Hi,
I've been reading both linux1394-devel and -user for the past month or so,
and have posted about the problems I'm having previously:
http://sourceforge.net/mailarchive/forum.php?thread_id=3633881&forum_id=5387
I've now done a lot more testing and reading up on ieee1394, as well as
trawling through the source and trying to make sense of it, maybe someone
has an idea now that I've found out some interesting facts. To recap, when
I connect the sbp2 device into my laptop's onboard OHCI controller, I can
usually mount the drive, but after a few seconds (about 5-60; varies) the
bus appears to reset itself, and SCSI errors ensue. The sbp2 device works
just fine on my desktop machine (nForce2 chipset, also OHCI). Also,
everything works fine under Windows 2000 as well, so it's not a
hardware-only issue.
This is what's new: I've now set up the laptop with 2.6.2-rc1-mm1 +
ieee1394 trunk revision 1105 (=yesterday). The problem still remains.
However, I have now also tried eth1394 to connect my laptop and
desktop(2.6.1). This works fine, and I can transfer quite a lot of data
over the firewire network, and everything seems perfectly stable. The next
thing I tried was the following: Keep the laptop plugged into the desktop,
with eth1394 working between the two. Desktop has two 1394 ports on the
same controller, so I plugged the sbp2 device into the second one.
Desktop's hotplug "grabs" the device, laptop can't log in. (makes sense).
After a rmmod sbp2 on both machines, and a modprobe sbp2 on the laptop, it
sees the device. I can mount it, transfer data, everything remains stable!
So far it's been connected for about 30 minutes, I've transferred about 2gb
of data (just my home dir - lots of small files as well as a few very large
ones). eth1394 is also still up and running and worked fine during the
transfer between laptop and sbp2-drive as well. My layman's conclusion is
that (as Stefan Richter has suggested before) the OHCI driver and my
laptop's controller don't play well, and the reason it works when I connect
it via the desktop is that the desktop's controller becomes the "master"(?)
of the bus and handles all the management, and the laptop's controller just
needs to log into whatever device(s) it wants without having to worry too
much about managing the bus. I haven't found very much in-depth info on the
1394 bus and can't get the ieee specs, so I don't actually know if this has
any resemblance of the actual situation. Unfortunately, I don't have my
desktop machine with me when I'm on the go (which is supposed to be the
purpose of the sbp2 drive) so pass-through is not a feasible solution to my
problem. I'm still trying to find out enough about ieee1394 and kernel
module debugging to track down the issue myself, so I'd really appreciate
pointers from people in the know, no matter how small.
By the way, can I assume that most important people on the -devel list will
also be reading -user, or should I move this thread?
Regards,
Phil Jordan
-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn
_______________________________________________
mailing list Linux1394-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux1394-user
Received on Mon Jan 26 18:03:24 2004
This archive was generated by hypermail 2.1.8 : Mon 02 May 2005 - 09:16:46 CEST