On 01/04/2018 06:18 AM, Wells, Roger K. wrote:
On 01/02/2018 01:18 PM, Al Stone wrote:
> On 12/26/2017 12:23 PM, Wells, Roger K. wrote:
>> Small inconvenience but new and annoying:
>> Machine is Thinkpad x260
>> uname -a: Linux rwells-x260 4.14.7-300.fc27.x86_64 #1 SMP Mon Dec 18 16:06:12
>> UTC 2017 x86_64 x86_64 x86_64 GNU/Linux
>> Desktop: Gnome 3.26.4-1.fc27.x86_64
>>
>> When the lid is opened on the suspended machine the screen saver appears and the
>> clock proceeds to update until
>> the enter key is hit. Then the clock stops updating for 27 seconds followed by
>> the login entry dialog appearing.
>> Then everything is back to normal.
>> This began right after the upgrade from F26 to F27 and has persisted through one
>> or two subsequent routine updates.
>> There are inconsistencies, sometimes it only does it when there is only wireless
>> connections and no wired options.
>> Sometimes it doesn't do it at all, no pattern here.
>> Sometimes the 27 second delay is much shorter but this is quite rare.
>> I have been running Fedora/Gnome for years and have never seen this.
>>
>> Any thoughts or things to try to help pin it down will be appreciated.
>>
>> thanks
> There have been a lot of recent changes in ACPI code for suspend and hibernate;
> it's always possible that something got tweaked in a recent kernel that could
> affect this particular model of machine.
>
> That being said, the 27 second delay sounds like the laptop hibernated (not
> suspended); if you waited a variable length of time to open the lid again, it
> might explain some of the variation -- sometimes it suspended, sometimes it was
> caught before hibernation was complete, sometime it was caught after it had
> fully hibernated. The other things that occur to me are that perhaps hibernate
> was not working before for this model of laptop and now it does; or, the power
> settings changed defaults, or did not retain settings when updating; or, some
> of the recent changes for lid notification aren't quite right for this laptop
> (there have been a few cases of that). Those are some of the places where I
> would start looking, at least...
>
It turns out that this delay only occurs when the suspend happened when an
external filesystem is mounted.
In this case a cifs mount, and IIRC there have been some issues related to
version changes that necessitated specifying "vers=1.0" in the fstab file.
I will do some experiments and report back if anything interesting develops.
Interesting. Could there be an ordering dependency somewhere on resume? E.g.,
I need to start file system but can't do that until network is ready but the
info I need to start the network is on the file system ... some weirdness like
that?
--
ciao,
al
-----------------------------------
Al Stone
Software Engineer
Red Hat, Inc.
ahs3(a)redhat.com
-----------------------------------