Re: [GIT PULL v2] PCI changes for v6.17

[Nam Cao <namcao@linutronix.de>]

Lukas Wunner <lukas@wunner.de> writes:

> On Thu, Aug 07, 2025 at 10:34:08AM +0700, Ammar Faizi wrote:
>> On Fri, 01 Aug 2025 21:37:28 +0000, pr-tracker-bot@kernel.org, wrote:
>> >
>> > The pull request you sent on Fri, 1 Aug 2025 09:22:54 -0500:
>> >
>> > > git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci.git tags/pci-v6.17-changes
>> >
>> > has been merged into torvalds/linux.git:
>> > https://git.kernel.org/torvalds/c/0bd0a41a5120f78685a132834865b0a631b9026a
>> 
>> Yesterday, I synced with Linus' tree, but couldn't boot. Crashed with
>> this call trace:
>> 
>>   https://gist.githubusercontent.com/ammarfaizi2/3ba41f13517be4bae70cde869347d259/raw/0ac09b3e1d90d51c3fed14ca9f837f45d7730f0a/crash.jpg
>> 
>> This morning, I synced with Linus' tree again, still the same result.
>> 
>> I suspect it's related to pci. I'm still bisecting. I've successfully
>> narrowed it down to this pci pull.
>> 
>>   0bd0a41a5120 (refs/bisect/bad) Merge tag 'pci-v6.17-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci
>>   db5f0c3e3e60 ring-buffer: Convert ring_buffer_write() to use guard(preempt_notrace)
>>   12d518961586 tracing: Use __free(kfree) in trace.c to remove gotos
>>   debe57fbe12c tracing: Add guard() around locks and mutexes in trace.c
>>   788fa4b47cdc tracing: Add guard(ring_buffer_nest)
>>   c89504a703fb tracing: Remove unneeded goto out logic
>>   877d94c74e4c (refs/bisect/good-877d94c74e4c6665d2af55c0154363b43b947e60) Merge tag 'linux-watchdog-6.17-rc1' of git://www.linux-watchdog.org/linux-watchd
>> 
>> Now, I am testing:
>> 
>>   769ce531faa6 (HEAD) Merge branch 'pci/controller/msi-parent'
>> 
>> I'll be back to it later. git bisect says:
>> 
>>   Bisecting: 65 revisions left to test after this (roughly 6 steps)
>
> Kenneth reports early-stage reboots caused by d7d8ab87e3e
> ("PCI: vmd: Switch to msi_create_parent_irq_domain()"):
>
> https://lore.kernel.org/all/dfa40e48-8840-4e61-9fda-25cdb3ad81c1@panix.com/
>
> Perhaps you're witnessing the same issue?

Thanks for the Cc. The backtrace does look like something that the
commit would cause.

Let me stare at it.

Nam