On Fri, 9 Sep 2005, Chris Clayton wrote:

> > It would help if you included _all_ the kernel log messages in that file, 
> > not just the debugging messages.
> > 
> That's all there was in the log for this particular boot. I've never looked 
> into
> it before, but it looks as if syslogging on my system is configured to send
> different stuff to different log files. I'll look into that this weekend.

Yes, sometimes distributions set things up that way.  If you simply change 
the appropriate line in /etc/syslog.conf to say something like:

kern.*:         -/var/log/kernel

that would be sufficient.

> > Those messages mean the keyboard sent back a reply packet that was longer
> > than it was supposed to be.  More accurately, the keyboard's internal hub
> > did this.  The computer asked it for a 1-byte packet, and the response was
> > more than 1 byte.  This was on the interrupt-in endpoint.  Note that 
> > according to the information you posted from lsusb, the maxpacket size for 
> > that endpoint is 1, so the hardware is definitely misbehaving.
> > 
> > > I've also backed out one or two at a time, all the uhci related patches 
> > > (plus any
> > > dependent patches to host or core) that went into -rc1 but that hasn't 
> > > made any
> > > difference. Bearing in mind the new diagnostics, does anyone have a 
> > > suggestion
> > > about where I should look next, please?
> > 
> > There are lots of other things you can try.  For example, what happens 
> > when you plug the keyboard into a different computer running the same 
> > versions of Linux?  What happens if you don't plug the keyboard in until 
> > well after the bootup procedure is finished?
> > 
> > Assuming you another means of getting in to your computer (logging in over 
> > a network, for instance), you can use the usbmon facility to trace the 
> > data being sent to/from the keyboard.  Read about usbmon in 
> > Documentation/usb/usbmon.txt in the kernel source.  This patch will help:
> > 
> > http://marc.theaimsgroup.com/?l=linux-usb-devel&m=112561058119237&w=2
> > 
> 
> Here's what I get from /sys/kernel/debug/usbmon/5t when I tried to log in 
> using
> the usb keyboard. I managed to type only the first four characters of my user 
> name
> before key depressions ceased to be echoed to the console.
> 
> d9f3dae0 323797083 C Ii:003:01 0 8 = 00000600 00000000
> d9f3dae0 323797121 S Ii:003:01 -115 8 = 00000600 00000000
> d9f3dae0 323909151 C Ii:003:01 0 8 = 00000000 00000000
> d9f3dae0 323909156 S Ii:003:01 -115 8 = 00000000 00000000
> d9f3dae0 323957185 C Ii:003:01 0 8 = 00000b00 00000000
> d9f3dae0 323957191 S Ii:003:01 -115 8 = 00000b00 00000000
> d9f3dae0 324037242 C Ii:003:01 0 8 = 00000000 00000000
> d9f3dae0 324037246 S Ii:003:01 -115 8 = 00000000 00000000
> d9f3dae0 324165333 C Ii:003:01 0 8 = 00001500 00000000
> d9f3dae0 324165338 S Ii:003:01 -115 8 = 00001500 00000000
> d9f3dae0 324245390 C Ii:003:01 0 8 = 00000000 00000000
> d9f3dae0 324245394 S Ii:003:01 -115 8 = 00000000 00000000
> d9f3dae0 324469550 C Ii:003:01 0 8 = 00000c00 00000000
> d9f3dae0 324469556 S Ii:003:01 -115 8 = 00000c00 00000000
> d9f3dae0 324549606 C Ii:003:01 0 8 = 00000000 00000000
> d9f3dae0 324549610 S Ii:003:01 -115 8 = 00000000 00000000
> d9ddc3e0 324701715 C Ii:002:01 -75 1 = 00
> d9ddc3e0 324701723 S Ii:002:01 -115 1 = 00
> 
> Does this help, please?

This tells us what we already knew.  The keyboard sent a few characters; 
they are the lines containing "C Ii:003:01".  Then the keyboard's internal 
hub sent an invalid response; that's the second-to-last line with the -75, 
and after that the keyboard stopped responding entirely.

A simple test would be to try a replacement keyboard.

Alan Stern



-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
Linux-usb-users@lists.sourceforge.net
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-users

Reply via email to