[PATCH] D125498: [AMDGPU] gfx11 scalar alu instructions

Matt Arsenault via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Mon May 16 14:18:43 PDT 2022


arsenm added inline comments.


================
Comment at: llvm/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp:6368
+  StringRef FieldName = getTokenStr();
+  if (!skipToken(AsmToken::Identifier, "expected a field name") ||
+      !skipToken(AsmToken::LParen, "expected a left parenthesis"))
----------------
Missing tests for all of these parser errors


================
Comment at: llvm/lib/Target/AMDGPU/Disassembler/AMDGPUDisassembler.cpp:500
+    Res = tryDecodeInst<uint64_t>(DecoderTableGFX1132, MI, DW, Address);
+    if (Res) break;
+
----------------
Joe_Nash wrote:
> arsenm wrote:
> > Seems weird that we aren’t guarding each table lookup with subtarget checks 
> It would be possible, but may be redundant with the more fine grained checks done in decodeInstruction via AMDGPUGenDisassemblerTables.inc::checkDecoderPredicate. In any case, I don't think we want that change for this patch.
Right this is following what's already done but it looks suspicious


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D125498/new/

https://reviews.llvm.org/D125498



More information about the llvm-commits mailing list