[Libdlo] Another kernel panic

[Bernie Thompson]

Hi Alexander,

I haven't seen this kernel panic before.

On Thu, Oct 21, 2010 at 11:03 AM, Alexander Todorov <atodorov at otb.bg> wrote:

> This happens when I remove the device and plug-it in several times (usually
> after the 2nd time). I remove the HUB cable not the DisplayLink video
> device.
>

Possibly a race condition, with removals/re-insertions. Does how quickly you
do the re-insertion make a difference?

Message from syslogd at sumu01 at Oct 21 21:00:39 ...
>  kernel:EIP: [<c06f577b>] hcd_buffer_free+0x1b/0x110 SS:ESP 0068:f70d9f14
>
>
This function is is in the core USB stack.  Within udlfb.c, the
usb_free_coherent() call when we free urbs as the driver is tearing down, is
the most likely caller that reaches hcd_buffer_free.

Do you have the rest of the call stack?  Maybe post more of the panic and
any surrounding udlfb output from kern.log?

So what we need to figure out is if this is a simple race where we pass
something bad to usb_free_coherent, or if it's something more complicated
related to other things that are getting torn down in the USB core stack
itself when a device has just had a hardware removal.

Also, can you post uname -a info?

I'll try to reproduce this problem, too - might take some time, as I've done
removal/arrival testing, and haven't caught this one yet.

Thanks for catching and reporting this,
Bernie
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/libdlo/attachments/20101021/1356ac24/attachment.html>