Re: include/linux/seqlock.h:1257:undefined reference to `__scoped_seqlock_bug'

[Oliver Sang <oliver.sang@intel.com>]

hi, Peter Zijlstra,

On Thu, Dec 04, 2025 at 12:05:04PM +0100, Peter Zijlstra wrote:
> On Wed, Dec 03, 2025 at 10:22:28PM +0800, kernel test robot wrote:
> > Hi Oleg,
> > 
> > FYI, the error/warning still remains.
> > 
> > tree:   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
> > head:   3f9f0252130e7dd60d41be0802bf58f6471c691d
> > commit: 488f48b32654dc6be04d9cc12f75ce030c9cb21b seqlock: Change thread_group_cputime() to use scoped_seqlock_read()
> > date:   6 weeks ago
> > config: arm-mvebu_v5_defconfig (https://download.01.org/0day-ci/archive/20251203/202512032230.5pqkcg8x-lkp@intel.com/config)
> > compiler: arm-linux-gnueabi-gcc (GCC) 15.1.0
> 
> >    arm-linux-gnueabi-ld: kernel/sched/build_policy.o: in function `__scoped_seqlock_next':
> > >> include/linux/seqlock.h:1257:(.text+0x96a0): undefined reference to `__scoped_seqlock_bug'
> 
> I cannot reproduce using the gcc-15.1 toolchain from kernel.org
> 
> $ make O=tmp-build/ ARCH=arm CROSS_COMPILE=/opt/cross-15.1/bin/arm-linux-gnueabi- -j64
> 
> completes using your .config. I've tried twice now. In any case, if this
> reliably fails for you, could you try:

yes, we can reproduce reliably with the reproducer.

and yes, the issue can be fixed by below patch. thanks

Tested-by: kernel test robot <oliver.sang@intel.com>

> 
>   https://lkml.kernel.org/r/20251204104332.GG2528459@noisy.programming.kicks-ass.net
> 
> perhaps it is that same problem. If not, could you perhaps hand me the
> kernel/sched/build_policy.s file?
>