<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
<br>
Patrick Boettcher wrote:
<blockquote cite="midPine.LNX.4.64.0702061008010.9842@pub1.ifh.de"
type="cite">
<pre wrap="">On Tue, 6 Feb 2007, Chris Johns wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Patrick Boettcher wrote:
</pre>
<blockquote type="cite">
<pre wrap="">I'm not exactly sure what needs to and what can be done.
</pre>
</blockquote>
<pre wrap="">I would like to stop the kernel oops. Having the kernel crash is a pain.
</pre>
</blockquote>
<pre wrap=""><!---->
My question was related to USB debugging - I'm not sure what can be done
on the USB side to see some more information (why is there a disconnect).
Maybe something can be done with the ehci-module.
</pre>
<blockquote type="cite">
<pre wrap="">Feb 6 17:23:20 localhost kernel: urb completition error -32.
Feb 6 17:23:20 localhost kernel: urb completition error -71.
Feb 6 17:23:20 localhost last message repeated 347 times
Feb 6 17:23:20 localhost kernel: usb 8-1: USB disconnect, address 2
Feb 6 17:23:20 localhost kernel: ehci_hcd 0000:02:02.2: qh f6902180 (#82)
</pre>
</blockquote>
</blockquote>
<br>
<br>
I asked a few questions on the linux-usb-dev list. They said the best
way to track down this problem is to run usbmon and log the USB traffic
on the device until the problem occurs. The quote:<br>
<br>
<blockquote type="cite"> I suppose you can run usbmon constantly and
have it dumping the traffic<br>
in a file, if your disk is big enough. Then if the device disconnects,<br>
we'll see the preceeding traffic and may be able to make deductions.<br>
Please look if your system has this file:<br>
/usr/share/doc/kernel-*/Documentation/usb/usbmon.txt<br>
<br>
This is not guaranteed to work, because sometimes devices just
disconnect<br>
for no good reason, e.g. if they have bad power or a cables are not
shielded<br>
well. But I don't have better ideas.</blockquote>
<br>
Another of the developers said they they just tracked down a kernel bug
in USB core that didn't make it into 2.6.20. The details and the patch
are here:<br>
<blockquote><a
href="http://sourceforge.net/mailarchive/message.php?msg_id=38138485">http://sourceforge.net/mailarchive/message.php?msg_id=38138485</a><br>
<br>
</blockquote>
I plan to compile this up and see if it improves things. But haven't
had a chance yet.<br>
<br>
Cheers,<br>
Matt<br>
</body>
</html>