[PATCH] D38906: AMDGPU/SI: Implement d16 support for buffer intrinsics

[Changpeng Fang via Phabricator via llvm-commits]

cfang added inline comments.


================
Comment at: lib/Target/AMDGPU/SIISelLowering.cpp:3565
+  if (SDValue Res = replaceResultType(Op, EquivLoadVT, DAG)) {
+    *Chain = Res.getOperand(0);
+    if (Unpacked) { // From v2i32/v4i32 back to v2f16/v4f16.
----------------
arsenm wrote:
> This is not the correct chain, this is the input chain, not the output result chain
I though "Res" is the output here, except that it has to be type-converted to the final result.

So how to get the  "output result chain" you mentioned? Thanks.


https://reviews.llvm.org/D38906