[llvm] r291025 - AMDGPU: Remove unneccessary intermediate vector

[Richard Smith via llvm-commits]

On 4 January 2017 at 14:54, Matt Arsenault via llvm-commits <
llvm-commits at lists.llvm.org> wrote:

> Author: arsenm
> Date: Wed Jan  4 16:54:10 2017
> New Revision: 291025
>
> URL: http://llvm.org/viewvc/llvm-project?rev=291025&view=rev
> Log:
> AMDGPU: Remove unneccessary intermediate vector
>

This is causing buildbot failures:


http://lab.llvm.org:8011/builders/clang-x86_64-linux-selfhost-modules/builds/2056

(These were masked by other breakage until now.) Reverted in r291067.


> Modified:
>     llvm/trunk/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp
>
> Modified: llvm/trunk/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp
> URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/
> AMDGPU/AsmParser/AMDGPUAsmParser.cpp?rev=291025&r1=291024&r2=291025&
> view=diff
> ============================================================
> ==================
> --- llvm/trunk/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp (original)
> +++ llvm/trunk/lib/Target/AMDGPU/AsmParser/AMDGPUAsmParser.cpp Wed Jan  4
> 16:54:10 2017
> @@ -822,6 +822,7 @@ public:
>    bool isForcedVOP3() const { return ForcedEncodingSize == 64; }
>    bool isForcedDPP() const { return ForcedDPP; }
>    bool isForcedSDWA() const { return ForcedSDWA; }
> +  ArrayRef<unsigned> getMatchedVariants() const;
>
>    std::unique_ptr<AMDGPUOperand> parseRegister();
>    bool ParseRegister(unsigned &RegNo, SMLoc &StartLoc, SMLoc &EndLoc)
> override;
> @@ -1630,31 +1631,34 @@ unsigned AMDGPUAsmParser::checkTargetMat
>    return Match_Success;
>  }
>
> +// What asm variants we should check
> +ArrayRef<unsigned> AMDGPUAsmParser::getMatchedVariants() const {
> +  if (getForcedEncodingSize() == 32)
> +    return {AMDGPUAsmVariants::DEFAULT};
> +
> +  if (isForcedVOP3())
> +    return {AMDGPUAsmVariants::VOP3};
> +
> +  if (isForcedSDWA())
> +    return {AMDGPUAsmVariants::SDWA};
> +
> +  if (isForcedDPP())
> +    return {AMDGPUAsmVariants::DPP};
> +
> +  return {AMDGPUAsmVariants::DEFAULT,
> +          AMDGPUAsmVariants::VOP3,
> +          AMDGPUAsmVariants::SDWA,
> +          AMDGPUAsmVariants::DPP};
>

You're returning ArrayRefs referring to local arrays here. Perhaps you
could add static local variables to hold the contents of these arrays
instead?


> +}
> +
>  bool AMDGPUAsmParser::MatchAndEmitInstruction(SMLoc IDLoc, unsigned
> &Opcode,
>                                                OperandVector &Operands,
>                                                MCStreamer &Out,
>                                                uint64_t &ErrorInfo,
>                                                bool MatchingInlineAsm) {
> -  // What asm variants we should check
> -  std::vector<unsigned> MatchedVariants;
> -  if (getForcedEncodingSize() == 32) {
> -    MatchedVariants = {AMDGPUAsmVariants::DEFAULT};
> -  } else if (isForcedVOP3()) {
> -    MatchedVariants = {AMDGPUAsmVariants::VOP3};
> -  } else if (isForcedSDWA()) {
> -    MatchedVariants = {AMDGPUAsmVariants::SDWA};
> -  } else if (isForcedDPP()) {
> -    MatchedVariants = {AMDGPUAsmVariants::DPP};
> -  } else {
> -    MatchedVariants = {AMDGPUAsmVariants::DEFAULT,
> -                       AMDGPUAsmVariants::VOP3,
> -                       AMDGPUAsmVariants::SDWA,
> -                       AMDGPUAsmVariants::DPP};
> -  }
> -
>    MCInst Inst;
>    unsigned Result = Match_Success;
> -  for (auto Variant : MatchedVariants) {
> +  for (auto Variant : getMatchedVariants()) {
>      uint64_t EI;
>      auto R = MatchInstructionImpl(Operands, Inst, EI, MatchingInlineAsm,
>                                    Variant);
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20170104/41845f9e/attachment.html>