Commit Graph

4762 Commits

Author SHA1 Message Date
Alexey Samsonov
860a1abf1d [TSan] Refactor/simplify ReportLocation structure.
# Make DataInfo (describing a global) a member of ReportLocation
    to avoid unnecessary copies and allocations.
  # Introduce a constructor and a factory method, so that
    all structure users don't have to go to internal allocator directly.
  # Remove unused fields (file/line).

No functionality change.

llvm-svn: 221302
2014-11-04 22:07:57 +00:00
Alexey Samsonov
0b622f1d64 Correct the usage of DataInfo structure in TSan
llvm-svn: 221297
2014-11-04 21:26:56 +00:00
Alexey Samsonov
e3754634c0 [TSan] Make ReportStack contain __sanitizer::AddressInfo object.
AddressInfo contains the results of symbolization. Store this object
directly in the symbolized stack, instead of copying data around and
making unnecessary memory allocations.

No functionality change.

llvm-svn: 221294
2014-11-04 20:52:37 +00:00
Kostya Serebryany
c5bd9810cc [asan] [mips] changed ShadowOffset32 for systems having 16kb PageSize; patch by Kumar Sukhani
llvm-svn: 221288
2014-11-04 19:46:15 +00:00
Alexey Samsonov
26ca05ad96 [Sanitizer] Get rid of unnecessary allocations in StripModuleName. NFC.
llvm-svn: 221287
2014-11-04 19:34:29 +00:00
Alexey Samsonov
3931dca7ec [TSan] Keep original function and filename in ReportStack.
TSan used to do the following transformations with data obtained
from the symbolizer:
1) Strip "__interceptor_" prefix from function name.
2) Use "strip_path_prefix" runtime flag to strip filepath.

Now these transformations are performed right before the stack trace
is printed, and ReportStack structure contains original information.

This seems like a right thing to do - stripping is a detail of report
formatting implementation, and should belong there. We should, for
example, use original path to source file when we apply suppressions.

This change also make "strip_path_prefix" flag behavior in TSan
consistent with all the other sanitizers - now it should actually
match *the prefix* of path, not some substring. E.g. earlier TSan
would turn "/usr/lib/libfoo.so" into "libfoo.so" even if strip_path_prefix
was "/lib/".

Finally, strings obtained from symbolizer come from internal allocator,
and "stripping" them early by incrementing a "char*" ensures they can
never be properly deallocated, which is a bug.

llvm-svn: 221283
2014-11-04 18:41:38 +00:00
Kuba Brecka
6c22aec23c Use @rpath as LC_ID_DYLIB for ASan dylib on OS X
Change the LC_ID_DYLIB of ASan's dynamic libraries on OS X to be set to "@rpath/libclang_rt.asan_osx_dynamic.dylib" and similarly for iossim. Clang driver then sets the "-rpath" to be the real path to where clang currently has the dylib (because clang uses the relative path to its current executable). This means if you move the compiler or install the binary release, -fsanitize=address will link to the proper library.

Reviewed at http://reviews.llvm.org/D6018

llvm-svn: 221278
2014-11-04 17:34:50 +00:00
Alexey Samsonov
8f1632d5c1 [TSan] Don't strip binary/library name until the moment we print it.
This commit changes the place where TSan runtime turns full path
to binary or shared library into its basename
(/usr/foo/mybinary -> mybinary). Instead of doing it as early as possible
(when we obtained the full path from the symbolizer), we now do it as
late as possible (right before printing the error report).

This seems like a right thing to do - stripping to basename is a detail
of report formatting implementation, and should belong there. Also, we
might need the full path at some point - for example, to match the
suppressions.

llvm-svn: 221225
2014-11-04 01:55:20 +00:00
Alexey Samsonov
7acb67607b [TSan] Remove bogus unused global variables
llvm-svn: 221195
2014-11-03 22:46:13 +00:00
Alexey Samsonov
40733a8024 [TSan] Use StackTrace from sanitizer_common where applicable
Summary:
This change removes `__tsan::StackTrace` class. There are
now three alternatives:
  # Lightweight `__sanitizer::StackTrace`, which doesn't own a buffer
  of PCs. It is used in functions that need stack traces in read-only
  mode, and helps to prevent unnecessary allocations/copies (e.g.
  for StackTraces fetched from StackDepot).
  # `__sanitizer::BufferedStackTrace`, which stores buffer of PCs in
  a constant array. It is used in TraceHeader (non-Go version)
  # `__tsan::VarSizeStackTrace`, which owns buffer of PCs, dynamically
  allocated via TSan internal allocator.

