On Mon, Jun 15, 2009 at 8:48 AM, Alexander Monakov<[email protected]> wrote: >> I will need some time to determine whether it improves the >> link quality (but at least txpower 1 seems not worse than txpower 20 >> or auto). > > Nah, testing with ftp transfers does not show any improvement or > degradation when changing txpower. Traffic is extremely uneven: 1-2 > seconds long high throughput bursts are followed by tens of seconds of > no ftp traffic. Btw, how does the driver decide whether it needs to > retransfer packets (can loss of TCP ACKs be the reason for such > behavior)? > > I forgot to mention that card lowers transmission rate to 1 Mbps by > default, but still exhibits the behavior described in the first mail. > Locking rate to 36, 48 or 54 Mbps does not seem to affect frequency of > receiving deauth packets from AP. I also tried playing with rts and > frag settings, to no avail. > > -- > Alexander Monakov > _______________________________________________ > Bcm43xx-dev mailing list > [email protected] > https://lists.berlios.de/mailman/listinfo/bcm43xx-dev >
Try using the new rt73usb for monitoring, as it gives radiotap headers, with signal strength in dBm, which is much more meaningful than the prism header's funky unit. I have the same card, and have not experienced this problem (though I did get PHY transmission error lockups last time I used it). It may be a good idea to turn on debugging in b43 to see if the disassociations coincide with a PHY error. -- Vista: [V]iruses, [I]ntruders, [S]pyware, [T]rojans and [A]dware. :-) _______________________________________________ Bcm43xx-dev mailing list [email protected] https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
