[PATCH] D158186: AMDGPU: Don't error for uniform dynamic alloca on DAG path

[Matt Arsenault via Phabricator via llvm-commits]

arsenm created this revision.
arsenm added reviewers: AMDGPU, scott.linder, cdevadas, yassingh, pravinjagtap.
Herald added subscribers: foad, kerbowa, hiraditya, tpr, dstuttard, yaxunl, jvesely, kzhuravl.
Herald added a project: All.
arsenm requested review of this revision.
Herald added a subscriber: wdng.
Herald added a project: LLVM.

This is more consistent with the globalisel handling. This still won't
work correctly in a function (it might work inside kernels).

      

Also add some tests to cover restore of SP (which is currently broken,
in most of the cases).


https://reviews.llvm.org/D158186

Files:
  llvm/lib/Target/AMDGPU/SIISelLowering.cpp
  llvm/test/CodeGen/AMDGPU/dynamic_stackalloc.ll
  llvm/test/CodeGen/AMDGPU/dynamic_stackalloc.r600.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D158186.551135.patch
Type: text/x-patch
Size: 44613 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20230817/d9795a31/attachment.bin>