c2s crashes for dead loop

[smoku]

This is the log , I had modify some code , so the line of sources in the log file is not same jabberd-s11.

sx (io.c:70) completed nad: <route
xmlns='http://jabberd.jabberstudio.org/ns/component/1.0' from='im800.net'
to='c2s'><sc:session xmlns:sc='http://jabberd.jabberstudio.org/ns/session/1.0'
action='replaced' sc:sm='f14317dae5afaf5bfa2384a7bddacc4437dde166'
sc:c2s='c1ff443bcf804faa3137fd50d5a48d745eb9b871'/></route>
sx (chain.c:119) calling nad read chain
sx (io.c:123) tag 6 event 6 data 0x8c97b90
sx (error.c:79) prepared error: <stream:error
xmlns:stream='http://etherx.jabber.org/streams'><conflict
xmlns='urn:ietf:params:xml:ns:xmpp-streams'/></stream:error>
sx (error.c:94) tag 10 event 1 data 0x0
Tue Jun 13 10:07:15 2006 c2s.c:38 want write
Tue Jun 13 10:07:15 2006 c2s.c:399 write action on fd 10
sx (io.c:292) 10 ready for writing
sx (io.c:311) handing app 602 bytes to write
sx (io.c:312) tag 10 event 3 data 0x8e97c60
Tue Jun 13 10:07:15 2006 c2s.c:104 writing to 10
sx (io.c:345) tag 10 event 0 data 0x0
Tue Jun 13 10:07:15 2006 c2s.c:33 want read
sx (io.c:439) 10 state change from 4 to 5
sx (io.c:450) tag 10 event 1 data 0x0
Tue Jun 13 10:07:15 2006 c2s.c:38 want write
Tue Jun 13 10:07:15 2006 c2s.c:399 write action on fd 10
sx (io.c:292) 10 ready for writing
sx (io.c:311) handing app 602 bytes to write
sx (io.c:312) tag 10 event 3 data 0x8e97c60
Tue Jun 13 10:07:15 2006 c2s.c:104 writing to 10
sx (io.c:345) tag 10 event 0 data 0x0
....
....
....
Tue Jun 13 10:07:48 2006 main.c:592 running time checks
sx (error.c:79) prepared error: <stream:error
xmlns:stream='http://etherx.jabber.org/streams'><host-gone
xmlns='urn:ietf:params:xml:ns:xmpp-streams'/><text
xmlns='urn:ietf:params:xml:ns:xmpp-streams'>connection timed
out</text></stream:error>
sx (error.c:94) tag 10 event 1 data 0x0
Tue Jun 13 10:07:48 2006 c2s.c:38 want write
Tue Jun 13 10:07:48 2006 c2s.c:399 write action on fd 10
sx (io.c:292) 10 ready for writing
sx (io.c:311) handing app 602 bytes to write
sx (io.c:312) tag 10 event 3 data 0x8e97c60
Tue Jun 13 10:07:48 2006 c2s.c:104 writing to 10
sx (io.c:345) tag 10 event 0 data 0x0
Tue Jun 13 10:07:48 2006 c2s.c:33 want read
sx (error.c:79) prepared error: <stream:error
xmlns:stream='http://etherx.jabber.org/streams'><host-gone
xmlns='urn:ietf:params:xml:ns:xmpp-streams'/><text
xmlns='urn:ietf:params:xml:ns:xmpp-streams'>connection timed
out</text></stream:error>
sx (error.c:94) tag 10 event 1 data 0x0
Tue Jun 13 10:07:48 2006 c2s.c:38 want write
Tue Jun 13 10:07:48 2006 c2s.c:399 write action on fd 10
sx (io.c:292) 10 ready for writing
sx (io.c:311) handing app 602 bytes to write
sx (io.c:312) tag 10 event 3 data 0x8e97c60
Tue Jun 13 10:07:48 2006 c2s.c:104 writing to 10
sx (io.c:345) tag 10 event 0 data 0x0
Tue Jun 13 10:07:48 2006 c2s.c:33 want read
sx (error.c:79) prepared error: <stream:error
xmlns:stream='http://etherx.jabber.org/streams'><host-gone
xmlns='urn:ietf:params:xml:ns:xmpp-streams'/><text
xmlns='urn:ietf:params:xml:ns:xmpp-streams'>connection timed
out</text></stream:error>
sx (error.c:94) tag 10 event 1 data 0x0
Tue Jun 13 10:07:48 2006 c2s.c:38 want write
Tue Jun 13 10:07:48 2006 c2s.c:399 write action on fd 10
sx (io.c:292) 10 ready for writing
sx (io.c:311) handing app 602 bytes to write
sx (io.c:312) tag 10 event 3 data 0x8e97c60
Tue Jun 13 10:07:48 2006 c2s.c:104 writing to 10
.... the following is the dead loop

I parses the c2s source and finds a problem in the c2s

static void _c2s_time_checks(c2s_t c2s) {
    sess_t sess;
    time_t now;
    union xhashv xhv;

    now = time(NULL);

    if(xhash_iter_first(c2s->sessions))
        do {
            xhv.sess_val = &sess;
            xhash_iter_get(c2s->sessions, NULL, xhv.val);

            if(c2s->io_check_idle > 0 && now > sess->last_activity +
c2s->io_check_idle) {
                log_write(c2s->log, LOG_NOTICE, "[%d] [%s, port=%d] timed out",
sess->fd, sess->ip, sess->port);

                sx_error(sess->s, stream_err_HOST_GONE, "connection timed
out");
                sx_close(sess->s);

                continue;
            }

            if(c2s->io_check_keepalive > 0 && now > sess->last_activity +
c2s->io_check_keepalive && sess->s->state >= state_STREAM) {
                log_debug(ZONE, "sending keepalive for %d", sess->fd);

                sx_raw_write(sess->s, " ", 1);
            }

        } while(xhash_iter_next(c2s->sessions));
}

it is a problem in calling the sx_close() when the state of sx is state_CLOSING. The sx_close() does not close the sx and the sx_close() return immediately . this will cause to the _c2s_time_checks() dead loop.

[smoku]

I believe this is fixed in 2.1. If not, please reopen the ticket.