This mask is defined in the <specifier>-map-mask property (see Section 2.5.1).
Similarly, when the specifier is mapped, some fields in the unit specifier may need to be kept unmodified and
passed through from the child node to the parent node. In this case, a <specifier>-map-pass-thru property
(see Section 2.5.1) may be specified to apply a mask to the child specifier and copy any bits that match to the
parent unit specifier.
<specifier>-map-mask
Property: <specifier>-map-mask
Value type: <prop-encoded-array> encoded as a bit mask
Description:
A <specifier>-map-mask property may be specified for a nexus node. This property specifies a mask that is
ANDed with the child unit specifier being looked up in the table specified in the <specifier>-map property.
If this property is not specified, the mask is assumed to be a mask with all bits set.
<specifier>-map-pass-thru
Property: <specifier>-map-pass-thru
Value type: <prop-encoded-array> encoded as a bit mask
Description:
A <specifier>-map-pass-thru property may be specified for a nexus node. This property specifies a mask that
is applied to the child unit specifier being looked up in the table specified in the <specifier>-map property.
Any matching bits in the child unit specifier are copied over to the parent specifier. If this property is not
specified, the mask is assumed to be a mask with no bits set.
#<specifier>-cells
Property: #<specifier>-cells
Value type: <u32>
Description:
The #<specifier>-cells property defines the number of cells required to encode a specifier for a domain.
The following shows the representation of a fragment of a devicetree with two GPIO controllers and a sample specifier
map for describing the GPIO routing of a few gpios on both of the controllers through a connector on a board to a device.
The expansion device node is on one side of the connector node and the SoC with the two GPIO controllers is on the other
side of the connector.
soc {
soc_gpio1: gpio-controller1 {
#gpio-cells = <2>;
};
soc_gpio2: gpio-controller2 {
#gpio-cells = <2>;
};
};
connector: connector {
#gpio-cells = <2>;
gpio-map = <0 0 &soc_gpio1 1 0>,
<1 0 &soc_gpio2 4 0>,
<2 0 &soc_gpio1 3 0>,
<3 0 &soc_gpio2 2 0>;
gpio-map-mask = <0xf 0x0>;
gpio-map-pass-thru = <0x0 0x1>;
(continues on next page)
expansion_device {
reset-gpios = <&connector 2 GPIO_ACTIVE_LOW>;
};
Each row in the gpio-map table consists of three parts: a child unit specifier, which is mapped to a gpio-controller node
with a parent specifier.
• For example, the first row of the specifier-map table specifies the mapping for GPIO 0 of the connector. The com-
ponents of that row are shown here
child specifier: 0 0
specifier parent: &soc_gpio1
parent specifier: 1 0
– The child specifier is <0 0>, which specifies GPIO 0 in the connector with a flags field of 0. This takes two
32-bit cells as specified by the #gpio-cells property of the connector node, which is the child specifier domain.
– The specifier parent is specified by a phandle which points to the specifier parent of the connector, the first
GPIO controller in the SoC.
– The parent specifier is <1 0>. The number of cells to represent the gpio specifier (two cells) is determined by
the #gpio-cells property on the specifier parent, the soc_gpio1 node.
∗ The value <1 0> is a value specified by the device binding for the GPIO controller. The value <1> specifies
the GPIO pin number on the GPIO controller to which GPIO 0 on the connector is wired. The value <0>
specifies the flags (active low, active high, etc.).
In this example, the gpio-map-mask property has a value of <0xf 0>. This mask is applied to a child unit specifier before
performing a lookup in the gpio-map table. Similarly, the gpio-map-pass-thru property has a value of <0x0 0x1>. This
mask is applied to a child unit specifier when mapping it to the parent unit specifier. Any bits set in this mask are cleared
out of the parent unit specifier and copied over from the child unit specifier to the parent unit specifier.
To perform a lookup of the connector’s specifier source number for GPIO 2 from the expansion device’s reset-gpios prop-
erty, the following steps would be performed:
• The child specifier forms the value <2 GPIO_ACTIVE_LOW>.
– The specifier is encoding GPIO 2 with active low flags per the GPIO binding.
• The gpio-map-mask value <0xf 0x0> is ANDed with the child specifier, giving a result of <0x2 0>.
• The result is looked up in the gpio-map table, which maps to the parent specifier <3 0> and &soc_gpio1 phandle.
• The gpio-map-pass-thru value <0x0 0x1> is inverted and ANDed with the parent specifier found in the gpio-
map table, resulting in <3 0>. The child specifier is ANDed with the gpio-map-pass-thru mask, form-
ing <0 GPIO_ACTIVE_LOW> which is then ORed with the cleared parent specifier <3 0> resulting in <3
GPIO_ACTIVE_LOW>.
