On Mon, Aug 20, 2012 at 1:55 PM, Maarten Lankhorst <maarten.lankhorst@xxxxxxxxxxxxx> wrote: > Hey, > > Op 20-08-12 17:15, Jerome Glisse schreef: >> On Mon, Aug 20, 2012 at 9:42 AM, Maarten Lankhorst >> <maarten.lankhorst@xxxxxxxxxxxxx> wrote: >>> How is this different from just calling with no_wait == false? >>> As far as I can tell, both paths end up with the same result.. >>> >>> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@xxxxxxxxxxxxx> >> NAK this seriously modify the behavior. The ttm_eu_del_from_lru_locked >> part is important. It must happen with lru lock held and without any >> dropping of this lock prior to wait for bo unreserve. >> > > You're right, I missed the part where it removed those, causing the later > patch to be wrong too. However I still think the code can be made more > readable. Wouldn't it be better if it used the unlocked variants instead? > It would save a lot of extra list traversals, and you could drop > removed, reserved and put_count from ttm_validate_buffer. > > ~Maarten > No, as i said the lock can not be drop, i don't see much simplification of this code. Also the path you trying to modify is taken only is some bo is reserved by some other process, which is supposed to be rare event. Cheers, Jerome _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/dri-devel
