Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 06 Sep 2003 12:44:17 -0600
From:      James Earl <jdearl@telus.net>
To:        Kevin Oberman <oberman@es.net>
Cc:        freebsd-mobile@freebsd.org
Subject:   Re: Another ThinkPad R40 setup
Message-ID:  <20030906184417.GC752@h24-64-142-219.lb.shawcable.net>

next in thread | raw e-mail | index | archive | help
> > The first problem was with the fxp device, and "device timeouts." I
> > got past that by commenting out sio and pcmcia related devices in
> > the kernel.  I still need to go back and see if it was sio or
> > pcmcia stuff, or part of the pcmcia devices.
>
> Just disable sio1 in /boot/device.hints. No reason to disable stuff
> that you actually have! Just add hint.sio.1.disabled="1". That will
> free up an IRQ.

I gave that a try, and re-enabled the pcmcia devices and it looks like  
disabling sio devices doesn't have any effect in solving the device  
timeouts.

I also tried enabling only some of the pcmcia devices - non of the  
configurations solved the problem either.  The only thing that fixed  
the device timeouts for me so far is completely commenting out pcmcia  
devices:

#device         cbb                     # cardbus (yenta) bridge
#device         pcic                    # ExCA ISA and PCI bridges
#device         pccard                  # PC Card (16-bit) bus
#device         cardbus                 # CardBus (32-bit) bus

Here's the output from when I do have cbb, pccard, and cardbus built:

cbb0: <TI1510 PCI-CardBus Bridge> irq 11 at device 0.0 on pci2
cardbus0: <CardBus bus> on cbb0
pccard0: <16-bit PCCard bus> on cbb0
cbb0: [MPSAFE]
cbb0: Unsupported card type detected

(I noticed TI1510 isn't listed in the pccbb(4) manpage - I'm not sure  
what this means?)

Commenting out pcmcia devices on a notebook isn't really a great  
solution, although I'll probably never use a pcmcia card! :)

Is there any other way I can free an IRQ and still have pcmcia devices  
enabled?  I tried playing with some of the PCI BIOS settings with no  
luck.  I can change the irq's of what's listed below (dmesg output) in  
the BIOS setup, or change it to "disabled" or "automatic":

pcib0: slot 29 INTA is routed to irq 11
pcib0: slot 29 INTB is routed to irq 11
pcib0: slot 29 INTC is routed to irq 11
pcib0: slot 29 INTD is routed to irq 11
pcib0: slot 31 INTB is routed to irq 11
pcib0: slot 31 INTB is routed to irq 11
pcib0: slot 31 INTB is routed to irq 11

Actually I believe in the BIOS setup it's more like INTA to INTF or  
something.  All are set to irq 11 currently.

> > This looks like 5.1-Release. These should be fixed in CURRENT and
> > can be ignored in most cases during RELEASE. (It was a chicken/egg  
> > issue that Nate fixed doing some rather "odd" things to get some of  
> > the ACPI code to run before it would normally be started.)

I updated my system to 5.1-CURRENT and the error messages are gone.   
Thanks for the tip.  It doesn't look like ACPI suspend states work too  
hot on this notebook!  S5 will almost turn off the notebook, but stops  
at "Powering system off using ACPI."  S4 just freezes the system.

Is it mainly the manufacturers that are having a hard time keeping to  
the ACPI standards?



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030906184417.GC752>