Also, I noticed that (prior to running kismet and the wifi freaking out) iwlist scanning does not work:
$ sudo iwlist eth1 scanning eth1 Interface doesn't support scanning : Operation not supported
IIRC this should be solved in CVS drivers.
Yes. David Gibson emailed me last night about the dragorn patch. He says that this is really old code that imports a lot of unnecessary things from linux-wlan-ng tree. He said a cleaned up version exists in CVS but that if you're using 2.6 you should check out the "for_linus" branch.
I checked it out and built the drivers last night and I'm running them right now. They seem stable and allow kismet to run as well as iwlist.
There is one catch however...
I had to disable a firmware version check because they claim that Agere 8.xx versions are buggy. I have 8.4 but I've used this thing in monitor mode for over a year now and haven't seen any problems.
Anyway I would suggest trying this code out because it's going to be merged into Linus's tree in the future.
I can provide a patch against the CVS snapshot because the code wont build on it's own. It uses some deprecated macro. As far as disabling the firmware check... well I suggest you try the code with the check. Maybe your firmware is considered stable. If not then you can decide whether or not you're comfortable disabling it.
Ultimately the decision now is whether we can get the CVS versions of this driver patched into the Debian PPC kernels. The sources are coming from upstream so I can't think of a good reason why this wouldn't be possible. Again, having a number of users out there testing this would be wise prior to attempting to making anything official.
Maybe we could put out a module package similar to nvidia's driver, where the user just builds the code outside the kernel tree. This way you could test drivers without having to install the entire kernel image.
This is what I did here... I just built the code outside my kernel tree and dropped the new drivers in place (saving the old ones of course).
At any rate Barry, perhaps should try this code and see whether or not it functions better than the original patch I posted. If nothing else you will get the same behavior with a working iwlist. Also, could you post your firmware version. I'd like to see if we can associate problems with a specific version/versions. There has obviously been problems with firmware or the check wouldn't exist.
--
"Imagination is more important than knowledge."
- Albert Einstein
signature.asc
Description: OpenPGP digital signature
