On Sat, Jun 16, 2012 at 03:59:16AM +0200, Guillem Jover wrote:
> Hi!
>
> [ Ccing Lucas because I saw a related post which *might* be related,
> but it's not clear. I've not trimmed the mail for your convenience.
> <http://www.mail-archive.com/e1000-devel@lists.sourceforge.net/msg04477.html>]
>
> On Sat, 2012-06-16 at 02:32:28 +0530, Ritesh Raj Sarraf wrote:
> > severity 677638 normal
> > tags 677638 +moreinfo
> > thanks
> >
> > Setting severity of normal because:
> > * You have an option to disable.
>
> That's right, but only as long as you know what needs disabling, it
> took me a while to spot what the culprit was, because the battery/ac
> stuff is a bit hairy, there's at least acpid, laptop-mode-tools,
> pm-utils, and the kernel messing with this stuff. Suddenly getting the
> network to stop working is pretty mysterious given all those layers,
> and as such (as said before) even if the real problem is with the driver
> or the kernel PM settings or whatever, if laptop-mode-tools triggers
> this (when it could avoid it), then that's a “problem” with it.
>
> > * This problem is not commonly reported.
>
> It could be that other poeple might not have been able to spot what
> triggered it, it's really not obvious. Checking google for similar
> errors I've got on my dmesg, I find quite some people reporting
> similar stuff on random forums and mailing lists.
>
> > * Not reproducible on my machine (with the same driver)
>
> Well, if it can affect other users then I'd say it justifies the
> severity. :)
I can confirm the problem on my Thinkpad X200 with
00:19.0 Ethernet controller: Intel Corporation 82567LF Gigabit Network
Connection (rev 03)
Subsystem: Lenovo Device 20ee
Flags: bus master, fast devsel, latency 0, IRQ 43
Memory at f2600000 (32-bit, non-prefetchable) [size=128K]
Memory at f2624000 (32-bit, non-prefetchable) [size=4K]
I/O ports at 1820 [size=32]
Capabilities: <access denied>
Kernel driver in use: e1000e
It's been a persistent problem for some time. I initially suspected
a kernel regression since it started when I updated from Squeeze to
testing (about the time when 2.6.39 was current). Switching back to
the 2.6.32 from Squeeze fixed it, but I hadn't had the chance to track
this down properly until today, when I found this bug.
Setting BATT_THROTTLE_ETHERNET=0 in /etc/laptop-mode/conf.d/ethernet.conf
fixed wired ethernet for me.
I agree with Guillem that this should be fixed for Wheezy by setting
BATT_THROTTLE_ETHERNET=0 by default.
While a workaround is available it is very difficult to find (who suspects
laptop power saving settings when network-manager or the kernel are culprits
much more likely?). Also e1000e is quite a common chipset.
Cheers,
Moritz
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
