Re: Continued problems with Aironet device and 5.2-RC

> From: James Earl <james_at_icionline.ca>
> Date: Mon, 22 Dec 2003 14:02:58 -0700
> Sender: owner-freebsd-current_at_freebsd.org
> 
> On Mon, 2003-12-22 at 11:21, M. Warner Losh wrote:
> > In message: <1072116630.576.4.camel_at_chero>
> >             James Earl <james_at_icionline.ca> writes:
> > : On Mon, 2003-12-22 at 02:30, M. Warner Losh wrote:
> > : > In message: <20031222092127.GA2355_at_cactus.homeunix.org>
> > : >             jqdkf_at_army.com writes:
> > : > : On Mon, Dec 22, 2003 at 01:20:14AM -0700, M. Warner Losh wrote:
> > : > : > In message: <20031222053002.GA872_at_cactus.homeunix.org>
> > : > : >             jqdkf_at_army.com writes:
> > : > : > : cbb0: Unsupported card type detected
> > : > : > 
> > : > : > This is typically a symptom of the resource allocation bugs I've been
> > : > : > fighting...
> > : > : > 
> > : > : 
> > : > : So are you suggesting that I should adjust the resource by myself, like
> > : > : irq numbers, for now?
> > : > 
> > : > I have the unhelpful advice "make sure two devices aren't using the
> > : > same address."  You'll likely have to decode all the BARs to make sure
> > : > that this is the case.
> > : 
> > : So, would this be the reason why I haven't been able to get my an0
> > : device to appear on it's own in the vmstat -i output?  It seems the
> > : uhci2 device follows the an0 device to whatever irq I assign to it (an0)
> > : in the bios setup.
> > 
> > an0 will share interrupts with the CardBus bridge and whatever it
> > shares with.
> 
> Do you know if my fxp0 device would also share interrupts with the
> cardbus bridge also?
> 
> Would you say my Aironet problems sound related to the resource
> allocation bugs that you're fighting?

As I have seen it, the problems are not generally IRQ related, as far as
I can tell.

As to what devices share interrupts, that depends on BIOS configuration
to some extent. cbb0, cbb1, and any devices connected to those will
always share an irq. On many systems other devices also share this
IRQ. On my T30, the fxp0 and wi0 devices both share this interrupt.

Most of the resource issues revolve around the addresses to use for I/O
devices. This is partly addressed by start_memory. The
allow_unsupported_io_range tunable also get into a different part of
this issue.
-- 
R. Kevin Oberman, Network Engineer
Energy Sciences Network (ESnet)
Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
E-mail: oberman_at_es.net			Phone: +1 510 486-8634