On Mon, Jun 29, 1998 at 12:08:07AM -0400, Alex deVries wrote:
> Alright. So I went through to see if I could locate anywhere in the
> kernel where there might be problems with psaux. I recompiled, stuck some
> printk's in, and I got: Nothing. Nadda. Zip. The whole machine hangs
> (including num lock).
> When I do a 'cat /dev/psaux', the printk's prove that the open works fine.
> Then in the first read, it attempts to do a schedule(). So far, no
> problem, the system is still up. But it never comes out of the
> schedule(), and a subsequent mouse input hangs the system (or the
> keyboard, anyway).
Mouse interrupts where looping. Fixed, gpm works now.
There I small uglyness in gpm when using it with certain types of mouse
like PS/2 mice which are available with two or three buttons. Until
the optional third middle button has been used for the first time gpm
doesn't know that it's using a three button mouse. So the right button
changes it's meaning when the middle one has been pressed. Just a
small thing and it affects all architectures anyway.