On Sun, Nov 11, 2012 at 1:10 PM, Chris Wilson <chris at chris-wilson.co.uk> wrote: > On Sat, 10 Nov 2012 21:17:05 +0100, Daniel Vetter <daniel at ffwll.ch> wrote: >> Hm, I've thought we could get away for the gpu activity tracking by >> unconditionally assuming that any passed-in object is in is in all >> read domains. For cache flushing it doesn't matter what we set anyway >> since we invalidate/flush unconditionally all caches before/after each >> batch. And for activity tracking we unconditionally put all objects in >> a given execbuf to the front of the active list (with updated >> ring/seqno). The only important thing is to keep the write tracking to >> not break our various read/read optimizations. > > One overlooked aspect is that the domain tracking helps with debugging > and identifying what the buffers are used for. It has paid dividends > many times when reading error-states, and certainly will prove to be > useful again in future. Hm, good point. Otoh why don't we allow userspace to simply pass in an arbitrary debug tag in that u64, carry the latest one around in the bo and dump that everywhere (in addition to the domains)? -Daniel -- Daniel Vetter Software Engineer, Intel Corporation +41 (0) 79 365 57 48 - http://blog.ffwll.ch
