[llvm] [AMDGPU] Split Dpp8FI and Dpp16FI operands (PR #82379)

[Jay Foad via llvm-commits]

================
@@ -474,13 +467,11 @@ DecodeStatus AMDGPUDisassembler::getInstruction(MCInst &MI, uint64_t &Size,
                         MI, DecW, Address, CS);
       if (Res && convertDPP8Inst(MI) == MCDisassembler::Success)
----------------
jayfoad wrote:

Now `convertDPP8Inst` always returns success. I have patches which will clean this up some more.

https://github.com/llvm/llvm-project/pull/82379