On Mon, Jul 13, 2015 at 2:44 PM, Emil Velikov <[email protected]> wrote: > On 13 July 2015 at 19:30, Ilia Mirkin <[email protected]> wrote: >> On Mon, Jul 13, 2015 at 2:28 PM, Emil Velikov <[email protected]> >> wrote: >>> On 13 July 2015 at 17:32, samuel.pitoiset <[email protected]> wrote: >>>> On 13/07/2015 18:08, Emil Velikov wrote: >>>>> >>>>> Is there some widely know issue with it ? mesa & envytools does not >>>>> mention anything (based of a quick scan). >>>> >>>> >>>> There is no know issue with NV50_3D_BIND_TSC2, but Ilia doesn't seem to >>>> want >>>> to use it. :-) >>>> That's why I followed his suggestion by reducing the max. number of >>>> samplers >>>> to 16 like for nvc0. >>>> >>> Ack, thanks for the confirmation. Currently it reads like a warning >>> sign "stay away from XXX" - so if you want to you can soften if a bit. >> >> "stay away from BIND_TSC2" is right. It's very much not to be used for >> regular applications from what I can tell. > I guess I'm having a dull moment and cannot justify the "hostility" > towards it. Is it solely based upon "the blob does not use it so we > shouldn't either" or is there something else ? Perhaps they cannot be > bothered to hook it up. Should that be a discouragement towards using > it in nouveau ?
It's an extra 2 (not 16 as the original patch thought!) samplers. Whole separate system for setting them. Seems like they may have been separate for a reason. There's no API reason to expose them either. -ilia _______________________________________________ mesa-dev mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/mesa-dev
