[PATCH] D117482: AMDGPU: Don't clobber source register for V_SET_INACTIVE_*

[Ruiling, Song via Phabricator via llvm-commits]

ruiling created this revision.
ruiling added reviewers: foad, arsenm, nhaehnle, critson, piotr.
Herald added subscribers: kerbowa, hiraditya, t-tye, tpr, dstuttard, yaxunl, jvesely, kzhuravl, qcolombet.
ruiling requested review of this revision.
Herald added subscribers: llvm-commits, wdng.
Herald added a project: LLVM.

The WWM register has unmodeled register liveness, For v_set_inactive_*,
clobberring source register is dangerous because it will overwrite the
inactive lanes. When the source vgpr is dead at v_set_inactive_lane,
the inactive lanes may be not really dead. This may make common
optimizations doing wrong.

For example in a simple if-then cfg in Machine IR:
bb.if:

  %src =

bb.then:

  %src1 = COPY %src
  %dst = V_SET_INACTIVE %src1(tied-def 0), %inactive

bb.end

  ... = PHI [0, %bb.then] [%src, %bb.if]

The register coalescer will think it is safe to optimize "%src1 = COPY %src"
in bb.then. And at the same time, there is no interference for the PHI in
bb.end. The source and destination values of the PHI will be assigned
the same register. The single PHI register will be overwritten by the
v_set_inactive, then we would get wrong value in bb.end.

With this change, we will copy the content of the source register before
setting inactive lanes after register allocation. Yes, this will sacrifice
the WWM code generation a little, but I don't have any better idea to do things
correctly.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D117482

Files:
  llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
  llvm/lib/Target/AMDGPU/SIInstructions.td
  llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.set.inactive.ll
  llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
  llvm/test/CodeGen/AMDGPU/atomic_optimizations_local_pointer.ll
  llvm/test/CodeGen/AMDGPU/atomic_optimizations_pixelshader.ll
  llvm/test/CodeGen/AMDGPU/llvm.amdgcn.set.inactive.ll
  llvm/test/CodeGen/AMDGPU/set-inactive-wwm-overwrite.ll
  llvm/test/CodeGen/AMDGPU/wwm-reserved-spill.ll
  llvm/test/CodeGen/AMDGPU/wwm-reserved.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D117482.400529.patch
Type: text/x-patch
Size: 56738 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20220117/c41d29d2/attachment.bin>