Comment 3 for bug 630748

Revision history for this message
Kimmo Jukarainen (kimju) wrote :

After upgrading from lucid to maverick I started having similar
problems with a 5300 card. First the transfer speeds start
dropping drastically and after a while (couple of minutes)
the card stops moving any data at all.

Currently running with: linux-image-2.6.35-22-generic 2.6.35-22.3.

The dmesg is flooded with messages like:

iwlagn 0000:03:00.0: BA scd_flow 0 does not match txq_id 10

This is probably an upstream bug, as reported in these bugzillas:

https://bugzilla.kernel.org/show_bug.cgi?id=16691
http://bugzilla.intellinuxwireless.org/show_bug.cgi?id=2214

The fix seems to require updated ucode from Intel, which they
haven't released yet.

The workaround is to disable .11n functionality, by giving to
the iwlagn module the parameter "11n_disable=1".

Also, rmmod + modprobe even without the 11n_disable parameter
seems to fix the problem for few minutes, but then it starts again.

Some more information can be seen for example in:
https://bugs.archlinux.org/task/20542

In launchpad, a quick look shows that these three bugs are
probably clones of each other:

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/646711
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/630748
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/604175

And maybe even these. Some of them are from older kernels, but
there seems to be reports that this ucode bug started manifesting
itself around kernel 2.6.31, but mostly since 2.6.35:

https://bugs.launchpad.net/ubuntu/+source/linux/+bug/634876
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/630714
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/602649
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/594889
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/583210
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/568756