[llvm] [AMDGPU][MC] Allow null where 128b or larger dst reg is expected (PR #115200)

[Matt Arsenault via llvm-commits]

================
@@ -9720,6 +9720,12 @@ unsigned AMDGPUAsmParser::validateTargetOperandClass(MCParsedAsmOperand &Op,
     // The following code enables it for SReg_64 operands
     // used as source and destination. Remaining source
     // operands are handled in isInlinableImm.
+    //
+    // Additionally, allow null where destination of 128-bit or larger is
+    // expected.
----------------
arsenm wrote:

This comment can merge with the above. It doesn't need to mention specific sizes, it "may be used anywhere scalar sources can normally be used"

https://github.com/llvm/llvm-project/pull/115200