Re: ✗ Fi.CI.BAT: failure for series starting with [1/3] drm/atomic-helper: Fix leak in disable_all

From: Daniel Vetter <daniel@ffwll.ch>
To: Chris Wilson <chris@chris-wilson.co.uk>
Cc: tomi.p.sarvela@intel.com, Daniel Vetter <daniel.vetter@ffwll.ch>,
	intel-gfx@lists.freedesktop.org
Subject: Re: ✗ Fi.CI.BAT: failure for series starting with [1/3]  drm/atomic-helper: Fix leak in disable_all
Date: Fri, 14 Jul 2017 23:50:27 +0200	[thread overview]
Message-ID: <20170714215027.a3o5mszxne7f7nen@phenom.ffwll.local> (raw)
In-Reply-To: <150006707204.13756.15862851816583435771@mail.alporthouse.com>

On Fri, Jul 14, 2017 at 10:17:52PM +0100, Chris Wilson wrote:
> Quoting Daniel Vetter (2017-07-14 22:03:05)
> > On Fri, Jul 14, 2017 at 09:32:13PM +0100, Chris Wilson wrote:
> > > Quoting Patchwork (2017-07-14 21:18:04)
> > > > == Series Details ==
> > > > 
> > > > Series: series starting with [1/3] drm/atomic-helper: Fix leak in disable_all
> > > > URL   : https://patchwork.freedesktop.org/series/27327/
> > > > State : failure
> > > > 
> > > > == Summary ==
> > > > 
> > > > Series 27327v1 Series without cover letter
> > > > https://patchwork.freedesktop.org/api/1.0/series/27327/revisions/1/mbox/
> > > > 
> > > > Test gem_exec_fence:
> > > >         Subgroup await-hang-default:
> > > >                 pass       -> INCOMPLETE (fi-blb-e6850)
> > > >                 pass       -> INCOMPLETE (fi-pnv-d510)
> > > > Test gem_exec_flush:
> > > >         Subgroup basic-batch-kernel-default-uc:
> > > >                 fail       -> PASS       (fi-snb-2600) fdo#100007
> > > > Test kms_busy:
> > > >         Subgroup basic-flip-default-a:
> > > >                 pass       -> INCOMPLETE (fi-ilk-650)
> > > >                 pass       -> INCOMPLETE (fi-snb-2520m)
> > > >                 pass       -> INCOMPLETE (fi-snb-2600)
> > > >                 pass       -> INCOMPLETE (fi-ivb-3520m)
> > > >                 pass       -> INCOMPLETE (fi-ivb-3770)
> > > >                 pass       -> INCOMPLETE (fi-byt-j1900)
> > > >                 pass       -> INCOMPLETE (fi-byt-n2820)
> > > >                 pass       -> INCOMPLETE (fi-hsw-4770)
> > > >                 pass       -> INCOMPLETE (fi-hsw-4770r)
> > > >                 pass       -> INCOMPLETE (fi-bdw-5557u)
> > > >                 pass       -> INCOMPLETE (fi-bdw-gvtdvm)
> > > >                 pass       -> INCOMPLETE (fi-skl-6260u)
> > > >                 pass       -> INCOMPLETE (fi-skl-6700hq) fdo#101144
> > > >                 pass       -> INCOMPLETE (fi-skl-6700k)
> > > >                 pass       -> INCOMPLETE (fi-skl-6770hq)
> > > >                 pass       -> INCOMPLETE (fi-skl-gvtdvm)
> > > >                 none       -> INCOMPLETE (fi-skl-x1585l)
> > > >                 pass       -> INCOMPLETE (fi-bxt-j4205)
> > > >                 pass       -> INCOMPLETE (fi-kbl-7500u)
> > > >                 pass       -> INCOMPLETE (fi-kbl-7560u)
> > > >                 pass       -> INCOMPLETE (fi-kbl-r)
> > > >                 pass       -> INCOMPLETE (fi-glk-2a)
> > > >         Subgroup basic-flip-default-c:
> > > >                 pass       -> INCOMPLETE (fi-bsw-n3050)
> > > 
> > > Looks like we decremented a refcount once too often (after it was freed,
> > > poison 0x6b--). That clean up plane->old_fb?
> > 
> > I can't even run that testcase without my gpu reset fixes, epic win :(
> 
> It's hard to tell, but considering that the error it is detecting is slab
> corruption, it is more likely an earlier test that is doing the dirty
> deed. This is kasan territory. Tomi!

I did fire up my full pile of stuff, and it indeed goes somehow horribly
wrong in the modeset path for the reset code. Looks like my
innocently-looking change to disable_all (we call that somewhere in there)
to properly refcount ->fb wasn't all that innocent :(

It's getting late, I'll try to figure out what's going tomorrow or on Mon.
Promises to be another intriguing wild goose chase at least. kasan didn't
say anything special, but somehow the atomic commit machinery died
completely.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx