On Fri, Jun 14, 2024 at 8:15 AM Tetsuo Handa
<penguin-kernel@xxxxxxxxxxxxxxxxxxx> wrote:
>
> On 2024/06/08 20:04, Tetsuo Handa wrote:
> > On 2024/06/08 19:53, Tetsuo Handa wrote:
> >> inconsistent {HARDIRQ-ON-W} -> {IN-HARDIRQ-W} usage.
> >
> > Oops, "inconsistent {SOFTIRQ-ON-W} -> {IN-SOFTIRQ-W} usage." example was
> > found at https://syzkaller.appspot.com/text?tag=CrashReport&x=14f0179a980000 .
> >
> > Then, do we want to
> >
> > - if (in_hardirq()) {
> > + if (!in_task()) {
> >
> > instead?
> >
>
> "inconsistent {HARDIRQ-ON-W} -> {IN-HARDIRQ-W} usage." upon unlock from IRQ work
> was reported at https://syzkaller.appspot.com/bug?extid=40905bca570ae6784745 .
imo the issue is elsewhere. syzbot reports:
local_lock_acquire include/linux/local_lock_internal.h:29 [inline]
__mmap_lock_do_trace_released+0x9c/0x620 mm/mmap_lock.c:243
__mmap_lock_trace_released include/linux/mmap_lock.h:42 [inline]
it complains about:
local_lock(&memcg_paths.lock);
in TRACE_MMAP_LOCK_EVENT.
which looks like a false positive.
