On 2 December 2013 10:01, Chris Forbes <[email protected]> wrote: > Paul, > > You had some patches a while ago that did CMS -> UMS conversion, but > ended up not being needed -- if we wanted ot resurrect the conversion > code itself, it would now be straightforward to get CMS for signed > integer surfaces as well, until we see the app try to mask off > channels. > > Do you think this is worthwhile? >
For now, I don't think it's worthwhile. My guess is that real-world apps using signed integer multisample surfaces are very rare (perhaps non-existent), so I'd prefer not to resurrect that hack until we find an app that needs it. > > -- Chris > > On Tue, Dec 3, 2013 at 5:23 AM, Paul Berry <[email protected]> > wrote: > > On 30 November 2013 15:29, Chris Forbes <[email protected]> wrote: > >> > >> This series enables the compressed multisample layout for multisample > >> textures. > >> Previously we would only use CMS for renderbuffers, since our > texelFetch() > >> implementation didn't understand it. > >> > >> Changes from V1: > >> - Include fix for Haswell's second copy of the sample mask > >> - Actually allocate an MCS surface for textures with this layout > >> - Emit the MCS info in surface state for textures > >> - Fix swizzling of MCS data in VS > >> > >> _______________________________________________ > >> mesa-dev mailing list > >> [email protected] > >> http://lists.freedesktop.org/mailman/listinfo/mesa-dev > > > > > > I sent comments on patches 3, 6, and 9. The rest are: > > > > Reviewed-by: Paul Berry <[email protected]> > > > > Thanks for working on this, Chris! >
_______________________________________________ mesa-dev mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/mesa-dev
