[PATCH] D105579: [AMDGPU] Refine -O0 and -O1 passes.

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUTargetMachine.cpp:982
+  // FIXME: We need to run a pass to propagate the attributes when calls are
+  // supported.
   if (TM->getTargetTriple().getArch() == Triple::amdgcn)
----------------
rampitec wrote:
> arsenm wrote:
> > The real fixme is we should not be relying on attribute propagation for correctness
> I have just moved lost comment. Do you want to change it?
I guess the problem exists beyond here and the comment is out of place. I guess just drop the comment?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D105579/new/

https://reviews.llvm.org/D105579