Test Plan: compiler-rt test suite

Reviewers: dvyukov, kcc

Reviewed By: kcc

Subscribers: llvm-commits, kcc

Differential Revision: http://reviews.llvm.org/D6004

llvm-svn: 221194
2014-11-03 22:23:44 +00:00
Alexey Samsonov
46ecdeecc4 [TSan} Build Go version with -std=c++11
llvm-svn: 221192
2014-11-03 22:18:12 +00:00
Alexey Samsonov
e6bbcfa3df [TSan] Fix signed-compare warning in the unit test
llvm-svn: 221190
2014-11-03 22:17:39 +00:00
Kostya Serebryany
40aa4a26d1 [asan] increase the initial buffer size in caller-callee dumper
llvm-svn: 220991
2014-10-31 19:49:46 +00:00
Kostya Serebryany
b6eae0c2bc [asan] run-time part of the caller-callee coverage instrumentation
llvm-svn: 220975
2014-10-31 17:19:11 +00:00
Peter Collingbourne
b190e1b74b Update __dfsw_s{,n}printf custom functions for new calling convention.
Differential Revision: http://reviews.llvm.org/D6029

llvm-svn: 220907
2014-10-30 13:23:01 +00:00
Aaron Ballman
d98e197386 Reverting r220517; it seems this broke check-asan.
llvm-svn: 220869
2014-10-29 19:25:20 +00:00
Adrian Prantl
44ed3d04bf Unbreak the darwin build.
Patch by Ismail Pazarbasi!

llvm-svn: 220683
2014-10-27 16:20:59 +00:00
Viktor Kutuzov
c8f3435f0e [Tsan] Enable thread sanitizer tests on FreeBSD
Differential Revision: http://reviews.llvm.org/D5995

llvm-svn: 220676
2014-10-27 11:28:53 +00:00
Viktor Kutuzov
adf81dbcc6 [Tsan] Fix references to renamed variables in DPrintf()
Differential Revision: http://reviews.llvm.org/D5993

llvm-svn: 220674
2014-10-27 11:22:24 +00:00
Viktor Kutuzov
9deb007191 [Tsan] Make calloc() to not track allocated space unless thread is completely initialized
Differential Revision: http://reviews.llvm.org/D5992

llvm-svn: 220673
2014-10-27 11:19:08 +00:00
Viktor Kutuzov
a58fed9ab7 [Tsan] Fix guessing data range on FreeBSD
Differential Revision: http://reviews.llvm.org/D5990

llvm-svn: 220672
2014-10-27 11:14:24 +00:00
Alexey Samsonov
e853b4f2e4 [Sanitizer] Return code that calculates hash for stacktrace back to StackDepot implementation
llvm-svn: 220663
2014-10-27 03:10:27 +00:00
Hans Wennborg
f89a3f8623 Try to fix the Windows build after r220637
llvm-svn: 220643
2014-10-26 19:27:02 +00:00
Alexey Samsonov
3741ab82ba Change StackDepot interface to use StackTrace more extensively
llvm-svn: 220637
2014-10-26 06:23:07 +00:00
Alexey Samsonov
6e7af8156f Fixup deadlock sanitizer after r220635
llvm-svn: 220636
2014-10-26 05:43:17 +00:00
Alexey Samsonov
9c85927012 [Sanitizer] Make StackTrace a lightweight reference to array of PCs, and
introduce a BufferedStackTrace class, which owns this array.

Summary:
This change splits __sanitizer::StackTrace class into a lightweight
__sanitizer::StackTrace, which doesn't own array of PCs, and BufferedStackTrace,
which owns it. This would allow us to simplify the interface of StackDepot,
and eventually merge __sanitizer::StackTrace with __tsan::StackTrace.

Test Plan: regression test suite.

Reviewers: kcc, dvyukov

Reviewed By: dvyukov

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D5985

llvm-svn: 220635
2014-10-26 03:35:14 +00:00
Saleem Abdulrasool
d888169d75 builtins: avoid endian.h in favour of compiler builtins
Use the __BYTE_ORDER__ macro which has been available in GCC since 4.6 and is
available in clang as well.

llvm-svn: 220626
2014-10-25 20:54:51 +00:00
Alexey Samsonov
3ca50c34ec Add runtime flag 'symbolize_inline_frames' to disable symbolization of inlined frames done in llvm-symbolizer
llvm-svn: 220582
2014-10-24 18:34:43 +00:00
Dmitry Vyukov
e9a5f03716 tsan: support mmap(MAP_32BIT)
Allow user memory in the first TB of address space.
This also enabled non-pie binaries and freebsd.
Fixes issue:
https://code.google.com/p/thread-sanitizer/issues/detail?id=5

llvm-svn: 220571
2014-10-24 17:07:29 +00:00
Viktor Kutuzov
a836924be4 [Tsan] Do not intercept non-FreeBSD functions on FreeBSD
Differential Revision: http://reviews.llvm.org/D5858

llvm-svn: 220554
2014-10-24 10:55:03 +00:00
Viktor Kutuzov
f266601ce9 [Tsan] Fix the thread_name tests to build on FreeBSD
Differential Revision: http://reviews.llvm.org/D5855

