On Fri, 14 Aug 2015 11:26:50 +0200, Takashi Iwai wrote:
On Fri, 14 Aug 2015 10:30:05 +0200, Michael Chang wrote:
On Fri, Aug 14, 2015 at 09:54:45AM +0200, Takashi Iwai wrote:
On Fri, 14 Aug 2015 09:37:59 +0200, Andrei Borzenkov wrote:
On Fri, Aug 14, 2015 at 10:07 AM, Takashi Iwai <tiwai@suse.de> wrote:
On Fri, 14 Aug 2015 08:50:00 +0200, Michael Chang wrote:
On Thu, Aug 13, 2015 at 11:55:42PM +0200, Lars Müller wrote: > Hi, > > with 4.1.4-1-default s2disk works again with an ATI graphics controller. > > But on suspend grub2 isn't prepared. This worked once with grub2.
I think s2disk(uswsusp) would not run suspend/resume prepare hooks. You have to run higher level interface like pm-hibernate or systemctl hibernate that could have provided the pm hooks.
pm-utils have been already dropped, so there should be neither s2disk nor pm-hibernate command. The grub2-once thing is handled now in grub2 package itself.
Something needs to call grub2-once though during hibernation. It was done by pm-utils previously; I was sure there was script as part of systemd but I do not see it now.
Right, it was dropped in openSUSE:FACTORY/systemd changeset 215:
% osc rdiff -c215 openSUSE:Factory/systemd .... +Wed Feb 18 05:01:38 UTC 2015 - crrodriguez@opensuse.org .... +- systemd-sleep-grub: moved to the grub2 package where it belongs as a + suspend/resume hook (SR#286533) (drops prepare-suspend-to-disk.patch)
However, SR#286533 was declined and it's lost since then. WTF.
(And more badly, from the nature of BS, I can't resubmit or fix it because the source repo got removed)
IMO, we should have some mechanism to make some SR dependent on other SR for avoiding such failure.
IMHO, systemd should maintain the hooks if it breaks without it. It should be as self contained as possible to avoid adding requirment to it's environment.
Well, grub2 is no requirement for systemd or hibernation itself. It can work (in theory) with other boot loaders. So, having the systemd hook in grub2 appears more natural to me.
Now thinking on this again, I'm afraid that systemd may have a bug there. It doesn't call resume hooks if the write to /sys/power/state fails. So, if you set grubenv in the hook but the hibernation itself failed by kernel, the faulty grubenv is left. There seems no bug report, so I'll create it now. Takashi -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org