acpi and acpid

Dagmar d'Surreal dagmar.wants at nospam.com
Fri Apr 30 12:39:48 PDT 2004


On Fri, 2004-04-30 at 04:16, Declan Moriarty wrote:
> Duane Cox enlightened us
> > Hello
> > 
> > I compiled ACPI into the kernel as well as the button and other various features.  I also compiled the acpid, following the hints and have it setup.
> > 
> > But when I hit the power button on this server and the other server (different hardware) neither one shuts down.
> > 
> > My question is, is there a file in /proc/acpi that I can monitor for the event change when I press the power button?
> > 
> > At this point Im not sure the ACPI is working.  Even though they are new 2004 servers and ACPI is detected and enabled on bootup of the kernel....
> > 
> 
> This could be hardware.
> 
> The VIA chipset ACPI is a disaster area. I have Via chipset with a
> recent m/b and everything is assigned irq10 or irq11 :-(.

Pshaw.  If you think that's bad, you should see how poorly it's handled
under Win32.  I'm constantly plagued with lame/gimpy performance under
Win32 because it does such an awful job of managing the IRQ sharing.  At
least under Linux it's handled with relative efficiency.

If you don't mind opening your case and looking inside to see which
cards are in which slots, you can at least usually assign specific
interrupts to individual PCI slots in the motherboard BIOS, and that
will separate about 3/4 of the little punters onto their own
interrupts... it's hard to get them all separated because much of the
time the first PCI slot shares with the AGP port, or PCI slots 5 and 6
share with PCI slots 2 and 3.  Hardcore gamers will be very familiar
with this song and dance, but it's basically part of what makes the
difference between a machine made by someone like Dell or Alienware, and
a machine where some guy just stuffed all the PCI cards in there without
thinking about it.
-- 
The email address above is phony because my penis is already large enough, kthx. 
              AIM: evilDagmar  Jabber: evilDagmar at jabber.org




More information about the blfs-support mailing list