OK, so, no reply on this one. Can someone tell me who wrote ssl3_shutdown, so I can just ask them directly? Unfortunately, the internal documentation in the code isn't sufficient to determine whether my proposed solution is correct.
Obviously, I'd prefer it if the code didn't get to the state where it was causing a processor fault, but that's the way it is right now, so I need to do something. I would think that this would be of general concern as well; maybe this change should even go back into the main code pool. Thanks in advance for any help. Allan Miller ----- Original Message ----- From: "Allan Miller" <[EMAIL PROTECTED]> To: <openssl-dev@openssl.org> Sent: Monday, May 08, 2006 6:00 PM Subject: Question about ssl3_shutdown Hi, I have a question about ssl3_shutdown in s3_lib.c. We are getting an occasional processor fault on the line that reads: else if (s->s3->alert_dispatch) where s is a valid pointer but s3 is NULL. I would like to just prevent the processor fault but leave things in a consistent state so that further connections will work, etc. It looks to me like I can just change the very first executable line of the procedure to read: /* Don't do anything much if we have not done the handshake or * we don't want to send messages :-) */ if ((s->quiet_shutdown) || (s->state == SSL_ST_BEFORE) || (s->s3 == NULL)) { s->shutdown=(SSL_SENT_SHUTDOWN|SSL_RECEIVED_SHUTDOWN); return(1); } (in other words, add the check "s->s3 == NULL" to the case for "not doing anything much"). Does that look right? Thanks, Allan Miller ______________________________________________________________________ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED] ______________________________________________________________________ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]