[PATCH] D74316: AMDGPU/GlobalISel: Start selecting image intrinsics

[Jay Foad via Phabricator via llvm-commits]

foad added inline comments.


================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUInstructionSelector.cpp:1266
+  const bool UseNSA = NumVAddrRegs != 1 && NumVAddrDwords == NumVAddrRegs;
+  if (UseNSA && !STI.hasFeature(AMDGPU::FeatureNSAEncoding)) {
+    LLVM_DEBUG(dbgs() << "Trying to use NSA on non-NSA target\n");
----------------
`STI.hasNSAEncoding()`?


================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUInstructionSelector.cpp:1322
+  MIB.addImm(IsA16 &&  // a16 or r128
+             STI.hasFeature(AMDGPU::FeatureR128A16) ? -1 : 0);
+  if (IsGFX10)
----------------
`STI.hasR128A16()`?


================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUInstructionSelector.cpp:1357
     return selectDSAppendConsume(I, false);
-  default:
+  default: {
     return selectImpl(I, *CoverageInfo);
----------------
Looks like a spurious change.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D74316/new/

https://reviews.llvm.org/D74316