c9db838267
[asan] enable LeakSanitizer (LSan) by default in asan. This only affects Linux x86_64. LSan has been used in various projects for more than half a year and we now consider it quite stable to be on by default.
...
llvm-svn: 206896
2014-04-22 13:56:56 +00:00
6e647c13e4
[ASan] Fixed the expectation for Darwin memset test.
...
Added a comment about keeping the Linux/Darwin test versions in sync.
llvm-svn: 206881
2014-04-22 10:53:57 +00:00
26c4e96455
[sanitizer] enable check_printf by default for asan and tsan. We've been using this flag for a while and it seems to be stable enough for the rest of the world.
...
llvm-svn: 206872
2014-04-22 09:17:08 +00:00
a7f9071f25
[sanitizer] Intercept tsearch.
...
llvm-svn: 206755
2014-04-21 14:21:51 +00:00
5e52f3a142
tsan: deflake test
...
ocasionally it fails with a slightly different report:
WARNING: ThreadSanitizer: signal handler spoils errno (pid=3674)
#0 MyHandler(int, siginfo*, void*)
...
#4 __tsan_free_hook
#5 main signal_errno.cc:40
llvm-svn: 206754
2014-04-21 14:21:24 +00:00
38bb53b2c5
[asan] add a run-time flag detect_container_overflow=true/false
...
llvm-svn: 206753
2014-04-21 14:18:45 +00:00
632751539e
[asan] add __asan_memset and friends
...
llvm-svn: 206748
2014-04-21 11:58:25 +00:00
74f8d65ae4
[asan] add one more mode to run asan tests with '-with-calls', i.e. using -mllvm -asan-instrumentation-with-call-threshold=0. This is very unfortunate since it effectively doubles the number of gunit asan tests that need to be built and run, but as long as we need out-lined calls (PR17409) I see no other way to test this workaround
...
llvm-svn: 206742
2014-04-21 10:31:35 +00:00
3d2629c514
[asan] implement __asan_loadN/__asan_storeN for out-lined asan checks
...
llvm-svn: 206733
2014-04-21 07:09:01 +00:00
561c4db707
[asan] Reenable tests that should pass since PR19207 is fixed.
...
llvm-svn: 206593
2014-04-18 13:24:03 +00:00
474011d55d
[msan] Add missing quotes.
...
llvm-svn: 206589
2014-04-18 13:03:54 +00:00
191ebd874f
[msan] Run msan_test in the new with-calls mode.
...
llvm-svn: 206587
2014-04-18 12:19:28 +00:00
83477cb93b
[msan] Missing declarations for the new interface functions.
...
llvm-svn: 206586
2014-04-18 12:18:00 +00:00
8f41674719
[msan] Add new MSan callbacks for instrumentation-with-calls mode.
...
llvm-svn: 206584
2014-04-18 12:15:24 +00:00
10d4b14a16
tsan: add benchmark that allows to investigate shadow memory consumption
...
llvm-svn: 206578
2014-04-18 10:37:39 +00:00
da1c510ea6
[ASan] Change AddressSanitizer.LoadStoreCallbacks to use asan_malloc and asan_free.
...
Interceptors don't really work on OSX in asan_noinst_test.cc (this is more or less intentional),
so one shouldn't call intercepted functions in this test -- added a comment about this.
llvm-svn: 206477
2014-04-17 17:29:07 +00:00
664b1e2f9b
[asan] fix a bug in outlined asan checks and in the corresponding test (ouch)
...
llvm-svn: 206470
2014-04-17 14:38:25 +00:00
aefbec9dfd
Move pthread_cond_* interceptors from sanitizer_common with all the ugly hacks to TSan
...
llvm-svn: 206423
2014-04-16 23:06:46 +00:00
ae401ca5e7
[asan] add __asan_load1/__asan_store1/... callbacks to asan-rt; together with -mllvm -asan-instrumentation-with-call-threshold=N this will be a workaround for PR17409
...
llvm-svn: 206387
2014-04-16 13:52:28 +00:00
498b18a01e
[asan] try to harden __tls_get_addr interceptor against signals (no good test still); update the comment in a test.
...
llvm-svn: 206367
2014-04-16 09:07:27 +00:00
d4b1b2068e
[asan] fix the alloctor code to not use opaque data structure, which was larger than needed. This was a leftover of the allocator1=>allocator2 migration; thanks Yuri Gribov for reminding
...
llvm-svn: 206280
2014-04-15 13:30:32 +00:00
65a6c57013
[ASan] Replace a CHECK for mmap_limit_mb with a RAW_CHECK.
...
In the case of a CHECK failure the program tries to fork and launch llvm-symbolizer,
but hangs in mz_force_lock because one of the allocator locks is already acquired.
llvm-svn: 206274
2014-04-15 11:36:18 +00:00
b1232474b2
[asan] try fixing the mmap_limit_mb failure on buildbot (tests pass locally)
...
llvm-svn: 206262
2014-04-15 08:35:43 +00:00
8f80ccc635
sanitizer_atomic_msvc.h: add atomic_fetch_{add,sub} overloads for uintptr_t
...
This should hopefully unbreak the MSVC build after r206178.
llvm-svn: 206200
2014-04-14 17:43:49 +00:00
ec77f619bb
sanitizer_deadlock_detector.h: MSVC 2012 doesn't like compound literals
...
llvm-svn: 206199
2014-04-14 17:43:45 +00:00
81d80f7578
[asan] Fix mmap_limit_mb test to not use C++11.
