[PATCH] D46596: [AMDGPU] Optimze old value of v_mov_b32_dpp

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: lib/Transforms/InstCombine/InstCombineCalls.cpp:3480-3482
+    if (cast<Constant>(II->getArgOperand(5))->isZeroValue() ||
+        cast<ConstantInt>(II->getArgOperand(3))->getZExtValue() != 0xF ||
+        cast<ConstantInt>(II->getArgOperand(4))->getZExtValue() != 0xF ||
----------------
I would prefer to use checked dyn_casts here rather than crash on invalid intrinsic uses


================
Comment at: test/CodeGen/AMDGPU/llvm.amdgcn.update.dpp.ll:2-3
+; RUN: llc -march=amdgcn -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck --check-prefixes=GCN,GCN-NOOPT,GFX8 %s
+; RUN: opt -S -O1 < %s | FileCheck --check-prefix=OPT %s
+; RUN: opt -S -O1 < %s | llc -march=amdgcn -mcpu=tonga -mattr=-flat-for-global | FileCheck --check-prefixes=GCN,GCN-OPT,GFX8 %s
 
----------------
These should go in the test/InstCombine/AMDGPU/amdgcn-intrinsics.ll


https://reviews.llvm.org/D46596