| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
| |
Differential Revision: http://reviews.llvm.org/D5210
llvm-svn: 217309
|
|
|
|
|
|
| |
Differential Revision: http://reviews.llvm.org/D5212
llvm-svn: 217308
|
|
|
|
|
|
| |
Differential Revision: http://reviews.llvm.org/D5206
llvm-svn: 217307
|
|
|
|
|
|
| |
Differential Revision: http://reviews.llvm.org/D5205
llvm-svn: 217306
|
|
|
|
|
|
|
| |
We are interested in verifying that -gline-tables-only provides enough
debug information for verbose error reports and symbolized stack traces.
llvm-svn: 217284
|
|
|
|
|
|
| |
Bug: https://code.google.com/p/address-sanitizer/issues/detail?id=336
llvm-svn: 217228
|
|
|
|
|
|
| |
reproducer where this assertion fails...
llvm-svn: 217215
|
|
|
|
| |
llvm-svn: 217194
|
|
|
|
| |
llvm-svn: 217177
|
|
|
|
|
|
|
|
|
|
|
|
| |
The naming scheme we're using for counters in profile data shares a
prefix with some fixed names we use for the runtime, notably
__llvm_profile_data_begin and _end. Embarrassingly, this means a
function called begin() can't be instrumented.
This modifies the runtime names so as not to collide with the
instrumentation.
llvm-svn: 217166
|
|
|
|
|
|
| |
Looks like the second crash doesn't happen on Mac.
llvm-svn: 217165
|
|
|
|
|
|
| |
Differential Revision: http://reviews.llvm.org/D5171
llvm-svn: 217156
|
|
|
|
|
|
| |
Differential Revision: http://reviews.llvm.org/D5166
llvm-svn: 217148
|
|
|
|
|
|
| |
Missing <atomic> header.
llvm-svn: 217142
|
|
|
|
|
|
|
|
|
| |
Chained origins make plain memory stores async-signal-unsafe.
We already disable it inside signal handlers.
This change grabs all origin-related locks before fork() and releases
them after fork() to avoid a deadlock in the child process.
llvm-svn: 217140
|
|
|
|
|
|
|
| |
(previously ASan would just crash upon the second SEGV)
Other tools do not use this code yet.
llvm-svn: 217137
|
|
|
|
| |
llvm-svn: 217111
|
|
|
|
| |
llvm-svn: 217101
|
|
|
|
|
|
|
|
| |
http://reviews.llvm.org/D4692
Patch by Maxim Ostapenko!
llvm-svn: 217079
|
|
|
|
|
|
|
| |
The second part of the fix of
https://code.google.com/p/thread-sanitizer/issues/detail?id=71
llvm-svn: 217031
|
|
|
|
|
|
| |
It disables the per-stack limit.
llvm-svn: 217030
|
|
|
|
|
|
| |
Fixes issue https://code.google.com/p/thread-sanitizer/issues/detail?id=74
llvm-svn: 216906
|
|
|
|
|
|
|
|
|
|
|
| |
There interceptors do not seem to be strictly necessary for tsan.
But we see cases where the interceptors consume 70% of execution time.
Memory blocks passed to fgetgrent_r are "written to" by tsan several times.
First, there is some recursion (getgrnam_r calls fgetgrent_r), and each
function "writes to" the buffer. Then, the same memory is "written to"
twice, first as buf and then as pwbufp (both of them refer to the same addresses).
llvm-svn: 216904
|
|
|
|
|
|
|
| |
Fixes issue
https://code.google.com/p/thread-sanitizer/issues/detail?id=71
llvm-svn: 216903
|
|
|
|
|
|
| |
See http://reviews.llvm.org/D4794
llvm-svn: 216900
|
|
|
|
| |
llvm-svn: 216885
|
|
|
|
| |
llvm-svn: 216884
|
|
|
|
|
|
|
| |
Guard the source with #if SANITIZER_POSIX to fix compilation on Windows.
use it only in
llvm-svn: 216882
|
|
|
|
|
|
|
| |
Users expect synchronization between sigaction and arrival
of the signal. See the test for details.
llvm-svn: 216878
|
|
|
|
|
|
|
|
|
| |
sanitizer_unwind_posix.cc
This allows the slow unwinder to be used on OSX (it's disabled for now, see
https://code.google.com/p/address-sanitizer/issues/detail?id=137 for details)
llvm-svn: 216877
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Currently it fails with:
# Testing race detector.
race_windows_amd64.syso:gotsan.cc:(.text+0x4d89): undefined reference to `__imp_SymGetModuleBase64'
race_windows_amd64.syso:gotsan.cc:(.text+0x4d90): undefined reference to `__imp_SymFunctionTableAccess64'
race_windows_amd64.syso:gotsan.cc:(.text+0x4dd8): undefined reference to `__imp_StackWalk64'
x86_64-w64-mingw32/bin/ld.exe: race_windows_amd64.syso
: bad reloc address 0x10 in section `.data'
collect2.exe: error: ld returned 1 exit status
llvm-svn: 216875
|
|
|
|
| |
llvm-svn: 216705
|
|
|
|
|
|
| |
address is properly poisoned and return the cookie value. If not, return 0 to avoid infinite loop of DTORs (in case of use-after-free). Calls to this function will be inserted by clang (separate change)
llvm-svn: 216692
|
|
|
|
|
|
| |
on chromium (I suspect that LLVM looses nosanitizer metadata; no test yet)
llvm-svn: 216684
|
|
|
|
| |
llvm-svn: 216665
|
|
|
|
|
|
|
|
| |
asan_win_uar_thunk object file
This directive results in 'symbol already defined' when linking DLLs that use -MD
llvm-svn: 216663
|
|
|
|
| |
llvm-svn: 216661
|
|
|
|
|
|
|
| |
to make the tests green.
Slow stack unwinding is disabled in libsanitizer on Darwin now.
llvm-svn: 216657
|
|
|
|
| |
llvm-svn: 216544
|
|
|
|
| |
llvm-svn: 216542
|
|
|
|
| |
llvm-svn: 216534
|
|
|
|
|
|
|
|
|
|
| |
There is no reason to have different library names for shared and static
cases on linux. It also breaks Android where we install the shared asan-rt
library into the system and should keep the old name.
This change reverts most of r216380 limiting it to win32 targets only.
llvm-svn: 216533
|
|
|
|
| |
llvm-svn: 216494
|
|
|
|
| |
llvm-svn: 216459
|
|
|
|
| |
llvm-svn: 216454
|
|
|
|
|
|
| |
This warning makes it a bit hard to define strlen and wcslen interceptors in an elegant yet portable way
llvm-svn: 216450
|
|
|
|
| |
llvm-svn: 216448
|
|
|
|
|
|
|
| |
With this patch, "check-asan" passes all the tests with both MT and MD ASan RTL if you set COMPILER_RT_BUILD_SHARED_ASAN to ON
(PR20214)
llvm-svn: 216447
|
|
|
|
|
|
| |
functionality change intended
llvm-svn: 216407
|
|
|
|
| |
llvm-svn: 216387
|