[Intel-gfx] [PATCH] [v2] drm/i915: Check for get_pages instead of shmem (filp)

[Ben Widawsky]

On Wed, Feb 10, 2016 at 04:23:08PM +0000, Chris Wilson wrote:
> On Wed, Feb 10, 2016 at 07:42:23AM -0800, Ben Widawsky wrote:
> > Do you guys get the CI mails? This version has regressions. v1 did not. I don't
> > know what to trust.
> 
> I didn't even see v2 itself!
>  
> > On Tue, Feb 09, 2016 at 11:44:12AM -0800, Ben Widawsky wrote:
> > > This behavior of checking for a shmem backed GEM object was introduced here:
> > > commit 4c914c0c7c787b8f730128a8cdcca9c50b0784ab
> > > Author: Brad Volkin <bradley.d.volkin at intel.com>
> > > Date:   Tue Feb 18 10:15:45 2014 -0800
> > > 
> > >     drm/i915: Refactor shmem pread setup
> > > 
> > > It is possible for an object to not be a shmem backed GEM object (for example
> > > userptr objects). An example of how we hit this failure can be found through
> > > copy_batch() in the command parser because we allocate a userptr object for the
> > > batch which contains privileged instructions. Userptr calls
> > > drm_gem_private_object_init() which explicitly sets the filp to none.
> > > 
> > > NOTE: I manually retyped this from a test machine. So I haven't even compiled
> > > this exact patch.
> > > 
> > > v2: Use same logic as from a2a4f916c2f (Kristian, Dave Gordon)
> > > 
> > > Cc: Chris Wilson <chris at chris-wilson.co.uk>
> > > Cc: Kristian Høgsberg <krh at bitplanet.net>
> > > Cc: Dave Gordon <david.s.gordon at intel.com>
> > > Signed-off-by: Ben Widawsky <benjamin.widawsky at intel.com>
> > > Tested-by: Jordan Justen <jordan.l.justen at intel.com> (v1)
> > > Reviewed-by: Jordan Justen <jordan.l.justen at intel.com> (v1)
> 
> > > ---
> > >  drivers/gpu/drm/i915/i915_gem.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > 
> > > diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
> > > index e9b19bc..7fd79b0 100644
> > > --- a/drivers/gpu/drm/i915/i915_gem.c
> > > +++ b/drivers/gpu/drm/i915/i915_gem.c
> > > @@ -489,7 +489,7 @@ int i915_gem_obj_prepare_shmem_read(struct drm_i915_gem_object *obj,
> > >  
> > >  	*needs_clflush = 0;
> > >  
> > > -	if (!obj->base.filp)
> > > +	if (WARN_ON((obj->ops->flags & I915_GEM_OBJECT_HAS_STRUCT_PAGE) == 0))
> 
> Don't use WARN_ON, there is code (or will be) where we use
> prepare_shmem_read/write to determine if we can use the shmem paths.
> 
> Also i915_gem_obj_prepare_shmem_write() requires the same treatment.
> 
> My apologies I had this patch but appear to have accidentally squashed
> it whilst rebasing. Thanks!
> -Chris
> 

So... is someone going to land this fix? We need it.

-- 
Ben Widawsky, Intel Open Source Technology Center