• The specifier <3 GPIO_ACTIVE_LOW> is appended to the mapped phandle &soc_gpio1 resulting in <&soc_gpio1
3 GPIO_ACTIVE_LOW>.
THREE
The sections that follow specify the requirements for the base set of device nodes required in a DTSpec-compliant device-
tree.
All devicetrees shall have a root node and the following nodes shall be present at the root of all devicetrees:
• One /cpus node
• At least one /memory node
The devicetree has a single root node of which all other device nodes are descendants. The full path to the root node is /.
26
Devicetree Specification, Release v0.4
Note: All other standard properties (Section 2.3) are allowed but are optional.
A devicetree may have an aliases node (/aliases) that defines one or more alias properties. The alias node shall be at
the root of the devicetree and have the node name /aliases.
Each property of the /aliases node defines an alias. The property name specifies the alias name. The property value
specifies the full path to a node in the devicetree. For example, the property serial0 = "/simple-bus@fe000000/
serial@llc500" defines the alias serial0.
Alias names shall be a lowercase text strings of 1 to 31 characters from the following set of characters.
An alias value is a device path and is encoded as a string. The value represents the full path to a node, but the path does
not need to refer to a leaf node.
A client program may use an alias property name to refer to a full device path as all or part of its string value. A client
program, when considering a string as a device path, shall detect and use the alias.
Example
aliases {
serial0 = "/simple-bus@fe000000/serial@llc500";
ethernet0 = "/simple-bus@fe000000/ethernet@31c000";
};
Given the alias serial0, a client program can look at the /aliases node and determine the alias refers to the device path
/simple-bus@fe000000/serial@llc500.
A memory device node is required for all devicetrees and describes the physical memory layout for the system. If a system
has multiple ranges of memory, multiple memory nodes can be created, or the ranges can be specified in the reg property
of a single memory node.
The unit-name component of the node name (see Section 2.2.1) shall be memory.
The client program may access memory not covered by any memory reservations (see Section 5.3) using any storage
attributes it chooses. However, before changing the storage attributes used to access a real page, the client program is
responsible for performing actions required by the architecture and implementation, possibly including flushing the real
page from the caches. The boot program is responsible for ensuring that, without taking any action associated with a change
in storage attributes, the client program can safely access all memory (including memory covered by memory reservations)
as WIMG = 0b001x. That is:
• not Write Through Required
• not Caching Inhibited
• Memory Coherence
• Required either not Guarded or Guarded
If the VLE storage attribute is supported, with VLE=0.
Note: All other standard properties (Section 2.3) are allowed but are optional.
When booting via [UEFI], the system memory map is obtained via the GetMemoryMap() UEFI boot time service as
defined in [UEFI] § 7.2, and if present, the OS must ignore any /memory nodes.
Given a 64-bit Power system with the following physical memory layout:
• RAM: starting address 0x0, length 0x80000000 (2 GB)
• RAM: starting address 0x100000000, length 0x100000000 (4 GB)
Memory nodes could be defined as follows, assuming #address-cells = <2> and #size-cells = <2>.
Example #1
memory@0 {
device_type = "memory";
reg = <0x000000000 0x00000000 0x00000000 0x80000000
0x000000001 0x00000000 0x00000001 0x00000000>;
};
Example #2
memory@0 {
device_type = "memory";
reg = <0x000000000 0x00000000 0x00000000 0x80000000>;
};
memory@100000000 {
device_type = "memory";
reg = <0x000000001 0x00000000 0x00000001 0x00000000>;
};
The reg property is used to define the address and size of the two memory ranges. The 2 GB I/O region is skipped. Note
that the #address-cells and #size-cells properties of the root node specify a value of 2, which means that two 32-bit
cells are required to define the address and length for the reg property of the memory node.
Reserved memory is specified as a node under the /reserved-memory node. The operating system shall exclude reserved
memory from normal usage. One can create child nodes describing particular reserved (excluded from normal use) memory
regions. Such memory regions are usually designed for the special usage by various device drivers.
Parameters for each memory region can be encoded into the device tree with the following nodes:
#address-cells and #size-cells should use the same values as for the root node, and ranges should be empty so
that address translation logic works correctly.
Each child of the reserved-memory node specifies one or more regions of reserved memory. Each child node may either
use a reg property to specify a specific range of reserved memory, or a size property with optional constraints to request
a dynamically allocated block of memory.
