On Fri, Dec 14, 2007 at 11:10:45AM -0500, Alex Iliadis wrote: > That's not it, the main problem is not that the device stops receiving > packets, it's the corruption of the packet content it does. As a note you > can control power management by issuing: > iwconfig wlan0 power off
The original packet corruption issue was that the driver reported all frames, even those with invalid FCS checksums. However, since the driver also provides valid FCS data, this was fixed by adding checksum validation to Kismet to throw out frames. This is enabled on the nokia packet sources in Kismet. The dropping dead, and other issues, are part of why I never released an official Kismet package. Having just got an 810, I'm inspired again to look at it (and was just working on adding navkit support in to use the 810 builtin gps, actually) - are you saying there is another packet corruption issue which is time-based? Something that might *look* like corruption (to users and to kismet itself) would be if something else changes the interface back to managed mode - then it'll be sending 802.3 frames on an interface opened with some 802.11 linktype variant (prism2avs if I recall?) I'd be very interested in being able to turn off the wlan connection manager programatically, ie throw the device into offline mode and reconfigure the wifi for rfmon inside kismet, then reactive the state when it's done - but haven't found anything, yet, on how to do so. With OS2008 having better OTG host mode support, I'm also considering building the common USB drivers as a package so that you can use kismet with a USB wireless card and avoid the issues with the binary drivers. Obviously something like the alfa 500mW usb card won't work with the low-power OTG chipset, but the handful of USB nics I've got handy at least enumerated properly. -m -- Mike Kershaw/Dragorn <[EMAIL PROTECTED]> GPG Fingerprint: 3546 89DF 3C9D ED80 3381 A661 D7B2 8822 738B BDB1 "A baby seal walks into a club..."
pgpyBwg3tjb7v.pgp
Description: PGP signature
_______________________________________________ maemo-developers mailing list [email protected] https://lists.maemo.org/mailman/listinfo/maemo-developers
