Quoting Matthew Auld (2019-06-14 12:15:27) > On Fri, 14 Jun 2019 at 08:10, Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> wrote: > > > > Throughout the code base we internally track vma (objects bound into > > a particular GTT), with the objects themselves being the common backing > > storage. By making the vma itself reference counted we can start > > operating on the vma concurrently, moving work into async threads. > > > > Just the conversion to making sure we keep track of the vma reference > > counts is not particularly pleasant. > > --- > > [snip] > > > @@ -2060,6 +2057,7 @@ static struct i915_vma *pd_vma_create(struct gen6_ppgtt *ppgtt, int size) > > if (!vma) > > return ERR_PTR(-ENOMEM); > > > > + kref_init(&vma->ref); > > i915_active_init(&vma->active, NULL, NULL); > > > > vma->vm = &ggtt->vm; > > Just a first pass. Do we need i915_vm_get(&ggtt->vm); so we match the > i915_vm_put() in __i915_vma_release()? Yup. That would have an interesting explosion at some point. Tempted to leave it to make sure CI does explode :) Now that I mention it... I think I'll will leave a comment with the fix. -Chris _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
