Hi,
On 3/22/2017 18:11, Axel Lin wrote:
I recently found it's easy to get connection abort when I test
connecting web server.
(test by just repeating send http request)
It's easy to get below debug message:
tcp_slowtmr: max SYN retries reached
Just a guess: could it be that your test is making so many connections
so fast that you're cycling through all local port numbers and thus end
up reusing them, while previous connections with those port numbers are
still in TIME_WAIT state on the web server side at the time that the
first SYN is sent for the new connection on the client side? It should
be fairly easy to check this with netstat on the server side, if you
have access to such a tool there of course.
If that is indeed the case, behavior would definitely differ between
1.4.1 and 2.0.x due to patch #9039 and bug #50498, although it would
then be a matter of perspective as to whether the issue is with lwIP or
your test..
Regards,
David
_______________________________________________
lwip-users mailing list
[email protected]
https://lists.nongnu.org/mailman/listinfo/lwip-users
