[PATCH] D73585: [AMDGPU][GlobalISel] Select 8-byte LDS Ops with 4-byte alignment

Matt Arsenault via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Tue Jan 28 14:43:59 PST 2020


arsenm added inline comments.


================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUInstructionSelector.cpp:2353
+  int64_t ConstAddr = 0;
+  if (isBaseWithConstantOffset(Root, *MRI)) {
+    const MachineOperand &LHS = RootDef->getOperand(1);
----------------
This can use getPtrBaseWithConstantOffset and avoid re-figuring out the offset below


================
Comment at: llvm/lib/Target/AMDGPU/AMDGPUInstructionSelector.cpp:2371
+    }
+  } else if (RootDef->getOpcode() == AMDGPU::G_SUB) {
+    // TODO
----------------
G_SUB isn't pointer typed, so this would never happen. This would require looking through G_INTTOPTR


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D73585/new/

https://reviews.llvm.org/D73585





More information about the llvm-commits mailing list