commit 95238eb9174f0cfee9d313ce15b4f9b471f3d0e5 Author: Nick Mathewson nickm@torproject.org Date: Thu Nov 16 11:45:15 2017 -0500
Fix a traceback when closing a blocked connection "immediately".
When we close a connection via connection_close_immediately, we kill its events immediately. But if it had been blocked on bandwidth read/write, we could try to re-add its (nonexistent) events later from connection_bucket_refill -- if we got to that callback before we swept the marked connections.
Fixes bug 24167. Fortunately, this hasn't been a crash bug since we introduced connection_check_event in 0.2.9.10, and backported it.
This is a bugfix on commit 89d422914a0c3cb, I believe, which appeared in Tor 0.1.0.1-rc. --- changes/bug24167 | 7 +++++++ src/or/connection.c | 4 ++++ 2 files changed, 11 insertions(+)
diff --git a/changes/bug24167 b/changes/bug24167 new file mode 100644 index 000000000..fd0d87eff --- /dev/null +++ b/changes/bug24167 @@ -0,0 +1,7 @@ + o Minor bugfixes (network layer): + - When closing a connection via close_connection_immediately(), we + mark it as "not blocked on bandwidth", to prevent later calls + from trying to unblock it, and give it permission to read. This + fixes a backtrace warning that can happen on relays under various + circumstances. Fixes bug 24167; bugfix on 0.1.0.1-rc. + diff --git a/src/or/connection.c b/src/or/connection.c index 276dca281..61f4f5aa6 100644 --- a/src/or/connection.c +++ b/src/or/connection.c @@ -721,6 +721,10 @@ connection_close_immediate(connection_t *conn)
connection_unregister_events(conn);
+ /* Prevent the event from getting unblocked. */ + conn->read_blocked_on_bw = + conn->write_blocked_on_bw = 0; + if (SOCKET_OK(conn->s)) tor_close_socket(conn->s); conn->s = TOR_INVALID_SOCKET;