MemorySanitizer (MSan) is a tool that detects use of uninitialized memory.
MSan is supported on x86_64 Linux. Additional info on the tool is available at http://clang.llvm.org/docs/MemorySanitizer.html.
MSan in Chromium is unlikely to be usable on systems other than Ubuntu Precise/Trusty - please see the note on instrumented libraries below.
MSan bots are running on chromium.memory.fyi, client.webrtc and chromium.webkit. There are also two LKGR builders for ClusterFuzz: no origins, chained origins (see below for explanation). V8 deployment is ongoing.
You can grab fresh Chrome binaries for Linux built with MSan here.
To set up an MSan build in GN:
gclient runhooks gn args out/msan
In the resulting editor, set the build variables:
is_msan = true
is_debug = false # Release build.
(Note: if you intend to run the Blink web tests with the MSan-instrumented content_shell binary, you must use out/Release instead of out/msan, because otherwise the test expectations will not apply correctly.)
(In older versions of Chromium you also had to explicitly set "use_prebuilt_instrumented_libraries = true". This is now the default if is_msan is set and can no longer be overridden.)
MSan requires using Instrumented system libraries. Note that instrumented libraries are supported on Ubuntu Precise/Trusty only. More information: instrumented-libraries-for-dynamic-tools.
The following flags are implied by is_
msan=true (i.e. you don't have to set
Some common flags may break a MSAN build. For example, don't set "dcheck_always_on = true" when using MSAN.
If you are trying to reproduce a test run from the Linux ChromiumOS MSan Tests build, other GN args may also be needed. You can look for them via your test run page, under the section "lookup builder GN args". Add all of them, except the goma_dir.
Running on gLinux locally
testing/xvfb.py out/msan/unit_tests --gtest_filter="<your test filter>" Running on Ubuntu Trusty
Run the resulting binaries as usual. Pipe both stderr and stdout through
tools/valgrind/asan/asan_symbolize.py to get symbolized reports:
./out/msan/browser_tests |& tools/valgrind/asan/asan_symbolize.py
Chrome must not use hardware OpenGL when running under MSan. This is because libgl.so is not instrumented and will crash the GPU process. SwANGLE can be used as a software OpenGL implementation, although it is extremely slow. There are several ways to proceed:
- --disable-gpu: This forces Chrome to use the software path for compositing and raster. WebGL will still work using SwANGLE.
--use-gl=angle --use-angle=swiftshader: This switches Chrome to use SwANGLE for compositing, (maybe) raster and WebGL.
--use-gl=angle --use-angle=swiftshader --disable-gl-drawing-for-tests: Use this if you don't care about the actual pixel output. This exercises the default code paths, however expensive SwANGLE calls are replaced with stubs (i.e. nothing actually gets drawn to the screen).
If neither flag is specified, Chrome will fall back to the first option after the GPU process crashes with an MSan report.
MSan allows the user to trade off execution speed for the amount of information
provided in reports. This is controlled by the GN/GYP flag
msan_track_origins=0: MSan will tell you where the uninitialized value was used, but not where it came from. This is the fastest mode.
msan_track_origins=1(deprecated): MSan will also tell you where the uninitialized value was originally allocated (e.g. which malloc() call, or which local variable). This mode is not significantly faster than
msan_track_origins=2, and its use is discouraged. We do not provide pre-built instrumented libraries for this mode.
msan_track_origins=2(default): MSan will also report the chain of stores that copied the uninitialized value to its final location. If there are more than 7 stores in the chain, only the first 7 will be reported. Note that compilation time may increase in this mode.
MSan does not support suppressions. This is an intentional design choice.
We have a blocklist file which is applied at compile time, and is used mainly to compensate for tool issues. Blocklist rules do not work the way suppression rules do - rather than suppressing reports with matching stack traces, they change the way MSan instrumentation is applied to the matched function. In addition, blocklist changes require a full clobber to take efffect. Please refrain from making changes to the blocklist file unless you know what you are doing.
Note also that instrumented libraries use separate blocklist files.
- Please keep in mind that simply reading/copying uninitialized memory will not cause an MSan report. Even simple arithmetic computations will work. To produce a report, the code has to do something significant with the uninitialized value, e.g. branch on it, pass it to a libc function or use it to index an array.
- When you examine a stack trace in an MSan report, all third-party
libraries you see in it (with the exception of libc and its
components) should reside under
out/Release/instrumented_libraries. If you see a DSO under a system-wide directory (e.g. /
lib/), then the report is likely bogus and should be fixed by simply adding that DSO to the list of instrumented libraries (please file a bug under
Stability-Memory-MemorySanitizerand/or ping eugenis@).
- Inline assembly is also likely to cause bogus reports. Consequently, assembly-optimized third-party code (such as libjpeg_turbo, libvpx, libyuv, ffmpeg) will have those optimizations disabled in MSan builds.
- If you're trying to debug a V8-related issue, please keep in mind that MSan builds run V8 in ARM64 mode, as explained below.
MSan reserves a separate memory region ("shadow memory") in which it tracks the
status of application memory. The correspondence between the two is bit-to-bit:
if the shadow bit is set to 1, the corresponding bit in the application memory
is considered "poisoned" (i.e. uninitialized). The header file
<sanitizer/msan_interface.h> declares interface functions which can be used to
examine and manipulate the shadow state without changing the application memory,
which comes in handy when debugging MSan reports.
Print the complete shadow state of a range of application memory, including the origins of all uninitialized values, if any. (Note: though initializedness is tracked on bit level, origins have 4-byte granularity.)
void __msan_print_shadow(const volatile void *x, size_t size);
The following prints a more minimalistic report which shows only the shadow memory:
void __msan_dump_shadow(const volatile void *x, size_t size);
To mark a memory range as fully uninitialized/initialized:
void __msan_poison(const volatile void *a, size_t size); void __msan_unpoison(const volatile void *a, size_t size); void __msan_unpoison_string(const volatile char *a);
The following forces an MSan check, i.e. if any bits in the memory range are uninitialized the call will crash with an MSan report.
void __msan_check_mem_is_initialized(const volatile void *x, size_t size);
This milder check returns the offset of the first (at least partially) poisoned byte in the range, or -1 if the whole range is good:
intptr_t __msan_test_shadow(const volatile void *x, size_t size);
Hint: sometimes to reduce log spam it makes sense to query
__msan_test_shadow() before calling
The complete interface can be found in
Functions such as
__msan_unpoison() can also be used to permanently annotate
your code for MSan, but please CC eugenis@ if you intend to do so.
Because MSan only supports Ubuntu Precise/Trusty and not Rodete, the ClusterFuzz reproduce tool cannot reproduce bugs found using MSan (on Rodete).
If you are on Rodete, you can try to reproduce them manually using docker to run MSan by following these instructions.