[PATCH] D123611: [AMDGPU][MC][GFX10] Assembler fails to detect unsupported 64bit DPP opcodes

[Dmitry Preobrazhensky via Phabricator via llvm-commits]

dp created this revision.
dp added reviewers: rampitec, foad.
Herald added subscribers: hsmhsm, kerbowa, hiraditya, t-tye, tpr, dstuttard, yaxunl, jvesely, kzhuravl, arsenm.
Herald added a project: All.
dp requested review of this revision.
Herald added subscribers: llvm-commits, wdng.
Herald added a project: LLVM.

64-bit DPP opcodes are supported on GFX90A and GFX940 only. However assembler fails to identify these opcodes as unsupported when they are used on other GPUs.

The reason for this issue is contamination of matcher tables with disabled opcodes.

For a detailed description of the issue, see bug 54690 <https://github.com/llvm/llvm-project/issues/54690>.


https://reviews.llvm.org/D123611

Files:
  llvm/lib/Target/AMDGPU/SIInstrInfo.td
  llvm/lib/Target/AMDGPU/VOP1Instructions.td
  llvm/lib/Target/AMDGPU/VOP2Instructions.td
  llvm/test/MC/AMDGPU/gfx10_unsupported.s

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D123611.422229.patch
Type: text/x-patch
Size: 10299 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20220412/7d109401/attachment.bin>