Bug#628444: iwlagn - "MAC is in deep sleep", cannot restore wifi operation

[Shannon Dealy <dealy@deatech.com>]

On Fri, 16 Mar 2012, Bjørn Mork wrote:

> Shannon Dealy <dealy@deatech.com> writes:
>
> > I created a file "/etc/modprobe.d/iwlagn.conf" and placed the
> > following line in it:
> >
> >   options iwlagn 11n_disable=1 11n_disable50=1
>
> Note that the 11n_disable50 options was removed in 3.0 and the iwlagn
> module was renamed to iwlwifi in 3.2.

Unfortunate, since further testing shows that it is the 11n_disable50=1 
that matters.  Disabling either of the above options makes my link stable, 
and since 11n_disable=1 would effectively cause "11n_disable50" to happen 
as well, the implication is that the problem is in the code which is/was 
controlled by the 11n_disable50.

Would be nice if anyone can confirm if this fixes the deep sleep problem 
as well.

> Which makes this workaround pretty much irrelevant to any current Debian
> kernel as noone(?) has seen the bug in 2.6.32.

While it may not be relevant to you, it is completely relevant in that it:

  - provides developers with a narrowed range of places to look for the problem
    in all versions of the kernel/driver code, including those were the
    option was removed (they can easily check what part of the code it
    previously disabled)

  - using just the 11n_disable option should still solve the stability
    problems I am seeing and possibly the deep sleep as well (need
    confirmation on that) for people running 3.x (though granted it is not
    a great workaround given the performance hit)

  - I assume you are speaking with regard to 2.6.32 being the default
    stable kernel so it won't affect people running stock Debian stable?
    I don't recall seeing confirmation from anyone that the problem went
    away by downgrading to 2.6.32.

> Are you still using the  2.6.39-1 kernel you originally opened this bug 
> against?
[snip]

Yes, though it has been rebuilt for debug tracing of the iwlagn driver. 
Unfortunately I can't afford the down time right now that a kernel 
upgrade beyond 2.6.39 would entail (other software would have to be 
upgraded and that might require that I write some code as well to 
deal with the kernel changes).  I also can't downgrade to 2.6.32 
(assuming that it would make any difference) as it doesn't support some of 
my hardware.

FWIW

Shannon C. Dealy      |               DeaTech Research Inc.
dealy@deatech.com     |          - Custom Software Development -
Phone: (800) 467-5820 |          - Natural Building Instruction -
   or: (541) 929-4089 |                  www.deatech.com