[llvm] [AMDGPU] Support TFE D16 format buffer loads on gfx8/10/11/12 (PR #211465)

[Matt Arsenault via llvm-commits]

================
@@ -520,6 +520,9 @@ class GCNSubtarget final : public AMDGPUGenSubtargetInfo,
   // Has V_PK_MOV_B32 opcode
   bool hasPkMovB32() const { return HasGFX90AInsts; }
 
+  // Has a hardware encoding for TFE D16 format buffer loads.
+  bool hasBufferTFEFormatD16() const { return !HasGFX90AInsts; }
----------------
arsenm wrote:

Should move this to a feature but I suppose that can be separate 

https://github.com/llvm/llvm-project/pull/211465