[llvm] [GlobalISel] prevent G_UNMERGE_VALUES for vectors with different elements (PR #133335)

[LLVM Continuous Integration via llvm-commits]

llvm-ci wrote:

LLVM Buildbot has detected a new failure on builder `ml-opt-dev-x86-64` running on `ml-opt-dev-x86-64-b1` while building `llvm` at step 6 "test-build-unified-tree-check-all".

Full details are available at: https://lab.llvm.org/buildbot/#/builders/137/builds/20578

<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/GlobalISel/insertelement.ll' FAILED ********************
Exit Code: 1

Command Output (stderr):
--
/b/ml-opt-dev-x86-64-b1/build/bin/llc -global-isel -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx900 -verify-machineinstrs < /b/ml-opt-dev-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/GlobalISel/insertelement.ll | /b/ml-opt-dev-x86-64-b1/build/bin/FileCheck -check-prefix=GPRIDX /b/ml-opt-dev-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/GlobalISel/insertelement.ll # RUN: at line 2
+ /b/ml-opt-dev-x86-64-b1/build/bin/FileCheck -check-prefix=GPRIDX /b/ml-opt-dev-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/GlobalISel/insertelement.ll
+ /b/ml-opt-dev-x86-64-b1/build/bin/llc -global-isel -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx900 -verify-machineinstrs
/b/ml-opt-dev-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/GlobalISel/insertelement.ll:6514:16: error: GPRIDX-NEXT: is not on the line after the previous match
; GPRIDX-NEXT: s_load_dwordx16 s[12:27], s[8:9], 0x0
               ^
<stdin>:6451:2: note: 'next' match was here
 s_load_dwordx16 s[12:27], s[8:9], 0x0
 ^
<stdin>:6448:15: note: previous match ended here
; %bb.0: ; %bb
              ^
<stdin>:6449:1: note: non-matching line after previous match is here
 s_add_u32 flat_scratch_lo, s12, s17
^

Input file: <stdin>
Check file: /b/ml-opt-dev-x86-64-b1/llvm-project/llvm/test/CodeGen/AMDGPU/GlobalISel/insertelement.ll

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

Input was:
<<<<<<
           .
           .
           .
        6446:  runtime_loader_kernel_symbol = 0 
        6447:  .end_amd_kernel_code_t 
        6448: ; %bb.0: ; %bb 
        6449:  s_add_u32 flat_scratch_lo, s12, s17 
        6450:  s_addc_u32 flat_scratch_hi, s13, 0 
        6451:  s_load_dwordx16 s[12:27], s[8:9], 0x0 
next:6514      !~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~  error: match on wrong line
        6452:  s_load_dwordx2 s[0:1], s[8:9], 0x40 
        6453:  s_waitcnt lgkmcnt(0) 
        6454:  s_lshr_b32 s2, s12, 1 
        6455:  s_and_b32 s2, s2, 1 
        6456:  s_lshl_b32 s2, s2, 1 
           .
           .
           .
>>>>>>

--

...

```

</details>

https://github.com/llvm/llvm-project/pull/133335