Re: [OE-core][scarthgap] linux-yocto/6.6: update to v6.6.157

["Yoann Congal" <yoann.congal@smile.fr>]

On Mon Oct 19, 2026 at 4:43 PM CEST, Bruce Ashfield wrote:
> On Thu, Oct 15, 2026 at 3:05 AM Yoann Congal <yoann.congal@smile.fr> wrote:
>
>> On Wed Oct 14, 2026 at 6:55 PM CEST, Bruce Ashfield wrote:
>> > On Wed, Oct 14, 2026 at 10:31 AM Yoann Congal <yoann.congal@smile.fr>
>> wrote:
>> >
>> >> On Tue Sep 22, 2026 at 10:08 AM CEST, Yoann Congal wrote:
>> >> > On Thu Sep 17, 2026 at 7:16 AM CEST, Bruce Ashfield via
>> >> lists.openembedded.org wrote:
>> >> >> From: Bruce Ashfield <bruce.ashfield@gmail.com>
>> >> >>
>> >> >> Updating linux-yocto/6.6 to the latest korg -stable release that
>> >> comprises
>> >> >> the following commits:
>> >> >>
>> >> >>     79643295eba17 Linux 6.6.157
>> >> >>     32e598324edc3 nvmet-rdma: fix queue leak when connect backlog is
>> >> exceeded
>> >> >>     [...]
>> >> >>     73b7a11814838 Revert "arm64: mm: Don't remap pgtables for
>> allocate
>> >> vs populate"
>> >> >>
>> >> >> Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com>
>> >> >> ---
>> >> >>  .../linux/linux-yocto-rt_6.6.bb               |  6 ++--
>> >> >>  .../linux/linux-yocto-tiny_6.6.bb             |  6 ++--
>> >> >>  meta/recipes-kernel/linux/linux-yocto_6.6.bb  | 28
>> +++++++++----------
>> >> >>  3 files changed, 20 insertions(+), 20 deletions(-)
>> >> >
>> >> > Hello Bruce,
>> >> >
>> >> > I think either this 6.6.157 upgrade or the 6.6.156 one triggers a
>> >> > lttng ptest failure:
>> >> >
>> https://autobuilder.yoctoproject.org/valkyrie/?#/builders/73/builds/4604
>> >> > qemux86-64-ptest debian11-vk-3 scarthgap completed at 2026-09-21
>> >> 11:38:31+00:00
>> >> >
>> https://autobuilder.yoctoproject.org/valkyrie/?#/builders/61/builds/4577
>> >> > qemuarm64-ptest ubuntu2404-vk-arm2 scarthgap completed at 2026-09-21
>> >> 14:57:16+00:00
>> >> >
>> >> >
>> >>
>> https://valkyrie.yocto.io/pub/non-release/20260921-71/testresults/qemux86-64-ptest/core-image-ptest-lttng-tools/log.do_testimage.311077.20260921110838
>> >> :
>> >> > (only one of each message):
>> >> > |FAIL: kernel/test_userspace_probe 91 - Load command with opts:
>> >>
>> --input-path=/tmp/tmp.test_userspace_probe_save_load_elf_trace_path.zSA9mh
>> >> > |FAIL: tools/base-path/test_ust 35 - Load command with opts: -i
>> >> ./tools/base-path//load-stream-extra-path.lttng
>> >> > |FAIL: tools/save-load/test_load 2 - Load command with opts: -i
>> >> ./tools/save-load//load-42.lttng
>> >> > |FAIL: tools/rotation/test_ust 101 - App is un-registered
>> >> > |FAIL: tools/save-load/test_autoload 2 - Lttng-tool list command with
>> >> option load-42
>> >> > |FAIL: tools/rotation/test_save_load_mi 17 - Lttng-tool list command
>> >> with option yTVkM2lxbFoDS7Nu
>> >> > |FAIL: tools/rotation/test_save_load_mi 8 - Valid lttng
>> enable-rotation
>> >> size XML
>> >> > |FAIL: tools/rotation/test_save_load_mi 5 - Valid lttng
>> enable-rotation
>> >> timer XML
>> >> > |FAIL: tools/relayd-grouping/test_ust 120 - Valid lttng list XML
>> >> > |FAIL: tools/rotation/test_save_load_mi 18 - Valid lttng list XML
>> after
>> >> load
>> >> > |FAIL: tools/base-path/test_ust 37 - Stop lttng tracing for session
>> >> load-stream-extra-path
>> >> > |FAIL: tools/trigger/test_add_trigger_cli 293 - add-trigger mi is
>> valid
>> >> > |FAIL: tools/trigger/test_remove_trigger_cli 20 - remove-trigger mi is
>> >> valid
>> >> > |FAIL: tools/base-path/test_ust 39 - UST per-uid network trace path is
>> >> valid
>> >> > |FAIL: tools/save-load/test_autoload 1 - Start session daemon
>> >> > |FAIL: tools/save-load/test_autoload 3 - No session daemon to kill
>> >> > |FAIL: tools/save-load/test_load 3 - Destroy session load-42
>> >> > |FAIL: tools/base-path/test_ust 38 - Destroy session
>> >> load-stream-extra-path
>> >> > |FAIL: tools/rotation/test_save_load_mi 21 - Destroy session
>> >> yTVkM2lxbFoDS7Nu
>> >> > |FAIL: tools/mi/test_mi 3 - MI test: Version xsd validation
>> >> > |FAIL: tools/rotation/test_save_load_mi 20 - Found the right rotation
>> >> size value in list XML after load
>> >> > |FAIL: tools/base-path/test_ust 40 - Validate trace for event
>> tp:tptest
>> >> > |FAIL: tools/base-path/test_ust 36 - Start tracing for session
>> >> load-stream-extra-path
>> >> > |FAIL: kernel/test_userspace_probe 92 - Validate userspace probe
>> listing
>> >> >
>> >> > Any ideas?
>> >> >
>> >> > Thanks!
>> >>
>> >> Hello Bruce,
>> >>
>> >> Gentle ping on the above upgrade failure.
>> >>
>> >> Do you have time to look at this? If you don't, we'll have organize to
>> >> find other resources, we can't not update the kernel.
>> >>
>> >
>> > I'll have time next week, but I can't run the test images locally, so
>> I'll
>> > need to come up with my
>> > own tests :)
>>
>> Awesome, thanks Bruce!
>>
>
>
> I bumped -stable again, and when I run the lttng ptests against that
> kernel, I have all tests passing.
>
> So either something in that kernel update fixed it, or I'm doing something
> wrong.
>
> Did you see this once ? or consistently ?

