[SRU][N][PATCH 0/3] Coresight fails to build on 6.8.0-102 due to missing function and arg definitions

Tue Feb 24 19:23:17 UTC 2026

On Mon, Feb 23, 2026 at 03:05:12PM +0100, Stefan Bader wrote:
> On 21/02/2026 02:01, Noah Wager wrote:
> > BugLink: https://bugs.launchpad.net/bugs/2142337
> >
> > SRU Justification:
> >
> > [Impact]
> >
> > When CONFIG_CORESIGHT is enabled, noble 6.8.0-102 fails to build with the following errors:
> >
> > ```
> > drivers/hwtracing/coresight/coresight-catu.c: In function 'catu_init':
> > drivers/hwtracing/coresight/coresight-catu.c:728:15: error: implicit declaration of function 'coresight_init_driver' [-Werror=implicit-function-declaration]
> >    728 | ret = coresight_init_driver("catu", &catu_driver, &catu_platform_driver);
> >        | ^~~~~~~~~~~~~~~~~~~~~
> > drivers/hwtracing/coresight/coresight-catu.c: In function 'catu_exit':
> > drivers/hwtracing/coresight/coresight-catu.c:736:9: error: implicit declaration of function 'coresight_remove_driver'; did you mean 'coresight_remove_links'? [-Werror=implicit-function-declaration]
> >    736 | coresight_remove_driver(&catu_driver, &catu_platform_driver);
> >        | ^~~~~~~~~~~~~~~~~~~~~~~
> >        | coresight_remove_links
> > drivers/hwtracing/coresight/coresight-tmc-core.c: In function '__tmc_probe':
> > drivers/hwtracing/coresight/coresight-tmc-core.c:500:65: error: 'id' undeclared (first use in this function); did you mean 'fd'?
> >    500 | coresight_get_uci_data(id));
> >        | ^~
> >        | fd
> > drivers/hwtracing/coresight/coresight-tmc-core.c:500:65: note: each undeclared identifier is reported only once for each function it appears in
> > drivers/hwtracing/coresight/coresight-tmc-core.c: In function 'tmc_init':
> > drivers/hwtracing/coresight/coresight-tmc-core.c:728:16: error: implicit declaration of function 'coresight_init_driver' [-Werror=implicit-function-declaration]
> >    728 | return coresight_init_driver("tmc", &tmc_driver, &tmc_platform_driver);
> >        | ^~~~~~~~~~~~~~~~~~~~~
> > drivers/hwtracing/coresight/coresight-tmc-core.c: In function 'tmc_exit':
> > drivers/hwtracing/coresight/coresight-tmc-core.c:733:9: error: implicit declaration of function 'coresight_remove_driver'; did you mean 'coresight_remove_links'? [-Werror=implicit-function-declaration]
> >    733 | coresight_remove_driver(&tmc_driver, &tmc_platform_driver);
> >        | ^~~~~~~~~~~~~~~~~~~~~~~
> >        | coresight_remove_links
> > ```
> >
> > These errors are a result of including the following patches in upstream stable updates, but missing some dependencies from the original patch sets:
> >
> >    Error description: __tmc_probe:id not defined
> >    Break commit: n/linux aaf260b9572e ("coresight: tmc: Move ACPI support from AMBA driver to platform driver")
> >    Missing commit: 852e9a32058a ("coresight: stm: Extract device name from AMBA pid based table lookup")
> >                    3ab210297c31 ("coresight: tmc: Extract device properties from AMBA pid based table lookup")
> >    Patch series: "coresight: Move remaining AMBA ACPI devices into platform driver"
> >                  (https://lists-ec2.linaro.org/archives/list/coresight@lists.linaro.org/thread/XGJAJMHSDE5P7YSX33CIW5DHKR6C343L/)
> >
> >    Error description: coresight_{init,remove}_driver not defined
> >    Break commit: n/linux aaf260b9572e ("coresight: tmc: Move ACPI support from AMBA driver to platform driver")
> >    Missing commit: 075b7cd7ad7d ("coresight: Add helpers registering/removing both AMBA and platform drivers")
> >    Patch series: "coresight: Move remaining AMBA ACPI devices into platform driver"
> >                  (https://lists-ec2.linaro.org/archives/list/coresight@lists.linaro.org/thread/XGJAJMHSDE5P7YSX33CIW5DHKR6C343L/)
> >
> > [Fix]
> >
> > Cherry-pick upstream commits:
> > 852e9a32058a ("coresight: stm: Extract device name from AMBA pid based table lookup")
> > 3ab210297c31 ("coresight: tmc: Extract device properties from AMBA pid based table lookup")
> > 075b7cd7ad7d ("coresight: Add helpers registering/removing both AMBA and platform drivers")
> > from the series:
> > "coresight: Move remaining AMBA ACPI devices into platform driver".
> >
> > NOTE: The series also has "Move remaining AMBA ACPI devices into platform driver"
> > commits for the remaining devices, but I elected not to port them. We have
> > the respective commits for tmc and catu applied because they were
> > dependencies of "{tmc,catu}: Support atclk", but since the commits for the
> > other devices are not necessary for the patches to apply, and as of now there
> > are no upstream stable candidates with "Fixes" for the other devices, I don't
> > think it's worth the increased change size. Fortunately, if they are needed
> > in the future, they are clean cherry-picks (besides simple context changes)
> > and we have all the dependencies.
>
> Hm, though this is minimal I wonder whether it would be another case of
> deciding whether the change that broke the build was one that is really
> required in 6.8. This driver seems to have seen a lot of rework since 6.8
> and 6.12. Which poses quite a bit of chances that anything coming via 6.12.y
> is because it got broken after 6.8...
>
> -Stefan
>

Paraphrasing Manuel, it looks like "atclk" is an optional clock for some
devices that wasn't properly enabled, so the original changes aimed to
fix that. I'm not entirely sure if that means this is a "bug fix" or
"feature add", as the feature apparently didn't work until now (and
still doesn't, due to the build error). There are derivatives that are
affected by this build issue, so maybe the revert is better so we can
unblock the issue. For reference, that would require reverting these
in-tree commits:

0b88e5c14c2f ("coresight: catu: Support atclk")
cf08211db65c ("coresight: catu: Move ACPI support from AMBA driver to
platform driver")
b29444d683eb ("coresight: tmc: Support atclk")
aaf260b9572e ("coresight: tmc: Move ACPI support from AMBA driver to
platform driver")

Do you think I should go ahead with submitting the revert as a v2?

- Noah

> >
> > [Test Plan]
> >
> > Ensure CONFIG_CORESIGHT and CONFIG_CORESIGHT_CATU are enabled, then compile test.
> >
> > [Where problems could occur]
> >
> > There is a low chance of regression because these commits are only small
> > refactors that don't change behaviour. Only static function definitions are
> > modified, and functionality is only added via new helpers, not removed.
> > The new patches have a potential to create conflicts for derivative kernels
> > and other downstream projects that rebase on this kernel.
> >
> > Anshuman Khandual (3):
> >    coresight: stm: Extract device name from AMBA pid based table lookup
> >    coresight: tmc: Extract device properties from AMBA pid based table
> >      lookup
> >    coresight: Add helpers registering/removing both AMBA and platform
> >      drivers
> >
> >   drivers/hwtracing/coresight/coresight-core.c  | 29 +++++++++++++++++++
> >   .../hwtracing/coresight/coresight-funnel.c    | 19 ++----------
> >   drivers/hwtracing/coresight/coresight-priv.h  | 10 +++++++
> >   .../coresight/coresight-replicator.c          | 20 ++-----------
> >   drivers/hwtracing/coresight/coresight-stm.c   | 12 +++++++-
> >   .../hwtracing/coresight/coresight-tmc-core.c  | 14 ++++++---
> >   include/linux/coresight.h                     |  7 +++++
> >   7 files changed, 72 insertions(+), 39 deletions(-)
> >
>