yamt commented on a change in pull request #5252:
URL: https://github.com/apache/incubator-nuttx/pull/5252#discussion_r792395337
##########
File path: net/tcp/tcp_send_buffered.c
##########
@@ -355,10 +355,39 @@ static uint16_t psock_send_eventhandler(FAR struct
net_driver_s *dev,
FAR void *pvconn, FAR void *pvpriv,
uint16_t flags)
{
- FAR struct tcp_conn_s *conn = (FAR struct tcp_conn_s *)pvconn;
+ /* FAR struct tcp_conn_s *conn = (FAR struct tcp_conn_s *)pvconn;
+ *
+ * Do not use pvconn argument to get the TCP connection pointer (the above
+ * commented line) because pvconn is normally NULL for some events like
+ * NETDEV_DOWN. Instead, the TCP connection pointer can be reliably
+ * obtained from the corresponding TCP socket.
+ */
+
FAR struct socket *psock = (FAR struct socket *)pvpriv;
+ FAR struct tcp_conn_s *conn;
bool rexmit = false;
+ DEBUGASSERT(psock != NULL);
+
+ /* Get the TCP connection pointer reliably from
+ * the corresponding TCP socket.
+ */
+
+ conn = psock->s_conn;
+ DEBUGASSERT(conn != NULL);
+
+ /* The TCP socket is connected and, hence, should be bound to a device.
+ * Make sure that the polling device is the one that we are bound to.
+ */
+
+ DEBUGASSERT(conn->dev != NULL);
Review comment:
> Could you also add this debug assert in your copy (before line 383)
just for debug?
i will consider. but it isn't that reproducible anyway.
if they are expected to be always true, why don't you just add them?
btw, what's the motivation of this PR?
it seems that it's proven the old way was actually more reliable than the
new way...
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
