- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I test on FreeRTOS+LwIP build and test using google chrome browser.
Initially the browser can connect to the https server.
Reload or press F5 to send request for a few times.
Then the https server no longer response anymore.
I add a debug print, the error is "Error starting TLS connection",
the wiced_tcp_start_tls() return 19.
(Sometimes I also got wiced_tcp_start_tls result = 5018/5003/5030/17/2,
I have no idea why so many different error code.)
- Labels:
-
SDK 3.x
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Same issue here with ThreadX/NetX_Duo in 3.5.2.
Sometimes I get errors like this:
LOG_DEBUG: [eci] Accepting new connection from 192.168.1.15 on 50033
Error starting TLS connection
LOG_ERROR: [eci] accept failed 1b66
LOG_DEBUG: [event][59] Executing event 20002700 (801c721) for 0
LOG_DEBUG: [eci] Accepting new connection from 192.168.1.15 on 50033
Error starting TLS connection
LOG_ERROR: [eci] accept failed 1b66
LOG_DEBUG: [event][67] Executing event 2000271c (801c721) for 0
And after a while, it works again ... But no idea why.
For HTTPS server, there is an issue with tcp_server_accept. If an accept fails (which is your case), it won't do the relisten. So you are facing a double issue I think.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Well after some more testing, I think that's the same issue I reported here: Unable to connect to device after wiced_tcp_accept returned an error
So it would be safe to say the issue is still present
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
This looks like a regression. It was working fine in SDK-3.1.2.