Am 24.02.2014 09:38, schrieb Dave Airlie: > I've taken a quick look at ARB_texture_query_levels, > > and the HW generally seems to implement this by returning it in the W > channel of RESINFO, which would map to the TXQ instruction, > > So my initial implementation added a LEVQ TGSI instruction but I do > think this isn't representative of the hw and I should add it to TXQ > with maybe a CAP to say it supports levels. > > Dave. > _______________________________________________ > mesa-dev mailing list > mesa-dev@lists.freedesktop.org > https://urldefense.proofpoint.com/v1/url?u=http://lists.freedesktop.org/mailman/listinfo/mesa-dev&k=oIvRg1%2BdGAgOoM1BIlLLqw%3D%3D%0A&r=F4msKE2WxRzA%2BwN%2B25muztFm5TSPwE8HKJfWfR2NgfY%3D%0A&m=5d7NG60G1rh6MNbfSxLGJ6BP279NlasrWRKYXmIL1rA%3D%0A&s=3d194a0f62820a267c7954a144cc78ebcca134ea4e9c1d3cb6e3c7aafe6678bb >
Doing this with TXQ sounds right to me. I'm not sure you even need a cap bit - this is typically dx10 hw which can do resinfo and dx10 is the reason all hw returns number of levels in the w channel anyway. Though it's possible there's some newer non-desktop hw out there not concerned about d3d (but OpenGL ES instead) which can't do it, so a cap is fine too. llvmpipe actually right now will only return this in the w channel for sviewinfo, not txq, but it could be changed. I think as long as you give it the right write mask it should be able to drop all the extra unneeded calculations. Though obviously for llvmpipe/softpipe it can be implemented easily either way. Roland _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
