ieee1394 sbp2 broken since 2.6.0-test11

ieee1394 sbp2 broken since 2.6.0-test11

[Daniel Robbins]

Ben,

I've been having to use the ieee1394 driver from 2.6.0-test10 in order
to burn CDs reliably. With 2.6.0-test11 and beyond, I experience
intermittent login failures and then reconnect failures to my firewire
CD burners. With 2.6.3, even when the login or reconnect goes OK, the
system can hard-lock while burning. The CD burners are using Oxford 911
chips, firmware 3.6 and 3.8. I've tested using a variety of firewire PCI
cards and mainboards, both SMP (uniprocessor hyper-threading) and UP,
and a couple of different CD drives.

On the linux1394.org site, the Oxford 911 chip has a "works great"
rating, but it does not indicate what kernel(s) provide this "great"
functionality -- maybe this information could be added to the site?

What concerns me is that the ieee1394 drivers seem to be becoming less
and less reliable and compatible with every 2.6 kernel release. The
changes being made to the drivers may be fixing issues for some users,
but they are also breaking things for other (previously happy) firewire
users. 

It appears that more QA is needed to make sure that one person's fix
doesn't become someone else's "b0rk," as seems to be happening now.
Without organized testing on a wider range of hardware, new driver
releases will probably just be differently compatible rather than more
compatible.

I do not have time to assist with ieee1394 development, but will gladly
send you a variety of firewire parts so that you are able to test the
ieee1394 drivers on a wider range of hardware, with the goal of having
driver compatibility issues identified and fixed before they can impact
Linux users. I will also gladly assist with QA as needed. But ideally,
it's probably best if a ieee1394 developer is doing QA, since then when
problems are found they can be immediately debugged and fixed. LKML "my
firewire is busted" threads are a time-consuming and semi-unproductive
ordeal for everyone involved, I think.

I am very interested in Linux having excellent firewire support. I think
firewire is great. If you or any other capable developer wants to take
the time to test on a wider variety of hardware to help ensure that the
mainline ieee1394 driver is as solid and compatible as humanly possible,
I'll send the hardware.

Please (someone?) take me up on this offer.

Sincerly,

Daniel

Re: ieee1394 sbp2 broken since 2.6.0-test11

[Ben Collins]

On Fri, Feb 27, 2004 at 12:38:48AM -0700, Daniel Robbins wrote:
> Ben,
> 
> I've been having to use the ieee1394 driver from 2.6.0-test10 in order
> to burn CDs reliably. With 2.6.0-test11 and beyond, I experience
> intermittent login failures and then reconnect failures to my firewire
> CD burners. With 2.6.3, even when the login or reconnect goes OK, the
> system can hard-lock while burning. The CD burners are using Oxford 911
> chips, firmware 3.6 and 3.8. I've tested using a variety of firewire PCI
> cards and mainboards, both SMP (uniprocessor hyper-threading) and UP,
> and a couple of different CD drives.

Have you tried the latest 2.6 bk? I did a lot of work with sbp2 and the
nodemgr.

-- 
Debian     - http://www.debian.org/
Linux 1394 - http://www.linux1394.org/
Subversion - http://subversion.tigris.org/
WatchGuard - http://www.watchguard.com/