[EMAIL PROTECTED] - Fri Jun 11 17:02:20 2004]: > On June 11, 2004 10:34 am, Richard Levitte - VMS Whacker wrote: > > rt> I see. The problem seems more about the model used by AEP though. > > rt> Ie. we could use CRYPTO_thread_id() instead of getpid() (because > > rt> unless CRYPTO_set_id_callback() is called, this devolves into > > rt> getpid() anyway). > > > > And that will not help one bit, as far as I can tell, since > > CRYPTO_thread_id() should return the thread id (which is the same as > > the pid with Linux Threads), so basically that solves the problem by > > reiterating the exact same problem... > > Of course this is not desirable long-term, but for now it would provide a > way to move forward and verify what's going on and how to fix it. In > particular, I've no idea about the AEP interface and what it is supposed > to do - eg. the leaks that were reported may or may not be cleaned up, so > getting the if(I've forked) handler to run at the run time would solve > this.
??????????????????????? Either you don't understand what I'm saying, or I don't understand the issue. I'm reiterating: on Linux with LinuxThreads, pthread_self() returns exactly the same as getpid(). The normal thing to do would be to have pthread_self() as callback for CRYPTO_thread_id(). And either way, replacing getpid(), which *normally* gives the same value for all threads in the same process, with something that's guaranteed to give different values for the different threads in the same process doesn't feel like a good test to see if the process has forked. The library is bound to think that each thread is a new process on all architectures if we use CRYPTO_thread_id() instead if getpid(), and that's the exact issue presented in this ticket. How on earth can replacing getpid() with CRYPTO_thread_id() be a solution to this issue? -- Richard Levitte [EMAIL PROTECTED] ______________________________________________________________________ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
