Bug#574555: BUG: scheduling while atomic: irq/11-b43/2018/0x00000101
Ben Hutchings wrote:
On Fri, 2010-06-18 at 16:40 +0200, Arne Nordmark wrote:
My problem seems related to those in Ubuntu report #533335
<https://bugs.launchpad.net/ubuntu/+source/linux/+bug/533335>. If I
unload the 3c59x module, I can no longer reproduce this problem, and
wireless is stable. With the 3c59x module loaded, b43 wireless is
essentially unusable, since the machine will reliably lock up.
I've attached two patches which together may fix this problem, but they
involve quite major changes to the driver. I do not have any hardware
of this type which I could use to test them.
Indeed, I can no longer reproduce the problem with these patches
applied. Both wired (3c59x) and wireless (b43) can now handle GB
transfers, which is about 100x more than was typically needed to trigger
the problem for wireless.
I now question my judgement in reporting my problem on an existing bug
report, as the resolution does not fit the original reporter's
description (no 3c59x module loaded).
Please could you follow the instructions at
to build a package with these changes included, then test whether these
changes fix the problem for you and keep the wired network card working.
Since that section is very explicit in mentioning dependencies, may I
point out that the "test-patches" script relies on commands ("dch" at
least) from the "devscripts" package.