[llvm] [RISCV] Lower fixed reverse vector_shuffles through vector_reverse (PR #104461)

[Philip Reames via llvm-commits]

================
@@ -229,10 +229,10 @@ define <32 x i8> @reverse_v32i8(<32 x i8> %a) {
 ; CHECK-LABEL: reverse_v32i8:
 ; CHECK:       # %bb.0:
 ; CHECK-NEXT:    li a0, 32
-; CHECK-NEXT:    lui a1, %hi(.LCPI12_0)
-; CHECK-NEXT:    addi a1, a1, %lo(.LCPI12_0)
 ; CHECK-NEXT:    vsetvli zero, a0, e8, m2, ta, ma
-; CHECK-NEXT:    vle8.v v12, (a1)
+; CHECK-NEXT:    vid.v v10
----------------
preames wrote:

I think I'm missing something, possibly two somethings.

1) It seems we already support fixed vector vector.reverse?  Do we current legalize to SHUFFLE_VECTOR or something?

2) Why is the prior lowering loading the constant from the constant pool?  The VID sequence here looks always profitable, even for generic shuffle lowering..

https://github.com/llvm/llvm-project/pull/104461