Following the generic-names recommended practice, node names should reflect the purpose of the node (ie. “framebuffer”
or “dma-pool”). Unit address (@<address>) should be appended to the name if the node is a static allocation.
A reserved memory node requires either a reg property for static allocations, or a size property for dynamics allocations.
Dynamic allocations may use alignment and alloc-ranges properties to constrain where the memory is allocated from.
If both reg and size are present, then the region is treated as a static allocation with the reg property taking precedence
and size is ignored.
Note: All other standard properties (Section 2.3) are allowed but are optional.
The no-map and reusable properties are mutually exclusive and both must not be used together in the same node.
Linux implementation notes:
• If a linux,cma-default property is present, then Linux will use the region for the default pool of the contiguous
memory allocator.
• If a linux,dma-default property is present, then Linux will use the region for the default pool of the consistent
DMA allocator.
Regions in the /reserved-memory node may be referenced by other device nodes by adding a memory-region property
to the device node.
When booting via [UEFI], static /reserved-memory regions must also be listed in the system memory map obtained
via the GetMemoryMap() UEFI boot time service as defined in [UEFI] § 7.2. The reserved memory regions need to be
included in the UEFI memory map to protect against allocations by UEFI applications.
Reserved regions with the no-map property must be listed in the memory map with type EfiReservedMemoryType. All
other reserved regions must be listed with type EfiBootServicesData.
Dynamic reserved memory regions must not be listed in the [UEFI] memory map because they are allocated by the OS
after exiting firmware boot services.
This example defines 3 contiguous regions are defined for Linux kernel: one default of all device drivers (named linux,
cma and 64MiB in size), one dedicated to the framebuffer device (named framebuffer@78000000, 8MiB), and one for
multimedia processing (named multimedia@77000000, 64MiB).
/ {
#address-cells = <1>;
#size-cells = <1>;
memory {
reg = <0x40000000 0x40000000>;
};
reserved-memory {
#address-cells = <1>;
#size-cells = <1>;
ranges;
display_reserved: framebuffer@78000000 {
(continues on next page)
multimedia_reserved: multimedia@77000000 {
compatible = "acme,multimedia-memory";
reg = <0x77000000 0x4000000>;
};
};
/* ... */
fb0: video@12300000 {
memory-region = <&display_reserved>;
/* ... */
};
scaler: scaler@12500000 {
memory-region = <&multimedia_reserved>;
/* ... */
};
codec: codec@12600000 {
memory-region = <&multimedia_reserved>;
/* ... */
};
};
The /chosen node does not represent a real device in the system but describes parameters chosen or specified by the
system firmware at run time. It shall be a child of the root node.
Note: All other standard properties (Section 2.3) are allowed but are optional.
Example
chosen {
bootargs = "root=/dev/nfs rw nfsroot=192.168.1.1 console=ttyS0,115200";
};
Older versions of devicetrees may be encountered that contain a deprecated form of the stdout-path property called
linux,stdout-path. For compatibility, a client program might want to support linux,stdout-path if a stdout-path property is
not present. The meaning and use of the two properties is identical.
A /cpus node is required for all devicetrees. It does not represent a real device in the system, but acts as a container for
child cpu nodes which represent the systems CPUs.
Note: All other standard properties (Section 2.3) are allowed but are optional.
The /cpus node may contain properties that are common across cpu nodes. See Section 3.8 for details.
For an example, see Section 3.8.4.
A cpu node represents a hardware execution block that is sufficiently independent that it is capable of running an operating
system without interfering with other CPUs possibly running other operating systems.
Hardware threads that share an MMU would generally be represented under one cpu node. If other more complex CPU
topographies are designed, the binding for the CPU must describe the topography (e.g. threads that don’t share an MMU).
CPUs and threads are numbered through a unified number-space that should match as closely as possible the interrupt
controller’s numbering of CPUs/threads.
Properties that have identical values across cpu nodes may be placed in the /cpus node instead. A client program must
first examine a specific cpu node, but if an expected property is not found then it should look at the parent /cpus node.
This results in a less verbose representation of properties which are identical across all CPUs.
The node name for every CPU node should be cpu.
The following table describes the general properties of cpu nodes. Some of the properties described in Table 3.9 are select
standard properties with specific applicable detail.
Note: All other standard properties (Section 2.3) are allowed but are optional.
Note: All other standard properties (Section 2.3) are allowed but are optional.
Older versions of devicetree may be encountered that contain a bus-frequency property on CPU nodes. For compatibility,
a client-program might want to support bus-frequency. The format of the value is identical to that of clock-frequency. The
recommended practice is to represent the frequency of a bus on the bus node using a clock-frequency property.