Twice (x86-64 and arm64) but during the same a-full build:
* https://autobuilder.yoctoproject.org/valkyrie/?#/builders/73/builds/4604
  qemux86-64-ptest debian11-vk-3 scarthgap completed at 2026-09-21 11:38:31+00:00
* https://autobuilder.yoctoproject.org/valkyrie/?#/builders/61/builds/4577
  qemuarm64-ptest ubuntu2404-vk-arm2 scarthgap completed at 2026-09-21 14:57:16+00:00

Then, I removed the patch from the branch.

> There are some AB-INT bugs that you (among others) have been tracking:
>
>   - 16309 — ACCEPTED: [scarthgap] AB-INT PTEST: lttng-tools ptest timeout
>   - 16139 — AB-INT PTEST: lttng-tools ptest failure:
> tools/config-directory/test_config.py
>   - 16358 — AB-INT PTEST: lttng-tools: Couldn't login into serial console
> as root using blank password
>
> Could it be one of those ?

I know I've looked through existing bugs and did not find a matching
one. But, given the fact that it happened twice during the same a-full
on 2 different architectures, I don't think it was an AB-INT.

> I'll send my stable updates later today, and maybe we can run them again to
> see if you get the same passes.

Happy to. I usually do 2-3 a-full per review cycles. If the
lttng-tools ptests are green, I'll consider this fixed and take the
kernel upgrade patch.

Thanks!
-- 
Yoann Congal
Smile ECS