[llvm] 3ed5c19 - [sancov] Add -diff and -union options to compute set difference and union of sancov files (#171191)
via llvm-commits
llvm-commits at lists.llvm.org
Tue Dec 16 02:11:43 PST 2025
Author: Manuel Carrasco
Date: 2025-12-16T10:11:39Z
New Revision: 3ed5c191c78359a2cb17528fc8724410415f47d1
URL: https://github.com/llvm/llvm-project/commit/3ed5c191c78359a2cb17528fc8724410415f47d1
DIFF: https://github.com/llvm/llvm-project/commit/3ed5c191c78359a2cb17528fc8724410415f47d1.diff
LOG: [sancov] Add -diff and -union options to compute set difference and union of sancov files (#171191)
Add two new actions for `.sancov` files:
`-diff`: Computes the difference between two `.sancov` files (A - B) and
writes the result to a new file. Requires exactly two inputs and an
`--output` option. The output preserves the binary format (magic number
and bitness) of the first input.
`-union`: Computes the union of one or more `.sancov` files and writes
the result to a new file. Requires at least one input and an `--output`
option. The output preserves the binary format of the first input.
A warning is shown if input files differ in bitness (32-bit vs 64-bit),
but the operation proceeds using the first file’s bitness.
Added:
llvm/test/tools/sancov/Inputs/dummy-32bits.0.sancov
llvm/test/tools/sancov/diff-different-bitness.test
llvm/test/tools/sancov/diff-different-files.test
llvm/test/tools/sancov/diff-same-file.test
llvm/test/tools/sancov/union-different-bitness.test
llvm/test/tools/sancov/union-different-files.test
llvm/test/tools/sancov/union-same-file.test
Modified:
clang/docs/SanitizerCoverage.rst
llvm/tools/sancov/Opts.td
llvm/tools/sancov/sancov.cpp
Removed:
################################################################################
diff --git a/clang/docs/SanitizerCoverage.rst b/clang/docs/SanitizerCoverage.rst
index 23720e542e4e9..4ab2d09366f4f 100644
--- a/clang/docs/SanitizerCoverage.rst
+++ b/clang/docs/SanitizerCoverage.rst
@@ -563,18 +563,39 @@ Sancov matches these files using module names and binaries file names.
.. code-block:: console
- USAGE: sancov [options] <action> (<binary file>|<.sancov file>)...
-
- Action (required)
- -print - Print coverage addresses
- -covered-functions - Print all covered functions.
- -not-covered-functions - Print all not covered functions.
- -symbolize - Symbolizes the report.
-
- Options
- -blocklist=<string> - Blocklist file (sanitizer blocklist format).
- -demangle - Print demangled function name.
- -strip_path_prefix=<string> - Strip this prefix from file paths in reports
+ USAGE: sancov [options] <action> <binary files...> <.sancov files...> <.symcov files...>
+
+ Action (required):
+ -covered-functions Print all covered funcions.
+ -
diff Compute
diff erence between two sancov files (A - B) and write to the new output sancov file
+ -html-report REMOVED. Use -symbolize & coverage-report-server.py.
+ -merge Merges reports.
+ -not-covered-functions Print all not covered funcions.
+ -print-coverage-pcs Print coverage instrumentation points addresses.
+ -print-coverage-stats Print coverage statistics.
+ -print Print coverage addresses
+ -symbolize Produces a symbolized JSON report from binary report.