The following properties of a cpu node describe the translate look-aside buffer in the processor’s MMU.
Note: All other standard properties (Section 2.3) are allowed but are optional.
The following properties of a cpu node describe the processor’s internal (L1) cache.
Note: All other standard properties (Section 2.3) are allowed but are optional.
Older versions of devicetrees may be encountered that contain a deprecated form of the next-level-cache property called
l2-cache. For compatibility, a client-program may wish to support l2-cache if a next-level-cache property is not present.
3.8.4 Example
cpus {
#address-cells = <1>;
#size-cells = <0>;
cpu@0 {
device_type = "cpu";
reg = <0>;
d-cache-block-size = <32>; // L1 - 32 bytes
i-cache-block-size = <32>; // L1 - 32 bytes
d-cache-size = <0x8000>; // L1, 32K
i-cache-size = <0x8000>; // L1, 32K
timebase-frequency = <82500000>; // 82.5 MHz
clock-frequency = <825000000>; // 825 MHz
};
};
Processors and systems may implement additional levels of cache hierarchy. For example, second-level (L2) or third-level
(L3) caches. These caches can potentially be tightly integrated to the CPU or possibly shared between multiple CPUs.
A device node with a compatible value of "cache" describes these types of caches.
The cache node shall define a phandle property, and all cpu nodes or cache nodes that are associated with or share the
cache each shall contain a next-level-cache property that specifies the phandle to the cache node.
A cache node may be represented under a CPU node or any other appropriate location in the devicetree.
Multiple-level and shared caches are represented with the properties in Table 3-9. The L1 cache properties are described
in Table 3-8.
Note: All other standard properties (Section 2.3) are allowed but are optional.
3.9.1 Example
See the following example of a devicetree representation of two CPUs, each with their own on-chip L2 and a shared L3.
cpus {
#address-cells = <1>;
#size-cells = <0>;
cpu@0 {
device_type = "cpu";
reg = <0>;
cache-unified;
cache-size = <0x8000>; // L1, 32 KB
cache-block-size = <32>;
timebase-frequency = <82500000>; // 82.5 MHz
next-level-cache = <&L2_0>; // phandle to L2
L2_0:l2-cache {
compatible = "cache";
cache-unified;
cache-size = <0x40000>; // 256 KB
cache-sets = <1024>;
cache-block-size = <32>;
cache-level = <2>;
next-level-cache = <&L3>; // phandle to L3
L3:l3-cache {
compatible = "cache";
cache-unified;
cache-size = <0x40000>; // 256 KB
cache-sets = <0x400>; // 1024
cache-block-size = <32>;
cache-level = <3>;
};
};
};
cpu@1 {
device_type = "cpu";
reg = <1>;
cache-unified;
cache-block-size = <32>;
cache-size = <0x8000>; // L1, 32 KB
timebase-frequency = <82500000>; // 82.5 MHz
clock-frequency = <825000000>; // 825 MHz
next-level-cache = <&L2_1>; // phandle to L2
L2_1:l2-cache {
compatible = "cache";
cache-unified;
cache-level = <2>;
cache-size = <0x40000>; // 256 KB
cache-sets = <0x400>; // 1024
cache-line-size = <32>; // 32 bytes
next-level-cache = <&L3>; // phandle to L3
};
};
};
FOUR
DEVICE BINDINGS
This chapter contains requirements, known as bindings, for how specific types and classes of devices are represented in
the devicetree. The compatible property of a device node describes the specific binding (or bindings) to which the node
complies.
Bindings may be defined as extensions of other each. For example a new bus type could be defined as an extension of the
simple-bus binding. In this case, the compatible property would contain several strings identifying each binding—from
the most specific to the most general (see Section 2.3.1, compatible).
When creating a new devicetree representation for a device, a binding should be created that fully describes the required
properties and value of the device. This set of properties shall be sufficiently descriptive to provide device drivers with
needed attributes of the device.
Some recommended practices include:
1. Define a compatible string using the conventions described in Section 2.3.1.
2. Use the standard properties (defined in Section 2.3 and Section 2.4) as applicable for the new device. This usage
typically includes the reg and interrupts properties at a minimum.
3. Use the conventions specified in Section 4 (Device Bindings) if the new device fits into one the DTSpec defined
device classes.
4. Use the miscellaneous property conventions specified in Section 4.1.2, if applicable.
5. If new properties are needed by the binding, the recommended format for property names is: "<company>,
<property-name>", where <company> is an OUI or short unique string like a stock ticker that identifies the creator
of the binding.
