cookdav wrote:
> janoholm;1822120 Wrote:
>> Thanks Cookdav for seeing that!
>> Do you think I shall try the b43legacy-driver again? Because the
>> ndiswrapper-solution is not without problems I have to do
>> several of the advices written on the opensuse/ndiswrapper-page:
>> - load ndiswrapper module from /etc/sysconfig/kernel
>> - use ifup instead of NM
>> - tried to switch of encrypt
>>
>> Jan
>
> I don't have that device, and even tho I'm a real ndiswrapper-lover,
> I'd recommend you try the native drivers first.
>
> [I'm using ifup instead of NM, because NM wouldn't work, but
> since then, there's been updates to NM. So, try NM first.]

You should use NM and b43legacy, rather than ndiswrapper.

I did the conversion of bcm43xx to create b43legacy, thus I know it
works. There are some problems for Ubuntu users that have rfkill
switches, but I'm not aware of any with openSUSE kernels. The main
reason not to use ndiswrapper is that I am now working with the
reverse engineering team that is writing the specifications for
extending b43 to work with the latest Broadcom devices. In that work,
I see how little error checking that Broadcom does relative to the
open-source Linux driver. I have not done any RE on the Windows
driver, but I would expect the same philosophy. Of course, any
undetected error leads to a "Blue Screen of Death" on Windows, and an
equally messy crash in Linux. To add further insult, no Linux
developer will even look at any crash from the "tainted" kernel that
results from loading a non-GPLed driver. Once you run arbitrary code
at the inner-most level, all bets are off.

To help understand why you are having a problem, please post the output of

sudo iwlist scan
ifconfig
iwconfig
dmesg | grep ssb
dmesg | grep b43

Larry