Ted Lemon <mellon@hoffman.vix.com> wrote:

>There is an actual pms driver fix that solves problems on some
>machines.   However, the thing that usually gets me is if I make the
>mistake of running a getty on the console _and_ running xdm.   Both
>processes wind up fighting for keyboard input, which inevitably
>results in heinous lossage.

I've seen the keyboard freeze problem going back to the early days of 1.2 
when I started running NetBSD/i386.  I mentioned it on this list back 
then and received a suggestion of not starting getty on the terminal 
where xdm runs.  Recently, today in fact, I installed 1.3 BETA from 
scratch on my Gateway system at work where I'd seen this problem 
previously.  I've replaced the mouse with a serial Logitech to eliminate 
the problems with the pms driver, some of which are screwing up my laptop 
running 1.3 BETA right now.

For X I snagged the 3.3.1 copy from Walnut Creek.  It fixes a lot of 
problems that I seemed to be having with the Mach64 card on my Gateway.  
One of the interesting things it seems to do is kill the getty process 
when you invoke X, and restart it when X exits.  Unfortunately the 
problem with the keyboard lockup is still there, and I've tried starting 
xdm both manually and automatically.  On my system I can get it to lockup 
in a matter of minutes. Killing xdm didn't seem to bring things back as I 
recall, but I may have missed that magic combination.

Does anyone have another idea about what this might be caused by?

Also, the pms driver in -current acts strange on startup on my laptop.  
The first time the device is accessed it refuses to respond to the mouse 
movements.  But the second time everything is A-OK.  On my laptop the 
track pointer is supposed to be disabled when the external mouse is 
plugged in and that's the way it works under Windows-95.  Under NetBSD it 
seems that either the external mouse is randomly being disabled as the 
track pointer is never disabled.

-bob