Example: "ibm,ppc-interrupt-server#s"
This section defines a list of helpful properties that might be applicable to many types of devices and device classes. They
are defined here to facilitate standardization of names and usage.
43
Devicetree Specification, Release v0.4
clock-frequency Property
reg-shift Property
label Property
The class of serial devices consists of various types of point to point serial line devices. Examples of serial line devices
include the 8250 UART, 16550 UART, HDLC device, and BISYNC device. In most cases hardware compatible with the
RS-232 standard fit into the serial device class.
I2 C and SPI (Serial Peripheral Interface) devices shall not be represented as serial port devices because they have their
own specific representation.
clock-frequency Property
current-speed Property
Serial devices compatible to the National Semiconductor 16450/16550 UART (Universal Asynchronous Receiver Trans-
mitter) should be represented in the devicetree using following properties.
Note: All other standard properties (Section 2.3) are allowed but are optional.
Network devices are packet oriented communication devices. Devices in this class are assumed to implement the data
link layer (layer 2) of the seven-layer OSI model and use Media Access Control (MAC) addresses. Examples of network
devices include Ethernet, FDDI, 802.11, and Token-Ring.
address-bits Property
local-mac-address Property
mac-address Property
max-frame-size Property
Network devices based on the IEEE 802.3 collections of LAN standards (collectively referred to as Ethernet) may be
represented in the devicetree using following properties, in addition to properties specified of the network device class.
The properties listed in this section augment the properties listed in the network device class.
max-speed Property
phy-connection-type Property
phy-handle Property
This section specifies the requirements for representing Open PIC compatible interrupt controllers. An Open PIC inter-
rupt controller implements the Open PIC architecture (developed jointly by AMD and Cyrix) and specified in The Open
Programmable Interrupt Controller (PIC) Register Interface Specification Revision 1.2 [b18].
Interrupt specifiers in an Open PIC interrupt domain are encoded with two cells. The first cell defines the interrupt number.
The second cell defines the sense and level information.
Sense and level information shall be encoded as follows in interrupt specifiers:
Note: All other standard properties (Section 2.3) are allowed but are optional.
System-on-a-chip processors may have an internal I/O bus that cannot be probed for devices. The devices on the bus can
be accessed directly without additional configuration required. This type of bus is represented as a node with a compatible
value of “simple-bus”.
FIVE
The Devicetree Blob (DTB) format is a flat binary encoding of devicetree data. It used to exchange devicetree data between
software programs. For example, when booting an operating system, firmware will pass a DTB to the OS kernel.
Note: IEEE1275 Open Firmware [IEEE1275] does not define the DTB format. On most Open Firmware compliant
platforms the devicetree is extracted by calling firmware methods to walk through the tree structure.
The DTB format encodes the devicetree data within a single, linear, pointerless data structure. It consists of a small header
(see Section 5.2), followed by three variable sized sections: the memory reservation block (see Section 5.3), the structure
block (see Section 5.4), and the strings block (see Section 5.5). These should be present in the flattened devicetree in that
order. Thus, the devicetree structure as a whole, when loaded into memory at address, will resemble the diagram in Fig.
5.1 (lower addresses are at the top of the diagram).
struct fdt_header
(free space)
memory reservation block
(free space)
structure block
(free space)
strings block
(free space)
The (free space) sections may not be present, though in some cases they might be required to satisfy the alignment con-
straints of the individual blocks (see Section 5.6).
50
Devicetree Specification, Release v0.4
5.1 Versioning
Several versions of the flattened devicetree structure have been defined since the original definition of the format. Fields in
the header give the version, so that the client program can determine if the devicetree is encoded in a compatible format.
This document describes only version 17 of the format. DTSpec compliant boot programs shall provide a devicetree of
version 17 or later, and should provide a devicetree of a version that is backwards compatible with version 16. DTSpec
compliant client programs shall accept devicetrees of any version backwards compatible with version 17 and may accept
other versions as well.
Note: The version is with respect to the binary structure of the device tree, not its content.
5.2 Header
The layout of the header for the devicetree is defined by the following C structure. All the header fields are 32-bit integers,
stored in big-endian format.