+ -union Compute union of multiple sancov files and write to the new output sancov file
+
+ Generic Options:
+ -help Display this help
+ -h Alias for --help
+ -version Display the version
+ -v Alias for --version
+
+ OPTIONS:
+ -demangle=0 Alias for --no-demangle
+ -demangle Demangle function names
+ -ignorelist=<string> Ignorelist file (sanitizer ignorelist format)
+ -no-demangle Do not demangle function names
+ -no-skip-dead-files List dead source files in reports
+ -output=<string> Output file for
diff and union actions
+ -skip-dead-files=0 Alias for --no-skip-dead-files
+ -skip-dead-files Do not list dead source files in reports
+ -strip_path_prefix=<string>
+ Strip this prefix from files paths in reports
+ -use_default_ignorelist=0
+ Alias for --no-use_default_ignore_list
Coverage Reports
diff --git a/llvm/test/tools/sancov/Inputs/dummy-32bits.0.sancov b/llvm/test/tools/sancov/Inputs/dummy-32bits.0.sancov
new file mode 100644
index 0000000000000..fee9a125ef08b
Binary files /dev/null and b/llvm/test/tools/sancov/Inputs/dummy-32bits.0.sancov
diff er
diff --git a/llvm/test/tools/sancov/
diff -
diff erent-bitness.test b/llvm/test/tools/sancov/
diff -
diff erent-bitness.test
new file mode 100644
index 0000000000000..506db06a552a5
--- /dev/null
+++ b/llvm/test/tools/sancov/
diff -
diff erent-bitness.test
@@ -0,0 +1,6 @@
+REQUIRES: x86-registered-target && host-byteorder-little-endian
+RUN: rm -f %t.out.sancov
+RUN: sancov -
diff --output=%t.out.sancov %p/Inputs/dummy-32bits.0.sancov %p/Inputs/test-linux_x86_64.0.sancov 2>&1 | FileCheck %s
+
+CHECK: Input files have
diff erent bitness
+CHECK: 64-bit addresses will be truncated to 32 bits. This may result in data loss.
diff --git a/llvm/test/tools/sancov/
diff -
diff erent-files.test b/llvm/test/tools/sancov/
diff -
diff erent-files.test
new file mode 100644
index 0000000000000..db46593099e74
--- /dev/null
+++ b/llvm/test/tools/sancov/
diff -
diff erent-files.test
@@ -0,0 +1,7 @@
+REQUIRES: x86-registered-target && host-byteorder-little-endian
+RUN: rm -f %t.out.sancov
+RUN: sancov -
diff --output=%t.out.sancov %p/Inputs/test-linux_x86_64.1.sancov %p/Inputs/test-linux_x86_64.0.sancov
+RUN: sancov -print %t.out.sancov | FileCheck %s
+
+CHECK: 0x4e14c2
+CHECK: 0x4e178c
diff --git a/llvm/test/tools/sancov/
diff -same-file.test b/llvm/test/tools/sancov/
diff -same-file.test
new file mode 100644
index 0000000000000..fd3b5de79d9a6
--- /dev/null
+++ b/llvm/test/tools/sancov/
diff -same-file.test
@@ -0,0 +1,6 @@
+REQUIRES: x86-registered-target && host-byteorder-little-endian
+RUN: rm -f %t.out.sancov
+RUN: sancov -
diff --output=%t.out.sancov %p/Inputs/test-linux_x86_64.0.sancov %p/Inputs/test-linux_x86_64.0.sancov
+RUN: sancov -print %t.out.sancov | FileCheck %s --allow-empty --check-prefix=EMPTY
+
+EMPTY-NOT: {{.}}
diff --git a/llvm/test/tools/sancov/union-
diff erent-bitness.test b/llvm/test/tools/sancov/union-
diff erent-bitness.test
new file mode 100644
index 0000000000000..fa37451a59b71
--- /dev/null
+++ b/llvm/test/tools/sancov/union-
diff erent-bitness.test
@@ -0,0 +1,6 @@
+REQUIRES: x86-registered-target && host-byteorder-little-endian
+RUN: rm -f %t.out.sancov
+RUN: sancov -union --output=%t.out.sancov %p/Inputs/dummy-32bits.0.sancov %p/Inputs/test-linux_x86_64.0.sancov 2>&1 | FileCheck %s
+
+CHECK: Input files have
diff erent bitness
+CHECK: 64-bit addresses will be truncated to 32 bits. This may result in data loss.
