[PATCH] D143293: AMDGPU: Use module flag to get code object version at IR level folow-up

[Ivan Kosarev via Phabricator via llvm-commits]

kosarev added a comment.

Do we still need `getHsaAbiVersion()` and the `ELFABIVERSION_AMDGPU_HSA_*` constants?



================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp:335
   CodeObjectVersion = AMDGPU::getCodeObjectVersion(M);
-
   if (TM.getTargetTriple().getOS() == Triple::AMDHSA) {
----------------
Is this an accidental change?


================
Comment at: llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUTargetStreamer.h:114-115
+  void initializeTargetID(const MCSubtargetInfo &STI, StringRef FeatureString,
+                          unsigned COV) {
+    initializeTargetID(STI, COV);
 
----------------
`COV` looks a bit cryptic here. What if `CodeObjectVersion`?


================
Comment at: llvm/lib/Target/AMDGPU/Utils/AMDGPUBaseInfo.h:123
   TargetIDSetting SramEccSetting;
+  unsigned CodeObjectVersion;
 
----------------
We probably want this initialised -- regardless of whether it is then assigned a value?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D143293/new/

https://reviews.llvm.org/D143293