[PATCH] D103334: [GISel] Eliminate redundant bitmasking

[Jessica Paquette via Phabricator via llvm-commits]

paquette added inline comments.


================
Comment at: llvm/include/llvm/Target/GlobalISel/Combine.td:650
   redundant_and, redundant_sext_inreg, redundant_or, urem_pow2_to_mask,
-  zext_trunc_fold, icmp_to_true_false_known_bits]>;
+  zext_trunc_fold, icmp_to_true_false_known_bits, overlapping_and]>;
 
----------------
Probably shouldn't be in this group since it doesn't use KnownBits?


================
Comment at: llvm/lib/CodeGen/GlobalISel/CombinerHelper.cpp:3008
+
+  // FIXME: This should be removed once GISelKnownBits supports vectors.
+  if (Ty.isVector())
----------------
This doesn't use KnownBits?

(If this was yoinked from somewhere else, then we probably need to update a combine. GISelKnownBits supports vectors nowadays last time I checked.)


================
Comment at: llvm/lib/CodeGen/GlobalISel/CombinerHelper.cpp:3025
+      B.buildCopy(Dst, B.buildConstant(Ty, 0));
+    }
+  };
----------------
Minor stylistic suggestion:

```
if (C1 & C2) {
  B.buildAnd(...);
  return;
}

B.buildCopy(...);
```

Is it actually necessary to emit the copy at all though? You could do

```
Dst = G_CONSTANT iN 0
```

right?


================
Comment at: llvm/test/CodeGen/AArch64/GlobalISel/opt-overlapping-and.ll:1
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc < %s -o - -verify-machineinstrs -global-isel=1 | FileCheck %s
----------------
Can this be a MIR test?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D103334/new/

https://reviews.llvm.org/D103334