The problem, described in my previous email,

http://www.pasta.cs.uit.no/pipermail/linux-irda/2001-May/002799.html

seems to be related to the qos->max_turn_time, which in qos.c becomes
set to 500ms since the baudrate is lower than 115200.
If i modify qos.c to omit this "correction" and instead set the turntime
to 100ms and an appropriate line capacity around 300 bytes I get
'normal' throughput to the phone and the latency is decreased
accordingly.

Am I right in assuming that the qos->max_turn_time setting to 500ms is
forced by the standard?, and in that case, why is the computer only
inviting the phone to communicate with an interval of
'qos->max_turn_time'?
I'm purely speculating here, but I assume what's happening is that the
phone cannot send data in chunks of 1600 bytes which would be optimal
here (the calculated line capacity at 38400bps & 500ms)  and since the
computer only 'invites' the phone to communicate every
'qos->max_turn_time' ms  the communication more or less stalls and
latency is greatly increased. Wouldn't qos->min_turn_time +
the_time_used_to_send_available_data be a better interval?

Any help or comments would be greatly appreciated

/Thomas


--
       Thomas Hellstr�m, Fyrm�stareg�ngen 8, S-413 18 G�teborg, Sweden
                       Email: [EMAIL PROTECTED]
    Tel: +46 31 244077, +46 31 663295, +46 704 976916 // Fax: +46 31 546710



_______________________________________________
Linux-IrDA mailing list  -  [EMAIL PROTECTED]
http://www.pasta.cs.UiT.No/mailman/listinfo/linux-irda

Reply via email to