[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

[tor-commits] [tor/master] restore the sensible part of ac268a83408e1450544db2f23f364dfa3



commit b82e166bec5fcc468424af1ff71e2e753ac534a2
Author: Nick Mathewson <nickm@xxxxxxxxxxxxxx>
Date:   Fri Aug 29 12:21:57 2014 -0400

    restore the sensible part of ac268a83408e1450544db2f23f364dfa3
    
    We don't want to call event_del() postfork, if cpuworkers are
    multiprocess.
---
 src/or/connection.c |    7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/src/or/connection.c b/src/or/connection.c
index 6d205d1..aedb29d 100644
--- a/src/or/connection.c
+++ b/src/or/connection.c
@@ -553,10 +553,17 @@ connection_free_(connection_t *conn)
     tor_free(control_conn->incoming_cmd);
   }
 
+#ifdef TOR_IS_MULTITHREADED
   /* Probably already freed by connection_free. */
+  /* We don't do these frees on the multiprocess case, since in that case we
+   * don't want to call event_del() postfork or it's likely to mess up.
+   * Multiprocess builds are deprecated, so let's just have a one-time memory
+   * leak here.
+   */
   tor_event_free(conn->read_event);
   tor_event_free(conn->write_event);
   conn->read_event = conn->write_event = NULL;
+#endif
   IF_HAS_BUFFEREVENT(conn, {
       /* This was a workaround to handle bugs in some old versions of libevent
        * where callbacks can occur after calling bufferevent_free().  Setting



_______________________________________________
tor-commits mailing list
tor-commits@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-commits