[PATCH] D26303: AMDGPU/SI: Fix visit order assumption in SIFixSGPRCopies

Matt Arsenault via llvm-commits llvm-commits at lists.llvm.org
Mon Nov 7 08:32:57 PST 2016


arsenm accepted this revision.
arsenm added a comment.

LGTM with nits



================
Comment at: lib/Target/AMDGPU/SIFixSGPRCopies.cpp:242
+
+  for (unsigned i = 1; i < PHI.getNumOperands(); i+=2) {
+    unsigned Reg = PHI.getOperand(i).getReg();
----------------
Spaces around +=


================
Comment at: lib/Target/AMDGPU/SIFixSGPRCopies.cpp:253
+
+  for (unsigned i = 1; i < PHI.getNumOperands(); i+=2) {
+    unsigned Reg = PHI.getOperand(i).getReg();
----------------
Ditto


================
Comment at: lib/Target/AMDGPU/SIFixSGPRCopies.cpp:262
+    assert(DefInstr);
+    switch(DefInstr->getOpcode()) {
+    default: break;
----------------
Space before (


================
Comment at: lib/Target/AMDGPU/SIFixSGPRCopies.cpp:263
+    switch(DefInstr->getOpcode()) {
+    default: break;
+    case AMDGPU::SI_BREAK:
----------------
break on new line


https://reviews.llvm.org/D26303





More information about the llvm-commits mailing list