On Fri, 13 Jul 2012 17:41:01 +0200, Daniel Vetter <daniel at ffwll.ch> wrote: > On Fri, Jul 13, 2012 at 02:14:07PM +0100, Chris Wilson wrote: > > As we always flush the GPU cache prior to emitting the breadcrumb, we no > > longer have to worry about the deferred flush causing the > > pending_gpu_write to be delayed. So we can instead utilize the known > > last_write_seqno to hopefully minimise the wait times. > > > > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk> > > I like this, but I'd prefer if this would be at the end of the series as a > neat improvement - I'd really prefer if we get together a somewhat minimal > fix to take care of the flushing list and intel_begin_ring interruptable > patch. > > The reason is that the merge window approaches fast, and if we're unlucky > the current -next cycle won't make it into 3.6, so I'd have to send Dave a > smaller pile of fixes. Which this doesn't really look like. > > Or do I again miss something? It's not really my brightest day ;-) This is just to enable removing the pending_gpu_write bit, and really if the flushing list removal wasn't to enable this patch, what was the point? :-p -Chris -- Chris Wilson, Intel Open Source Technology Centre