[Intel-gfx] [PATCH 1/4] drm/i915: Reduce the time we hold struct mutex in intel_pipe_set_base()

Daniel Vetter daniel at ffwll.ch
Wed Mar 12 16:16:24 CET 2014


On Wed, Mar 12, 2014 at 08:32:41AM +0000, Chris Wilson wrote:
> On Tue, Mar 11, 2014 at 07:37:33PM +0200, ville.syrjala at linux.intel.com wrote:
> > From: Ville Syrjälä <ville.syrjala at linux.intel.com>
> > 
> > We don't need to hold struct_mutex all through intel_pipe_set_base(),
> > just need to hold it while pinning/unpinning the buffers.
> > 
> > So reduce the struct_mutext usage in intel_pipe_set_base() just like we
> > did for the sprite code in:
> >  commit 82284b6becdbef6d8cd3fb43e8698510833a5129
> >  Author: Ville Syrjälä <ville.syrjala at linux.intel.com>
> >  Date:   Tue Oct 1 18:02:12 2013 +0300
> > 
> >     drm/i915: Reduce the time we hold struct mutex in sprite update_plane code
> > 
> > The FBC and PSR locking is still entirely fubar. That stuff was
> > previouly done while holding struct_mutex, so leave it there for now.
> 
> Yup. I am amazed we enabled FBC when it has known deadlocks...

Well I occasional decide to retreat from certain fronts to fighther the
bigger war. But I've resolved that any psr/fbc stuff will only be merged
enabled by default now. We might need to postpone the locking fixes past
merging byt psr though I fear ...

> Reviewed-by: Chris Wilson <chris at chris-wilson.co.uk>

Merged this one here for now, I'll let you two duke it out over the
remaining patches ;-)

Thanks, Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch



More information about the Intel-gfx mailing list