[llvm] AMDGPU: Replace copy-to-mov-imm folding logic with class compat checks (PR #154501)

[LLVM Continuous Integration via llvm-commits]

llvm-ci wrote:

LLVM Buildbot has detected a new failure on builder `ml-opt-rel-x86-64` running on `ml-opt-rel-x86-64-b2` while building `llvm` at step 6 "test-build-unified-tree-check-all".

Full details are available at: https://lab.llvm.org/buildbot/#/builders/185/builds/24117

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

```
Step 6 (test-build-unified-tree-check-all) failure: test (failure)
******************** TEST 'LLVM :: CodeGen/AMDGPU/llvm.amdgcn.rcp.bf16.ll' FAILED ********************
Exit Code: 1

Command Output (stderr):
--
/b/ml-opt-rel-x86-64-b1/build/bin/llc -global-isel=0 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -mattr=+real-true16 < /b/ml-opt-rel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rcp.bf16.ll | /b/ml-opt-rel-x86-64-b1/build/bin/FileCheck -check-prefix=SDAG-TRUE16 /b/ml-opt-rel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rcp.bf16.ll # RUN: at line 2
+ /b/ml-opt-rel-x86-64-b1/build/bin/FileCheck -check-prefix=SDAG-TRUE16 /b/ml-opt-rel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rcp.bf16.ll
+ /b/ml-opt-rel-x86-64-b1/build/bin/llc -global-isel=0 -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -mattr=+real-true16
/b/ml-opt-rel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rcp.bf16.ll:38:21: error: SDAG-TRUE16-NEXT: expected string not found in input
; SDAG-TRUE16-NEXT: v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, 0x3e80
                    ^
<stdin>:103:32: note: scanning from here
 s_load_b64 s[0:1], s[4:5], 0x0
                               ^
<stdin>:104:2: note: possible intended match here
 v_mov_b16_e32 v0.l, 0x3e80
 ^
/b/ml-opt-rel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rcp.bf16.ll:59:21: error: SDAG-TRUE16-NEXT: expected string not found in input
; SDAG-TRUE16-NEXT: v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, 0x3c24
                    ^
<stdin>:197:32: note: scanning from here
 s_load_b64 s[0:1], s[4:5], 0x0
                               ^
<stdin>:198:2: note: possible intended match here
 v_mov_b16_e32 v0.l, 0x3c24
 ^
/b/ml-opt-rel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rcp.bf16.ll:80:21: error: SDAG-TRUE16-NEXT: expected string not found in input
; SDAG-TRUE16-NEXT: v_dual_mov_b32 v0, 0 :: v_dual_mov_b32 v1, 0x7fc0
                    ^
<stdin>:291:32: note: scanning from here
 s_load_b64 s[0:1], s[4:5], 0x0
                               ^
<stdin>:292:2: note: possible intended match here
 v_mov_b16_e32 v0.l, 0x7fc0
 ^

Input file: <stdin>
Check file: /b/ml-opt-rel-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/llvm.amdgcn.rcp.bf16.ll

-dump-input=help explains the following input dump.

Input was:
<<<<<<
           .
           .
           .
          98:  .globl rcp_bf16_constant_4 ; -- Begin function rcp_bf16_constant_4 
          99:  .p2align 8 
         100:  .type rcp_bf16_constant_4, at function 
         101: rcp_bf16_constant_4: ; @rcp_bf16_constant_4 
...

```

</details>

https://github.com/llvm/llvm-project/pull/154501