[PATCH] D148623: [MachineCSE] Extend the scope of multi block processing.

[Noah Goldstein via Phabricator via llvm-commits]

goldstein.w.n added inline comments.


================
Comment at: llvm/lib/CodeGen/MachineCSE.cpp:369
+
+  MachineBasicBlock::const_reverse_iterator I = MI; ++I;
+  MachineBasicBlock::const_reverse_iterator E = CSMI;
----------------
Why reverse iterator?


================
Comment at: llvm/test/CodeGen/Thumb2/thumb2-cbnz.ll:9
-; CHECK:      cmp	r0, #0
-  %cmp1 = icmp ne i32 %c, 0
   br i1 %cmp1, label %bb3, label %bb1
----------------
why change the IR in this test?


================
Comment at: llvm/test/CodeGen/X86/ins_subreg_coalesce-3.ll:63
+; CHECK-NEXT:  .LBB0_11: # %bb4897
 ; CHECK-NEXT:    retq
 entry:
----------------
Seems to cause a regression here?


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D148623/new/

https://reviews.llvm.org/D148623