Qualcomm Atheros Bigfoot Killer E2200 Ethernet working in linux

[Mahler122]

Hi All,

I have found a way to get the Qualcomm Atheros Bigfoot Killer E2200 ethernet card to work with existing Atheros linux drivers. Following in the footsteps of this post: http://ubuntuforums.org/showthread.php?p=12028162. I have basically 'stolen' the driver of a related card with existing linux drivers.

I have created a patch to be applied to the compat-wireless kernel modules. After application of this patch and procedure, my ethernet card worked.

The module which will be modified is alx, and after I told the kernel to load it explicitly by editing /etc/modules, My ethernet card was discovered at startup and properly suspended and hibernated. I can verify that gigabit speeds are possible with this 'stolen' driver.

I haven't had any probles yet, but I will post a reply to this thread if a problem appears at some later date, or Atheros releases a proper driver for this card.

You will need to install a few packages

sudo apt-get install build-essential linux-headers

You will need to download the compat-wireless-2012-05-10-p tarball from here: http://linuxwireless.org/download/co...5-10-p.tar.bz2

And download the attached compat-patch.txt.

Untar the tarball into a convenient location, and apply the patch making sure to use the --dry-run option to make sure there are no errors.

tar -xvf /path/to/compat-wireless-2012-05-10-p.tar.bz2
cd ./compat-wireless-2012-05-10-p.tar.bz2
patch --dry-run -p2 < /path/to/compat-patch.txt
patch -p2 < /path/to/compat-patch.txt

Now make and assuming no errors arise, install the package.

make
sudo make install

Now test that the module works by inserting it into the running kernel.

sudo modprobe alx

Your ethernet card should be discovered, and assuming you have the ethernet plugged into a working jack with dhcp, you should become connected to ethernet.

Assuming this is the case, modify your /etc/modules file so that this module is loaded at startup by adding the following lines at the end.

#E2200 support
alx

And you're good to go. Let me know how this works out for you.

[Paddy Landau]

Thank you for posting this for others to see.

If you mark the thread as solved, it will help others find it.

[Mahler122]

No problem!, Done.

[belnac]

Awesome, Mahler! Many thanks for this - will try it out at once.

[belnac]

Using Kernel 3.5.0-rc3, compilation fails on ./net/bluetooth/rfcomm/tty.c

Attached is a simple patch I've just created that fixes the compilation error, which is caused by different function definitions for tty_unlock and tty_lock; current version of tty.c expects the two (mutex) functions to accept an argument, which they now don't anymore.

Mahler, thanks again for this - the NIC has now been brought to life!

[Mahler122]

Just so I'm clear on this, this patch is meant to be applied after my patch correct?

I tested my patch with the 3.2.0 kernel. It's good to know about this compilation error with the 3.5-rc's.

[belnac]

Just so I'm clear on this, this patch is meant to be applied after my patch correct?

Yes, correct. Applying just your patch results in the compilation error mentioned above - and that was whilst compiling with kernel 3.5.0-rc3; haven't tested with other kernels. Applying the second patch fixes it and compilation proceeds nicely.

[Mahler122]

Just so everyone knows, I suspect that I'm experiencing the first issues related to this 'stolen' driver. It seems that everything simply freezes after ~ 2.5 days or so of use. The only route to recovery seemed to be a hard boot. However it might just be the graphical interface. Next time it happens I'm going to see whether my machine responds to ssh. (It's happened twice so far)

Please let me know if you experience this as well, and try not to have essential documents and what not loaded into ram to avoid losing them if this happens and you need to hard boot.

Of course, Atheros needs to release a legitimate driver which won't crash the kernel, but in the meantime, for me the NIC working is worth worrying about the crashing.

[chili555]

It seems that everything simply freezes after ~ 2.5 days or so of use. The only route to recovery seemed to be a hard boot.

What does /var/log/syslog suggest the origin of the lock-up is?

[Mahler122]

I'd say nothing. There is just a time gap between when the last message was issued, and when the first kenel boot message appears.

Here are the lines surrounding the gap which I suspect is the lock-up.

Jun 26 16:31:07 sibelius AptDaemon.Worker: INFO: Processing transaction /org/debian/apt/transaction/0b6a845712d54ef2854d86ddae157393
Jun 26 16:31:08 sibelius AptDaemon.PackageKit: INFO: Get updates()
Jun 26 16:31:09 sibelius AptDaemon.Worker: INFO: Finished transaction /org/debian/apt/transaction/0b6a845712d54ef2854d86ddae157393
Jun 26 16:37:08 sibelius AptDaemon: INFO: Quitting due to inactivity
Jun 26 16:37:08 sibelius AptDaemon: INFO: Quitting was requested
Jun 26 16:52:29 sibelius kernel: imklog 5.8.6, log source = /proc/kmsg started.
Jun 26 16:52:29 sibelius rsyslogd: [origin software="rsyslogd" swVersion="5.8.6" x-pid="1017" x-info="http://www.rsyslog.com"] start
Jun 26 16:52:29 sibelius rsyslogd: rsyslogd's groupid changed to 103
Jun 26 16:52:29 sibelius rsyslogd: rsyslogd's userid changed to 101
Jun 26 16:52:29 sibelius rsyslogd-2039: Could not open output pipe '/dev/xconsole' [try http://www.rsyslog.com/e/2039 ]
Jun 26 16:52:29 sibelius kernel: [    0.000000] Initializing cgroup subsys cpuset

Thanks for letting me know about this log. I was looking for something like this. I'll post another if I see it happen again.