llvm-svn: 220552
2014-10-24 09:27:34 +00:00
Viktor Kutuzov
6a9dec2aa2 [Tsan] Intercept pthread_set_name_np on FreeBSD
Differential Revision: http://reviews.llvm.org/D5932

llvm-svn: 220551
2014-10-24 09:23:06 +00:00
Viktor Kutuzov
9003426e2f [Tsan] Fix the signal_recursive.cc test to build on FreeBSD
Differential Revision: http://reviews.llvm.org/D5841

llvm-svn: 220550
2014-10-24 09:20:20 +00:00
Evgeniy Stepanov
7900202e14 [sanitizer] Add a bunch of sanity checks.
llvm-svn: 220528
2014-10-23 22:29:51 +00:00
Aaron Ballman
e927a17022 Cmake variables are global, which is why we would get crud like /machine:X86 in the list of compiler options for MSVC. Clear out the variable before attempting to enumerate arguments and set them.
llvm-svn: 220522
2014-10-23 22:13:52 +00:00
Aaron Ballman
29eec9a520 Appeasing an MSVC compile warning about "and"; NFC.
llvm-svn: 220519
2014-10-23 21:59:34 +00:00
Aaron Ballman
5f1fdd7d90 Do not set linker flags for MSVC; they are not the same thing as compiler flags. Note, this is already done in the add_compiler_rt_test test_suite function.
llvm-svn: 220517
2014-10-23 21:58:36 +00:00
Aaron Ballman
1d1f232041 Disabling some MSVC warnings that are of questionable value. Note, these are disabled by default in LLVM as well, so there is precedence.
C4146: 'unary minus operator applied to unsigned type, result still unsigned'
C4291: ''declaration' : no matching operator delete found; memory will not be freed if initialization throws an exception'
C4800: ''type' : forcing value to bool 'true' or 'false' (performance warning)'
llvm-svn: 220507
2014-10-23 20:39:58 +00:00
Aaron Ballman
7c7be2fa42 Change all of the build flags to switch /MD and /MDd into /MT, instead of just some of the targets. Fixes PR21241.
llvm-svn: 220506
2014-10-23 20:24:00 +00:00
Kostya Serebryany
2bd7631520 [tsan] add a stub of another dynamic annotation
llvm-svn: 220369
2014-10-22 05:02:20 +00:00
Evgeniy Stepanov
35eb265421 [msan] Handle param-tls overflow.
ParamTLS (shadow for function arguments) is of limited size. This change
makes all arguments that do not fit unpoisoned, and avoids writing
past the end of a TLS buffer.

llvm-svn: 220351
2014-10-22 00:12:40 +00:00
Evgeniy Stepanov
5efbbca818 [sanitizer] Fix build with _FILE_OFFSET_BITS=64.
Sanitizer source is not affected by _FILE_OFFSET_BITS in general,
but this one file must be built with 32-bit off_t. More details in the code.

llvm-svn: 220328
2014-10-21 21:08:13 +00:00
Viktor Kutuzov
cdcc46dc84 [Tsan] Add FreeBSD support to longjmp-related definitions
Differential Revision: http://reviews.llvm.org/D5857

llvm-svn: 220292
2014-10-21 13:38:05 +00:00
Viktor Kutuzov
6acdc50267 [Tsan] Fix sigaction_t to match system definition on FreeBSD
Differential Revision: http://reviews.llvm.org/D5856

llvm-svn: 220291
2014-10-21 13:31:05 +00:00
Alexey Samsonov
3eb69f5482 Remove unused local variable
llvm-svn: 220247
2014-10-20 23:58:08 +00:00
Dan Albert
c5d4bc7581 Fix checks for Android.
__ANDROID__ is a define that comes from the toolchain when building
for Android targets. ANDROID has a different meaning. ANDROID is
defined for _every_ Android build, including those done for host
modules. For host modules, we want to build the regular Linux
sanitizers and builtins, not the one for Android devices. This hasn't
been a problem until now because we only just started building the
sanitizers for the host.

llvm-svn: 220203
2014-10-20 15:35:01 +00:00
Kostya Serebryany
c9855da6d9 [asan] the run-time part of intra-object-overflow detector (-fsanitize-address-field-padding=1). Note that all of this is still experimental; don't use unless you are brave.
llvm-svn: 220013
2014-10-17 01:22:37 +00:00
Kostya Serebryany
87b70fc823 [asan] make sure coverage is dumped even if leaks are reported
llvm-svn: 219946
2014-10-16 18:59:07 +00:00
Renato Golin
6d41f7ae22 UBSAN stable runtime more generic with arm* targets
llvm-svn: 219924
2014-10-16 12:43:10 +00:00
Renato Golin
45abb8c969 Adds stable-runtime to ubsan to avoid broken ARM tests with asan
llvm-svn: 219906
2014-10-16 07:48:27 +00:00