[PATCH] D37124: AMDGPU: Correct operand types for v_mad_mix*

Dmitry Preobrazhensky via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Sat Aug 26 02:01:29 PDT 2017


dp accepted this revision.
dp added a comment.
This revision is now accepted and ready to land.

Looks good, except for minor issues.

Regarding gfx9_asm_all.s: you should not update it,  just delete all offending tests. The file will be updated by a test generator later.



================
Comment at: lib/Target/AMDGPU/AMDGPU.td:718
+def HasMadMix : Predicate<"Subtarget->hasMadMixInsts()">,
+  AssemblerPredicate<"FeatureGFX9Insts">;
+
----------------
unused


================
Comment at: lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp:4275
   if (OpSelHiIdx != -1) {
-    addOptionalImmOperand(Inst, Operands, OptIdx, AMDGPUOperand::ImmTyOpSelHi, -1);
+    // TODO: Should we change the printing to match?
+    int DefaultVal = IsPacked ? -1 : 0;
----------------
That would be nice. Maybe by a separate change?


https://reviews.llvm.org/D37124





More information about the llvm-commits mailing list