[llvm] [ScalarizeMaskedMemIntrin] Use pointer alignment from pointer of masked.compressstore/expandload. (PR #83519)

[Craig Topper via llvm-commits]

================
@@ -0,0 +1,986 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 4
+; RUN: llc -mtriple=riscv32 -mattr=+m,+v -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,RV32
+; RUN: llc -mtriple=riscv64 -mattr=+m,+v -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,RV64
+
+declare void @llvm.masked.compressstore.v1i8(<1 x i8>, ptr, <1 x i1>)
+define void @compressstore_v1i8(ptr align 2 %base, <1 x i8> %v, <1 x i1> %mask) {
+; CHECK-LABEL: compressstore_v1i8:
+; CHECK:       # %bb.0:
+; CHECK-NEXT:    vsetvli a1, zero, e8, mf8, ta, ma
+; CHECK-NEXT:    vfirst.m a1, v0
+; CHECK-NEXT:    bnez a1, .LBB0_2
+; CHECK-NEXT:  # %bb.1: # %cond.store
+; CHECK-NEXT:    vsetivli zero, 1, e8, mf8, ta, ma
+; CHECK-NEXT:    vse8.v v8, (a0)
+; CHECK-NEXT:  .LBB0_2: # %else
+; CHECK-NEXT:    ret
+  call void @llvm.masked.compressstore.v1i8(<1 x i8> %v, ptr %base, <1 x i1> %mask)
----------------
topperc wrote:

The alignment attribute should be on parameter in the intrinsic call. See for example, `unaligned_vpload_nxv1i64_allones_mask` in vpload.ll

https://github.com/llvm/llvm-project/pull/83519