On Sunday 16 November 2008 13:02:54 Florian Philipp wrote: > Can anyone say: "Yeah, I use the iwl3945 driver with 2.6.26 and it works > flawlessly for me." If so, could you please post your exact kernel > version (for example gentoo-sources-2.6.26-r2) and your kernel config?
No-one can say that, the driver is buggy. Something to do with waking the card up or bringing it out of deep sleep. Personally, I think it's along the lines of a race condition in how the card is reliably powered up (but that's just be thinking of what's most likely). Intel has a bugzilla, which shows that virtually every distro on all platforms using the 3945 have the identical problem in every kernel version that contains this code: http://intellinuxwireless.org/bugzilla/show_bug.cgi?format=multiple&id=1651 We have two options: - get involved, fix it ourselves, submit patch to upstream (intel) - wait for intel to come up with reliable code All the evidence points to intel taking this seriously, it's a popular card and intel wants their stuff to work nicely on linux. So chances are it's a deeper problem that it appears on the surface -- alan dot mckinnon at gmail dot com