diff --git a/llvm/test/tools/sancov/union-
diff erent-files.test b/llvm/test/tools/sancov/union-
diff erent-files.test
new file mode 100644
index 0000000000000..47dafc97ad3c2
--- /dev/null
+++ b/llvm/test/tools/sancov/union-
diff erent-files.test
@@ -0,0 +1,11 @@
+REQUIRES: x86-registered-target && host-byteorder-little-endian
+RUN: rm -f %t.out.sancov
+RUN: sancov -union --output=%t.out.sancov %p/Inputs/test-linux_x86_64.1.sancov %p/Inputs/test-linux_x86_64.0.sancov
+RUN: sancov -print %t.out.sancov | FileCheck %s
+
+CHECK: 0x4e1472
+CHECK: 0x4e14c2
+CHECK: 0x4e1520
+CHECK: 0x4e1553
+CHECK: 0x4e1586
+CHECK: 0x4e178c
diff --git a/llvm/test/tools/sancov/union-same-file.test b/llvm/test/tools/sancov/union-same-file.test
new file mode 100644
index 0000000000000..e7991d24fb7ac
--- /dev/null
+++ b/llvm/test/tools/sancov/union-same-file.test
@@ -0,0 +1,10 @@
+REQUIRES: x86-registered-target && host-byteorder-little-endian
+RUN: rm -f %t.out.sancov
+RUN: sancov -union --output=%t.out.sancov %p/Inputs/test-linux_x86_64.0.sancov %p/Inputs/test-linux_x86_64.0.sancov
+RUN: sancov -print %t.out.sancov | FileCheck %s
+
+CHECK: 0x4e132b
+CHECK: 0x4e1472
+CHECK: 0x4e1520
+CHECK: 0x4e1553
+CHECK: 0x4e1586
diff --git a/llvm/tools/sancov/Opts.td b/llvm/tools/sancov/Opts.td
index 2e8af81b2a40d..411d08a033050 100644
--- a/llvm/tools/sancov/Opts.td
+++ b/llvm/tools/sancov/Opts.td
@@ -22,6 +22,10 @@ def : Flag<["-"], "v">, Alias<version>, HelpText<"Alias for --version">, Group<g
def action_grp : OptionGroup<"Action">, HelpText<"Action (required)">;
def print : F<"print", "Print coverage addresses">,
Group<action_grp>;
+def
diff : F<"
diff ", "Compute
diff erence between two sancov files (A - B) and write to the new output sancov file">,
+ Group<action_grp>;
+def union_files : F<"union", "Compute union of multiple sancov files and write to the new output sancov file">,
+ Group<action_grp>;
def printCoveragePcs : F<"print-coverage-pcs", "Print coverage instrumentation points addresses.">,
Group<action_grp>;
def coveredFunctions : F<"covered-functions", "Print all covered funcions.">,
@@ -56,3 +60,7 @@ defm stripPathPrefix
defm ignorelist
: Eq<"ignorelist", "Ignorelist file (sanitizer ignorelist format)">,
MetaVarName<"<string>">;
+
+defm output
+ : Eq<"output", "Output file for
diff and union actions">,
+ MetaVarName<"<string>">;
diff --git a/llvm/tools/sancov/sancov.cpp b/llvm/tools/sancov/sancov.cpp
index a0585fad024c7..f2e71f97dda34 100644
--- a/llvm/tools/sancov/sancov.cpp
+++ b/llvm/tools/sancov/sancov.cpp
@@ -92,13 +92,15 @@ class SancovOptTable : public opt::GenericOptTable {
enum ActionType {
CoveredFunctionsAction,
+ DiffAction,
HtmlReportAction,
MergeAction,
NotCoveredFunctionsAction,
PrintAction,
PrintCovPointsAction,
StatsAction,
- SymbolizeAction
+ SymbolizeAction,
+ UnionAction
};
static ActionType Action;
@@ -108,6 +110,7 @@ static bool ClSkipDeadFiles;
static bool ClUseDefaultIgnorelist;
static std::string ClStripPathPrefix;
static std::string ClIgnorelist;
+static std::string ClOutputFile;
static const char *const DefaultIgnorelistStr = "fun:__sanitizer_.*\n"
"src:/usr/include/.*\n"
@@ -132,14 +135,19 @@ static const Regex SymcovFileRegex(".*\\.symcov");
// Contents of .sancov file: list of coverage point addresses that were
// executed.
struct RawCoverage {
- explicit RawCoverage(std::unique_ptr<std::set<uint64_t>> Addrs)
- : Addrs(std::move(Addrs)) {}
+ explicit RawCoverage(std::unique_ptr<std::set<uint64_t>> Addrs,
+ FileHeader Header)
+ : Addrs(std::move(Addrs)), Header(Header) {}
// Read binary .sancov file.
static ErrorOr<std::unique_ptr<RawCoverage>>
read(const std::string &FileName);
+ // Write binary .sancov file.
+ static void write(const std::string &FileName, const RawCoverage &Coverage);
+
std::unique_ptr<std::set<uint64_t>> Addrs;
+ FileHeader Header;
};
// Coverage point has an opaque Id and corresponds to multiple source locations.
