Bug#240656: xserver-xfree86: 5 second lock when switching from console to X with XFree 4.3
On Wed, 28 Apr 2004, Branden Robinson wrote:
> From: Branden Robinson <branden@debian.org>
> Subject: Re: Bug#240656: xserver-xfree86: 5 second lock when switching
> from console to X with XFree 4.3
>
> On Wed, Apr 28, 2004 at 05:16:10PM +0200, Chipzz wrote:
> > On Wed, 28 Apr 2004, Branden Robinson wrote:
> > > Can you guys do a couple of things for this bug?:
> > >
> > > 1) Report your gpm configuration by mailing your /etc/gpm.conf (I think
> > > that's what it's called -- I don't have it installed) to this bug.
> > >
> > > 2) Disable GPM and reconfigure XFree86 to talk directly to /dev/psaux or
> > > whatever port GPM was fastened to, and see if that causes the delay to
> > > go away.
> > >
> > > Thanks!
> >
> > This solves the problem partially for me, ie, the hang time goes down,
> > but does not completely disappear. It goes down from about 5 seconds to
> > about 1.5 seconds.
>
> Well, it's almost always going to take *some* amount of time to switch
> VT's. I've only seen something approaching lightning-fast VT switches
> on machines using fbcon and with video modes matched between the X
> server and virtual console.
>
> > Disabling gpm is really not an option for me though, as I am a quite
> > heavy console user.
>
> Okay. See below.
>
> > gpm.conf is attached.
> [...]
>
> Content-Description: gpm.conf
> [...]
> > device=/dev/psaux
> > responsiveness=
> > repeat_type=raw
^^^^^^^^^^^^^^^
> > type=imps2
> > append=""
> > sample_rate=
>
> I see you're repeating as "imps2". The following item was recently
Well, actually as "raw", which would amount to imps2.
> added to the xfree86 package NEWS.Debian file, and will appear in the
> next release. Can you please try out its advice?
Hmm, strange, I have apt-listchanges installed and never saw the notice.
Maybe I just looked over it.
> * GPM users only: those who use the the GPM repeater as the XFree86 X
> server's mouse device should note that the mouse driver in XFree86 4.3.0
> has been rewritten in such a way that using any protocol other than
> "IntelliMouse" on the XFree86 side and "ms3" on the GPM side does not work
> well. If you have configured GPM as a repeater, are using it with the
> XFree86 X server, and are using anything other than "ms3" as the repeating
> protocol, you will likely want to change it to "ms3". If your
> XF86Config-4 file is automatically handled (see the news entry for xfree86
> 4.2.1-11) and uses /dev/gpmdata as the port for the configured mouse, the
> protocol will automatically be migrated to "IntelliMouse" if necessary.
> Otherwise, you will likely want to edit the /etc/X11/XF86Config-4 file
> after installing this package to change any configured input devices using
> the "mouse" driver and "/dev/gpmdata" as the "Device" option to set the
> "Protocol" option to "IntelliMouse".
>
> If you go back to using /dev/gpmdata for your XFree86 input device, but
> have GPM repeat as "ms3" instead of "imps2", does the situation improve?
That worked. Maybe this should be better documented, as in the past I
tried various combinations and my previous configuration was the only
one that worked.
But still, I think the change is somewhat incorrect. A setting you expli-
citly specify gets overridden by X without any notice.
Anyway, thanks for the advice!
kr,
Chipzz AKA
Jan Van Buggenhout
--
------------------------------------------------------------------------
UNIX isn't dead - It just smells funny
Chipzz@ULYSSIS.Org
------------------------------------------------------------------------
Reply to: