Fallback to ARMC5 when ARMC6 is not configured #10193

bridadan · 2019-03-22T04:14:35Z

Description

This PR makes it so:

if -t ARM is being supplied as the toolchain
and if ARMC6 is not configured in the user's environment
and ARMC5 is configured
and the the target supports building with both ARMC5 and ARMC6
Then the tools will select ARMC5 as the toolchain for building. Previously, this was always set to ARMC6 (with a few targets being the exception).

If the tools fallback to ARMC5, the following warning will be printed at the bottom of the log:

------------------------------------------------------------
Warning: We noticed that you are using Arm Compiler 5. We are deprecating the use of Arm Compiler 5 soon. Please upgrade your environment to Arm Compiler 6 which is free to use with Mbed OS. For more information, please visit https://os.mbed.com/docs/mbed-os/latest/tools/index.html
------------------------------------------------------------

This feature has been documented here (thanks @SenRamakri): ARMmbed/mbed-os-5-docs#1024

Normally I'd write out all the implications of the changes in this PR, however the implementation details are fairly extensive. Reading the docs above is a good place to start, probably followed by the commit history.

FYI @screamerbg

TODO

The ARMC5 warning needs to be updated with a link to the proper doc.
~~I would really prefer to bring this in with a unit test. The implementation is quite complex and the details can easily be lost if verifying manually.~~ Most likely this will have to come in in a later release due to time constraints.

Pull request type

[ ] Fix
[ ] Refactor
[ ] Target update
[x] Functionality change
[ ] Docs update
[ ] Test update
[ ] Breaking change

Reviewers

@SenRamakri @theotherjimmy @bulislaw

Release Notes

The -t ARM argument defaults to using the Arm Compiler 6 when it is configured in a user's enviroment. However, In cases where a target supports building with both Arm Compiler 5 and Arm Compiler 6 and Arm Compiler 6 is not properly configured, the tools will fallback to compiling with Arm Compiler 5.

ciarmcom · 2019-03-22T06:00:30Z

@bridadan, thank you for your changes.
@SenRamakri @theotherjimmy @bulislaw @ARMmbed/mbed-os-maintainers @ARMmbed/mbed-os-tools please review.

0xc0170 · 2019-03-22T06:33:51Z

Started CI to get early test run

mbed-ci · 2019-03-22T08:12:52Z

Test run: FAILED

Summary: 2 of 9 test jobs failed
Build number : 1
Build artifacts

Failed test jobs:

jenkins-ci/mbed-os-ci_build-IAR8
jenkins-ci/mbed-os-ci_build-GCC_ARM

mbed-ci · 2019-03-22T08:15:10Z

Test run: FAILED

Summary: 2 of 9 test jobs failed
Build number : 2
Build artifacts

Failed test jobs:

jenkins-ci/mbed-os-ci_build-GCC_ARM
jenkins-ci/mbed-os-ci_build-IAR8

0xc0170 · 2019-03-22T09:23:34Z

Looks like this has affect on one target and its unsupported toolchains (build failures)

0xc0170 · 2019-03-22T09:25:34Z

Setting to rc4

cc @adbridge @bulislaw

bulislaw

Looks good. Please remember to add a link to docs to the warning.

bridadan · 2019-03-22T12:58:41Z

@0xc0170 The PR is failing because the CI is attempting to build the examples for the NUMAKER_PFM_M2351 target using IAR and GCC_ARM, both of which the target does not support (and that is the error we get in CI: "Could not compile for NUMAKER_PFM_M2351: Target NUMAKER_PFM_M2351 is not supported by toolchain IAR").

In my opinion there are two fixes for this:

Change CI so it doesn't attempt the build in the first place
Change the example builder script its invoking so it double checks supported toolchains and does not raise an error

I'd prefer to go with choice 1. Any thoughts @OPpuolitaival?

bridadan · 2019-03-22T13:34:44Z

Thanks to @OPpuolitaival, I believe the last commit will fix the CI failure.

adbridge · 2019-03-22T13:48:08Z

ci started

bulislaw

Looks good (here's a 🍩 token that can be redeemed next time I'm around. Don't loose it though there's no replacement policy!)

bridadan · 2019-03-22T18:42:54Z

I'm looking into the exporter issue. Since it failed on #10131 as well I doubt my changes caused this issue.

bridadan · 2019-03-22T19:07:50Z

The exporter issue is unrelated to this change, however I've pushed a fix (hopefully) for it here: #10199

mbed-ci · 2019-03-22T19:27:59Z

Test run: FAILED

Summary: 1 of 13 test jobs failed
Build number : 3
Build artifacts

Failed test jobs:

jenkins-ci/mbed-os-ci_exporter

mbed-ci · 2019-03-22T21:52:41Z

Test run: FAILED

Summary: 1 of 13 test jobs failed
Build number : 4
Build artifacts

Failed test jobs:

jenkins-ci/mbed-os-ci_exporter

SenRamakri

Changes looks good to me, although uvision6 exporter seems to be failing.

adbridge · 2019-03-23T23:08:22Z

@bridadan could you resolve the conflict please, this is the last PR needing to get through ci now ?

The affected targets are Renesas targets, USI_WM_BN_BM_22 based targets, and the MTB_MXCHIP_EMW3166.

These functions will be used to handle some of the error state and warning messages produced when the scripts attempt to select a valid toolchain.

