2003-06-14 01:28:28

by Ryan C. Underwood

[permalink] [raw]
Subject: mga dualhead console + gpm = instant reboot


Hello,

I run the mga dualhead console. It works ok for the most part (some
strange behavior on the second head happens that can be noticed in e.g.
lynx when the cursor is blinking). However, if I move the gpm mouse on
the first head, switch to a console on the second head, move gpm mouse
again, then switch back to a console on the first head, moving the mouse
thereafter results in an instant reboot of the system.

Since there does not appear to be any kernel panic or oops, I am at a
loss how to track the problem down. Any ideas?

Thanks,
--
Ryan Underwood, <nemesis at icequake.net>, icq=10317253


2003-06-14 21:27:29

by Petr Vandrovec

[permalink] [raw]
Subject: Re: mga dualhead console + gpm = instant reboot

On Fri, Jun 13, 2003 at 08:42:12PM -0500, Ryan Underwood wrote:
> Hello,
>
> I run the mga dualhead console. It works ok for the most part (some
> strange behavior on the second head happens that can be noticed in e.g.
> lynx when the cursor is blinking). However, if I move the gpm mouse on
> the first head, switch to a console on the second head, move gpm mouse
> again, then switch back to a console on the first head, moving the mouse
> thereafter results in an instant reboot of the system.
>
> Since there does not appear to be any kernel panic or oops, I am at a
> loss how to track the problem down. Any ideas?

Kernel version? And if it is 2.4.x, did you boot with
'video=scrollback:0' ? If not, then please do so...
Petr

2003-06-14 22:36:14

by Ryan C. Underwood

[permalink] [raw]
Subject: Re: mga dualhead console + gpm = instant reboot

On Sat, Jun 14, 2003 at 11:41:14PM +0200, Petr Vandrovec wrote:
> On Fri, Jun 13, 2003 at 08:42:12PM -0500, Ryan Underwood wrote:
> > Hello,
> >
> > I run the mga dualhead console. It works ok for the most part (some
> > strange behavior on the second head happens that can be noticed in e.g.
> > lynx when the cursor is blinking). However, if I move the gpm mouse on
> > the first head, switch to a console on the second head, move gpm mouse
> > again, then switch back to a console on the first head, moving the mouse
> > thereafter results in an instant reboot of the system.
> >
> > Since there does not appear to be any kernel panic or oops, I am at a
> > loss how to track the problem down. Any ideas?
>
> Kernel version? And if it is 2.4.x, did you boot with
> 'video=scrollback:0' ? If not, then please do so...

Sorry, fresh 2.4.21, but it has happened on previous 2.4 kernels too. I
will look into the scrollback option. What does that accomplish with
respect to the mouse cursor?

thanks,
--
Ryan Underwood, <nemesis at icequake.net>, icq=10317253

2003-06-14 22:47:42

by Petr Vandrovec

[permalink] [raw]
Subject: Re: mga dualhead console + gpm = instant reboot

On Sat, Jun 14, 2003 at 05:50:01PM -0500, Ryan Underwood wrote:
> On Sat, Jun 14, 2003 at 11:41:14PM +0200, Petr Vandrovec wrote:
> > On Fri, Jun 13, 2003 at 08:42:12PM -0500, Ryan Underwood wrote:
> > > Hello,
> > >
> > > I run the mga dualhead console. It works ok for the most part (some
> > > strange behavior on the second head happens that can be noticed in e.g.
> > > lynx when the cursor is blinking). However, if I move the gpm mouse on
> > > the first head, switch to a console on the second head, move gpm mouse
> > > again, then switch back to a console on the first head, moving the mouse
> > > thereafter results in an instant reboot of the system.
> > >
> > > Since there does not appear to be any kernel panic or oops, I am at a
> > > loss how to track the problem down. Any ideas?
> >
> > Kernel version? And if it is 2.4.x, did you boot with
> > 'video=scrollback:0' ? If not, then please do so...
>
> Sorry, fresh 2.4.21, but it has happened on previous 2.4 kernels too. I
> will look into the scrollback option. What does that accomplish with
> respect to the mouse cursor?

It stops code from using software scrollback, which is common for both
heads, and which can corrupt memory if screen size on both /dev/fb* differs.

Maybe it wont help, but scrollback is known problem, so lets rule this
one out first.

I'll try to build some 2.4.x kernel on monday, but I assume that I would
notice such reboot during regular usage of dualhead system, as in the past
I was using 2.4.x...
Petr