Comment # 30 on bug 905988 from Benjamin Poirier

(In reply to Paul Tannington from comment #27)
[...]
> To try to eliminate my own (software) setup I booted, from a "power off"
> condition, with the "live" version of openSUSE 13.2
> (openSUSE-13.2-KDE-Live-x86_64.iso Last modified: Thu, 30 Oct 2014 15:16:32
> GMT). Ethtool reported wol as "ug". There was an immediate restart after
> suspend to ram, although for some reason this did not complete.
> 
> I repeated that using a USB stick on which I still had a copy of V13.1
> (openSUSE-13.1-KDE-Live-x86_64.iso Last modified: Wed, 06 Nov 2013 15:56:17
> GMT). Ethtool reported wol as also being "ug"! When suspended to ram the nic
> (as indicated by the connected router) had dropped down to 10M and was still
> active, but there was no (unwanted) restart. Manual resume also failed to
> complete, I'm guessing that has something to do with the live environment
> and can be ignored.

That is an interesting test. So, the desired suspend behavior does not have to
do with the way the controller is initialized at boot but rather with what is
done before suspend. Perhaps 13.1 had some script executed before suspend to
adjust wol settings?

I would still appreciate if you could test the 13.1 kernel with 13.2.

> I appreciate you efforts on this, but I don't see any other bug reports
> about this behaviour. 13.2 has been released for about a month now and I
> believe the realtek nic is fairly common still. Whilst I'm quite happy to
> provide information where able, please don't forget that I have a
> work-around for this problem coutesy of Takashi Iwai from Comment # 10 - I
> don't know that it is worth pursuing, but you are the judge of that :)

Yes, I'd like to find the root cause of this problem. I've asked Al to try and
find a machine with an r8169 controller so that we could try to reproduce the
problem as well.