ASan: add new interface functions - __asan_(un)poison_stack_memory. Calls to these functions are inserted by the instrumentation pass in use-after-scope mode

llvm-svn: 169201

llvm-svn: 168508

[asan/tsan] do not use __WORDSIZE macro, as it is glibc-private thing. Instead, define our own SANITIZER_WORDSIZE

llvm-svn: 168424

[ASan] use raw Exit() to kill the program in case ASan finds an error while reporting the error in the same thread

llvm-svn: 168308

[ASan] Revert r168040 and r168043 and take a cleaner solution suggested by Kostya: return the known frame name for fake stack instead of looking it up.

llvm-svn: 168046

the beginning of the fake frame when reporting an use-after-return error.
Fixes http://code.google.com/p/address-sanitizer/issues/detail?id=126

llvm-svn: 168040

[asan] change the asan output slightly to avoid user confusion: a) add ':' after 'AddressSanitizer' and b) changed 'crashed' to 'SEGV' 

llvm-svn: 165932

[ASan] Change __asan_set_on_error_callback to weak overridable __asan_on_error, so that ASan would call the latter even if it finds the error early (i.e. during module initialization)

llvm-svn: 165008

[ASan] Move printing descriptions of heap addresses in error report from allocator internals to asan_report.cc.

llvm-svn: 164112

[ASan] increase sleep time if ASan finds two bugs simultaneously to make sure full error report is printed

llvm-svn: 164018

In code review, it looked like these warnings do not actually fire with
modern Clang or GCC. However, the pragma is flat out rejected by GCC in
many configurations (-Werror) so it ended up causing more problems than
it fixed.

Daniel agreed with reverting this, but I think didn't have time to get
to it.

llvm-svn: 163583

llvm-svn: 163411

[ASan] Add print_full_thread_history runtime option (on by default) that prints all full thread creation paths for threads involved in ASan error report

llvm-svn: 163200

llvm-svn: 162754

llvm-svn: 162747

llvm-svn: 162746

[asan] run-time part of the initialization order checker. Patch by Reid Watson with some bits from kcc. The sub-pass is off by default for now. On simple tests it works fine. 

llvm-svn: 162278

[ASan] Add __asan_set_on_error_callback() interface function that allows user to set a callback to be called right when ASan detects an error

llvm-svn: 161754

[ASan] If ASan finds second error report, wait for some time and die (instead of running in a busy loop) to make sure ASan won't hang if it finds error while reporting an error in the same thread

llvm-svn: 161749

llvm-svn: 161666

llvm-svn: 161576

[ASan] Small fix to please tests on Windows, where stack unwinding using provided pc/bp works too bad

llvm-svn: 161575

llvm-svn: 161574

llvm-svn: 161572

llvm-svn: 161571

llvm-svn: 161570

llvm-svn: 161569

[ASan] Create new files asan_report.{h,cc} as a preparation for refactoring of ASan error reporting code.
Currently ASan reports many kinds of errors, and the code that actually prints error messages can
be found inside allocator, OS-specific files, interceptors code etc.

An example of maintenance troubles this situation causes:
There is currently an ASan interface function that registers
callback which should take the char buffer with error report printed by ASan.
This function is now broken, as one has to insert callback calls to all the places in
ASan code where the error reports are printed, surprisingly it is not only
"__asan_report_error" function...

llvm-svn: 161568