On Thu, 2015-01-15 at 16:18 +0100, Ondrej Zary wrote: > On Thursday 15 January 2015, Jeff Kirsher wrote: > > On Thu, 2015-01-15 at 14:40 +0100, Ondrej Zary wrote: > > > On Thursday 13 November 2014, Jeff Kirsher wrote: > > > > On Wed, 2014-11-12 at 18:18 -0500, David Miller wrote: > > > > > From: Ondrej Zary <li...@rainbow-software.org> > > > > > Date: Wed, 12 Nov 2014 23:47:25 +0100 > > > > > > > > > > > Enabling WoL on some Toshiba laptops (such as Portege R100) > > > > > > causes > > > > > > > > > battery drain after shutdown (WoL is active even on battery). > > > > > > These > > > > > > > > > laptops have the WoL bit set in EEPROM ID, causing e100 driver > > > > > > to > > > > > > > > > enable WoL by default. > > > > > > > > > > > > Check subsystem vendor ID and if it's Toshiba, don't enable WoL > > > > > > by > > > > > > > > > default from EEPROM settings. > > > > > > > > > > > > Fixes > > > > > > https://bugs.launchpad.net/ubuntu/+source/linux/+bug/110784 > > > > > > > > > Signed-off-by: Ondrej Zary <li...@rainbow-software.org> > > > > > > > > > > Jeff, are you gonna pick this up? > > > > > > > > Yes, sorry I did not catch it earlier. > > > > > > What happened to this patch? I don't see it in net.git or > > > net-next.git > > > (checked both davem's and jkirsher's) > > > > Sorry, I thought I had replied with a NAK on this patch after further > > review of the changes. > > > > We don't fix BIOS issues in the driver especially regarding feature > > enablement like WoL. We would end up with dozens of these kinds of fixes > > if we to allow this. > > > > You should go back to the OEM and ask for a BIOS update to resolve this or > > configure udev so that ethtool disables WoL. > > This is not a BIOS bug. When the machine is powered off in BIOS (or GRUB), > everything is OK. > > The bug is that e100 driver enables WoL based on some bit in EEPROM that > happens to be set on at least some Toshiba laptops and user has no way to > change it.
Yes, the EEPROM can be modified/updated through the BIOS update I suggested earlier. So again, a BIOS issue. OR you can configure udev so that ethtool disables WoL if you do not want to pursue a EEPROM update through a BIOS update. > Windows driver does not do this. Other Linux ethernet drivers > don't do this. When user wants WoL, (s)he enables it in BIOS and OS. Maybe > this (mis)feature should be removed from the driver. >
signature.asc
Description: This is a digitally signed message part