...
llvm-svn: 206182
2014-04-14 15:37:02 +00:00
681016b2ea
[asan] Do not use lambda in sanitizer lit tests.
...
We may be building with a very old C++ library.
llvm-svn: 206180
2014-04-14 15:19:07 +00:00
9ce9a6cdf6
[msan] Intercept wcsftime().
...
llvm-svn: 206179
2014-04-14 14:59:42 +00:00
bcfbea6d4e
[asan] added internal flag mmap_limit_mb
...
llvm-svn: 206178
2014-04-14 14:51:01 +00:00
90527cb324
[asan] don't use bool in public interface, make sure the interface headers are usable in plain C
...
llvm-svn: 206160
2014-04-14 11:16:53 +00:00
f694ab1f18
[asan] provide better reports for cases where memcpy/etc get negative size parameter. Also fix a typo found by Tetsuo Kiso
...
llvm-svn: 206158
2014-04-14 09:50:52 +00:00
51f5b5fd73
tsan: serialize report printing in standalone deadlock detector
...
otherwise reports get intermixed
llvm-svn: 206043
2014-04-11 17:54:27 +00:00
b5eb8f0212
tsan: fix vector clocks
...
the new optimizations break when thread ids gets reused (clocks go backwards)
add the necessary tests as well
llvm-svn: 206035
2014-04-11 15:38:03 +00:00
fa1d56c31f
tsan: ignore interceptors from symbolizer
...
ocassionally we see races coming from symbolizer interceptors (e.g. memcmp)
reproducible only only on large complex programs
llvm-svn: 206034
2014-04-11 15:36:54 +00:00
77d8793fa6
[tsan] Fix output tests.
...
llvm-svn: 206031
2014-04-11 13:01:20 +00:00
652cbd7c15
[tsan] Fix false positive on xdr*_create.
...
llvm-svn: 206030
2014-04-11 12:29:24 +00:00
d326c81117
[msan] mprotect() more memory to detect user code outside of application range.
...
llvm-svn: 206028
2014-04-11 12:04:29 +00:00
7004b8c07c
Enable building of sanitizers on FreeBSD
...
llvm-svn: 205919
2014-04-09 18:45:12 +00:00
82185a2088
Declare _DYNAMIC and dl_phdr_info in asan_linux.cc on FreeBSD
...
llvm-svn: 205852
2014-04-09 13:37:19 +00:00
0ee8505800
Update arc config to new domain.
...
llvm-svn: 205702
2014-04-07 10:22:41 +00:00
24c8d92fec
[sanitizer] Intercept a subset of sunrpc interface (xdr_*).
...
llvm-svn: 205627
2014-04-04 14:51:23 +00:00
7dc3c0fb2b
[ASan] Explicitly specify -ldl/-lpthread in RUN-lines where needed
...
llvm-svn: 205623
2014-04-04 12:24:46 +00:00
1d66c4d5b7
tsan: improve error message in test
...
we've seen a flake on this test
next time it happens we will be able to gather some info
llvm-svn: 205619
2014-04-04 09:52:41 +00:00
f653cda269
[msan] Introduce MsanThread. Move thread-local allocator cache out of TLS.
...
This reduces .tbss from 109K down to almost nothing.
llvm-svn: 205618
2014-04-04 09:47:41 +00:00
9f20c9b17c
[asan] fix a leak in __tls_get_addr handler; introduce a run-time flag to disable this handler completely; remove a workaround for a bug fixed in glibc
...
llvm-svn: 205617
2014-04-04 09:10:58 +00:00
6d8f71f928
Revert r205613.
...
llvm-svn: 205614
2014-04-04 08:58:16 +00:00
fe82d5f849
[msan] Fix compilation of a disabled test.
...
llvm-svn: 205613
2014-04-04 08:39:50 +00:00
425314a65f
[TSan] Fix a rare deadlock on multithreaded fork.
...
If a multi-threaded program calls fork(), TSan ignores all memory accesses
in the child to prevent deadlocks in TSan runtime. This is OK, as child is
probably going to call exec() as soon as possible. However, a rare deadlocks
could be caused by ThreadIgnoreBegin() function itself.
ThreadIgnoreBegin() remembers the current stack trace and puts it into the
StackDepot to report a warning later if a thread exited with ignores enabled.
Using StackDepotPut in a child process is dangerous: it locks a mutex on
a slow path, which could be already locked in a parent process.
The fix is simple: just don't put current stack traces to StackDepot in
ThreadIgnoreBegin() and ThreadIgnoreSyncBegin() functions if we're
running after a multithreaded fork. We will not report any
"thread exited with ignores enabled" errors in this case anyway.
Submitting this without a testcase, as I believe the standalone reproducer
is pretty hard to construct.
llvm-svn: 205534
2014-04-03 12:51:26 +00:00
b4525218d4
[TSan] Fix a typo ThreadIgnoreSyncEnd. Found by inspection
...
llvm-svn: 205531
2014-04-03 12:28:16 +00:00
11ff0a26a4
[ASan] Fix incompatible runtimes check: don't iterate /proc/self/maps on every call to __asan_init
...
llvm-svn: 205418
2014-04-02 13:09:22 +00:00
Kostya Serebryany