[llvm] [AMDGPU][True16][CodeGen] optimize codegen for mad-mix in true16 (PR #124995)

[Matt Arsenault via llvm-commits]

================
@@ -5878,6 +5878,14 @@ AMDGPUInstructionSelector::selectVOP3PMadMixModsImpl(MachineOperand &Root,
       CheckAbsNeg();
     }
 
+    // Since we looked through FPEXT and removed it, we must also remove
+    // G_TRUNC. G_TRUNC to 16-bits would have a destination in RC VGPR_16, which
+    // is not compatible with MadMix instructions
+    Register PeekSrc = Src;
+    if (mi_match(PeekSrc, *MRI, m_GTrunc(m_Reg(PeekSrc))) &&
+        MRI->getType(PeekSrc).getSizeInBits() == 32)
----------------
arsenm wrote:

Can you add a mir test which stresses this 

https://github.com/llvm/llvm-project/pull/124995