On Wednesday 01 August 2007 20:54:47 Michael Buesch wrote: > On Wednesday 01 August 2007 20:40:59 Larry Finger wrote: > > Michael Buesch wrote: > > > > > > That's a known bug and I am currently trying to find out why it > > > fails to tune the power to higher levels. > > > Even adjusting power manually doesn't have any effect. So it's not > > > a bug in the power scaling algorithm, but deeper in the attenuation > > > adjustment. > > > Might be some broken table upload. > > > > > > > It is something in the PHY initialization. I don't get sparking performance > > with a 4318 using the > > new port from softmac to mac80211, but I do get usable performance of at > > least 5 Mbs for all rates, > > even at 15-20 m from the AP. One thing is that the signal level from > > iwconfig is very low at -61 > > dBm; whereas, the 4311 gives -37 at the same position relative to the AP. > > The noise values are both > > -72 dBm. > > Sure, you are absolutely right. If I define has_hardware_pctl() to > always be 0, I get better performance. So we somewhere blow up > the hwpctl init. I'm debugging this at the moment.
No I don't. Hum. But today I get rather good performance (well compared to yesterday) from my tree. Seems to depend on the moonphase or something. I hate benchmarking. -- Greetings Michael. _______________________________________________ Bcm43xx-dev mailing list [email protected] https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
