Bug#779412: /sys/power/pm_async defaulting to 1 : unreliable resume event execution

[Chris]

Am Tue, 28 Apr 2015 22:46:50 +0530
schrieb Ritesh Raj Sarraf <rrs at debian.org>:

> You should have reported that on the LMT bug.

It was archived now, and does not accept new messages.

(Fullqoting your message here now to the new udev bug,
I'd suggest to concentrate all further related info there as well.)






> AI to self: Check on when the resumption of userspace is granted.
> 
> @Ondrej: That's interesting. I get similar link reset on my SATA HDD,
> but I never paid attention if the power savings settings were
> changed. Actually so did powertop, which ideally should have detected
> it even if I did not.
> 
> On Tue, Apr 28, 2015 at 10:17 PM, Ondřej Grover
> <ondrej.grover at gmail.com> wrote:
> 
> > Hi,
> >
> > I'd like to point out that my usage of echo 0 > ... pm_async in
> > 780956 was related more to an apparent race issue in the HDD driver
> > than a udev event. The udev trigger works correctly on my system,
> > but with async pm often after it was triggered and applied
> > requested settings on resume the HDD driver suddenly timed out and
> > the SATA link was reset and lost the settings state.
> >
> > Kind regards,
> > Ondřej Grover
> >
> >
> > On Tue, Apr 28, 2015 at 6:00 PM, Chris <email.bug at arcor.de> wrote:
> >
> >>
> >> There is now a specific spin-off from the closed
> >> https://bugs.debian.org/780956 (laptop-mode-tools hooks not
> >> triggerd) in the udev binary package: (reassing to src:systemd did
> >> not succeed)
> >>
> >>
> >> To ensure packages have a reliable resume event hook available,
> >> udev may echo 0 > /sys/power/pm_async, or some device blacklist
> >> may be needed (in the kernel?) to disable async pm?.
> >> https://bugs.debian.org/783638
> >> (original reporters may subscribe)
> >>
> >> The relevant "trigger on resume" mechanism from udev seems to be:
> >> ACTION=="add|remove", SUBSYSTEM=="machinecheck", RUN+="lmt-udev
> >> auto force"
> >>
> >
> >
> 
>