Re: ACPI call for testing

[David Huang <khym%azeotrope.org@localhost>]

On Nov 10, 2009, at 4:40 PM, Matthew Mondor wrote:

> On Tue, 10 Nov 2009 22:49:30 +0100
> Geert Hendrickx <ghen%telenet.be@localhost> wrote:
>
> > may remember helping me debug the problem).  It only works when  
> > disabling
> > either one of MP, ehci, or acpi, and I have used the latter  
> > workaround for
> > a long time now.  Your patched kernel boots, although it seems much  
> > slower.
> > Also keyboard repeats (when keeping a key presed) are much slower.   
> > This is
> > an USB keyboard connected via ehci.
>
> Is the system performance really affected or is it a keyboard input
> latency issue?  Also, does tweaking the software repeat rate allow you
> to have a decent result (using wsconsctl for console and xset for  
> X11)?


I have what may be a related problem on my Tyan K8WE system: An amd64 - 
current kernel with ACPI and MP (is disabling that even an option  
these days? I know it's not on i386; never checked on amd64) works  
fine. However, if I add ehci, it seems to work fine until I press a  
key on the PS/2 keyboard. Remotely ssh-ing in and doing things seems  
fine, a "systat vm" shows a usual amount of interrupts. However, once  
I press a key on the keyboard, the system gets incredibly slow. Even  
kernel output looks like I'm connected with a 300 baud modem. E.g., If  
I Ctrl-Alt-Esc to break into DDB, then do a "ps", I can see it output  
the characters one at a time. I have no idea what the interrupt rate  
is when it gets slow; it's slow enough that I can't log in on the  
console (maybe I could if I were willing to wait for a really long  
time), and a ssh session into the machine is effectively hung too.

--
Name: Dave Huang         |  Mammal, mammal / their names are called /
INet: khym%azeotrope.org@localhost |  they raise a paw / the bat, the cat /
FurryMUCK: Dahan         |  dolphin and dog / koala bear and hog -- TMBG
Dahan: Hani G Y+C 33 Y++ L+++ W- C++ T++ A+ E+ S++ V++ F- Q+++ P+ B+ PA 
+ PL++