[llvm] [AMDGPU] Remove GFX12 encoding hack (PR #78702)

[via llvm-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-backend-amdgpu

Author: Jay Foad (jayfoad)

<details>
<summary>Changes</summary>

This is no longer needed now that we have implemented GFX12 encoding for
all instructions.


---
Full diff: https://github.com/llvm/llvm-project/pull/78702.diff


1 Files Affected:

- (modified) llvm/lib/Target/AMDGPU/SIInstrInfo.cpp (-6) 


``````````diff
diff --git a/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp b/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
index 48fdb803e9d7346..5a9401103da8228 100644
--- a/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
+++ b/llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
@@ -9136,12 +9136,6 @@ int SIInstrInfo::pseudoToMCOpcode(int Opcode) const {
 
   int MCOp = AMDGPU::getMCOpcode(Opcode, Gen);
 
-  // TODO-GFX12: Remove this.
-  // Hack to allow some GFX12 codegen tests to run before all the encodings are
-  // implemented.
-  if (MCOp == (uint16_t)-1 && Gen == SIEncodingFamily::GFX12)
-    MCOp = AMDGPU::getMCOpcode(Opcode, SIEncodingFamily::GFX11);
-
   // -1 means that Opcode is already a native instruction.
   if (MCOp == -1)
     return Opcode;

``````````

</details>


https://github.com/llvm/llvm-project/pull/78702