Closed GoogleCodeExporter closed 9 years ago
I would prefer a solution that detects this error and calls set_buf_size()
with, say, double the current buffer size.
Are you sure it's not saying "UDP Error 234 : More data is available"? That's
ERROR_MORE_DATA.
Could you please test this patch?
http://dpaste.com/0W85QND/
Original comment by arvid.no...@gmail.com
on 10 May 2014 at 3:50
I've committed this to trunk in version [9950]. Please let me know if there's
any issue with this patch.
Original comment by arvid.no...@gmail.com
on 10 May 2014 at 6:38
not working, the code is not functional as it is too far up.
i moved it down and also had to increase the internal ASIO values:
http://upload.beamng.com/patches/udp_socket_fix.patch
are you sure that no packet is lost - i am not sure about the "continue" in
there
Original comment by tho...@thomasfischer.biz
on 10 May 2014 at 7:20
Sorry if i wasn't clear enough. The buffer increase must come after
"receive_from" as this spawns the error message. On my test machine it seems to
have increased the memory 3 times via this to a final bufsize of 16384.
I am not sure if it relates to the bandwidth I am achieving with this: 8MB/s
download
Original comment by tho...@thomasfischer.biz
on 10 May 2014 at 7:26
I imagine you either are transferring over loopback or have jumbo frames
enabled on your local network.
I made some slight modifications to your patch, does this work? I removed
setting the send buffer size. It may be necessary, but I would hope to set it
in response to sending a packet failing. do you see any errors from send_to()
if you don't set it?
http://dpaste.com/0JV8J5R/
Original comment by arvid.no...@gmail.com
on 10 May 2014 at 9:23
checked-in
Original comment by arvid.no...@gmail.com
on 10 May 2014 at 11:49
Original issue reported on code.google.com by
tho...@thomasfischer.biz
on 9 May 2014 at 7:13