Re: [PATCH 7.3 0000/1257] 7.3.4-rc1 review

["Barry K. Nathan" <barryn@pobox.com>]

On 11/7/26 3:49 PM, Barry K. Nathan wrote:
> On 11/7/26 5:38 AM, Greg Kroah-Hartman wrote:
>> This is the start of the stable review cycle for the 7.3.4 release.
>> There are 1257 patches in this series, all will be posted as a response
>> to this one.  If anyone has any issues with these being applied, please
>> let me know.
>>
>> Responses should be made by Mon, 09 Nov 2026 13:46:08 +0000.
>> Anything received after that time might be too late.
>>
>> The whole patch series can be found in one patch at:
>>     https://www.kernel.org/pub/linux/kernel/v7.x/stable-review/patch-7.3.4-rc1.gz
>> or in the git tree and branch at:
>>     git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git linux-7.3.y
>> and the diffstat can be found below.
>>
>> thanks,
>>
>> greg k-h
> 
> 7.3.4-rc1, 6.18.63-rc1, and 6.12.117-rc1 are all giving me the same build
> error with clang that I previously had with 7.3.2-rc1 and 6.18.61-rc1,
> due to the same xfs patch as before ("xfs: cross-reference the primary
> superblock"):
> 
> fs/xfs/scrub/agheader.c:123:6: error: variable 'error' is used uninitialized
>        whenever 'if' condition is true [-Werror,-Wsometimes-uninitialized]
>    123 |         if (agno == 0)
>        |             ^~~~~~~~~
> fs/xfs/scrub/agheader.c:435:9: note: uninitialized use occurs here
>    435 |         return error;
>        |                ^~~~~
> fs/xfs/scrub/agheader.c:123:2: note: remove the 'if' if its condition is always
>        false
>    123 |         if (agno == 0)
>        |         ^~~~~~~~~~~~~~
>    124 |                 goto out_xref;
>        |                 ~~~~~~~~~~~~~
> fs/xfs/scrub/agheader.c:111:13: note: initialize the variable 'error' to silence
>        this warning
>    111 |         int                     error;
>        |                                      ^
>        |                                       = 0
> 1 error generated.
> make[4]: *** [scripts/Makefile.build:287: fs/xfs/scrub/agheader.o] Error 1
> 
> I sent an email about this earlier today, before the -rc was sent out:
> https://lore.kernel.org/stable/7840e515-e7d5-4ef9-8c1d-6390374a5b33@pobox.com/
> 
> Reverting this patch fixes the build, or the minimal fix I proposed in
> that email also fixes the build:
> 
> --- a/fs/xfs/scrub/agheader.c    2026-11-06 22:42:57.147299404 -0800
> +++ b/fs/xfs/scrub/agheader.c    2026-11-06 22:45:59.898908897 -0800
> @@ -108,7 +108,7 @@ xchk_superblock(
>       xfs_agnumber_t        agno = sc->sm->sm_agno;
>       uint32_t        v2_ok;
>       __be32            features_mask;
> -    int            error;
> +    int            error = 0;
>       __be16            vernum_mask;
> 
>       /*
> 
> So far I've only tried building a minimal test config. Later today or
> tomorrow I'll do testing with my normal configs, but (to make a long
> story short) I will probably experience temporary health problems
> for the next few days, so that will probably slow down my testing.

I took 7.3.4-rc1, reverted "sched_ext: Fix the stall watchdog testing
and blaming the wrong sched under bypass", and added the fix to
initialize error to 0 in xchk_superblock() in fs/xfs/scrub/agheader.c
(quoted above).

The resulting kernel is running well on 3 different amd64 systems and
an arm64 virtual machine, with no regressions observed. (This includes
a successful XFS scrub on one of the amd64 systems.)
-- 
-Barry K. Nathan  <barryn@pobox.com>