Hi, I'm debugging my server and I found that behind SSL_write() there are ssl3_write_bytes() and ssl3_write_pending(). I assume that the last function has a queue of pending messages that has to be sent. Before SIGSEGV I was having SIGPIPE. When I ignored SIGPIPE (should I?) sometimes I fall on a SIGSEGV. The problem arises that if a client has left, the ssl3_write_pending() (is it works as I think, there's no man page on that) will try to write to an unspecified BIO. Assuming that we are having exclusive access to our client structure when we issue SSL_write(), if the result is queued then when ssl3_write_pending() really sends those bytes will be on a NULL BIO.
My question is: Does this SSL3_write_pending() works on the life of SSL_write() ? That is, SSL_write() do not return untill ssl3_write_pending() return. If ssl3_write_pending() is "forked" or "threaded", how can we assure that ssl3_write_pending() will write on an existing BIO? Regards, -- If you want freedom, compile the source. Get gentoo. Sebastián Treu http://labombiya.com.ar ______________________________________________________________________ OpenSSL Project http://www.openssl.org User Support Mailing List openssl-users@openssl.org Automated List Manager majord...@openssl.org
