Re: [PATCH] drm/i915: Move flags describing VMA mappings into the VMA

[Daniel Vetter <daniel@ffwll.ch>]

On Fri, Oct 24, 2014 at 04:43:08PM +0200, Daniel Vetter wrote:
> On Fri, Oct 24, 2014 at 01:09:08PM +0100, Chris Wilson wrote:
> > On Fri, Oct 24, 2014 at 12:42:33PM +0100, Tvrtko Ursulin wrote:
> > > From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> > > 
> > > If these flags are on the object level it will be more difficult to allow
> > > for multiple VMAs per object.
> > > 
> > > v2: Simplification and cleanup after code review comments (Chris Wilson).
> > > 
> > > Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> > > Cc: Chris Wilson <chris@chris-wilson.co.uk>
> > Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
> > 
> > > diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c
> > > index 053d99e..80f0c80 100644
> > > --- a/drivers/gpu/drm/i915/i915_gpu_error.c
> > > +++ b/drivers/gpu/drm/i915/i915_gpu_error.c
> > > @@ -567,6 +567,7 @@ i915_error_object_create(struct drm_i915_private *dev_priv,
> > >  			 struct i915_address_space *vm)
> > >  {
> > >  	struct drm_i915_error_object *dst;
> > > +	struct i915_vma *vma = NULL;
> > >  	int num_pages;
> > >  	bool use_ggtt;
> > >  	int i = 0;
> > > @@ -587,16 +588,17 @@ i915_error_object_create(struct drm_i915_private *dev_priv,
> > >  		dst->gtt_offset = -1;
> > >  
> > >  	reloc_offset = dst->gtt_offset;
> > > +	if (i915_is_ggtt(vm))
> > > +		vma = i915_gem_obj_to_ggtt(src);
> > >  	use_ggtt = (src->cache_level == I915_CACHE_NONE &&
> > > -		    i915_is_ggtt(vm) &&
> > > -		    src->has_global_gtt_mapping &&
> > > -		    reloc_offset + num_pages * PAGE_SIZE <= dev_priv->gtt.mappable_end);
> > > +		   vma && (vma->bound & GLOBAL_BIND) &&
> > > +		   reloc_offset + num_pages * PAGE_SIZE <= dev_priv->gtt.mappable_end);
> > >  
> > >  	/* Cannot access stolen address directly, try to use the aperture */
> > >  	if (src->stolen) {
> > >  		use_ggtt = true;
> > >  
> > > -		if (!src->has_global_gtt_mapping)
> > > +		if (!(vma && vma->bound & GLOBAL_BIND))
> > >  			goto unwind;
> > 
> > This looked odd as I have the vma being passed in:
> > 
> > static struct drm_i915_error_object *
> > i915_error_object_create(struct drm_i915_private *dev_priv,
> >                          struct i915_vma *vma);
> > 
> > but I presume the ppgtt error-capture is still outstanding.
> 
> Oh meh, this is probably one of those tasks that I've asked Jesse to close
> in one of the "fuck this all" rage-quits in the past few months ... Indeed
> both code and testcase seem to not have materialized afaics.
> 
> Jesse?

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