[PATCH] D37504: [x86] Fix PR34377 by disabling cmov conversion when we relied on it performing a zext of a register.

[Amjad Aboud via Phabricator via llvm-commits]

aaboud added inline comments.


================
Comment at: llvm/trunk/lib/Target/X86/X86CmovConversion.cpp:297
+                [&](MachineInstr &UseI) {
+                  return UseI.getOpcode() == X86::SUBREG_TO_REG;
+                }))
----------------
Thanks Chandler for the quick fix.
However, we can reduce the restriction to only the case where:
1. We are compiling for 64bit
2. The CMOV destination is 32bit.

All other cases has no issue with the CMOV behavior.

Do you agree?


Repository:
  rL LLVM

https://reviews.llvm.org/D37504