[PATCH] D33860: [AMDGPU] Untangle SDWA pass from SIShrinkInstructions

[Stanislav Mekhanoshin via Phabricator via llvm-commits]

rampitec added inline comments.


================
Comment at: lib/Target/AMDGPU/SIPeepholeSDWA.cpp:261
     Mods |= Abs ? SISrcMods::ABS : 0;
-    Mods |= Neg ? SISrcMods::NEG : 0;
+    Mods ^= Neg ? SISrcMods::NEG : 0;
   } else if (Sext) {
----------------
SamWot wrote:
> Why do you use XOR here?
You can set neg modifier yoyrself, but it also can be present in the original instruction on the same operand. The right effective modifier is exactly xor of them.


Repository:
  rL LLVM

https://reviews.llvm.org/D33860