@@ -264,7 +272,7 @@ RawCoverage::read(const std::string &FileName) {
// to compactify the data.
Addrs->erase(0);
- return std::make_unique<RawCoverage>(std::move(Addrs));
+ return std::make_unique<RawCoverage>(std::move(Addrs), *Header);
}
// Print coverage addresses.
@@ -277,6 +285,34 @@ raw_ostream &operator<<(raw_ostream &OS, const RawCoverage &CoverageData) {
return OS;
}
+// Write coverage addresses in binary format.
+void RawCoverage::write(const std::string &FileName,
+ const RawCoverage &Coverage) {
+ std::error_code EC;
+ raw_fd_ostream OS(FileName, EC, sys::fs::OF_None);
+ failIfError(EC);
+
+ OS.write(reinterpret_cast<const char *>(&Coverage.Header),
+ sizeof(Coverage.Header));
+
+ switch (Coverage.Header.Bitness) {
+ case Bitness64:
+ for (auto Addr : *Coverage.Addrs) {
+ uint64_t Addr64 = Addr;
+ OS.write(reinterpret_cast<const char *>(&Addr64), sizeof(Addr64));
+ }
+ break;
+ case Bitness32:
+ for (auto Addr : *Coverage.Addrs) {
+ uint32_t Addr32 = static_cast<uint32_t>(Addr);
+ OS.write(reinterpret_cast<const char *>(&Addr32), sizeof(Addr32));
+ }
+ break;
+ default:
+ fail("Unsupported bitness: " + std::to_string(Coverage.Header.Bitness));
+ }
+}
+
static raw_ostream &operator<<(raw_ostream &OS, const CoverageStats &Stats) {
OS << "all-edges: " << Stats.AllPoints << "\n";
OS << "cov-edges: " << Stats.CovPoints << "\n";
@@ -1015,6 +1051,88 @@ static void readAndPrintRawCoverage(const std::vector<std::string> &FileNames,
}
}
+static const char *bitnessToString(uint32_t Bitness) {
+ switch (Bitness) {
+ case Bitness64:
+ return "64-bit";
+ case Bitness32:
+ return "32-bit";
+ default:
+ fail("Unsupported bitness: " + std::to_string(Bitness));
+ return nullptr;
+ }
+}
+
+// Warn if two file headers have
diff erent bitness.
+static void warnIfDifferentBitness(const FileHeader &Header1,
+ const FileHeader &Header2,
+ const std::string &File1Desc,
+ const std::string &File2Desc) {
+ if (Header1.Bitness != Header2.Bitness) {
+ errs() << "WARNING: Input files have
diff erent bitness (" << File1Desc
+ << ": " << bitnessToString(Header1.Bitness) << ", " << File2Desc
+ << ": " << bitnessToString(Header2.Bitness)
+ << "). Using bitness from " << File1Desc << ".\n";
+
+ if (Header1.Bitness == Bitness32 && Header2.Bitness == Bitness64) {
+ errs() << "WARNING: 64-bit addresses will be truncated to 32 bits. "
+ << "This may result in data loss.\n";
+ }
+ }
+}
+
+// Compute
diff erence between two coverage files (A - B) and write to output
+// file.
+static void
diff RawCoverage(const std::string &FileA, const std::string &FileB,
+ const std::string &OutputFile) {
+ auto CovA = RawCoverage::read(FileA);
+ failIfError(CovA);
+
+ auto CovB = RawCoverage::read(FileB);
+ failIfError(CovB);
+
+ const FileHeader &HeaderA = CovA.get()->Header;
+ const FileHeader &HeaderB = CovB.get()->Header;
+
+ warnIfDifferentBitness(HeaderA, HeaderB, FileA, FileB);
+
+ // Compute A - B
+ auto DiffAddrs = std::make_unique<std::set<uint64_t>>();
+ std::set_
diff erence(CovA.get()->Addrs->begin(), CovA.get()->Addrs->end(),
+ CovB.get()->Addrs->begin(), CovB.get()->Addrs->end(),
+ std::inserter(*DiffAddrs, DiffAddrs->end()));
+
+ RawCoverage DiffCov(std::move(DiffAddrs), HeaderA);
+ RawCoverage::write(OutputFile, DiffCov);
+}
+
+// Compute union of multiple coverage files and write to output file.
