List of SPIR-V Translator patches that were reverted during pulldowns from Khronos repo

[vmaksimo]

Should be enabled back once CPU RT is updated to new version:

• Encode debug info producer in SPIR-V KhronosGroup/SPIRV-LLVM-Translator@601e4a5. Revert patch: 6822b42
  Follow-up patch to be applied too: KhronosGroup/SPIRV-LLVM-Translator@8f59b42
  Relanded here: [SPIR-V] Reland Encode debug info producer in SPIR-V #4082
• Implement SPV_INTEL_memory_access_aliasing extension KhronosGroup/SPIRV-LLVM-Translator@974749c. Revert patch: 4f637b0
  Already reverted back - no work is needed here d6caceb
• Promote SPV_INTEL_optimization_hints to SPV_KHR_expect_assume Promote SPV_INTEL_optimization_hints to SPV_KHR_expect_assume KhronosGroup/SPIRV-LLVM-Translator#991. Revert patches:
  • 290baf6
  • f7d6423
  • 733c4fc
    Relanded in Re-land "Promote SPV_INTEL_optimization_hints to SPV_KHR_expect_assume" #3960
• Add SPIR-V 1.4 checks Add SPIR-V 1.4 checks KhronosGroup/SPIRV-LLVM-Translator#1177
• Fix the collection of entry point interfaces KhronosGroup/SPIRV-LLVM-Translator#1334

Should be enabled once GPU driver is updated to new version:

• add an entry point wrapper around functions (llvm pass) KhronosGroup/SPIRV-LLVM-Translator#1149
• Fix for the change above Translate function attributes (incl. parameter attrs) for entry point kernels KhronosGroup/SPIRV-LLVM-Translator#1409

[bader]

WRT extensions: doesn't 2398c20 allows us to keep patches in?

[AlexeySachkov]

WRT extensions: doesn't 2398c20 allows us to keep patches in?

Yes, it does, we can now re-apply patches with reverted extensions.

[vmaksimo]

Re-applied in #3687

[vmaksimo]

Unfortunately, re-applying of these patches caused regressions that were caught by llvm-test-suite. Will wait for a proper investigation and fix (maybe from @MrSidims)

[vmaksimo]

This gap is tracked in #7592. Closing as a duplicate