Flattened Devicetree Header Fields
struct fdt_header {
uint32_t magic;
uint32_t totalsize;
uint32_t off_dt_struct;
uint32_t off_dt_strings;
uint32_t off_mem_rsvmap;
uint32_t version;
uint32_t last_comp_version;
uint32_t boot_cpuid_phys;
uint32_t size_dt_strings;
uint32_t size_dt_struct;
};
magic
This field shall contain the value 0xd00dfeed (big-endian).
totalsize
This field shall contain the total size in bytes of the devicetree data structure. This size shall encompass all sections
of the structure: the header, the memory reservation block, structure block and strings block, as well as any free
space gaps between the blocks or after the final block.
off_dt_struct
This field shall contain the offset in bytes of the structure block (see Section 5.4) from the beginning of the header.
off_dt_strings
This field shall contain the offset in bytes of the strings block (see Section 5.5) from the beginning of the header.
off_mem_rsvmap
This field shall contain the offset in bytes of the memory reservation block (see Section 5.3) from the beginning of
the header.
version
This field shall contain the version of the devicetree data structure. The version is 17 if using the structure as defined
in this document. An DTSpec boot program may provide the devicetree of a later version, in which case this field
shall contain the version number defined in whichever later document gives the details of that version.
last_comp_version
This field shall contain the lowest version of the devicetree data structure with which the version used is backwards
compatible. So, for the structure as defined in this document (version 17), this field shall contain 16 because version
5.1. Versioning 51
Devicetree Specification, Release v0.4
17 is backwards compatible with version 16, but not earlier versions. As per Section 5.1, a DTSpec boot program
should provide a devicetree in a format which is backwards compatible with version 16, and thus this field shall
always contain 16.
boot_cpuid_phys
This field shall contain the physical ID of the system’s boot CPU. It shall be identical to the physical ID given in the
reg property of that CPU node within the devicetree.
size_dt_strings
This field shall contain the length in bytes of the strings block section of the devicetree blob.
size_dt_struct
This field shall contain the length in bytes of the structure block section of the devicetree blob.
5.3.1 Purpose
The memory reservation block provides the client program with a list of areas in physical memory which are reserved; that
is, which shall not be used for general memory allocations. It is used to protect vital data structures from being overwritten
by the client program. For example, on some systems with an IOMMU, the TCE (translation control entry) tables initialized
by a DTSpec boot program would need to be protected in this manner. Likewise, any boot program code or data used
during the client program’s runtime would need to be reserved (e.g., RTAS on Open Firmware platforms). DTSpec does
not require the boot program to provide any such runtime components, but it does not prohibit implementations from doing
so as an extension.
More specifically, a client program shall not access memory in a reserved region unless other information provided by
the boot program explicitly indicates that it shall do so. The client program may then access the indicated section of the
reserved memory in the indicated manner. Methods by which the boot program can indicate to the client program specific
uses for reserved memory may appear in this document, in optional extensions to it, or in platform-specific documentation.
The reserved regions supplied by a boot program may, but are not required to, encompass the devicetree blob itself. The
client program shall ensure that it does not overwrite this data structure before it is used, whether or not it is in the reserved
areas.
Any memory that is declared in a memory node and is accessed by the boot program or caused to be accessed by the
boot program after client entry must be reserved. Examples of this type of access include (e.g., speculative memory reads
through a non-guarded virtual page).
This requirement is necessary because any memory that is not reserved may be accessed by the client program with arbitrary
storage attributes.
Any accesses to reserved memory by or caused by the boot program must be done as not Caching Inhibited and Memory
Coherence Required (i.e., WIMG = 0bx01x), and additionally for Book III-S implementations as not Write Through Re-
quired (i.e., WIMG = 0b001x). Further, if the VLE storage attribute is supported, all accesses to reserved memory must
be done as VLE=0.
This requirement is necessary because the client program is permitted to map memory with storage attributes specified
as not Write Through Required, not Caching Inhibited, and Memory Coherence Required (i.e., WIMG = 0b001x), and
VLE=0 where supported. The client program may use large virtual pages that contain reserved memory. However, the
client program may not modify reserved memory, so the boot program may perform accesses to reserved memory as Write
Through Required where conflicting values for this storage attribute are architecturally permissible.
5.3.2 Format
The memory reservation block consists of a list of pairs of 64-bit big-endian integers, each pair being represented by the
following C structure.
struct fdt_reserve_entry {
uint64_t address;
uint64_t size;
};
Each pair gives the physical address and size in bytes of a reserved memory region. These given regions shall not overlap
each other. The list of reserved blocks shall be terminated with an entry where both address and size are equal to 0. Note
that the address and size values are always 64-bit. On 32-bit CPUs the upper 32-bits of the value are ignored.
Each uint64_t in the memory reservation block, and thus the memory reservation block as a whole, shall be located at an
8-byte aligned offset from the beginning of the devicetree blob (see Section 5.6).
As with the /reserved-memory node (Section 3.5.4), when booting via [UEFI] entries in the Memory Reservation Block
must also be listed in the system memory map obtained via the GetMemoryMap() to protect against allocations by UEFI
applications. The memory reservation block entries should be listed with type EfiReservedMemoryType.
