[llvm] fcacda8 - [RISCV] Remove constant_fold_cast_op from RISCVPostLegalizerCombiner.

[Craig Topper via llvm-commits]

Author: Craig Topper
Date: 2024-11-12T23:28:48-08:00
New Revision: fcacda899fcd812251a44a5b01548d7bb74d0481

URL: https://github.com/llvm/llvm-project/commit/fcacda899fcd812251a44a5b01548d7bb74d0481
DIFF: https://github.com/llvm/llvm-project/commit/fcacda899fcd812251a44a5b01548d7bb74d0481.diff

LOG: [RISCV] Remove constant_fold_cast_op from RISCVPostLegalizerCombiner.

This is no longer tested after other recent changes. AArch64 does
have this in their PostLegalizerCombiner.

Added: 
    

Modified: 
    llvm/lib/Target/RISCV/RISCVCombine.td

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/RISCV/RISCVCombine.td b/llvm/lib/Target/RISCV/RISCVCombine.td
index 22a447c2649b29..030613a7d89040 100644
--- a/llvm/lib/Target/RISCV/RISCVCombine.td
+++ b/llvm/lib/Target/RISCV/RISCVCombine.td
@@ -24,6 +24,6 @@ def RISCVO0PreLegalizerCombiner: GICombiner<
 def RISCVPostLegalizerCombiner
     : GICombiner<"RISCVPostLegalizerCombinerImpl",
                  [sub_to_add, combines_for_extload, redundant_and,
-                  identity_combines, shift_immed_chain, commute_constant_to_rhs,
-                  constant_fold_cast_op]> {
+                  identity_combines, shift_immed_chain,
+                  commute_constant_to_rhs]> {
 }