[PATCH] D84643: AMDGPU: Don't assume call targets are registers

[Matt Arsenault via Phabricator via llvm-commits]

arsenm created this revision.
arsenm added reviewers: kerbowa, foad, rampitec, madhur13490.
Herald added subscribers: hiraditya, t-tye, tpr, dstuttard, yaxunl, nhaehnle, wdng, jvesely, kzhuravl.
Herald added a project: LLVM.

GlobalISel let through a call to null, which would then fold into the
source operand like any other inline immediate. The SelectionDAG
lowering deletes calls to null and undef as a workaround from before
calls were supported. We should probably drop the special handling
case in the DAG lowering now, since the middle end optimizers delete
null calls anyway.


https://reviews.llvm.org/D84643

Files:
  llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp
  llvm/test/CodeGen/AMDGPU/call-constant.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D84643.280868.patch
Type: text/x-patch
Size: 3818 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20200727/d9b6426f/attachment.bin>