Dave,
That's good information. The reason you see
the mouse problem with version 5.0.3 of our
server, and perhaps not with XFree, is because
our PS/2 code up through 5.0.3 used active
PS/2 probing, and the screwy clock really messed
that up. Because of that, we disabled PS/2
probing in the new 3D server, and I do have a
beta 5.0.4 with it disabled as well. If you
would like me to send that to you, just send
me your serial number. It sounds like you have
it fixed, but if you're not happy with your hack,
that will fix it as well. Any other registered
user of version 5.x.x is welcome to that offer
as well. If you see your mouse erratically jumping
to the top, right corner, you probably have that
problem. Thanks, Dave, for your input!
Bryan
dave madden wrote:
>
> Hi Bryan.
>
> =>From: Bryan Chambers <[EMAIL PROTECTED]>
> =>...
> =>The Accelerated-X server 5.0.3 has not been tested
> =>on and does not support 2.3.x kernels. Have you
> =>seen this problem on a supported 2.2.x kernel?
>
> Yes, I recently reinstalled the system with RH6.2, and it happened
> even with the stock 2.2.14-5.0 kernel. (The symptoms -- erratic mouse
> movement -- were the same, anyway. I don't know for sure that it was
> the 8253 time constant getting whacked.)
>
> =>Also,
> =>have you tried disabling one of your processors; I've
> =>seen SMP systems mess with the clock before.
>
> That's a good idea. [...save buffer, reboot nosmp ... continue] I'm
> now running on a single CPU, and the problem hasn't exhibited yet
> (kernel is 2.3.99pre9).
>
> =>Finally, did you not have this problem with XFree; is that why
> =>you claim it's related to our server?
>
> I haven't run XFree for a year or two; when I installed your server, I
> started having clock problems (you may remember my messages from a
> while back :-). At the time, I tried reverting to XFree and still had
> trouble, but I didn't know enough about the libraries to be sure I was
> running 100% XFree, and not some portions of your code through the
> shared objects. Besides, Xaccel was so much faster that I was
> motivated to fix the time problem in the kernel so that I could keep
> using it. (And I did, in fact, perpetrate a kernel hack that allowed
> me to run with Xaccel.)
>
> Anyway, I don't believe the problem is specifically with the Xaccel
> server, but only that the server drives the kernel and/or hardware in
> a way that causes the problem to exhibit. I mentioned it only because
> it's the one really unusual thing that might make the difference
> between an SMP guru being able to repro my bug or not. (Many people
> have Voodoo3 graphics cards, some have dual-proc boxes, a few have
> N440BX motherboards, but how many have all that and Xaccel, too?
> Probably only me!)
>
> So I'll live with my kernel hack for another few linux releases (it
> really works fine!) and then probably be inspired to fiddle with it
> again. Meantime, maybe somebody else will find this message in an
> archive and get the clue he needs to solve his own time problem.
>
> Regards,
> d.
-
Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/dmentre/smp-howto/
To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
