[PATCH] D130763: [AMDGPU] Account for VData size increase from tfe bit for image instructions

[Mirko Brkusanin via Phabricator via llvm-commits]

mbrkusanin created this revision.
mbrkusanin added reviewers: foad, arsenm.
mbrkusanin added a project: LLVM.
Herald added subscribers: kosarev, kerbowa, hiraditya, t-tye, tpr, dstuttard, yaxunl, jvesely, kzhuravl.
Herald added a project: All.
mbrkusanin requested review of this revision.
Herald added a subscriber: wdng.

TFE bit effectively increases size of VData by 32bit so we need _V* variants in
tablegen to account for that.

Note that this change for cmpswap also renames _V1_V*_ into _V2_V*_ and _V2_V*_
into _V4_V*_ for VReg_64 and VReg_128 respectively.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D130763

Files:
  llvm/lib/Target/AMDGPU/MIMGInstructions.td
  llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.image.atomic.dim.mir
  llvm/test/CodeGen/AMDGPU/release-vgprs.mir
  llvm/test/MC/AMDGPU/gfx10_asm_mimg.s
  llvm/test/MC/AMDGPU/gfx11_asm_mimg.s
  llvm/test/MC/Disassembler/AMDGPU/gfx10_mimg.txt
  llvm/test/MC/Disassembler/AMDGPU/gfx11_mimg.txt
  llvm/test/MC/Disassembler/AMDGPU/mimg_vi.txt

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D130763.448596.patch
Type: text/x-patch
Size: 28461 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20220729/f87a1aaa/attachment-0001.bin>