Xpra: Ticket #2587: can't peek at some http traffic

Fedora 31 server running trunk (v4).

Instead of detecting the http traffic, peek returns no data and we fail with:

io_in_cb(<socket.socket fd=4, family=AddressFamily.AF_INET, type=SocketKind.SOCK_STREAM, proto=0, laddr=('0.0.0.0', 10000)>, <flags G_IO_IN of type GLib.IOCondition>)
(..)
socket tcp socket: 192.168.0.10:10000 <- 192.168.0.106:47438 peek: got 0 bytes
may_wrap_socket: no data, not wrapping
(..)
New tcp connection received
 from '192.168.0.106:47438'
 on '0.0.0.0:10000'

That's despite waiting for 1 second polling the socket...

wireshark shows that the TCP packet does contain the HTTP GET request:

GET / HTTP/1.1
Host: 192.168.0.10:10000
Connection: keep-alive
Cache-Control: max-age=0
Upgrade-Insecure-Requests: 1
Save-Data: on
User-Agent: Mozilla/5.0 (Linux; Android 10; PH-1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/80.0.3987.87 Mobile Safari/537.36
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,image/apng,*/*;q=0.8,application/signed-exchange;v=b3;q=0.9
Accept-Encoding: gzip, deflate
Accept-Language: en-GB,en;q=0.9,fr-FR;q=0.8,fr;q=0.7,th-TH;q=0.6,th;q=0.5,en-US;q=0.4

This happens with android clients, but not with desktop browsers!?



Wed, 12 Feb 2020 16:45:48 GMT - Antoine Martin: status changed

Adding a recv call immediately after the empty peek shows that the data is there waiting to be read!?


Thu, 13 Feb 2020 05:45:05 GMT - Antoine Martin:

This is a kernel / distro bug. Every other distro works just fine. Tested: Ubuntu 18.04, Ubuntu 19.10, Debian Stretch, Debian Buster, etc..

The problem only occurs on Fedora 30 and 31. (Not tested Fedora 29 or earlier). And also with MS Windows servers!


Thu, 13 Feb 2020 10:08:03 GMT - Antoine Martin:

Things I've tried:

Wireshark shows:

... Why do we reset the connection!? We're actually doing a recv when that happens!?

Also happens with v3.


Fri, 14 Feb 2020 03:26:42 GMT - Antoine Martin: attachment set

make all sockets peekable


Fri, 14 Feb 2020 03:57:42 GMT - Antoine Martin:

The patch above uses the same workaround that is used with ssl sockets: emulate peek by actually reading from the socket (without PEEK) but cache the data and return it later. (also uncomment enable_peek() to use it) Does not always help. No idea why.


Tue, 17 Mar 2020 17:22:19 GMT - Antoine Martin: attachment set

wireshark capture of a failure


Fri, 08 May 2020 14:45:01 GMT - Antoine Martin: milestone changed


Mon, 12 Oct 2020 14:21:30 GMT - Antoine Martin: attachment set

updated patch


Mon, 12 Oct 2020 15:17:52 GMT - Antoine Martin: status changed; resolution set

r27660 logs more details of connection timeouts.

I really don't understand why this makes any difference, but r27662 fixes things. (at least I haven't been able to reproduce the bug with the Android tablet or mswindows browsers)


Sat, 23 Jan 2021 05:55:21 GMT - migration script:

this ticket has been moved to: https://github.com/Xpra-org/xpra/issues/2587