Re: help with firewire on shuttle SS51G - follow-up

Since I've been able to get it working (though I'm not quite sure how),
I figured I'd post what I did so the solution is in the archive.

Following Stefan's advice, I focused on what was happening at startup.

1. My first idea was to try to get linux to fully boot with the
firewire drive on and connected, thinking that this would make sure the
ports were "awake". To do this without hanging, I turned off kudzu
(Redhat's Hardware checker). Result: No changes -- the drive was still
not recognized.

2. Next, I remembered that I had an Adaptec host adapter card that I'd
tried but given up on because it had caused a kernel panic on startup.
Now that I'd disabled the automatic hardware checking, I thought I'd
give it another try. Sure enough, this worked fine.

3. So this morning, just for kicks, I tried unplugging the drive from
the Adaptec card and plugging it into the on-board firewire port. And:
success! After running the rescan bus script, the drive was recognized.

4. Because I couldn't resist trying one more configuration, I next
removed the Adaptec card and tried #1 again. This time, though, it
worked. Unfortunately, I don't know why. The only difference in terms
of configuration between #1 and #4 is that I'd added a line to fstab
for /dev/sda2 (for use with my iPod). Not for automounting, but so a
regular user could mount it. Could this have made the difference? Are
all devices in fstab "probed" upon startup?

I'd like to thank Stefan for the quick response and valuable advice.

-Joshua Kuritzky

On Feb 16, 2004, at 5:42 PM, Stefan Richter wrote:

> On 16 Feb, Joshua Kuritzky wrote:
>> I'm trying to access an external firewire hard drive on my Shuttle
>> SS51G (SiS651 chipset), running RH 9 (kernel: 2.4.20-28), but cannot
>> get linux to recognize the drive (I've also tried it with my iPod).
>
> There were some useful bug fixes in the 1394 drivers after the
> 2.4.20 release. The 2.4.21 flavour is certainly better, and the
> most recent 1394 drivers for 2.4 can't hurt.
>
> Though I am definitely not sure whether on of these fixes
> addresses your problem.
>
>> The output from dmesg when I do a 'modprobe ohci1394 && modprobe sbp2'
>> looks like this:
>>
>>> ohci1394: $Rev: 693 $ Ben Collins <bcollins@debian.org>
>>> PCI: Found IRQ 12 for device 00:10.0
>>> ohci1394_0: OHCI-1394 1.0 (PCI): IRQ=[12] MMIO=[ec105000-ec1057ff]
>>> Max Packet=
>>> [2048]
>>> scsi1 : IEEE-1394 SBP-2 protocol driver (host: ohci1394)
>>> $Rev: 707 $ James Goodwin <jamesg@filanet.com>
>>> SBP-2 module load options:
>>> - Max speed supported: S400
>>> - Max sectors per I/O supported: 255
>>> - Max outstanding commands supported: 64
>>> - Max outstanding commands per lun supported: 1
>>> - Serialized I/O (debug): no
>>> - Exclusive login: yes
>
> There is a line missing like
> ieee1394: Host added: ID:BUS[0-00:1023] GUID[00d0f5200800613d]
>
>> Unplugging/plugging the device yields nothing in dmesg,
>> /var/log/messages, or the devices file.
>
> The host is not completely initialized. Hence all external nodes
> cannot be recognized even at the lowest level.
>
>> If, however, I
>> attempt to boot with the firewire drive connected and turned on, then
>> startup hangs right when it says "Looking for new hardware..."
>
> The presence of other nodes at the bus during power-on obviously
> changes the way the OHCI chip comes up.
>
> A while ago there was a discussion here or on linux1394-devel
> about certain OHCI chips that initialize with deactivated ports.
> A special register(?) manipulation could wake those ports up.
> Unfortunately I have no link to that thread handy.
>
> I believe that manipulation was not incorporated into the 1394
> driver sources because it might have side-effects on chips
> without that deaf ports flaw.
> --
> Stefan Richter
> -=====-=-=-- --=- =----
> http://arcgraph.de/sr/
>

-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click
_______________________________________________
mailing list Linux1394-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux1394-user
Received on Tue Feb 17 16:52:45 2004