Re: [PATCH] Revert "exec: make de_thread() freezable (was: Re: Linux 4.20-rc4)

[Oleg Nesterov <oleg@redhat.com>]

On 12/03, Michal Hocko wrote:
>
> On Mon 03-12-18 13:31:49, Oleg Nesterov wrote:
> > On 12/03, Michal Hocko wrote:
> > >
> > > Now, I wouldn't mind to revert this because the code is really old and
> > > we haven't seen many bug reports about failing suspend yet. But what is
> > > the actual plan to make this work properly?
> >
> > I don't see a simple solution...
> >
> > But we need to fix exec/de_thread anyway, then we can probably reconsider
> > this patch.
>
> My concern is that de_thread fix might be too disruptive for stable
> kernels while we might want to have a simple enough fix for the the
> suspend issue in the meantime. That was actually the primary reason I've
> acked the hack even though I didn't like it.
>
> So can we find a way to shut the lockdep up

You have already mentioned freezable_schedule_unsafe(), but I agree with
"DO NOT ADD ANY NEW CALLERS OF THIS FUNCTION" above it ...

> when this is not really a
> deadlock?

not a deadlock, but lockdep is obviously right, suspend still can fail if
another task needs the same mutex.

Again, we have already discussed this, in my opinion we should blame exec/
de_thread for this and other problems.

> Or maybe this really is one and then we need a real fix for
> stable as well.

Well, strace -f can hang if it races with mt exec, it would be nice to have
a fix for stable.

Oleg.