+static void unionRawCoverage(const std::vector<std::string> &InputFiles,
+ const std::string &OutputFile) {
+ failIf(InputFiles.empty(), "union action requires at least one input file");
+
+ // Read the first file to get the header and initial coverage
+ auto UnionCov = RawCoverage::read(InputFiles[0]);
+ failIfError(UnionCov);
+
+ const FileHeader &UnionHeader = UnionCov.get()->Header;
+
+ for (size_t I = 1; I < InputFiles.size(); ++I) {
+ auto Cov = RawCoverage::read(InputFiles[I]);
+ failIfError(Cov);
+
+ const FileHeader &CurHeader = Cov.get()->Header;
+
+ warnIfDifferentBitness(UnionHeader, CurHeader, InputFiles[0],
+ InputFiles[I]);
+
+ UnionCov.get()->Addrs->insert(Cov.get()->Addrs->begin(),
+ Cov.get()->Addrs->end());
+ }
+
+ RawCoverage::write(OutputFile, *UnionCov.get());
+}
+
static std::unique_ptr<SymbolizedCoverage>
merge(const std::vector<std::unique_ptr<SymbolizedCoverage>> &Coverages) {
if (Coverages.empty())
@@ -1153,6 +1271,9 @@ static void parseArgs(int Argc, char **Argv) {
" Depending on chosen action the tool expects
diff erent input files:\n"
" -print-coverage-pcs - coverage-instrumented binary files\n"
" -print-coverage - .sancov files\n"
+ " -
diff - two .sancov files & --output option\n"
+ " -union - one or more .sancov files & --output "
+ "option\n"
" <other actions> - .sancov files & corresponding binary "
"files, .symcov files\n");
std::exit(0);
@@ -1176,6 +1297,12 @@ static void parseArgs(int Argc, char **Argv) {
case OPT_print:
Action = ActionType::PrintAction;
break;
+ case OPT_
diff :
+ Action = ActionType::DiffAction;
+ break;
+ case OPT_union_files:
+ Action = ActionType::UnionAction;
+ break;
case OPT_printCoveragePcs:
Action = ActionType::PrintCovPointsAction;
break;
@@ -1209,6 +1336,7 @@ static void parseArgs(int Argc, char **Argv) {
ClStripPathPrefix = Args.getLastArgValue(OPT_stripPathPrefix_EQ);
ClIgnorelist = Args.getLastArgValue(OPT_ignorelist_EQ);
+ ClOutputFile = Args.getLastArgValue(OPT_output_EQ);
}
int sancov_main(int Argc, char **Argv, const llvm::ToolContext &) {
@@ -1223,6 +1351,26 @@ int sancov_main(int Argc, char **Argv, const llvm::ToolContext &) {
readAndPrintRawCoverage(ClInputFiles, outs());
return 0;
}
+ if (Action == DiffAction) {
+ // -
diff requires exactly 2 input files and an output file.
+ failIf(ClInputFiles.size() != 2,
+ "
diff action requires exactly 2 input sancov files");
+ failIf(
+ ClOutputFile.empty(),
+ "
diff action requires --output option to specify output sancov file");
+
diff RawCoverage(ClInputFiles[0], ClInputFiles[1], ClOutputFile);
+ return 0;
+ }
+ if (Action == UnionAction) {
+ // -union requires at least 1 input file and an output file.
+ failIf(ClInputFiles.empty(),
+ "union action requires at least one input sancov file");
+ failIf(
+ ClOutputFile.empty(),
+ "union action requires --output option to specify output sancov file");
+ unionRawCoverage(ClInputFiles, ClOutputFile);
+ return 0;
+ }
if (Action == PrintCovPointsAction) {
// -print-coverage-points doesn't need coverage files.
for (const std::string &ObjFile : ClInputFiles) {
@@ -1257,6 +1405,8 @@ int sancov_main(int Argc, char **Argv, const llvm::ToolContext &) {
errs() << "-html-report option is removed: "
"use -symbolize & coverage-report-server.py instead\n";
return 1;
+ case DiffAction:
+ case UnionAction:
case PrintAction:
case PrintCovPointsAction:
llvm_unreachable("unsupported action");
More information about the llvm-commits
mailing list