[llvm] [AMDGPU] Bugfix for scaled MFMA parsing FP literals (PR #142493)

[Shilei Tian via llvm-commits]

================
@@ -8826,6 +8826,7 @@ void AMDGPUAsmParser::cvtScaledMFMA(MCInst &Inst,
   OptionalImmIndexMap OptionalIdx;
   unsigned Opc = Inst.getOpcode();
   unsigned I = 1;
+  int CbszInsIdx = AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::cbsz);
----------------
shiltian wrote:

nit:
```suggestion
  int CbszOpdx = AMDGPU::getNamedOperandIdx(Opc, AMDGPU::OpName::cbsz);
```

https://github.com/llvm/llvm-project/pull/142493