sio problem in -current (COM1)

[Cagle, John (ISS-Houston)]

Thanks!  It was an ACPI-related problem.  I disabled ACPI
(hint.acpi.0.disabled="1" in /boot/device.hints) and rebooted and now
both com ports show up properly as they did with FreeBSD 4.8.  (These
are just standard com ports, btw.)

I didn't realize ACPI was involved in legacy com port detection.  That's
weird.

Thanks,
John

> -----Original Message-----
> From: Bruce Evans [mailto:bde at zeta.org.au] 
> Sent: Wednesday, April 02, 2003 9:17 PM
> To: Cagle, John (ISS-Houston)
> Cc: freebsd-current at freebsd.org
> Subject: Re: sio problem in -current (COM1)
> 
> 
> On Wed, 2 Apr 2003, Cagle, John (ISS-Houston) wrote:
> 
> > I'm having a problem with -current on a ProLiant BL10e 
> blade server.  
> > On the blade server, we use a serial console on sio0/COM1.  
> This works 
> > perfectly with 4.8, but for some reason, the sio driver doesn't see 
> > COM1 at all, and assigns COM2 resources to sio0.  Any pointers to 
> > where I should look would be greatly appreciated.
> >
> > I've attached dmesg output for 4.8 and -current on this blade.
> 
> 4.8:
> > ...
> > config> di sio2
> 
> Maybe this helps.  sio2's irq is often the same as sio0's, 
> and this can sometimes cause problems.  Workarounds for some 
> of these problems have been broken and turned off since FreeBSD-2.
> 
> > ...
> > sio0 at port 0x3f8-0x3ff irq 4 flags 0x10 on isa0
> > sio0: type 16550A, console
> > sio1 at port 0x2f8-0x2ff irq 3 on isa0
> > sio1: type 16550A
> 
> -current:
> > ...
> > sio0 port 0x2f8-0x2ff irq 3 on acpi0
> > sio0: type 16550A, console
> 
> acpi sometimes does this for reasons that no one understands 
> AFAIK. Booting with -v should show the usual sio0 being 
> probed and something about why the probe failed, provided 
> sio0 is in hints.  The probe order may be relevant.
> 
> Bruce