[gpm] GPM Problems - more information

[Troy Schultz]

On Thu, 2002-04-04 at 20:22, Peter Berg Larsen wrote:
> On 4 Apr 2002, Troy Schultz wrote:
> 
> > I have tried running gpm using the ps2 protocol and still have the same
> > problems when switching between console and X.  Also using the -D
> > options does not seem to change the behavior in any way.  
> 
> Good (or bad) then it is probably not related to synaptic pad. Do you have
> access to an external mouse? What happens if you use that?
> 
I set gpm to use ps2 protocol but when I connected external mouse I lost
the touchpad and the mouse did nothing.  I had to disconnect the mouse
and restart gpm to get the touchpad working.  The mouse I used may be
the problem, I will try to get a better mouse and try things out and let
you know what happens.

> 
> > I am able to ssh into the notebook from an external computer and can
> > still do many things, but the only way I have found to get X running
> > properly again is to restart the computer.  
> 
> Does this mean you can kill X, but it does not help?
> 
If I manage to kill the X process and the Xinit process I still cannot
start another X session.  I have done some more testing and found the
problem may be related to the Radeon driver in XFree86 version 4.2.0. 
After killing the hung X server if I try to start another X session I
get an error from X indicating "(EE) RADEON(0): Idle timed out,
resetting engine" and this just keeps repeating.  After this I have to
restart the computer.

Interestingly everytime the X session hangs it is part way through
redrawing the screen.

> > If I try to force a new startX I get an error indicating that there is
> > already a server running on display 0.
> 
> Yes, try startx -- :1, this will start on display 1 (at ctrl-alt-f8).
After I leave the X session running on display 0 to enter console mode I
am able to start another X session on display 1 fine.  I can do
everything I want on that session, including change to console and back,
and everything is fine.  However if I change to the X session on display
0 at any time it hangs and I lose keyboard and mouse and am unable to do
anything else but ssh into the notebook and restart.

> 
> 
> > I went back through my logs and here is what was being dumped to
> > /var/log/messages when the mouse stopped responding after a screen
> > saver;
> 
> > 08 AA 55 C0 00 00 03 AA 55 C0 00 00 00 AA 55 C0 00 00 28 AA 55 C0
> 
> hmm, hmm. This is not from any protocol I know.
At this point gpm was configured to use the synps2 protocol and was
working fine until after the screen saver.  I went back through my log
files to when this happened with gpm 1.19.6 and found the same type of
data repeated.

> > If you want to review some of the information I have gathered so far
> > from my notebook it can be found on my personal web site at
> > http://www.schultzcentral.org/notebook/notebook.html
> 
> Does the 4-way button work with the patch? when using the button is
> pressure then reported as 0? how about fingersize?
> 
Using gpm 1.20.0 and your latest patches the 4-way scroll button
responds, but does not provide anything usefull.  Here is what I get;

- Press UP, mouse pointer moves right at a slow speed for approx. 2
seconds
- Press RIGHT, mouse pointer moves up at a faster speed for approx. 2
seconds
- Press DOWN, mouse pointer moves up at a slow speed for approx. 2
seconds
- Press LEFT, mouse pointer moves right at a faster speed for approx. 2
seconds

Prior to using gpm 1.20.0 and your patches I was using gpm 1.19.6 with
Daniel Wagenaar's hack from here
http://www.its.caltech.edu/%7Ewagenaar/compaq_linux/gpm_synaptics_hack.html

His hack was to use the 4-way scroll button on the Compaq to act as
normal mouse buttons.  He used LEFT for the left button, UP for the
middle button, and RIGHT for the right button.  His hack also provided a
bit better milti-finger detection

> 
> Peter

Best Regards
- Troy Schultz