On Thu, Jul 31, 2014 at 9:39 AM, Daniel Vetter <daniel.vetter@xxxxxxxx> wrote: > On Thu, Jul 31, 2014 at 8:52 AM, Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> wrote: >> On Wed, Jul 30, 2014 at 09:41:59PM +0200, Daniel Vetter wrote: >>> This essentially unbreaks non-ppgtt operation where we'd scribble over >>> random memory. >>> >>> While at it give the vm_to_ppgtt function a proper prefix and make it >>> a bit more paranoid. >> >> Wrong direction. If you make ggtt/ppgtt more similar we can loose >> having to write fragile code. > > So moving the kref into the address space? I guess that would work > too, but I'm off for my extended w/e now ;-) Ok got bored and tried to implement this, but it doesn't work out too well. It needs a fair bit of reorg so that we can handle the global gtt and the ppgtt with the same refcounting scheme and clean them up properly. I think as a stop-gap this one here is better and the refcounting unification needs to be done on top. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
