[llvm] [AMDGPU] Handle natively unsupported types in addrspace(7) lowering (PR #110572)

[Matt Arsenault via llvm-commits]

================
@@ -0,0 +1,10 @@
+; Note: The exact error messages aren't important here, but are included to catch
+; anything changing.
+; RUN: not --crash llc -global-isel=0 -mtriple=amdgcn -mcpu=gfx900 < %s 2>&1 \
+; RUN: | grep "LLVM ERROR: Scalarization of scalable vectors is not supported."
+; RUN: not --crash llc -global-isel=1 -mtriple=amdgcn -mcpu=gfx900 < %s 2>&1 \
+; RUN: | grep "LLVM ERROR: Invalid size request on a scalable vector."
+define void @buffer_store_nxv2i32(ptr addrspace(8) inreg %rsrc, i32 %offset) {
----------------
arsenm wrote:

Use FileCheck and not grep. Also -filetype=null 

https://github.com/llvm/llvm-project/pull/110572