Re: [Intel-gfx] [PATCH 5/5] drm/i915: Solve the GPU reset vs. modeset deadlocks with an rw_semaphore

2017-07-03 Thread Daniel Vetter
On Mon, Jul 03, 2017 at 12:30:48PM +0300, Ville Syrjälä wrote: > I've pretty much decided that I'll be happy if I can solve this for > future kernels. If someone else wants to try and cook up some kind > of simple regression fix I don't have any real objections. Imo fixing the regression isn't

Re: [Intel-gfx] [PATCH 5/5] drm/i915: Solve the GPU reset vs. modeset deadlocks with an rw_semaphore

2017-07-03 Thread Ville Syrjälä
On Mon, Jul 03, 2017 at 09:55:48AM +0200, Daniel Vetter wrote: > On Fri, Jun 30, 2017 at 09:46:36PM +0300, Ville Syrjälä wrote: > > On Fri, Jun 30, 2017 at 08:23:58PM +0200, Daniel Vetter wrote: > > > On Fri, Jun 30, 2017 at 5:44 PM, Ville Syrjälä > > > wrote: > > >

Re: [Intel-gfx] [PATCH 5/5] drm/i915: Solve the GPU reset vs. modeset deadlocks with an rw_semaphore

2017-07-03 Thread Daniel Vetter
On Fri, Jun 30, 2017 at 09:46:36PM +0300, Ville Syrjälä wrote: > On Fri, Jun 30, 2017 at 08:23:58PM +0200, Daniel Vetter wrote: > > On Fri, Jun 30, 2017 at 5:44 PM, Ville Syrjälä > > wrote: > > >> And if the GEM folks insist the old behavior can't be restored, then

Re: [Intel-gfx] [PATCH 5/5] drm/i915: Solve the GPU reset vs. modeset deadlocks with an rw_semaphore

2017-06-30 Thread Ville Syrjälä
On Fri, Jun 30, 2017 at 08:23:58PM +0200, Daniel Vetter wrote: > On Fri, Jun 30, 2017 at 5:44 PM, Ville Syrjälä > wrote: > >> And if the GEM folks insist the old behavior can't be restored, then we > >> just need a tailor-made get-out-of-jail card for gen4 gpu reset

Re: [Intel-gfx] [PATCH 5/5] drm/i915: Solve the GPU reset vs. modeset deadlocks with an rw_semaphore

2017-06-30 Thread Daniel Vetter
On Fri, Jun 30, 2017 at 5:44 PM, Ville Syrjälä wrote: >> And if the GEM folks insist the old behavior can't be restored, then we >> just need a tailor-made get-out-of-jail card for gen4 gpu reset somewhere >> in i915_sw_fence. Force-completing all render requests

Re: [Intel-gfx] [PATCH 5/5] drm/i915: Solve the GPU reset vs. modeset deadlocks with an rw_semaphore

2017-06-30 Thread Ville Syrjälä
On Fri, Jun 30, 2017 at 03:30:33PM +0200, Daniel Vetter wrote: > On Fri, Jun 30, 2017 at 03:25:58PM +0200, Daniel Vetter wrote: > > On Thu, Jun 29, 2017 at 04:49:48PM +0300, ville.syrj...@linux.intel.com > > wrote: > > > From: Ville Syrjälä > > > > > > Introduce

Re: [Intel-gfx] [PATCH 5/5] drm/i915: Solve the GPU reset vs. modeset deadlocks with an rw_semaphore

2017-06-30 Thread Daniel Vetter
On Fri, Jun 30, 2017 at 03:25:58PM +0200, Daniel Vetter wrote: > On Thu, Jun 29, 2017 at 04:49:48PM +0300, ville.syrj...@linux.intel.com wrote: > > From: Ville Syrjälä > > > > Introduce an rw_semaphore to protect the display commits. All normal > > commits use

Re: [Intel-gfx] [PATCH 5/5] drm/i915: Solve the GPU reset vs. modeset deadlocks with an rw_semaphore

2017-06-30 Thread Daniel Vetter
On Thu, Jun 29, 2017 at 04:49:48PM +0300, ville.syrj...@linux.intel.com wrote: > From: Ville Syrjälä > > Introduce an rw_semaphore to protect the display commits. All normal > commits use down_read() and hence can proceed in parallel, but GPU reset > will use

Re: [Intel-gfx] [PATCH 5/5] drm/i915: Solve the GPU reset vs. modeset deadlocks with an rw_semaphore

2017-06-29 Thread Chris Wilson
Quoting ville.syrj...@linux.intel.com (2017-06-29 14:49:48) > @@ -2640,15 +2600,13 @@ static void i915_reset_device(struct drm_i915_private > *dev_priv) > char *error_event[] = { I915_ERROR_UEVENT "=1", NULL }; > char *reset_event[] = { I915_RESET_UEVENT "=1", NULL }; >

[Intel-gfx] [PATCH 5/5] drm/i915: Solve the GPU reset vs. modeset deadlocks with an rw_semaphore

2017-06-29 Thread ville . syrjala
From: Ville Syrjälä Introduce an rw_semaphore to protect the display commits. All normal commits use down_read() and hence can proceed in parallel, but GPU reset will use down_write() making sure no other commits are in progress when we have to pull the plug on the