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

[LLVM Continuous Integration via llvm-commits]

llvm-ci wrote:

LLVM Buildbot has detected a new failure on builder `premerge-monolithic-windows` running on `premerge-windows-1` while building `llvm` at step 8 "test-build-unified-tree-check-all".

Full details are available at: https://lab.llvm.org/buildbot/#/builders/35/builds/3687

<details>
<summary>Here is the relevant piece of the build log for the reference</summary>

```
Step 8 (test-build-unified-tree-check-all) failure: test (failure)
******************** TEST 'LLVM :: CodeGen/AMDGPU/function-resource-usage.ll' FAILED ********************
Exit Code: 1

Command Output (stdout):
--
# RUN: at line 1
c:\ws\buildbot\premerge-monolithic-windows\build\bin\llc.exe -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -enable-ipra=0 < C:\ws\buildbot\premerge-monolithic-windows\llvm-project\llvm\test\CodeGen\AMDGPU\function-resource-usage.ll | c:\ws\buildbot\premerge-monolithic-windows\build\bin\filecheck.exe -check-prefix=GCN C:\ws\buildbot\premerge-monolithic-windows\llvm-project\llvm\test\CodeGen\AMDGPU\function-resource-usage.ll
# executed command: 'c:\ws\buildbot\premerge-monolithic-windows\build\bin\llc.exe' -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -enable-ipra=0
# executed command: 'c:\ws\buildbot\premerge-monolithic-windows\build\bin\filecheck.exe' -check-prefix=GCN 'C:\ws\buildbot\premerge-monolithic-windows\llvm-project\llvm\test\CodeGen\AMDGPU\function-resource-usage.ll'
# .---command stderr------------
# | C:\ws\buildbot\premerge-monolithic-windows\llvm-project\llvm\test\CodeGen\AMDGPU\function-resource-usage.ll:597:8: error: GCN: expected string not found in input
# | ; GCN: .set multi_call_with_multi_stage_recurse.numbered_sgpr, max(43, use_stack0.numbered_sgpr, use_stack1.numbered_sgpr, multi_stage_recurse1.numbered_sgpr)
# |        ^
# | <stdin>:2381:132: note: scanning from here
# |  .set multi_call_with_multi_stage_recurse.num_agpr, max(0, use_stack0.num_agpr, use_stack1.num_agpr, multi_stage_recurse1.num_agpr)
# |                                                                                                                                    ^
# | <stdin>:2382:2: note: possible intended match here
# |  .set multi_call_with_multi_stage_recurse.numbered_sgpr, max(45, use_stack0.numbered_sgpr, use_stack1.numbered_sgpr, multi_stage_recurse1.numbered_sgpr)
# |  ^
# | 
# | Input file: <stdin>
# | Check file: C:\ws\buildbot\premerge-monolithic-windows\llvm-project\llvm\test\CodeGen\AMDGPU\function-resource-usage.ll
# | 
# | -dump-input=help explains the following input dump.
# | 
# | Input was:
# | <<<<<<
# |              .
# |              .
# |              .
# |           2376:  .text 
# |           2377: .Lfunc_end31: 
# |           2378:  .size multi_call_with_multi_stage_recurse, .Lfunc_end31-multi_call_with_multi_stage_recurse 
# |           2379:  ; -- End function 
# |           2380:  .set multi_call_with_multi_stage_recurse.num_vgpr, max(41, use_stack0.num_vgpr, use_stack1.num_vgpr, multi_stage_recurse1.num_vgpr) 
# |           2381:  .set multi_call_with_multi_stage_recurse.num_agpr, max(0, use_stack0.num_agpr, use_stack1.num_agpr, multi_stage_recurse1.num_agpr) 
# | check:597'0                                                                                                                                        X error: no match found
# |           2382:  .set multi_call_with_multi_stage_recurse.numbered_sgpr, max(45, use_stack0.numbered_sgpr, use_stack1.numbered_sgpr, multi_stage_recurse1.numbered_sgpr) 
# | check:597'0     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# | check:597'1      ?                                                                                                                                                        possible intended match
# |           2383:  .set multi_call_with_multi_stage_recurse.private_seg_size, 0+(max(use_stack0.private_seg_size, use_stack1.private_seg_size, multi_stage_recurse1.private_seg_size)) 
# | check:597'0     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# |           2384:  .set multi_call_with_multi_stage_recurse.uses_vcc, or(1, use_stack0.uses_vcc, use_stack1.uses_vcc, multi_stage_recurse1.uses_vcc) 
# | check:597'0     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# |           2385:  .set multi_call_with_multi_stage_recurse.uses_flat_scratch, or(1, use_stack0.uses_flat_scratch, use_stack1.uses_flat_scratch, multi_stage_recurse1.uses_flat_scratch) 
# | check:597'0     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# |           2386:  .set multi_call_with_multi_stage_recurse.has_dyn_sized_stack, or(0, use_stack0.has_dyn_sized_stack, use_stack1.has_dyn_sized_stack, multi_stage_recurse1.has_dyn_sized_stack) 
# | check:597'0     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# |           2387:  .set multi_call_with_multi_stage_recurse.has_recursion, or(1, use_stack0.has_recursion, use_stack1.has_recursion, multi_stage_recurse1.has_recursion) 
# | check:597'0     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
...

```

</details>

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