Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 11 Dec 2012 15:36:16 -0800
From:      Adrian Chadd <adrian@freebsd.org>
To:        John Baldwin <jhb@freebsd.org>
Cc:        husyh@hush.com, freebsd-current@freebsd.org
Subject:   Re: ath0: unable to attach hardware
Message-ID:  <CAJ-VmonvJ4M5aJC9Mm5g%2B06Gfe=aoTg_p0NStj6gZtjkJ77Btw@mail.gmail.com>
In-Reply-To: <201212111549.49942.jhb@freebsd.org>
References:  <20121123213551.C2CB9E6739@smtp.hushmail.com> <201212101437.54825.jhb@freebsd.org> <CAJ-VmomM965-=QByhJZMPUn_PZAnSjafcm3cy3ojRaPc5fbWPg@mail.gmail.com> <201212111549.49942.jhb@freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On 11 December 2012 12:49, John Baldwin <jhb@freebsd.org> wrote:


> Look, it's up to you to look at more registers if you want to debug this
> further.  PCI says everything is ok, so the ball is in your court.

Right, that's why I've asked for those two above registers.

There are other things that could be wrong - eg, the device may
actually not have reset correctly.

This isn't the first time that someone's come to me with a "linux
works, freebsd doesn't" for an AR5212 era NIC. ath5k and FreeBSD do
the same thing at probe/attach time. I believe they do the same thing
during device power-on time too. There's some corner cases where the
chip doesn't reset right because the BIOS PCI bus reset code does
things in a brain dead manner (eg doing two PCI bus resets back to
back with not enough time in between for the MAC to settle.)

There may be PCI code differences in how Linux and FreeBSD does things
like "reset the PCI bus."



Adrian



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJ-VmonvJ4M5aJC9Mm5g%2B06Gfe=aoTg_p0NStj6gZtjkJ77Btw>