Andreas Mohr wrote: > Hi, > > On Mon, Jun 02, 2008 at 10:01:09PM +0100, Felix Oxley wrote: >> Thank you. >> >> setting the power level has fixed the problem! >> I am now running at 60 wakeups per second instead of 5000. :-) > > Why the "%"$! would adapting the powerlevel have anything to do with > the number of wakeups generated!?! > I hope someone can explain this reasonably, 'cause I sure as hell > wouldn't believe this to be related...
it's reasonably simple: most wireless adapters use the 'power setting' to throttle generated interrupts to the bus/cpu. This saves power on the bus, the CPU and parts of the adapter as well since internally the wireless adapter can buffer packets for a while just fine on it's own. > (or, in other words, why is the driver/hardware this incredibly B0RKEN > and how can it be fixed?) it's unsure if something is broken, since you don't know how often beacon interval goes off on the access point. And if the wireless card does not buffer these but sends them right away to the host... you get a cpu interrupt for each of them. Of course, we don't know for sure if this is the case in this situation. > Could it be that the card firmware is severely misbehaving in case there > was no initial powerlevel call done by the driver on startup? absolutely, I'm not ruling this out at all. Cheers, Auke _______________________________________________ Power mailing list [email protected] http://www.bughost.org/mailman/listinfo/power
