Re: [PATCH 2/2] perf clang: Fixes for more recent LLVM/clang

["Fāng-ruì Sòng" <maskray@google.com>]

On Mon, Oct 11, 2021 at 4:35 PM Ian Rogers <irogers@google.com> wrote:
>
> On Mon, Oct 11, 2021 at 4:24 PM Ian Rogers <irogers@google.com> wrote:
> >
> > The parameters to two functions and the location of a variable have
> > changed in more recent LLVM/clang releases.
> >
> > Tested with LLVM 6, 8, 9, 10 and 11.
> >
> > Signed-off-by: Ian Rogers <irogers@google.com>
> > ---
> >  tools/perf/util/c++/clang.cpp | 20 +++++++++++++-------
> >  1 file changed, 13 insertions(+), 7 deletions(-)
> >
> > diff --git a/tools/perf/util/c++/clang.cpp b/tools/perf/util/c++/clang.cpp
> > index c8885dfa3667..756200cb80b1 100644
> > --- a/tools/perf/util/c++/clang.cpp
> > +++ b/tools/perf/util/c++/clang.cpp
> > @@ -44,7 +44,6 @@ createCompilerInvocation(llvm::opt::ArgStringList CFlags, StringRef& Path,
> >                 "-triple", "bpf-pc-linux",
> >                 "-fsyntax-only",
> >                 "-ferror-limit", "19",
> > -               "-fmessage-length", "127",
>
> Sorry, missed from the commit message. I removed this flag to format
> output down to 127 columns as with clang 11 I see in perf test:
>
> 58: builtin clang support                                           :
> 58.1: builtin clang compile C source to IR                          :
> --- start ---
> test child forked, pid 279307
> error: unknown argument: '-fmessage-length'
> 1 error generated.
> test child finished with -1
>
> The flag isn't necessary for correct operation.

Right, I changed the -cc1 option from Separate-form -fmessage-length
to -fmessage-length= in 2020-03.
-ferror-limit can be deleted as well.

> Thanks,
> Ian
>
> >                 "-O2",
> >                 "-nostdsysteminc",
> >                 "-nobuiltininc",
> > @@ -55,7 +54,11 @@ createCompilerInvocation(llvm::opt::ArgStringList CFlags, StringRef& Path,
> >                 "-x", "c"};
> >
> >         CCArgs.append(CFlags.begin(), CFlags.end());
> > -       CompilerInvocation *CI = tooling::newInvocation(&Diags, CCArgs);
> > +       CompilerInvocation *CI = tooling::newInvocation(&Diags, CCArgs
> > +#if CLANG_VERSION_MAJOR >= 11
> > +                                                        ,/*BinaryName=*/nullptr
> > +#endif
> > +                                                        );
> >
> >         FrontendOptions& Opts = CI->getFrontendOpts();
> >         Opts.Inputs.clear();
> > @@ -151,13 +154,16 @@ getBPFObjectFromModule(llvm::Module *Module)
> >
> >         legacy::PassManager PM;
> >         bool NotAdded;
> > -#if CLANG_VERSION_MAJOR < 7
> > -       NotAdded = TargetMachine->addPassesToEmitFile(PM, ostream,
> > -                                                     TargetMachine::CGFT_ObjectFile);
> > +       NotAdded = TargetMachine->addPassesToEmitFile(PM, ostream
> > +#if CLANG_VERSION_MAJOR >= 7
> > +                                                      , /*DwoOut=*/nullptr
> > +#endif
> > +#if CLANG_VERSION_MAJOR < 10
> > +                                                      , TargetMachine::CGFT_ObjectFile
> >  #else
> > -       NotAdded = TargetMachine->addPassesToEmitFile(PM, ostream, nullptr,
> > -                                                     TargetMachine::CGFT_ObjectFile);
> > +                                                      , llvm::CGFT_ObjectFile
> >  #endif
> > +                                                      );
> >         if (NotAdded) {
> >                 llvm::errs() << "TargetMachine can't emit a file of this type\n";
> >                 return std::unique_ptr<llvm::SmallVectorImpl<char>>(nullptr);
> > --
> > 2.33.0.882.g93a45727a2-goog
> >

Related, does it make sense to drop CLANG_VERSION_MAJOR<7 support?