Bug#665881: [3.1 -> 3.2.12 regression] module ath5k is blocking wlan-card
Am Montag, 26. März 2012 schrieben Sie:
> found 665881 linux-2.6/3.2.12-1
> quit
>
> Hi Hans,
>
Hi Jonathan,
> Hans wrote:
> > Package: linux-2.6
> > Severity: important
>
> Thanks. Please specify a version number next time.
>
It is 3.2.0-2-amd64.
> [...]
>
> > [ 175.461616] ath5k phy0: gain calibration timeout (2462MHz)
> > [ 175.462671] ADDRCONF(NETDEV_UP): wlan0: link is not ready
>
> Interesting. I'd never seen this message before, but a search of
> the linux-wireless yielded some hits.
>
> No great ideas from me. So:
>
> - please boot without the virtualbox drivers to rule them out as
> a cause
>
I did, but I got no success.
> - does unloading the acer_wmi driver help? (I wouldn't expect it
> to, but it seems worth a check.)
>
No, it does not help.
> - is "rfkill list" output as expected?
>
This is the output:
rfkill list
2: phy1: Wireless LAN
Soft blocked: no
Hard blocked: no
Seems ok for me.
> - you mentioned that with some earlier 3.2.y kernel you had similar
> problems, though there was a workaround. May I assume that
> earlier kernels worked fine? Of the kernels listed at
> <http://snapshot.debian.org/package/linux-2.6/>, which is the
> newest version that works without problems?
>
Yes, with kernel 3.1.X everything went fine. I did not downgrade at the moment,
as I need this machine all the time and can not do much experiences during the
day.
> - please file a separate bug for the laptop with different hardware.
>
Yes, I will do, when this problem is fixed and is still on the other machine. I
just wanted to mention it, as it might also be, there is still a problem
besides of ath5k.
> - does this always happen right away or only in some circumstances?
> (e.g., after a suspend/resume)
>
The problem is always happening.
> - can you reproduce this without using wicd? i.e.
>
> install iw and iproute
> stop wicd
> ip link set wlan0 up
> iw dev wlan0 connect -w <SSID> key d:0:<WEP key>
> dhclient wlan0
>
> with
>
> iw event -t
>
> running in another terminal to see if anything interesting happens.
>
Yes, same problem without wicd! My system is configured, that I also can run it
without wicd, but I am using wicd as it is more comfortable. But wicd indeed
is not really necessary.
> Vaguely reminiscent of https://bugzilla.kernel.org/show_bug.cgi?id=16436
> but that bug has a different regression range which makes it not a great
> match. So I look forward to some clues to narrow this down.
>
I checked the bugreport, and yes, those error messages are very similar to
mine, except mine are starting at boot time. So the reason for my bug can not
be powermanagement (I think)
> Hope that helps,
> Jonathan
I hope it, too. :))
Hans
Reply to: