Skip to content
Snippets Groups Projects

vkd3d-shader/hlsl: Fixes for sm1 semantics.

Merged Elizabeth Figura requested to merge zfigura/vkd3d:pr7 into master
All threads resolved!

Merge request reports

Merge request pipeline #9738 skipped

Merge request pipeline skipped for 0654d88e

Approval is optional

Merged by Alexandre JulliardAlexandre Julliard 2 years ago (Apr 18, 2023 8:34pm UTC)

Merge details

  • Changes merged into with 0654d88e.
  • Did not delete the source branch.

Activity

Filter activity
  • Approvals
  • Assignees & reviewers
  • Comments (from bots)
  • Comments (from users)
  • Commits & branches
  • Edits
  • Labels
  • Lock status
  • Mentions
  • Merge request status
  • Tracking
  • Approving this because nothing screams of being wrong and I'm not uncomfortable with anything in this MR, but my knowledge of the low level bits of SM1 is pretty weak, so I can't really give an independent confirmation that we want these changes. But I have a general trust that Zeb knows what she's doing! :-)

  • Giovanni Mascellani approved this merge request

    approved this merge request

    • Resolved by Matteo Bruni

      Approving this because nothing screams of being wrong and I'm not uncomfortable with anything in this MR, but my knowledge of the low level bits of SM1 is pretty weak, so I can't really give an independent confirmation that we want these changes. But I have a general trust that Zeb knows what she's doing! :-)

      Sanity-checking the code changes is always appreciated even if you have to trust that the functional change is correct :-)

      1/4 and 4/4 were both things I came across while trying to get sm1 tests to work, so in a sense we already have tests for those, we just can't quite run them yet.

      2/4 and 3/4 are things related to ps_1_, which I think we originally dismissed as not worth implementing, on the basis that native doesn't support emitting them and implicitly promotes them to ps_2_. But native d3dx9_31 does support emitting ps_1_*, and perhaps more importantly, we want to be able to test that our sm1 -> glsl/spirv translator can correctly handle those versions.

  • Matteo Bruni approved this merge request

    approved this merge request

  • Matteo Bruni resolved all threads

    resolved all threads

  • 2/4 and 3/4 are things related to ps_1_, which I think we originally dismissed as not worth implementing, on the basis that native doesn't support emitting them and implicitly promotes them to ps_2_.

    I've certainly heard that view before, but I don't quite agree. That particular fact may certainly make it a bit harder to justify spending much time on these older shader models in a more corporate environment, but I'll happily take the patches if someone writes them.

  • Henri Verbeet approved this merge request

    approved this merge request

  • added 47 commits

    • c726e139...8ed74377 - 43 commits from branch wine:master
    • 98e24613 - vkd3d-shader/hlsl: Sort only uniforms by name in the sm1 backend.
    • 99d41320 - vkd3d-shader/hlsl: Map ps_1_* color and texcoord semantics to registers.
    • b19105ea - vkd3d-shader/hlsl: Do not write DCL instructions for ps_1_* shaders.
    • 0654d88e - vkd3d-shader/hlsl: Allow VPOS and VFACE to be specified in sm3 pixel shaders.

    Compare with previous version

  • Alexandre Julliard approved this merge request

    approved this merge request

Please register or sign in to reply
Loading