Matthew Dharm wrote:
Ooooh... a good test case. This is repeateable, right?
It's happened multiple times, but I don't think it's deterministic.
I suspect "pull out the USB cable" would repeat it nicely though! :)
I've been waiting for some additions to the scsi core for hot unplugging.
At least some have just been added, so expect patches soonish. I'd love to
have you test them.
I'll probably end up doing that as they get merged. Let me know
when that's ready and I'll see if this happens "on request".
- Dave
Matt
On Sun, Feb 16, 2003 at 01:00:02PM -0800, David Brownell wrote:
Hi Matt,
When running some storage tests, a couple of times I've gotten
oopses. Attached is 'dmesg' output for one sequence ... another
was in_irq (dead box), with the usb_storage_queuecommand() null
pointer getting triggered on the khubd usb_device_remove() path.
Seems like for some reason the host would get driven off-line
(electrical loading? VIA roothub wierdness?) and the cleanup
(new code) didn't happen as it should ... this is basically
that "unplug during disk i/o" case, except that I didn't do
anything to the cable.
I thought you should know about this ...
- Dave
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
