kern/48441: ums(4) flapping with recent logitech mice

[dholland%eecs.harvard.edu@localhost]

>Number:         48441
>Category:       kern
>Synopsis:       ums(4) flapping with recent logitech mice
>Confidential:   no
>Severity:       non-critical
>Priority:       medium
>Responsible:    kern-bug-people
>State:          open
>Class:          change-request
>Submitter-Id:   net
>Arrival-Date:   Wed Dec 11 06:30:00 +0000 2013
>Originator:     David A. Holland
>Release:        NetBSD 6.99.19 (20130404)
>Organization:
>Environment:
System: NetBSD valkyrie 6.99.19 NetBSD 6.99.19 (VALKYRIE) #13: Sun Nov  3 
17:07:36 EST 2013  dholland@valkyrie:/usr/src/sys/arch/amd64/compile/VALKYRIE 
amd64
Architecture: x86_64
Machine: amd64
>Description:

I recently got a new Logitech usb mouse, the most basic kind with a
wheel but nothing else. Now, when X isn't running, I get this over and
over again slightly less than once per minute:

Dec 10 07:12:23 valkyrie /netbsd: wsmouse0: detached
Dec 10 07:12:23 valkyrie /netbsd: wsmouseetached
Dec 10 07:12:23 valkyrie /netbsd: uhidev0: detached
Dec 10 07:12:23 valkyrie /netbsd: whidev0: at uhub2 port 3 (addr 2) disconnected
Dec 10 07:12:26 valkyrie /netbsd: uhidev0 at uhub2 port 3 configuration 1 
interface 0
Dec 10 07:12:26 valkyrie /netbsd: uhidev0: Logitech USB Optical Mouse, rev 
2.00/63.00, addr 2, iclass 3/1
Dec 10 07:12:26 valkyrie /netbsd: ums0 at uhidev0: 3 buttons and Z dir
Dec 10 07:12:26 valkyrie /netbsd: usmouse0 at ums0 mux 0

(yes, it says "wsmouseetached" in the log most times; it also has other
variants, such as "wsmousdetached".)

When X is running, this doesn't happen.

A couple other people have mentioned seeing this with new mice
recently too.

(On the plus side, it was flapping thus for several hours this morning
and the USB stack didn't panic.)

>How-To-Repeat:

as above

>Fix:

This is presumably a bug workaround for Windows; I guess if you don't
talk to the mouse for 60 seconds it assumes the OS's driver has wedged
and reconnects itself.

Probably we ought to try to figure out either how to disable this
behavior, or how to detect it and add something to ums to pat the
mouse regularly so it doesn't do this.

If anyone knows of a useful channel, it might also help in the long
run to send Logitech a cluebat labelled "exponential backoff".