Re: INFO: task hung in do_read_cache_page (3)

[Dmitry Vyukov]

On Mon, May 11, 2020 at 7:31 PM Mike Christie <mchristi@xxxxxxxxxx> wrote:
>
> On 5/11/20 8:19 AM, syzbot wrote:
> > syzbot has bisected this bug to:
> >
> > commit 2da22da573481cc4837e246d0eee4d518b3f715e
> > Author: Mike Christie <mchristi@xxxxxxxxxx>
> > Date:   Tue Aug 13 16:39:52 2019 +0000
> >
> >     nbd: fix zero cmd timeout handling v2
> >
> > bisection log:  https://syzkaller.appspot.com/x/bisect.txt?x=11d6ab14100000
> > start commit:   e99332e7 gcc-10: mark more functions __init to avoid secti..
> > git tree:       upstream
> > final crash:    https://syzkaller.appspot.com/x/report.txt?x=13d6ab14100000
> > console output: https://syzkaller.appspot.com/x/log.txt?x=15d6ab14100000
> > kernel config:  https://syzkaller.appspot.com/x/.config?x=8a96cf498e199d8b
> > dashboard link: https://syzkaller.appspot.com/bug?extid=518c54e255b5031adde4
> > syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=146e45ec100000
> > C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=164444a4100000
> >
> > Reported-by: syzbot+518c54e255b5031adde4@xxxxxxxxxxxxxxxxxxxxxxxxx
> > Fixes: 2da22da57348 ("nbd: fix zero cmd timeout handling v2")
> >
> > For information about bisection process see: https://goo.gl/tpsmEJ#bisection
> >
>
> How do you adjust/modify what is expected from the test or what is
> reported as an error?
>
> The patch added back behavior that got removed. With the patch we expect
> the hung task warnings, because we specifically want to hold onto
> running commands for longer than 120/hung_task_timeout seconds

Hi Mike,

All task hung messages are considered a kernel bug. So far this was
always true as far as I can tell. Tasks that are blocked for any
prolonged period should be killable.