There are two new functions: get_valid_toolchain_names and find_valid_toolchain. These functions are used to figure out if a fallback is possible and necessary. find_valid_toolchain is expected to be used by the front-end scripts. get_toolchain_name was updated with some different logic and comments.

Some unused imports were removed as well as some general clean up.

This removes the arugment help from the output, making the error much easier to find. This solves ARMmbed#10090.

bridadan · 2019-03-23T23:27:25Z

Rebased!

adbridge · 2019-03-23T23:29:19Z

ci started

mbed-ci · 2019-03-24T01:40:17Z

Test run: FAILED

Summary: 1 of 13 test jobs failed
Build number : 5
Build artifacts

Failed test jobs:

jenkins-ci/mbed-os-ci_exporter

0xc0170 · 2019-03-24T04:56:33Z

exporters restarted, license server error

ciarmcom requested review from bulislaw, SenRamakri, theotherjimmy and a team March 22, 2019 06:00

ciarmcom added the needs: review label Mar 22, 2019

0xc0170 added needs: CI and removed needs: review labels Mar 22, 2019

0xc0170 added needs: work release-version: 5.12.0-rc4 and removed needs: CI labels Mar 22, 2019

bulislaw approved these changes Mar 22, 2019

View reviewed changes

bridadan marked this pull request as ready for review March 22, 2019 12:59

0xc0170 added needs: review and removed needs: work labels Mar 22, 2019

0xc0170 approved these changes Mar 22, 2019

View reviewed changes

adbridge added needs: CI and removed needs: review labels Mar 22, 2019

theotherjimmy approved these changes Mar 22, 2019

View reviewed changes

bulislaw approved these changes Mar 22, 2019

View reviewed changes

bridadan force-pushed the arm_tc_with_fallback branch from 6f96d01 to bbcc25c Compare March 22, 2019 16:17

This was referenced Mar 22, 2019

Using Mbed Studio's ARMC6 with Mbed CLI ARMmbed/mbed-os-5-docs#1028

Merged

Add baremetal example to testing #10131

Merged

bridadan mentioned this pull request Mar 22, 2019

Fix the removal of Mbed-added core flags in uvision exporter #10199

Merged

SenRamakri approved these changes Mar 23, 2019

View reviewed changes

bridadan added 11 commits March 23, 2019 18:24

Whitespace clean up

a8032db

Only enable ARMC6 for a few targets

2b75dfd

The affected targets are Renesas targets, USI_WM_BN_BM_22 based targets, and the MTB_MXCHIP_EMW3166.

Add utility functions and Exceptions

08d9e32

These functions will be used to handle some of the error state and warning messages produced when the scripts attempt to select a valid toolchain.

ARMC5 is supported if a target lists ARM as a supported_toolchain

04e90c2

Front-end scripts now use the ARM toolchain fallback.

2b4960b

Some unused imports were removed as well as some general clean up.

Make args_erroor less verbose.

b809ec6

This removes the arugment help from the output, making the error much easier to find. This solves ARMmbed#10090.

Add ARMC5 as an option for the -t argument

1c31d36

Updating supported toolchain printing with new options

b94c418

Add link for ARMC5 warning

0f1bd76

get_mbed_official_release respects supported_toolchains

b95732a

bridadan force-pushed the arm_tc_with_fallback branch from bbcc25c to b95732a Compare March 23, 2019 23:27

NirSonnenschein added ready for merge and removed needs: CI labels Mar 24, 2019

NirSonnenschein merged commit b29b55a into ARMmbed:master Mar 24, 2019

NirSonnenschein removed the ready for merge label Mar 24, 2019

bridadan mentioned this pull request Mar 25, 2019

Error is too verbose when a compiler executable is not found #10090

Closed

Fallback to ARMC5 when ARMC6 is not configured #10193

Fallback to ARMC5 when ARMC6 is not configured #10193

Uh oh!

Conversation

bridadan commented Mar 22, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Pull request type

Reviewers

Release Notes

Uh oh!

ciarmcom commented Mar 22, 2019

Uh oh!

0xc0170 commented Mar 22, 2019

Uh oh!

mbed-ci commented Mar 22, 2019

Test run: FAILED

Uh oh!

mbed-ci commented Mar 22, 2019

Test run: FAILED

Uh oh!

0xc0170 commented Mar 22, 2019

Uh oh!

0xc0170 commented Mar 22, 2019

Uh oh!

bulislaw left a comment

Choose a reason for hiding this comment

Uh oh!

bridadan commented Mar 22, 2019

Uh oh!

bridadan commented Mar 22, 2019

Uh oh!

adbridge commented Mar 22, 2019

Uh oh!

bulislaw left a comment

Choose a reason for hiding this comment

Uh oh!

bridadan commented Mar 22, 2019

Uh oh!

bridadan commented Mar 22, 2019

Uh oh!

mbed-ci commented Mar 22, 2019

Test run: FAILED

Uh oh!

mbed-ci commented Mar 22, 2019

Test run: FAILED

Uh oh!

SenRamakri left a comment

Choose a reason for hiding this comment

Uh oh!

adbridge commented Mar 23, 2019

Uh oh!

bridadan commented Mar 23, 2019

Uh oh!

adbridge commented Mar 23, 2019

Uh oh!

mbed-ci commented Mar 24, 2019

Test run: FAILED

Uh oh!

0xc0170 commented Mar 24, 2019

Uh oh!

Uh oh!

bridadan commented Mar 22, 2019 •

edited

Loading