Re: [Intel-gfx] [PATCH] drm/i915: fix infinite recursion on unbind due to ilk vt-d w/a

2011-12-10 Thread Bobby Powers
lists.freedesktop.org/archives/intel-gfx/2011-October/012984.html >>> >>> >>> Did someone prove that doesn't work? >> >> This patch caused hard lockups for me after ~35 minutes of casual use >> (twice).  I've attached the oopses.  I'm runn

Re: [Intel-gfx] [PATCH] drm/i915: fix infinite recursion on unbind due to ilk vt-d w/a

2011-12-10 Thread Bobby Powers
caused hard lockups for me after ~35 minutes of casual use > (twice).  I've attached the oopses.  I'm running a Fedora 16 machine, > Lenovo T420 (i5-2540M w/ VT-d enabled), and at each time had a Windows > 7 KVM guest idling (not sure if that is relevant).  With this patch > reverte

Re: [Intel-gfx] [PATCH] drm/i915: fix infinite recursion on unbind due to ilk vt-d w/a

2011-12-10 Thread Bobby Powers
On Tue, Dec 6, 2011 at 12:43 PM, Ben Widawsky wrote: > On Tue, Dec 06, 2011 at 12:12:33PM +0100, Daniel Vetter wrote: >> The recursion loop goes retire_requests->unbind->gpu_idle->retire_reqeusts. >> >> Every time we go through this we need a >> - active object that can be retired >> - and there a

[Intel-gfx] [PATCH] drm/i915: fix infinite recursion on unbind due to ilk vt-d w/a

2011-12-09 Thread Bobby Powers
://lists.freedesktop.org/archives/intel-gfx/2011-October/012984.html >>> >>> >>> Did someone prove that doesn't work? >> >> This patch caused hard lockups for me after ~35 minutes of casual use >> (twice). ?I've attached the oopses. ?I'm r

[Intel-gfx] [PATCH] drm/i915: fix infinite recursion on unbind due to ilk vt-d w/a

2011-12-09 Thread Bobby Powers
caused hard lockups for me after ~35 minutes of casual use > (twice). ?I've attached the oopses. ?I'm running a Fedora 16 machine, > Lenovo T420 (i5-2540M w/ VT-d enabled), and at each time had a Windows > 7 KVM guest idling (not sure if that is relevant). ?With this patch > reverte

[Intel-gfx] [PATCH] drm/i915: fix infinite recursion on unbind due to ilk vt-d w/a

2011-12-08 Thread Bobby Powers
On Tue, Dec 6, 2011 at 12:43 PM, Ben Widawsky wrote: > On Tue, Dec 06, 2011 at 12:12:33PM +0100, Daniel Vetter wrote: >> The recursion loop goes retire_requests->unbind->gpu_idle->retire_reqeusts. >> >> Every time we go through this we need a >> - active object that can be retired >> - and there a

[PATCH] drm/i915: fix infinite recursion on unbind due to ilk vt-d w/a

2011-12-06 Thread Daniel Vetter
The recursion loop goes retire_requests->unbind->gpu_idle->retire_reqeusts. Every time we go through this we need a - active object that can be retired - and there are no other references to that object than the one from the active list, so that it gets unbound and freed immediately. Otherwise t

Re: [Intel-gfx] [PATCH] drm/i915: fix infinite recursion on unbind due to ilk vt-d w/a

2011-12-06 Thread Ben Widawsky
On Tue, Dec 06, 2011 at 12:12:33PM +0100, Daniel Vetter wrote: > The recursion loop goes retire_requests->unbind->gpu_idle->retire_reqeusts. > > Every time we go through this we need a > - active object that can be retired > - and there are no other references to that object than the one from >

[Intel-gfx] [PATCH] drm/i915: fix infinite recursion on unbind due to ilk vt-d w/a

2011-12-06 Thread Ben Widawsky
On Tue, Dec 06, 2011 at 12:12:33PM +0100, Daniel Vetter wrote: > The recursion loop goes retire_requests->unbind->gpu_idle->retire_reqeusts. > > Every time we go through this we need a > - active object that can be retired > - and there are no other references to that object than the one from >

[PATCH] drm/i915: fix infinite recursion on unbind due to ilk vt-d w/a

2011-12-06 Thread Daniel Vetter
The recursion loop goes retire_requests->unbind->gpu_idle->retire_reqeusts. Every time we go through this we need a - active object that can be retired - and there are no other references to that object than the one from the active list, so that it gets unbound and freed immediately. Otherwise t