On Wed, 17 Apr 2024, Peter Zijlstra wrote:
> On Wed, Apr 17, 2024 at 07:49:17PM +0200, Mikulas Patocka wrote:
> > Index: linux-2.6/kernel/sched/completion.c
> > ===================================================================
> > --- linux-2.6.orig/kernel/sched/completion.c 2024-04-17 19:41:14.000000000 +0200
> > +++ linux-2.6/kernel/sched/completion.c 2024-04-17 19:41:14.000000000 +0200
> > @@ -290,6 +290,26 @@ wait_for_completion_killable_timeout(str
> > EXPORT_SYMBOL(wait_for_completion_killable_timeout);
> >
> > /**
> > + * wait_for_completion_long_io - waits for completion of a task
> > + * @x: holds the state of this particular completion
> > + *
> > + * This is like wait_for_completion_io, but it doesn't warn if the wait takes
> > + * too long.
> > + */
> > +void wait_for_completion_long_io(struct completion *x)
> > +{
> > + /* Prevent hang_check timer from firing at us during very long I/O */
> > + unsigned long timeout = sysctl_hung_task_timeout_secs * HZ / 2;
> > +
> > + if (timeout)
> > + while (!wait_for_completion_io_timeout(x, timeout))
> > + ;
> > + else
> > + wait_for_completion_io(x);
> > +}
> > +EXPORT_SYMBOL(wait_for_completion_long_io);
>
> Urgh, why is it a sane thing to circumvent the hang check timer?
The block layer already does it - the bios can have arbitrary size, so
waiting for them takes arbitrary time.
Mikulas
