[clang] [AArch64][SME] Remove immediate argument restriction for svldr and svstr (PR #68908)

[Sander de Smalen via cfe-commits]

================
@@ -38,3 +38,18 @@ void test_svstr_vnum_za_1(uint32_t slice_base, void *ptr) {
 void test_svstr_za(uint32_t slice_base, void *ptr) {
   svstr_za(slice_base, ptr);
 }
+
+// CHECK-C-LABEL: @test_svstr_vnum_za_var(
+// CHECK-CXX-LABEL: @_Z22test_svstr_vnum_za_varjPvm(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[SVLB:%.*]] = tail call i64 @llvm.aarch64.sme.cntsb()
+// CHECK-NEXT:    [[MULVL:%.*]] = mul i64 [[SVLB]], [[VNUM:%.*]]
+// CHECK-NEXT:    [[TMP0:%.*]] = getelementptr i8, ptr [[PTR:%.*]], i64 [[MULVL]]
+// CHECK-NEXT:    [[TMP1:%.*]] = trunc i64 [[VNUM:%.*]] to i32
+// CHECK-NEXT:    [[TILESLICE:%.*]] = add i32 [[TMP1]], [[SLICE_BASE:%.*]]
+// CHECK-NEXT:    tail call void @llvm.aarch64.sme.str(i32 [[TILESLICE]], ptr [[TMP0]])
+// CHECK-NEXT:    ret void
+//
+void test_svstr_vnum_za_var(uint32_t slice_base, void *ptr, uint64_t vnum) {
----------------
sdesmalen-arm wrote:

```suggestion
void test_svstr_vnum_za_var(uint32_t slice_base, void *ptr, int64_t vnum) {
```

https://github.com/llvm/llvm-project/pull/68908