[llvm] Reapply [AMDGPU] Avoid resource propagation for recursion through multiple functions (PR #112251)

[Matt Arsenault via llvm-commits]

================
@@ -0,0 +1,82 @@
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx90a < %s | FileCheck %s
+
+; CHECK-LABEL: {{^}}qux
+; CHECK: .set qux.num_vgpr, 0
+; CHECK: .set qux.num_agpr, 0
+; CHECK: .set qux.numbered_sgpr, 32
+; CHECK: .set qux.private_seg_size, 0
+; CHECK: .set qux.uses_vcc, 0
+; CHECK: .set qux.uses_flat_scratch, 0
+; CHECK: .set qux.has_dyn_sized_stack, 0
+; CHECK: .set qux.has_recursion, 0
+; CHECK: .set qux.has_indirect_call, 0
+define void @qux() {
+entry:
----------------
arsenm wrote:

Needs some register usage content in this file 

https://github.com/llvm/llvm-project/pull/112251