[llvm] [RISCV] Remove vmv.s.x and vmv.x.s lmul pseudo variants (PR #71501)

[Philip Reames via llvm-commits]

================
@@ -534,7 +534,7 @@ define void @masked_load_v2i32_align1(ptr %a, <2 x i32> %m, ptr %res_ptr) nounwi
 ; RV64-SLOW:       # %bb.0:
 ; RV64-SLOW-NEXT:    vsetivli zero, 2, e32, mf2, ta, ma
 ; RV64-SLOW-NEXT:    vmseq.vi v8, v8, 0
-; RV64-SLOW-NEXT:    vsetvli zero, zero, e8, mf8, ta, ma
+; RV64-SLOW-NEXT:    vsetivli zero, 1, e8, m1, ta, ma
----------------
preames wrote:

This looks like a missing vsetvli insertion optimization.  

https://github.com/llvm/llvm-project/pull/71501