On Sun, Apr 13, 2014 at 10:05:08AM -0400, Sam Varshavchik wrote: > Suvayu Ali writes: > > >On Sun, Apr 13, 2014 at 08:38:11AM -0500, Ranjan Maitra wrote: > >> On Sun, 13 Apr 2014 09:15:04 -0400 Rahul Sundaram <metherid@xxxxxxxxx> > >> wrote: > >> > >> > On Sun, Apr 13, 2014 at 6:23 AM, Timothy Murphy wrote: > >> > > >> > > Roger wrote: > >> > > > >> > > > It happened. It was known for years. > >> > > > >> > > Everything I have seen says it has been known for about 1 week. > >> > > > >> > > Incidentally, I am no programmer but I would have thought > >> > > it would be relatively simple to set up a test > >> > > to see if a "malloc"-ed space could be transgressed. > >> > > > >> > > >> > Not in this case. openssl uses a custom malloc > >> > > >> > >> So, a valgrind -tool=memcheck --leak-check=yes --show-reachable=yes > >> --track-fds=yes --track-origins=yes would not have helped? > > > >AFAIU this is not a memory leak; it is a buffer overflow: lack of bounds > >checking. I do not think valgrind (or any other tool) can help with > >that. Feel free to correct me if I am wrong. > > For this particular vulnerability, it is not technically a buffer overflow. > The vulnerable code merely reads past the end of an allocated buffer. A > buffer overflow gets exploited by writing past the allocated buffer. > > If you want to classify this, the best classification here would be an > information leak, and not a buffer overflow. Indeed! I wasn't accurate there. Thanks for pointing it out. > > The chances of valgrind catching this exploit are quite small. The only > possible way I see valgrind could've complained would be if the original > buffer that was allocated for the TLS ping message got padded by malloc() by > a non-significant amount, and valgrind would've complained about reading > uninitialized memory. > > And ven in that case, valgrind() tends to complain about uninitialized > memory only if its value is used to branch on. There's lots of code that > ends up reading uninitialized memory without resulting in undefined > behavior, for one reason or another. valgrind() doesn't typically complain > reading a few uninitialized bytes after a range that's been legally > malloc()ed, unless its value is used for branching, or something. There is a nice analysis by a cloudflare engineer here: <https://blog.cloudflare.com/answering-the-critical-question-can-you-get-private-ssl-keys-using-heartbleed> Although the conclusion of that article about the likelihood of such an information leak was proven wrong by the results of their challenge, I think the analysis still seems very valid. Thanks to the other posters for their responses. Quite a nice read. Cheers, -- Suvayu Open source is the future. It sets us free. -- users mailing list users@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org
