Re: WARNING: suspicious RCU usage in count

[Paul E. McKenney]

On Thu, Dec 17, 2020 at 09:26:44AM +0100, Dmitry Vyukov wrote:
> On Wed, Dec 16, 2020 at 9:55 PM Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote:
> >
> > On Wed, Dec 16, 2020 at 11:34:10AM -0800, syzbot wrote:
> > > Unfortunately, I don't have any reproducer for this issue yet.
> > >
> > > IMPORTANT: if you fix the issue, please add the following tag to the commit:
> > > Reported-by: syzbot+51ce7a5794c3b12a70d1@xxxxxxxxxxxxxxxxxxxxxxxxx
> > >
> > > =============================
> > > WARNING: suspicious RCU usage
> > > 5.10.0-rc7-syzkaller #0 Not tainted
> > > -----------------------------
> > > kernel/sched/core.c:7270 Illegal context switch in RCU-bh read-side critical section!
> > >
> > > other info that might help us debug this:
> > >
> > >
> > > rcu_scheduler_active = 2, debug_locks = 0
> > > no locks held by udevd/9038.
> > >
> > > stack backtrace:
> > > CPU: 3 PID: 9038 Comm: udevd Not tainted 5.10.0-rc7-syzkaller #0
> > > Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.12.0-59-gc9ba5276e321-prebuilt.qemu.org 04/01/2014
> > > Call Trace:
> > >  __dump_stack lib/dump_stack.c:77 [inline]
> > >  dump_stack+0x107/0x163 lib/dump_stack.c:118
> > >  ___might_sleep+0x220/0x2b0 kernel/sched/core.c:7270
> > >  count.constprop.0+0x164/0x270 fs/exec.c:449
> > >  do_execveat_common+0x2fd/0x7c0 fs/exec.c:1893
> > >  do_execve fs/exec.c:1983 [inline]
> > >  __do_sys_execve fs/exec.c:2059 [inline]
> > >  __se_sys_execve fs/exec.c:2054 [inline]
> > >  __x64_sys_execve+0x8f/0xc0 fs/exec.c:2054
> > >  do_syscall_64+0x2d/0x70 arch/x86/entry/common.c:46
> >
> > This must be the victim of something else.  There's no way this call
> > trace took the RCU read lock.
> 
> +lockdep maintainers for lockdep false positive then and +Paul for rcu

Note that this was "RCU-bh" rather than "RCU", so it might be something
like local_bh_disable() rather than rcu_read_lock() that might_sleep()
is complaining about.

> There is another recent claim of a false "suspicious RCU usage":
> https://lore.kernel.org/lkml/CAKMK7uEiS5SrBYv-2w2wWL=9G4ByoHvtiWVsPqekswZzOGmzjg@xxxxxxxxxxxxxx/

That one does look familiar.  ;-)

							Thanx, Paul