Improve wild-free detection

[GoogleCodeExporter]

On a wild free, ASan usually crashes in random parts of the allocator. Ex.:
    #0 0xb69c44ff in bool __sanitizer::atomic_compare_exchange_strong<__sanitizer::atomic_uint8_t>(__sanitizer::atomic_uint8_t volatile*, __sanitizer::atomic_uint8_t::Type*, __sanitizer::atomic_uint8_t::Type, __sanitizer::memory_order) lib/sanitizer_common/sanitizer_atomic_clang.h:81
    #1 0xb69c44ff in __asan::Allocator::AtomicallySetQuarantineFlag(__asan::AsanChunk*, void*, __sanitizer::BufferedStackTrace*) [asan_rtl]
    #2 0xb69c44ff in __asan::Allocator::Deallocate(void*, unsigned long, __sanitizer::BufferedStackTrace*, __asan::AllocType) [asan_rtl]
    #3 0xb69c44ff in __asan::asan_free(void*, __sanitizer::BufferedStackTrace*, __asan::AllocType) [asan_rtl]

We can do slighly better.
 * For the small allocator, we always know the valid mapped range for the size class. A fast check.
 * For the large allocator, we can touch a byte in the memory region that is being freed, crash, and detect this situation in the SEGV handler either by looking at the crashing code address, or by setting a global (thread-local) flag immediately before touching memory and clearing it immediately after.

Original issue reported on code.google.com by euge...@google.com on 17 Jul 2015 at 5:03

[GoogleCodeExporter]

Adding Project:AddressSanitizer as part of GitHub migration.

Original comment by ramosian.glider@gmail.com on 30 Jul 2015 at 9:06

• Added labels: ProjectAddressSanitizer