[llvm] AMDGPU GlobalISel Add64 support (PR #124763)

[Matt Arsenault via llvm-commits]

================
@@ -46,7 +63,13 @@ define amdgpu_kernel void @lshl_add_u64_s2v(i64 %v) {
 
 define amdgpu_kernel void @lshl_add_u64_v2s(i64 %a) {
 ; GCN-LABEL: lshl_add_u64_v2s:
-; GCN: v_lshl_add_u64 v[{{[0-9:]+}}], v[{{[0-9:]+}}], 2, s[{{[0-9:]+}}]
+; GCN:       ; %bb.0:
+; GCN-NEXT:    flat_load_dwordx2 v[0:1], v[0:1]
+; GCN-NEXT:    s_load_dwordx2 s[0:1], s[4:5], 0x24
+; GCN-NEXT:    s_waitcnt vmcnt(0) lgkmcnt(0)
+; GCN-NEXT:    v_lshl_add_u64 v[0:1], v[0:1], 2, s[0:1]
+; GCN-NEXT:    flat_store_dwordx2 v[0:1], v[0:1] sc0 sc1
+; GCN-NEXT:    s_endpgm
   %v = load i64, ptr undef
----------------
arsenm wrote:

This test should be fixed to avoid loading from undef 

https://github.com/llvm/llvm-project/pull/124763