[asan] asan_allocator2: fix the memalign handling (clear the first word of allocated chunk in case there is an old memalign magic)

llvm-svn: 172603

[asan] enable asan_allocator2 by default on Linux. Will enable it on other platforms and remove the old one later, after additional testing. The new allocator is much faster and uses less memory

llvm-svn: 172531

llvm-svn: 172526

Move my_rand() to the common header.

This lets us avoid the use of rand_r in sanitizer_common tests.
There is no rand_r on Android.

llvm-svn: 172421

llvm-svn: 172398

[asan] use the slow CFI-based unwinder when reporting an error. Still use the fast unwinder for malloc/free. Linux-x86-only for now.

llvm-svn: 172397

ASan: Disable alloc/dealloc-mismatch checker on Mac for now (it produces weird false positives on googletest)

llvm-svn: 172394

llvm-svn: 172392

llvm-svn: 172390

llvm-svn: 172380

llvm-svn: 172226

llvm-svn: 172210

llvm-svn: 172194

llvm-svn: 172193

llvm-svn: 172192

llvm-svn: 172184

llvm-svn: 172183

llvm-svn: 172059

[asan] asan_allocator2: do less work under the quarantine lock; make the strcasecmp test more resistant to the contents of unaddressable memory 

llvm-svn: 172048

llvm-svn: 172045

llvm-svn: 171223

[asan] add a new interface function __asan_describe_address, useful for running asan-ified binary under a debugger (e.g. gdb)

llvm-svn: 171217

llvm-svn: 171200

[asan] implement more strict checking for memset/etc parameters. Instead of checking the first and the last byte, we check the entire shadow region. This costs ~10 slowdown for the instrumented functions. Motivated by a nasty memset-buffer-overflow-by-140-bytes in chrome which was reported as a use-after-free or not at all

llvm-svn: 171198

llvm-svn: 171188

llvm-svn: 171181

llvm-svn: 171161

Define COMPILER_RT_CAN_EXECUTE_TESTS variable on platforms where we can produce working binaries and use it in build rules for sanitizers tests

llvm-svn: 171160

llvm-svn: 171147

[asan] relax asan/lit_tests/malloc_delete_mismatch.cc to make it pass on Mac 10.7 (where extra frames creep in between malloc and main)

llvm-svn: 171145

[asan/tsan] when unmapping a chunk of user memory, apply madvise(MADV_DONTNEED) to the corresponding chunk of shadow memory. Also update sanitizer_allocator64_testlib.cc

llvm-svn: 171144

llvm-svn: 171112

[asan] asan_allocator2: do not align the requested size to the redzone size (saves a bit more memory)

llvm-svn: 171111

[asan] asan_allocator2: implement adaptive redzones. Now with asan_allocator2 allocations <= 48 bytes have 16 byte redzone, allocations of 48-96 bytes -- have 32 bytes redzone, etc (max redzone is 2048). If ASAN_OPTIONS=redzone=n is set, it changes the minimal redzone size

llvm-svn: 171107

[asan] asan_allocator2: by default use the StackDepot to store the stack traces instead of storing them in the redzones

llvm-svn: 171099

[asan] asan_allocator2: store the user requested size in just 32 bits (larger sizes use metadata). This allows to put one more 32-bit field into the chunk header

llvm-svn: 171089

[asan] add a test for right OOB with special large sizes. Fix this test in asan_allocator2. More test tweaking for allocator2

llvm-svn: 171058

[asan] properly define ASAN_LOW_MEMORY. Due to a mistake it has been always true causing redzone and quarantine sizes to be smaller than it was some time before (and out of sync with the documentation). Also make one test less greedy

llvm-svn: 171052

llvm-svn: 171036

[asan] asan_allocator2 fix two asserts that happen on full chrome: a) memalign called with 0 size and large alignment and b) malloc called after TSD has been destructed

llvm-svn: 170900