In message <20051014103128.GA1187@asim.lip6.fr>, Manuel Bouyer writes:
> On Fri, Oct 14, 2005 at 11:52:01AM +0200, Igor Sobrado wrote:
> > 
> > But now there are some "lost interrupt" messages related with
> > piixide0:1:0.
> 
> So the interrupt for the second channel isn't working any more. Maybe
> ACPI_PCI_FIXUP cause problems on this system ...

Indeed, ACPI_PCI_FIXUP can be the source of all these interrupt
related problems...

> > I2O is missing from the output of dmesg yet...
> 
> Did you try MPACPI_SCANPCI ? If the I2O devices are not in the PCI config
> space nor in the ACPI tables, I can't see how an OS could find them.
> Is there something showing up in BIOS ?

Done... it is not working either.  This is my current GENERIC.local file:

-----------------------------------
#	$NetBSD: GENERIC.local,v 1.1 1996/09/09 16:37:08 mycroft Exp $
#
#	GENERIC.local -- local additions to the GENERIC configuration
#

options 	CONSDEVNAME="\"com\"",CONADDR=0x2f8,CONSPEED=9600
options 	CONS_OVERRIDE	# Always use above! independent of boot info

acpi0 		at mainbus0
#options 	ACPI_PCI_FIXUP		# PCI interrupt routing via ACPI
options 	ACPI_ACTIVATE_DEV	# If set, activate inactive devices
#options 	ACPICA_PEDANTIC		# force strict conformance to the Spec.
#options	ACPI_DISABLE_ON_POWEROFF	# disable acpi on power off

acpiacad* 	at acpi?		# ACPI AC Adapter
#acpibat* 	at acpi?		# ACPI Battery
acpibut* 	at acpi?		# ACPI Button
acpiec* 	at acpi?		# ACPI Embedded Controller
#acpilid* 	at acpi?		# ACPI Lid Switch
acpitz* 	at acpi?		# ACPI Thermal Zone

#com* 		at acpi?		# Serial communications interface
#fdc* 		at acpi?		# Floppy disk controller
#lpt* 		at acpi?		# Parallel port
#npx*		at acpi?		# Math coprocessor
#pckbc*  	at acpi?		# PC keyboard controller
#wss* 		at acpi?		# NeoMagic 256AV in wss mode

ioapic* at mainbus?
options 	MPACPI		# configure CPUs and APICs using ACPI
options 	MPACPI_SCANPCI	# find PCI roots using MPACPI
-----------------------------------

(no ACPI_PCI_FIXUP, MPACPI_SCANPCI enabled)... same results
(lost interrupt messages)...  :-(

I suspect that there are some problems related with the machine
itself.  Perhaps it has not implemented ACPI as it should.  I asked
on a Dell community forum about I2O support.  From the specs I see
that this server is "I2O ready", but I am not sure what "ready" means.
Does it include I2O processors?  Are these processors an optional
component?  There is not a lot of information on the documentation
provided by Dell.  I will post an update here if someone replies to my
post in the community forum.  (no replies yet)

On the other hand, envstat(8) does not provide information from
environmental sensors (sensors not found...).  I never had these
problems with NetBSD on my laptop (that has a motherboard with
nearly the same components).

I will try an APM enabled kernel using this GENERIC.local file:

-----------------------------------
#	$NetBSD: GENERIC.local,v 1.1 1996/09/09 16:37:08 mycroft Exp $
#
#	GENERIC.local -- local additions to the GENERIC configuration
#

options 	CONSDEVNAME="\"com\"",CONADDR=0x2f8,CONSPEED=9600
options 	CONS_OVERRIDE	# Always use above! independent of boot info

apm0	at mainbus0			# Advanced power management

# Tuning for power management, see apm(4) for more details.
#options 	APM_NO_IDLE		# Don't call BIOS CPU idle function
#options 	APM_V10_ONLY		# Use only the APM 1.0 calls
#options 	APM_NO_POWEROFF		# Don't power off on halt(8)
#options 	APM_POWER_PRINT		# Print stats on the console
#options 	APM_DISABLE_INTERRUPTS=0 # Don't disable interrupts
-----------------------------------

It would be great supporting some ACPI features on that server:

  - environmental monitoring can be really useful.

  - power management will help saving energy when the machine is idle
    (I believe that APM will do a good work in this matter too)

  - ACPI buttons can help turning off this machine (it is a headless
    1U rack computer).

I will wait until a more recent ACPI CA subsystem is available in
a stable release of NetBSD (it will be available very soon!) and try
again.  But I am having too many problems to believe that it is
a consequence of bugs in current ACPI CA.

As soon as I am in front of that machine I will connect my laptop to
its serial port to play with BIOS setup.  But the system documentation
does not show power management or I2O related entries in the BIOS setup
options.

Thanks a lot for the time you spend on this matter on the last days!

Have a nice day,

Igor.