Timothy Arceri <tarc...@itsqueeze.com> writes: > On 10/10/17 09:31, Eric Anholt wrote: >> Timothy Arceri <tarc...@itsqueeze.com> writes: >> >>> We won't split varyings marked as always active because there >>> is no point in doing so. This means we need to mark both >>> sides of the interface as always active otherwise we will have >>> a mismatch and start removing things we shouldn't. >> >> Is this just needed because the next patch only updates a single bit in >> the read/write arrays for multi-channel variables? Or is there >> something else? > > We want to keep both sides the same. We don't want to have: > > vec4 on one side and float * 4 on the other. We might be able to handle > it but just for the simple readability of NIR dumps I don't think we > should do it. Thoughts?
Having them mix-and-matched seems OK to me, though I guess I can imagine similar issues popping up in other linking passes. Could you add a comment explaining the motivation here?
signature.asc
Description: PGP signature
_______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev
