[Intel-gfx] [PATCH] drm/i915/gt: Fix -EDEADLK handling regression

[Rodrigo Vivi]

On Tue, Jul 13, 2021 at 09:59:18PM +0200, Daniel Vetter wrote:
> On Tue, Jul 13, 2021 at 9:58 PM Daniel Vetter <daniel at ffwll.ch> wrote:
> >
> > On Thu, Jul 1, 2021 at 9:07 AM Maarten Lankhorst
> > <maarten.lankhorst at linux.intel.com> wrote:
> > > Op 30-06-2021 om 18:44 schreef Ville Syrjala:
> > > > From: Ville Syrjälä <ville.syrjala at linux.intel.com>
> > > >
> > > > The conversion to ww mutexes failed to address the fence code which
> > > > already returns -EDEADLK when we run out of fences. Ww mutexes on
> > > > the other hand treat -EDEADLK as an internal errno value indicating
> > > > a need to restart the operation due to a deadlock. So now when the
> > > > fence code returns -EDEADLK the higher level code erroneously
> > > > restarts everything instead of returning the error to userspace
> > > > as is expected.
> > > >
> > > > To remedy this let's switch the fence code to use a different errno
> > > > value for this. -ENOBUFS seems like a semi-reasonable unique choice.
> > > > Apart from igt the only user of this I could find is sna, and even
> > > > there all we do is dump the current fence registers from debugfs
> > > > into the X server log. So no user visible functionality is affected.
> > > > If we really cared about preserving this we could of course convert
> > > > back to -EDEADLK higher up, but doesn't seem like that's worth
> > > > the hassle here.
> > > >
> > > > Not quite sure which commit specifically broke this, but I'll
> > > > just attribute it to the general gem ww mutex work.
> > > >
> > > > Cc: stable at vger.kernel.org
> > > > Cc: Maarten Lankhorst <maarten.lankhorst at linux.intel.com>
> > > > Cc: Thomas Hellström <thomas.hellstrom at intel.com>
> > > > Testcase: igt/gem_pread/exhaustion
> > > > Testcase: igt/gem_pwrite/basic-exhaustion
> > > > Testcase: igt/gem_fenced_exec_thrash/too-many-fences
> > > > Fixes: 80f0b679d6f0 ("drm/i915: Add an implementation for i915_gem_ww_ctx locking, v2.")
> > > > Signed-off-by: Ville Syrjälä <ville.syrjala at linux.intel.com>
> > > > ---
> > > >  drivers/gpu/drm/i915/gt/intel_ggtt_fencing.c | 2 +-
> > > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > >
> > > > diff --git a/drivers/gpu/drm/i915/gt/intel_ggtt_fencing.c b/drivers/gpu/drm/i915/gt/intel_ggtt_fencing.c
> > > > index cac7f3f44642..f8948de72036 100644
> > > > --- a/drivers/gpu/drm/i915/gt/intel_ggtt_fencing.c
> > > > +++ b/drivers/gpu/drm/i915/gt/intel_ggtt_fencing.c
> > > > @@ -348,7 +348,7 @@ static struct i915_fence_reg *fence_find(struct i915_ggtt *ggtt)
> > > >       if (intel_has_pending_fb_unpin(ggtt->vm.i915))
> > > >               return ERR_PTR(-EAGAIN);
> > > >
> > > > -     return ERR_PTR(-EDEADLK);
> > > > +     return ERR_PTR(-ENOBUFS);
> > > >  }
> > > >
> > > >  int __i915_vma_pin_fence(struct i915_vma *vma)
> > >
> > > Makes sense..
> > >
> > > Reviewed-by: Maarten Lankhorst <maarten.lankhorst at linux.intel.com>
> > >
> > > Is it a slightly more reent commit? Might probably be the part that converts execbuffer to use ww locks.
> >
> > - please cc: dri-devel on anything gem/gt related.
> > - this should probably be ENOSPC or something like that for at least a
> > seeming retention of errno consistentcy:
> >
> > https://dri.freedesktop.org/docs/drm/gpu/drm-uapi.html#recommended-ioctl-return-values
> 
> Other option would be to map that back to EDEADLK in the execbuf ioctl
> somewhere, so we retain a distinct errno code.

I'm about to push this patch to drm-intel-fixes... I'm assuming if there's any fix it will
be a follow-up patch and not a revert or force push, right?!

> -Daniel
> -- 
> Daniel Vetter
> Software Engineer, Intel Corporation
> http://blog.ffwll.ch
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx