Hi PJ peeps,
Please apply this patch which removes the pj_thread_sleep() call from
pjsip_endpt_handle_events2(). I think it makes more sense just to
return the error to the user immediately and let them decide if they
want to sleep (like pjsua-lib does).
One concrete case when this sleep is problematic is when a signal
occurs. In this case, the call to pj_ioqueue_poll() will return
because of EINTR which will cause the main loop to additionally sleep
for the full timeout. This means, if pjsip_endpt_handle_events2() is
called with a large timeout, the signal/interrupt will block the app
for the full timeout interval if a signal happens. This is a problem.
Thanks,
Kal
diff --git a/pjsip/src/pjsip/sip_endpoint.c b/pjsip/src/pjsip/sip_endpoint.c
index 107ec125..48065562 100644
--- a/pjsip/src/pjsip/sip_endpoint.c
+++ b/pjsip/src/pjsip/sip_endpoint.c
@@ -742,7 +742,6 @@ PJ_DEF(pj_status_t) pjsip_endpt_handle_events2(pjsip_endpoint *endpt,
c = pj_ioqueue_poll( endpt->ioqueue, &timeout);
if (c < 0) {
pj_status_t err = pj_get_netos_error();
- pj_thread_sleep(PJ_TIME_VAL_MSEC(timeout));
if (p_count)
*p_count = count;
return err;
_______________________________________________
Visit our blog: http://blog.pjsip.org
pjsip mailing list
pjsip@xxxxxxxxxxxxxxx
http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org
