kern/117200: ugen0 prints strange string on attach if detached quickly

[Nick Hibma]

>Number:         117200
>Category:       kern
>Synopsis:       ugen0 prints strange string on attach if detached quickly
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    freebsd-bugs
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Mon Oct 15 09:10:02 UTC 2007
>Closed-Date:
>Last-Modified:
>Originator:     Nick Hibma
>Release:        FreeBSD 6.2-STABLE i386
>Organization:
AnyWi Technologies
>Environment:
System: FreeBSD hind.van-laarhoven.org 6.2-STABLE FreeBSD 6.2-STABLE #17: Thu Aug 16 12:47:43 CEST 2007 root at hind.van-laarhoven.org:/usr/src/sys/i386/compile/HIND i386

>Description:

The dmesg output (I noticed it in the security run output, hence the '+'):

+ugen0: at uhub4 port 2 (addr 4) disconnected
+All threads purged from ugen0
+ugen0: detached
+usbd_get_string: getting lang failed, using 0
+ugen0: ???????????????????????????????????????????????????????????????????????\^Y???\^D product 0xfffe, rev 1.00/1.00, addr 4
+ugen0: at uhub4 port 2 (addr 4) disconnected
+All threads purged from ugen0
+ugen0: detached

It seems that when detaching the device before it has completely attached
creates rubbish in the strin gdescriptor.

>How-To-Repeat:

Plug in a device and remove quickly. Most probably this can be produced with a
bad cable as well. Or now I come to think of it, perhaps with a header that you
pull off by jiggling it from left to right.

>Fix:

Scanned the source code, but couldn't find it straightaway. Strange that a
string descriptor is fetched while fetching the language id failed. Or perhaps
the bug is in the subr_bus code that displays the device attachment string?
>Release-Note:
>Audit-Trail:
>Unformatted: