[prev in list] [next in list] [prev in thread] [next in thread] 

List:       freebsd-stable
Subject:    Re: Clock issues and crash on resume on 10-Stable r263062M
From:       Kevin Oberman <rkoberman () gmail ! com>
Date:       2014-03-28 21:32:17
Message-ID: CAN6yY1t7+6RpKim_7nKGd6C=cx6AF=j13bHpEHmPq3q16ewHRQ () mail ! gmail ! com
[Download RAW message or body]

On Fri, Mar 28, 2014 at 2:39 AM, Aleksandr Rybalko <ray@ddteam.net> wrote:

> On Fri, 28 Mar 2014 00:11:49 -0700
> Kevin Oberman <rkoberman@gmail.com> wrote:
>
> > On Thu, Mar 27, 2014 at 11:53 PM, Aleksandr Rybalko <ray@ddteam.net>
> wrote:
> >
> > > On Thu, 27 Mar 2014 17:52:06 -0700
> > > Kevin Oberman <rkoberman@gmail.com> wrote:
> > >
> > > > On Thu, Mar 27, 2014 at 7:16 AM, Aleksandr Rybalko <ray@ddteam.net>
> > > > wrote:
> > > >
> > > > > On Wed, 19 Mar 2014 08:54:42 -0700
> > > > > Kevin Oberman <rkoberman@gmail.com> wrote:
> > > > >
> > > > > > On Wed, Mar 19, 2014 at 5:19 AM, Aleksandr Rybalko
> > > > > > <ray@ddteam.net>
> > > > > wrote:
> > > > > >
> > > > > > > On Tue, 18 Mar 2014 21:07:57 -0700
> > > > > > > Adrian Chadd <adrian@freebsd.org> wrote:
> > > > > > >
> > > > > > > > Well, changing the default is cool, but there's still some
> > > > > > > > odd bug there that could do with chasing down. :)
> > > > > > >
> > > > > > > Sorry, I currently can't work on investigation why GIANT lock
> > > > > > > is NULL sometime on resume. (IIRC it happen because I use
> > > > > > > callout_init w/o mtx, so callout try to lock using GIANT)
> > > > > > >
> > > > > > > But I will glad to get any help on that :)
> > > > > > >
> > > > > > >
> > > > > > Turned off vt switch and found a couple issues:
> > > > > >
> > > > > > 1. Audio was muted  (by pressing the mute button) when I
> > > > > > suspended the system. The suspend operation caused the system to
> > > > > > beep. When I resumed,
> > > > > it
> > > > > > did so again and I noticed that the audio was no longer muted.
> > > > > > This can
> > > > > be
> > > > > > an issue when in a meeting or in some other place where a loud
> > > > > > noise is
> > > > > not
> > > > > > desirable, but my be an issue with the Lenovo BIOS or EC.
> > > > > >
> > > > > > 2. When the system is resumed, the display is not refreshed and
> > > > > > is full
> > > > > of
> > > > > > artifacts.I moved to another desktop and back (I use gnome2) and
> > > > > > fixed everything except the panel which never gets redrawn. This
> > > > > > is possibly
> > > > > an X
> > > > > > issue (Intel 3000 graphics) or something with the KMS handling of
> > > > > > a
> > > > > resume.
> > > > > >
> > > > > > Probably neither of these is really a vt(4) issue, though, an I
> > > > > > really do appreciate having a working vty system after starting
> X.
> > > > > > --
> > > > > > R. Kevin Oberman, Network Engineer, Retired
> > > > > > E-mail: rkoberman@gmail.com
> > > > >
> > > > > Kevin,
> > > > >
> > > > > I was already describe possible culprits of Xorg screen puzzling
> > > > > after resume, there is two possible candidates:
> > > > > 1. mishandling of system PM events by Xorg.
> > > > > 2. something wrong inside drm2+drm/kms drivers.
> > > > >
> > > > > but things #1 in your list points more to mishandling of PM events.
> > > > >
> > > > > btw, crash with suspendswitch=1 looks like fixed :)
> > > > >
> > > > > Thanks!
> > > > >
> > > > > WBW
> > > > > --
> > > > > Aleksandr Rybalko <ray@ddteam.net>
> > > > >
> > > >
> > > >
> > > > Thanks for getting this running.
> > > >
> > > > FWIW, I realized that switching to a vty and back to X takes care of
> > > > he messed up display. Do you know when the crash was fixed? I am
> > > > running r263692 and it still crashes. (Typo in sysctl.conf.)
> > >
> > > it was in r263809
> > >
> > > >
> > > > --
> > > > R. Kevin Oberman, Network Engineer, Retired
> > > > E-mail: rkoberman@gmail.com
> > >
> >
> > Missed it by THAT much! (Possibly humor only recognized in American
> > culture.) Guess I need to rebuild my kernel!
>
> Maybe it depend on type of humor? :))))
>
> >
> > Thanks again! It's great to have vtys back.
> >
> > By the way, is there a way to increase the size of the scrollback buffer?
>
>
> Currently only by modify following row in the sys/dev/vt/vt.h:
> #define VBF_DEFAULT_HISTORY_SIZE        500
>
> >
> > --
> > > Aleksandr Rybalko <ray@ddteam.net>
> > >
> >
> >
> >
> > --
> > R. Kevin Oberman, Network Engineer, Retired
> > E-mail: rkoberman@gmail.com
>
>
> --
> Aleksandr Rybalko <ray@ddteam.net>
>

Thanks again, Aleksander!

I have manually added the fix to the latest r263878 10-Stable vt_core.c and
rebuilt the kernel. I have suspended and resumed repeatedly with
suspendswitch=1 with no problems at all. Also, with suspendswitch=1, the
audio retains mute state just fine. I have not increased the scrollback
buffer yet, but will soon. I'd love to see the scrollback size as a kenel
option for vt, just as it is for the old sc device. A  loadable would be
even better as it would allow increasing it to support a verbose boot
without rebuilding the kernel.

Looking forward to having the resume fix in 10!
-- 
R. Kevin Oberman, Network Engineer, Retired
E-mail: rkoberman@gmail.com
_______________________________________________
freebsd-stable@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscribe@freebsd.org"
[prev in list] [next in list] [prev in thread] [next in thread]