The structure block describes the structure and contents of the devicetree itself. It is composed of a sequence of tokens
with data, as described below. These are organized into a linear tree structure, as described below.
Each token in the structure block, and thus the structure block itself, shall be located at a 4-byte aligned offset from the
beginning of the devicetree blob (see Section 5.6).
The structure block is composed of a sequence of pieces, each beginning with a token, that is, a big-endian 32-bit integer.
Some tokens are followed by extra data, the format of which is determined by the token value. All tokens shall be aligned
on a 32-bit boundary, which may require padding bytes (with a value of 0x0) to be inserted after the previous token’s data.
The five token types are as follows:
FDT_BEGIN_NODE (0x00000001)
The FDT_BEGIN_NODE token marks the beginning of a node’s representation. It shall be followed by the node’s
unit name as extra data. The name is stored as a null-terminated string, and shall include the unit address (see Section
2.2.1), if any. The node name is followed by zeroed padding bytes, if necessary for alignment, and then the next
token, which may be any token except FDT_END.
FDT_END_NODE (0x00000002)
The FDT_END_NODE token marks the end of a node’s representation. This token has no extra data; so it is followed
immediately by the next token, which may be any token except FDT_PROP.
FDT_PROP (0x00000003)
The FDT_PROP token marks the beginning of the representation of one property in the devicetree. It shall be
followed by extra data describing the property. This data consists first of the property’s length and name represented
as the following C structure:
struct {
uint32_t len;
uint32_t nameoff;
}
The devicetree structure is represented as a linear tree: the representation of each node begins with an FDT_BEGIN_NODE
token and ends with an FDT_END_NODE token. The node’s properties and subnodes (if any) are represented before the
FDT_END_NODE, so that the FDT_BEGIN_NODE and FDT_END_NODE tokens for those subnodes are nested within
those of the parent.
The structure block as a whole consists of the root node’s representation (which contains the representations for all other
nodes), followed by an FDT_END token to mark the end of the structure block as a whole.
More precisely, each node’s representation consists of the following components:
• (optionally) any number of FDT_NOP tokens
• FDT_BEGIN_NODE token
– The node’s name as a null-terminated string
– [zeroed padding bytes to align to a 4-byte boundary]
• For each property of the node:
– (optionally) any number of FDT_NOP tokens
– FDT_PROP token
∗ property information as given in Section 5.4.1
∗ [zeroed padding bytes to align to a 4-byte boundary]
• Representations of all child nodes in this format
• (optionally) any number of FDT_NOP tokens
• FDT_END_NODE token
Note that this process requires that all property definitions for a particular node precede any subnode definitions for that
node. Although the structure would not be ambiguous if properties and subnodes were intermingled, the code needed to
process a flat tree is simplified by this requirement.
The strings block contains strings representing all the property names used in the tree. These null terminated strings are
simply concatenated together in this section, and referred to from the structure block by an offset into the strings block.
The strings block has no alignment constraints and may appear at any offset from the beginning of the devicetree blob.
5.6 Alignment
The devicetree blob shall be located at an 8-byte-aligned address. To maintain backwards compatibilty for 32-bit machines,
4-byte alignment is supported by some software, but this is not DTSpec-compliant.
For the data in the memory reservation and structure blocks to be used without unaligned memory accesses, they shall lie
at suitably aligned memory addresses. Specifically, the memory reservation block shall be aligned to an 8-byte boundary
and the structure block to a 4-byte boundary.
Furthermore, the devicetree blob as a whole can be relocated without destroying the alignment of the subblocks.
As described in the previous sections, the structure and strings blocks shall have aligned offsets from the beginning of the
devicetree blob. To ensure the in-memory alignment of the blocks, it is sufficient to ensure that the devicetree as a whole
is loaded at an address aligned to the largest alignment of any of the subblocks, that is, to an 8-byte boundary. A DTSpec
compliant boot program shall load the devicetree blob at such an aligned address before passing it to the client program. If
an DTSpec client program relocates the devicetree blob in memory, it should only do so to another 8-byte aligned address.
SIX
The Devicetree Source (DTS) format is a textual representation of a devicetree in a form that can be processed by dtc into
a binary devicetree in the form expected by the kernel. The following description is not a formal syntax definition of DTS,
but describes the basic constructs used to represent devicetrees.
The name of DTS files should end with “.dts”.
Other source files can be included from a DTS file. The name of include files should end with “.dtsi”. Included files can
in turn include additional files.
