[PATCH] D33212: AMDGPU/GlobalISel: Mark 32-bit float constants as legal

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: lib/Target/AMDGPU/SIInstrInfo.cpp:2118-2120
+
+  if (SIInstrInfo::isGenericOpcode(MI.getOpcode()))
+    return true;
----------------
arsenm wrote:
> tstellar wrote:
> > arsenm wrote:
> > > Why is this necessary? It looks like a separate change
> > The verifier rejects G_FCONSTANT instructions,  because they have FPImm operands.
> Does the base verifier do anything for these? This should call the base implementation
I guess the description is it's for target instructions, and the verifier itself could handle generic ones so nevermind


https://reviews.llvm.org/D33212