[llvm] [AMDGPU] Accept extractelement of a widening cast when folding image ops to a16 (PR #208207)

[Barbara Mitic via llvm-commits]

================
@@ -4018,6 +4018,55 @@ define amdgpu_kernel void @image_load_a16_mip_2d_const_noopt(ptr addrspace(1) %o
   ret void
 }
 
+define amdgpu_kernel void @image_load_a16_mip_2d_sext_vec(ptr addrspace(1) %out, <8 x i32> inreg %rsrc, <2 x i16> %coord) {
----------------
barbara-amd wrote:

These are already auto-generated by update_test_checks.py, not hand-written - double checked. The [[TMP1]]/[[TMP2]] names come from the fold creating fresh anonymous extractelements (and deleting the original named %s32/%t32), so the script auto-numbers the new unnamed values rather than reusing source names I think.

https://github.com/llvm/llvm-project/pull/208207