[llvm] [AMDGPU][True16][CodeGen] update waitcnt for true16 (PR #128927)

[Ivan Kosarev via llvm-commits]

================
@@ -748,27 +748,33 @@ RegInterval WaitcntBrackets::getRegInterval(const MachineInstr *MI,
 
   RegInterval Result;
 
-  unsigned Reg = TRI->getEncodingValue(AMDGPU::getMCReg(Op.getReg(), *ST)) &
-                 AMDGPU::HWEncoding::REG_IDX_MASK;
+  MCRegister MCReg = AMDGPU::getMCReg(Op.getReg(), *ST);
+  unsigned RegIdx = TRI->getHWRegIndex(MCReg);
+  assert(isUInt<8>(RegIdx));
+  unsigned Reg = (RegIdx << 1) | (AMDGPU::isHi16Reg(MCReg, *TRI) ? 1 : 0);
----------------
kosarev wrote:

Looks putting this under `useRealTrue16Insts()` can save time iterating through the larger ranges where 16-bit registers are not employed? Could be a separate patch?

https://github.com/llvm/llvm-project/pull/128927