msan and gcc ?

[Kostya Serebryany]

+Evgeniy

On Wed, Oct 1, 2014 at 9:24 AM, Xinliang David Li <davidxl@google.com> wrote:
> It may be helpful to document the following in msan's official page:
> 1) success stories (chrome land?)

The page https://code.google.com/p/memory-sanitizer/wiki/FoundBugs
may need some updates and it should be linked from
https://code.google.com/p/memory-sanitizer/wiki/MemorySanitizer

> 2) runtime overhead comparison with valgrind

Yep. We have the fresh data, need to publish in wiki.

--kcc


>
> David
>
> On Wed, Oct 1, 2014 at 9:07 AM, Kostya Serebryany <kcc@google.com> wrote:
>> [as text for real this time]
>> Sanitizer compiler module sizes in LLVM (in lines):
>>   1823 AddressSanitizer.cpp
>>   2780 MemorySanitizer.cpp
>>    564 ThreadSanitizer.cpp
>> Also note, that msan is the hardest to deploy among others sanitizers
>> because it requires to compile *everything*,
>> including libc++/libstdc++ and other system libs.
>> We've managed to do that for large projects like Chromium, LLVM, GCC,
>> and a few even larger ones,
>> and it was certainly worth it. Having msan in GCC would be nice, but
>> it is lots of work.
>>
>> --kcc
>>
>> On Wed, Oct 1, 2014 at 12:42 AM, Dmitry Vyukov <dvyukov@google.com> wrote:
>>> On Wed, Oct 1, 2014 at 11:30 AM, VandeVondele  Joost
>>> <joost.vandevondele@mat.ethz.ch> wrote:
>>>> Hi,
>>>>
>>>> I've noticed that gcc includes a msan_interface.h file, and I'm wondering if this implies that memory sanitizer is already part of gcc. If not, are there plans to port this useful looking tool to gcc during the current stage 1 ?
>>>
>>> Hi,
>>>
>>> No, msan is not part of gcc. And I am not aware of any plans to port
>>> msan to gcc.
>>> Note that msan's compiler pass is the most involved one as compared to
>>> asan/tsan.