Skip to content

anv: optimize dynamic state emission with blorp

What does this MR do and why?

There are bunch of fixes in this MR to deal with allocation failures (this is reproducable with vkoverhead if you run tests for long enough).

Then we try to allocate blorp states only once since a bunch of them never change.

And finally we try to avoid repacking some states in Anv too.

Edited by Lionel Landwerlin

Merge request reports