linux-mips
[Top] [All Lists]

Tulip driver broken on Cobalt RaQ1 in 2.6

To: linux-mips@linux-mips.org
Subject: Tulip driver broken on Cobalt RaQ1 in 2.6
From: Martin Michlmayr <tbm@cyrius.com>
Date: Thu, 23 Aug 2007 22:37:57 +0200
Original-recipient: rfc822;linux-mips@linux-mips.org
Sender: linux-mips-bounce@linux-mips.org
User-agent: Mutt/1.5.13 (2006-08-11)
We have Debian users who happily used 2.4.27 on their Cobalt Raq1 and
Qube 2700.  However, since we moved to 2.6 these machines stopped
working.  I found out that the network driver (tulip) is no longer
working on these machines.  Today I tried to track down when this
started to happen but I couldn't find a 2.6 release where it actually
worked.

The 2.4.27 release we have is based on Peter Horton's patches from
http://www.colonel-panic.org/cobalt-mips/  Today I tested current git,
and 2.6.18 (which work out of the box), as well as 2.6.12-rc2 and
2.6.16-rc1 with Peter's patches.  In all of these releases, network
would work fine on a RaQ2, but not on a RaQ1.  I'm not sure what
information to report because I found nothing obvious.  In 2.4.27, we
get:

PCI: Enabling device 00:07.0 (0045 -> 0047)
tulip0: Old format EEPROM on 'Cobalt Microserver' board.  Using substitute 
media control info.
tulip0:  EEPROM default media type Autosense.
tulip0:  Index #0 - Media MII (#11) described by a 21142 MII PHY (3) block.
tulip0:  MII transceiver #1 config 1000 status 7809 advertising 01e1.
eth0: Digital DS21143 Tulip rev 65 at 0x100000, 00:10:E0:00:27:5C, IRQ 4.

wheras 2.6.16-rc1 has:

PCI: Enabling device 0000:00:07.0 (0041 -> 0043)
tulip0: Old format EEPROM on 'Cobalt Microserver' board.  Using substitute 
media control info.
tulip0:  EEPROM default media type Autosense.
tulip0:  Index #0 - Media MII (#11) described by a 21142 MII PHY (3) block.
tulip0:  MII transceiver #1 config 1000 status 7809 advertising 01e1.
eth0: Digital DS21143 Tulip rev 65 at b0001000, 00:10:E0:00:27:5C, IRQ 20.

The address is different but I doubt this makes a difference because the
RaQ2 shows the same difference and here networking works.  I also noticed
the following in the boot logs of the RaQ1 with 2.6 that doesn't happen
with 2.4, but that's because 2.4 doesn't have such a warning:

Galileo: revision 2
Galileo: PCI retry count exceeded (06.0)

Does anyone who knows about Cobalt hardware have any idea where to look?
I'm happy to send boot logs and test patches if someone wants to
investigate this problem.

-- 
Martin Michlmayr
http://www.cyrius.com/

Attachment: raq1-git
Description: Text document

<Prev in Thread] Current Thread [Next in Thread>