Re: ACPI/APM suspend errors related to yds

[David Young <dyoung%pobox.com@localhost>]

On Thu, Jul 31, 2008 at 01:48:37PM +0200, Bernd Ernesti wrote:
> On Sun, Jul 27, 2008 at 07:11:51PM +0200, Bernd Ernesti wrote:
> > On Sun, Jul 27, 2008 at 12:47:55PM -0400, Jared D. McNeill wrote:
> > > Bernd Ernesti wrote:
> > > >Hi,
> > > >
> > > >I need a kernel without any ACPI support, because ACPI is broken for me
> > > >on a specific notebook.
> > > 
> > > Did you raise a PR on it?
> > 
> > No, because I don't expect that it can be fixed, because of the age of the 
> > BIOS,
> > which is 8 years old. IMHO the ACPI implmentation is too broken to be 
> > fixed, but
> > I would like to be proven wrong.
> 
> I just tried a 4.99.71 and it seems to work better.
> Now I remember my problem which was related to suspend to disk.
> 
> > I need apm too for suspend to disk too.
> 
> Fn + F12 didn't supsend with ACPI. Nothing happens so I tried it without
> ACPI and with apm and the first message was this (both devicies are attached
> to yds0):
>   Devices without power management support: opl0 mpu0
> 
> Then I got some BIOS message:
> 
> PCI Parity Error On Bus/Device/Function 0048h
> PCI System Error On Bus/Device/Function 0060h
> PCI Parity Error On Bus/Device/Function 0060h

Taking a guess based on the width of bus/device/function in "PCI
configuration mode 1"---see from pci_make_tag(9)---I think that we
extract the bus/device/function from the 16-bit hex number x in this way,

bus = (x & 0xff00) >> 8
device = (x & 0x00f8) >> 3
function = x & 0x0007

That yields PCI devices 9 and 12, yds0 and cbb0, respectively.  Perhaps
the APM BIOS thinks that it can read/write registers on those devices,
but it cannot because NetBSD has powered them down?  That may confuse
the BIOS.

Dave

-- 
David Young             OJC Technologies
dyoung%ojctech.com@localhost      Urbana, IL * (217) 278-3933 ext 24