Re: [Bugme-new] [Bug 12886] New: skge wake on lan

[Tomi Orava <tomimo@ncircle.nullnet.fi>]

Hi,

> Hm, are you sure the commit in question _is_ in 2.6.28.y?  I can't see it in
> there.  It is in the current Linus' tree, but not in v2.6.28 and surely not in
> -stable.  Do you use a distro kernel?

Ok, I've now verified that I did indeed mess up with the commit
mentioned previously. In other words the stable series 2.6.28.x
doesn't support the suspend feature on my Asus A7V880 at all.
The bios option is not respected at all and the command
'ethtool -s eth0 wol g' doesn't enable the suspend any better.

However, the commit taken from main branch actually makes also
the stable series to work with this motherboard. No additional
changes are needed, it just works out of the box.

Sorry about the mess up.

Regards,
Tomi Orava

>>> Also, please check if the appended patch helps (of course, without
>>> reverting
>>> the commit).
>>>
>>> ---
>>>  drivers/net/skge.c |    5 +++--
>>>  1 file changed, 3 insertions(+), 2 deletions(-)
>>>
>>> Index: linux-2.6/drivers/net/skge.c
>>> ===================================================================
>>> --- linux-2.6.orig/drivers/net/skge.c
>>> +++ linux-2.6/drivers/net/skge.c
>>> @@ -3856,8 +3856,9 @@ static struct net_device *skge_devinit(s
>>>  	skge->speed = -1;
>>>  	skge->advertising = skge_supported_modes(hw);
>>>
>>> -	if (device_may_wakeup(&hw->pdev->dev))
>>> -		skge->wol = wol_supported(hw) & WAKE_MAGIC;
>>> +	skge->wol = wol_supported(hw) & WAKE_MAGIC;
>>> +	if (skge->wol && device_can_wakeup(&hw->pdev->dev))
>>> +		device_set_wakeup_enable(&hw->pdev->dev, true);
>>>
>>>  	hw->dev[port] = dev;
>> Hmm, looks like this patch is against a 2.6.28.8 tree with the commit
>> reverted. I merged the couple lines as following.
>>
>> line 3859:
>>
>>         // if (pci_wake_enabled(hw->pdev))
>>         //      skge->wol = wol_supported(hw) & WAKE_MAGIC;
>>         skge->wol = wol_supported(hw) & WAKE_MAGIC;
>>         if (skge->wol && device_can_wakeup(&hw->pdev->dev))
>>                 device_set_wakeup_enable(&hw->pdev->dev, true);
>>
>> and this really seems to fix the problem. The system is able to wake
>> up via etherwake just fine.
> 
> Well, that's because commit 5177b3240a6608fc0c9c05cc32f4855c6540f8d5 changed
> the default from "wake-up enabled by default" to "wake-up disabled by default".
> The latter is safer IMO, but I'll prepare a (better) patch to change it back to
> the old behavior.
> 
> You should be able to set up WoL by running 'ethtool -s eth0 wol g' from your
> init scripts even without reverting the patch.  Please verify if that works.