On Wed 2015-05-27 13:43:49 -0400, Ritesh Raj Sarraf wrote:
> On Friday 22 May 2015 12:16 AM, Daniel Kahn Gillmor wrote:
>> In practice, it looks like the manpage is correct, and the comments in
>> ethernet.conf are wrong.
>>
>> This is too bad, because it would be nice to avoid disabling the link
>> when the carrier is still active.
> IIRC, the configuration comments are correct. This change was proposed
> by someone later, and the justification was similar. It should be in the
> git commit logs.
>
>> Also, with this flag set (as it is by default, i think), the ethernet
>> does not get re-enabled after moving back from Battery to AC power.
> By deafult, we wouldn't disable the eth device.
>
> Regarding the behavior, I'll look into this later. Meanwhile, if you
> root-cause it, please do comment on this bug.
>
>> Perhaps there's a cleanup step that needs to happen as well here?
> I need to verify first. And my current laptop has no eth devices. But
> I'll be able to figure out. Just need some time. My gut feel is that the
> manpage is outdated.
Ok, looking into it a bit further, it's even more complicated by a
couple other options.
DISABLE_ETHERNET_ON_BATTERY=1 is supposed to re-enable the ethernet when
the AC is plugged back in.
But if ENABLE_LAPTOP_MODE_ON_AC=0, then i think the
/usr/share/laptop-mode-tools/modules/ethernet never gets run at all
after the power has been plugged back in, so it doesn't know to bring
the NIC back up.
Also, while the test for active carrier appears to actually work (as
seen by set -x within .../modules/ethernet, it avoids bringing down eth0
if there is no NO-CARRIER present from "ip link show eth0"), something
else (also in laptop-mode-tools, because it doesn't happen when the
package is purged) is bringing down the NIC.
So this is still pretty confusing to me, and it looks like there is at
least one bug besides the documentation.
--dkg
--
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
