[Intel-gfx] Xorg[9132]: segfault at 0 ip 00007fbc84d6fb0d sp 00007ffca3765610 error 4 in intel_drv.so[7fbc84d4d000+18b000]

[Chris Wilson]

On Tue, Dec 15, 2015 at 11:42:34AM +0100, Daniel Vetter wrote:
> On Mon, Dec 14, 2015 at 02:11:43PM -0800, Marc MERLIN wrote:
> > On Tue, Dec 01, 2015 at 08:27:32AM -0800, Marc MERLIN wrote:
> > > On Sat, Nov 28, 2015 at 09:54:50AM -0800, Marc MERLIN wrote:
> > > > On Tue, Nov 17, 2015 at 05:11:05PM +0200, Jani Nikula wrote:
> > > > > On Tue, 17 Nov 2015, Marc MERLIN <marc_xorg at merlins.org> wrote:
> > > > > > So, this is probably the 3rd time I send such a report with different
> > > > > > kernels and get 0 response.
> > > > > > Is this a write only list and no one is really seeing any of them, or is
> > > > > > this an unknown/known problem that no one can work on?
> > > > > 
> > > > > Apologies; nothing wrong with your report, our fail.
> > > > > 
> > > > > Given all your attempts I hate to ask, but please do give v4.3 or later
> > > > > a try. If that fails too, please file a bug at [1]. Add drm.debug=14
> > > > > module parameter, and attach dmesg from boot to the bug.
> > > > 
> > > > I'm now running 4.3.0 and haven't had the same problem yet. But for the
> > > > first time, X has crashed after a few days of use:
> > > > 
> > > > Sadly I don't have a backtrace on this, but if you need me to dump symbols
> > > > on /usr/lib/xorg/modules/drivers/intel_drv.so, let me know.
> > > 
> > > And it crashed again:
> > > vlc[15647]: segfault at 7f73e28ed300 ip 0000 7f73e28ed300 sp 00007f74244d9778 error 14
> > > [drm:drm_mode_addfb2] [FB:58]
> > > [drm:i915_gem_open] 
> > > [drm:i915_gem_context_create_ioctl] HW context 1 created
> > > [drm:drm_mode_addfb2] [FB:65]
> > > [drm:i915_gem_context_create_ioctl] HW context 2 created
> > > [drm:i915_gem_context_destroy_ioctl] HW context 1 destroyed
> > > Xorg[4945]: segfault at 7f7a69f5a000 ip 00007f7a6c9ba517 sp 00007fff0abbcc00 error 6 in intel_drv.so[7f7a6c95c000+18b000]
> > 
> > Ok, it took a lot of time, but there you go, I have a crash with
> > backtrace.
> > Could someone now looking into this and hopefully fix it?
> 
> Looking at dmesg I don't see the stuck pageflip anymore. It looks like the
> DDX is simply falling over itself, with now blame to the kernel here. Or
> if there is it's a new/different bug. So we need to restart triage and
> check whether this was introduced by a ddx or kernel upgrade.

There was a resource leak that triggered an invalid object to be placed
on the request list after an error path. That error path is fixed, the
only question is who's resource leak was it..
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre