CMake cores: fix cpu/fpu flags #14196

0xc0170 · 2021-01-26T14:23:53Z

Summary of changes

Couple of cores did not have common flag --mcpu set. I also aligned fpu.

This one goes along with #14190 - both are needed to have both toolchains fixed for couple of cores.

Impact of changes

Migration actions required

Documentation

Pull request type

[x] Patch update (Bug fix / Target update / Docs update / Test update / Refactor)
[] Feature update (New feature / Functionality change / New API)
[] Major update (Breaking change E.g. Return code change / API behaviour change)

Test results

[x] No Tests required for this change (E.g docs only update)
[] Covered by existing mbed-os tests (Greentea or Unittest)
[] Tests / results supplied as part of this PR

Reviewers

Couple of cores did not have common flag `--mcpu` set

ciarmcom · 2021-01-26T14:30:26Z

@0xc0170, thank you for your changes.
@ARMmbed/mbed-os-tools @ARMmbed/mbed-os-maintainers please review.

kjbracey · 2021-01-27T10:08:39Z

tools/cmake/cores/Cortex-A9.cmake

@@ -9,6 +9,7 @@ if(${MBED_TOOLCHAIN} STREQUAL "GCC_ARM")
        "-mfpu=vfpv3"
        "-mfloat-abi=hard"
        "-mno-unaligned-access"
+        ""-mcpu=cortex-a9""


Excessive quoting. Shouldn't this replace the -march=armv7-a?

Good point, checked https://community.arm.com/developer/tools-software/tools/b/tools-software-ides-blog/posts/compiler-flags-across-architectures-march-mtune-and-mcpu

On Arm, -mcpu combines -march and -mtune.4 On x86, -mcpu and -mtune are the same.5 .

-mcpu=X: On Arm, this flag is a combination of -march and -mtune. It simultaneously specifies the target architecture and optimizes for a given microarchitecture. On x86 this flag is a deprecated synonym for -mtune.

I'll remove it

This is for Gcc:
This specifies the name of the target ARM processor. GCC uses this name to derive the name of the target ARM architecture

kjbracey · 2021-01-27T10:11:08Z

tools/cmake/cores/Cortex-M33F.cmake

@@ -8,6 +8,7 @@ if(${MBED_TOOLCHAIN} STREQUAL "GCC_ARM")
        "-mfpu=fpv5-sp-d16"
        "-mfloat-abi=softfp"
        "-march=armv8-m.main"
+        "-mcpu=cortex-m33+nodsp"


Again, this should be replacing the -march. That would possibly have been used due to older toolchains not knowing "cortex-m33". I assume that's valid with whatever we require now?

fixed now (I also fixed dsp options as there were again errors between march or toolchains).

I realized we specified nodsp for M33 core but had march for FE parts enabled with dsp. This fixes it. M33F has no dsp. M33FE cores have dsp enabled.

Pull request has been modified.

0xc0170 · 2021-01-27T10:32:51Z

I added 2 commits, addressing march and dsp options.

kjbracey · 2021-01-27T11:19:49Z

tools/cmake/cores/Cortex-M33FE-NS.cmake

@@ -7,11 +7,11 @@ if(${MBED_TOOLCHAIN} STREQUAL "GCC_ARM")
        "-mthumb"
        "-mfpu=fpv5-sp-d16"
        "-mfloat-abi=softfp"
-        "-march=armv8-m.main+dsp"
+        "-mcpu=cortex-m33+dsp"


Is +dsp valid here? Maybe, but don't see it in the docs. They just say cortex-m33 DSP is assumed, and you can specify +nodsp. (Note we're also not specifying +fp).

https://gcc.gnu.org/onlinedocs/gcc/ARM-Options.html

-mcpu works differently to -march in this regard - you're always enabling extensions in -march, but usually disabling them in -mcpu.

I'll remove dsp from this one

kjbracey · 2021-01-27T11:27:19Z

tools/cmake/cores/Cortex-M33FE-NS.cmake

    )
 elseif(${MBED_TOOLCHAIN} STREQUAL "ARM")
    list(APPEND common_options
-        "-mcpu=cortex-m33"
+        "-mcpu=cortex-m33+dsp"


Unlike for GCC, this is perfectly fine according to the docs, as +X and +noX are always available. But if you do want to say +dsp to be explicit, why not +fp too?

I would imagine ARMC6's defaults are the same as GCC's so might be best to keep it the same as GCC.

fixed, I only followed for now what we had (separate fpu settings for both toolchains). To keep both toolchains the same.

Yeah, it's not clear from docs how +fp/+nofp (or lack of and hence default) and -mfpu interact. Maybe don't mess with whatever we know/think works from original python stuff.

Move the fpu settings across all cores so they are aligned. Some cores did not have fpu set for ARMClang. I matched here both toolchains to set up the fpu the same (Only exception is soft vs hard - that is separate issue I would not like to touch in this series).

0xc0170 · 2021-01-27T12:34:23Z

I aligned fpu settings for all cores. I matched both toolchains. There still might be issues, please review.

0xc0170 · 2021-01-27T13:20:54Z

@jeromecoutant I requested review as you also experienced issues with the core flags in CMake.

jeromecoutant

I could verify with blinky application on:

NUCLEO_L552ZE_Q ("Cortex-M33FE")
NUCLEO_F767ZI ("Cortex-M7F")
NUCLEO_H743ZI2 ("Cortex-M7FD")
Build and execution OK with ARM and GCC

ciarmcom · 2021-01-30T12:50:27Z

This pull request has automatically been marked as stale because it has had no recent activity. @ARMmbed/mbed-os-maintainers, please start CI to get the PR merged.

0xc0170 · 2021-02-02T10:53:46Z

Ci started

mbed-ci · 2021-02-02T13:39:08Z

Jenkins CI Test : ✔️ SUCCESS

Build Number: 1 | 🔒 Jenkins CI Job | 🌐 Logs & Artifacts

CLICK for Detailed Summary

jobs	Status
jenkins-ci/mbed-os-ci_unittests	✔️
jenkins-ci/mbed-os-ci_cmake-example-GCC_ARM	✔️
jenkins-ci/mbed-os-ci_cmake-example-ARM	✔️
jenkins-ci/mbed-os-ci_build-greentea-ARM	✔️
jenkins-ci/mbed-os-ci_build-greentea-GCC_ARM	✔️
jenkins-ci/mbed-os-ci_build-example-GCC_ARM	✔️
jenkins-ci/mbed-os-ci_build-cloud-example-ARM	✔️
jenkins-ci/mbed-os-ci_build-example-ARM	✔️
jenkins-ci/mbed-os-ci_build-cloud-example-GCC_ARM	✔️
jenkins-ci/mbed-os-ci_cmake-example-test	✔️
jenkins-ci/mbed-os-ci_greentea-test	✔️

CMake cores: fix cpu flag

3006590

Couple of cores did not have common flag `--mcpu` set

0xc0170 requested a review from hugueskamba January 26, 2021 14:23

0xc0170 changed the title ~~CMake cores: fix cpu flag~~ CMake cores: fix cpu flag for Gcc Arm Jan 26, 2021

ciarmcom added the release-type: patch Indentifies a PR as containing just a patch label Jan 26, 2021

ciarmcom requested review from a team January 26, 2021 14:30

ciarmcom added the needs: review label Jan 26, 2021

kjbracey previously requested changes Jan 27, 2021

View reviewed changes

mergify bot added needs: work and removed needs: review labels Jan 27, 2021

0xc0170 added 2 commits January 27, 2021 10:25

CMake cores: remove march, not needed if we use mcpu

15acc00

CMake cores: fix dsp for M33FE (enabling it)

5d8d2ba

I realized we specified nodsp for M33 core but had march for FE parts enabled with dsp. This fixes it. M33F has no dsp. M33FE cores have dsp enabled.

kjbracey reviewed Jan 27, 2021

View reviewed changes

0xc0170 added 2 commits January 27, 2021 12:20

CMake cores: remove dsp as it is on by default

8e2c957

0xc0170 changed the title ~~CMake cores: fix cpu flag for Gcc Arm~~ CMake cores: fix cpu/fpu flags Jan 27, 2021

0xc0170 requested a review from jeromecoutant January 27, 2021 13:20

jeromecoutant approved these changes Jan 27, 2021

View reviewed changes

hugueskamba requested a review from kjbracey January 27, 2021 14:23

kjbracey approved these changes Jan 28, 2021

View reviewed changes

mergify bot added needs: CI and removed needs: work labels Jan 28, 2021

ciarmcom added the stale Stale Pull Request label Jan 30, 2021

mergify bot added ready for merge and removed needs: CI labels Feb 2, 2021

0xc0170 merged commit 2e96400 into ARMmbed:master Feb 2, 2021

mergify bot removed the ready for merge label Feb 2, 2021

mbedmain added release-version: 6.8.0 Release-pending labels Feb 16, 2021

LDong-Arm requested a review from rajkan01 February 22, 2021 10:04

0xc0170 deleted the fix-cores-gcc branch February 22, 2021 14:54

0xc0170 mentioned this pull request Feb 22, 2021

CMake: Fix Cortex-M33 GCC build #14313

Merged

mbedmain removed release-type: patch Indentifies a PR as containing just a patch Release-pending labels Feb 22, 2021

CMake cores: fix cpu/fpu flags #14196

CMake cores: fix cpu/fpu flags #14196

Uh oh!

Conversation

0xc0170 commented Jan 26, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary of changes

Impact of changes

Migration actions required

Documentation

Pull request type

Test results

Reviewers

Uh oh!

ciarmcom commented Jan 26, 2021

Uh oh!

Choose a reason for hiding this comment

Uh oh!

0xc0170 Jan 27, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

0xc0170 commented Jan 27, 2021

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

0xc0170 commented Jan 27, 2021

Uh oh!

0xc0170 commented Jan 27, 2021

Uh oh!

jeromecoutant left a comment

Choose a reason for hiding this comment

Uh oh!

ciarmcom commented Jan 30, 2021

Uh oh!

0xc0170 commented Feb 2, 2021

Uh oh!

mbed-ci commented Feb 2, 2021

Jenkins CI Test : ✔️ SUCCESS

Build Number: 1 | 🔒 Jenkins CI Job | 🌐 Logs & Artifacts

Uh oh!

Uh oh!

0xc0170 commented Jan 26, 2021 •

edited

Loading

0xc0170 Jan 27, 2021 •

edited

Loading