diff options
-rw-r--r-- | llvm/docs/CommandGuide/llvm-profdata.rst | 9 | ||||
-rw-r--r-- | llvm/test/tools/llvm-profdata/input-filenames.test | 17 | ||||
-rw-r--r-- | llvm/tools/llvm-profdata/llvm-profdata.cpp | 65 |
3 files changed, 85 insertions, 6 deletions
diff --git a/llvm/docs/CommandGuide/llvm-profdata.rst b/llvm/docs/CommandGuide/llvm-profdata.rst index 12f2771bd00..2742fd3d75d 100644 --- a/llvm/docs/CommandGuide/llvm-profdata.rst +++ b/llvm/docs/CommandGuide/llvm-profdata.rst @@ -44,6 +44,9 @@ interpreted as relatively more important than a shorter run. Depending on the nature of the training runs it may be useful to adjust the weight given to each input file by using the ``-weighted-input`` option. +Profiles passed in via ``-weighted-input``, ``-input-files``, or via positional +arguments are processed once for each time they are seen. + OPTIONS ^^^^^^^ @@ -65,6 +68,12 @@ OPTIONS Input files specified without using this option are assigned a default weight of 1. Examples are shown below. +.. option:: -input-files=path, -f=path + + Specify a file which contains a list of files to merge. The entries in this + file are newline-separated. Lines starting with '#' are skipped. Entries may + be of the form <filename> or <weight>,<filename>. + .. option:: -instr (default) Specify that the input profile is an instrumentation-based profile. diff --git a/llvm/test/tools/llvm-profdata/input-filenames.test b/llvm/test/tools/llvm-profdata/input-filenames.test new file mode 100644 index 00000000000..da0c47bf82a --- /dev/null +++ b/llvm/test/tools/llvm-profdata/input-filenames.test @@ -0,0 +1,17 @@ +# Create an input file. +RUN: echo '# comment 1' > %t.input +RUN: echo ' # comment 2' >> %t.input +RUN: echo 'bar' >> %t.input +RUN: echo ' baz' >> %t.input +RUN: echo "2,%t.weighted" >> %t.input + +# Create the weighted file, since these actually need to exist. +RUN: echo ' ' > %t.weighted + +RUN: llvm-profdata merge -f %t.input -dump-input-file-list -o /dev/null foo | FileCheck %s +RUN: llvm-profdata merge -input-files %t.input -dump-input-file-list -o /dev/null foo | FileCheck %s + +CHECK: 1,foo +CHECK-NEXT: 1,bar +CHECK-NEXT: 1,baz +CHECK-NEXT: 2,{{.*}}.weighted diff --git a/llvm/tools/llvm-profdata/llvm-profdata.cpp b/llvm/tools/llvm-profdata/llvm-profdata.cpp index 20a167226a3..064f36a8dc9 100644 --- a/llvm/tools/llvm-profdata/llvm-profdata.cpp +++ b/llvm/tools/llvm-profdata/llvm-profdata.cpp @@ -223,11 +223,53 @@ static WeightedFile parseWeightedFile(const StringRef &WeightedFilename) { return WeightedFile(FileName, Weight); } +static std::unique_ptr<MemoryBuffer> +getInputFilenamesFileBuf(const StringRef &InputFilenamesFile) { + if (InputFilenamesFile == "") + return {}; + + auto BufOrError = MemoryBuffer::getFileOrSTDIN(InputFilenamesFile); + if (!BufOrError) + exitWithErrorCode(BufOrError.getError(), InputFilenamesFile); + + return std::move(*BufOrError); +} + +static void parseInputFilenamesFile(MemoryBuffer *Buffer, + WeightedFileVector &WFV) { + if (!Buffer) + return; + + SmallVector<StringRef, 8> Entries; + StringRef Data = Buffer->getBuffer(); + Data.split(Entries, '\n', /*MaxSplit=*/-1, /*KeepEmpty=*/false); + for (const StringRef &FileWeightEntry : Entries) { + StringRef SanitizedEntry = FileWeightEntry.trim(" \t\v\f\r"); + // Skip comments. + if (SanitizedEntry.startswith("#")) + continue; + // If there's no comma, it's an unweighted profile. + else if (SanitizedEntry.find(',') == StringRef::npos) + WFV.emplace_back(SanitizedEntry, 1); + else + WFV.emplace_back(parseWeightedFile(SanitizedEntry)); + } +} + static int merge_main(int argc, const char *argv[]) { cl::list<std::string> InputFilenames(cl::Positional, cl::desc("<filename...>")); cl::list<std::string> WeightedInputFilenames("weighted-input", cl::desc("<weight>,<filename>")); + cl::opt<std::string> InputFilenamesFile( + "input-files", cl::init(""), + cl::desc("Path to file containing newline-separated " + "[<weight>,]<filename> entries")); + cl::alias InputFilenamesFileA("f", cl::desc("Alias for --input-files"), + cl::aliasopt(InputFilenamesFile)); + cl::opt<bool> DumpInputFileList( + "dump-input-file-list", cl::init(false), cl::Hidden, + cl::desc("Dump the list of input files and their weights, then exit")); cl::opt<std::string> OutputFilename("output", cl::value_desc("output"), cl::init("-"), cl::Required, cl::desc("Output file")); @@ -249,15 +291,26 @@ static int merge_main(int argc, const char *argv[]) { cl::ParseCommandLineOptions(argc, argv, "LLVM profile data merger\n"); - if (InputFilenames.empty() && WeightedInputFilenames.empty()) - exitWithError("No input files specified. See " + - sys::path::filename(argv[0]) + " -help"); - WeightedFileVector WeightedInputs; for (StringRef Filename : InputFilenames) - WeightedInputs.push_back(WeightedFile(Filename, 1)); + WeightedInputs.emplace_back(Filename, 1); for (StringRef WeightedFilename : WeightedInputFilenames) - WeightedInputs.push_back(parseWeightedFile(WeightedFilename)); + WeightedInputs.emplace_back(parseWeightedFile(WeightedFilename)); + + // Make sure that the file buffer stays alive for the duration of the + // weighted input vector's lifetime. + auto Buffer = getInputFilenamesFileBuf(InputFilenamesFile); + parseInputFilenamesFile(Buffer.get(), WeightedInputs); + + if (WeightedInputs.empty()) + exitWithError("No input files specified. See " + + sys::path::filename(argv[0]) + " -help"); + + if (DumpInputFileList) { + for (auto &WF : WeightedInputs) + outs() << WF.Weight << "," << WF.Filename << "\n"; + return 0; + } if (ProfileKind == instr) mergeInstrProfile(WeightedInputs, OutputFilename, OutputFormat, |