/include/ "FILE"
6.2 Labels
The source format allows labels to be attached to any node or property value in the devicetree. Phandle and path references
can be automatically generated by referencing a label instead of explicitly specifying a phandle value or the full path to a
node. Labels are only used in the devicetree source format and are not encoded into the DTB binary.
A label shall be between 1 to 31 characters in length, be composed only of the characters in the set Table 6.1, and must not
start with a number.
Labels are created by appending a colon (‘:’) to the label name. References are created by prefixing the label name with
an ampersand (’&’).
56
Devicetree Specification, Release v0.4
Devicetree nodes are defined with a node name and unit address with braces marking the start and end of the node definition.
They may be preceded by a label.
[label:] node-name[@unit-address] {
[properties definitions]
[child nodes]
};
Nodes may contain property definitions and/or child node definitions. If both are present, properties shall come before
child nodes.
Previously defined nodes may be deleted.
/delete-node/ node-name;
/delete-node/ &label;
except for properties with empty (zero length) value which have the form:
[label:] property-name;
/delete-property/ property-name;
Property values may be defined as an array of 32-bit integer cells, as null-terminated strings, as bytestrings or a combination
of these.
• Arrays of cells are represented by angle brackets surrounding a space separated list of C-style integers. Example:
Arithmetic operators
+ add
- subtract
* multiply
/ divide
% modulo
Bitwise operators
& and
| or
^ exclusive or
~ not
<< left shift
>> right shift
Logical operators
&& and
(continues on next page)
Relational operators
Ternary operators
• A null-terminated string value is represented using double quotes (the property value is considered to include the
terminating NULL character). Example:
compatible = "simple-bus";
• A bytestring is enclosed in square brackets [ ] with each byte represented by two hexadecimal digits. Spaces between
each byte are optional. Example:
or equivalently:
local-mac-address = [000012345678];
• Values may have several comma-separated components, which are concatenated together. Example:
• In a cell array a reference to another node will be expanded to that node’s phandle. References may be & followed
by a node’s label. Example:
• Outside a cell array, a reference to another node will be expanded to that node’s full path. Example:
ethernet0 = &EMAC0;
• Labels may also appear before or after any component of a property value, or between cells of a cell array, or between
bytes of a bytestring. Examples:
/dts-v1/;
[memory reservations]
/ {
[property definitions]
[child nodes]
};
/dts-v1/; shall be present to identify the file as a version 1 DTS (dts files without this tag will be treated by dtc as being
in the obsolete version 0, which uses a different format for integers in addition to other small but incompatible changes).
Memory reservations (see Section 5.3) are represented by lines in the form:
The / { ... }; section defines the root node of the devicetree, and all the device tree data is contained within it.
C style (/* ... \*/) and C++ style (//) comments are supported.
[IEEE1275] Boot (Initialization Configuration) Firmware: Core Requirements and Practices, 1994, This is the core stan-
dard (also known as IEEE 1275) that defines the devicetree concept adopted by the DTSpec and ePAPR. It
is available from Global Engineering (http://global.ihs.com/).
[b7] Open Firmware Recommended Practice: Interrupt Mapping, Version 0.9, Open Firmware Working Group,
1996 (http://devicetree.org/open-firmware/practice/imap/imap0_9d.pdf)
[CHRP] PowerPC Microprocessor Common Hardware Reference Platform (CHRP) Binding, Version 1.8, Open
Firmware Working Group, 1998 (http://devicetree.org/open-firmware/bindings/chrp/chrp1_8a.ps). This doc-
ument specifies the properties for Open PIC-compatible interrupt controllers.
[PAPR] Power.org Standard for Power Architecture Platform Requirements, power.org
[b18] The Open Programmable Interrupt Controller (PIC) Register Interface Specification Revision 1.2, AMD and
Cyrix, October 1995
[EPAPR] Power.org Standard for Embedded Power Architecture Platform Requirements, power.org, 2011, https://www.
power.org/documentation/power-org-standard-for-embedded-power-architecture-platform-requirements-epapr-v1-1-2/
[UEFI] Unified Extensable Firmware Interface Specification v2.8 Errata A, February 2020, UEFI Forum
60
INDEX
A
AMP, 4
B
Book III-E, 4
boot CPU, 4
boot program, 4
C
cell, 5
client program, 5
D
DMA, 5
DTB, 5
DTC, 5
DTS, 5
E
effective address, 5
I
interrupt specifier, 5
P
physical address, 5
Power ISA, 5
Q
quiescent CPU, 5
S
secondary CPU, 5
SMP, 5
SoC, 5
U
unit address, 5
61