Re: Recent USB mouse regression

Building a kernel & userland just before USB2 from svn change 184609
seems to have fixed this for now.  It seems that even when USB2 is not
loaded it affects a bit the way ums(4) works.

What is the best way of troubleshooting this?

On Sat, 08 Nov 2008 00:23:53 +0200, Giorgos Keramidas <keramida_at_ceid.upatras.gr> wrote:
> A recent 8.0-CURRENT installation _without_ USB2 enabled or loaded from
> `loader.conf' seems to have regressed a bit from a couple of weeks ago.
>
> When I attach an old wired mouse I have at home:
>
>   ums0: <Genius NetScroll + Mini Traveler, class 0/0, rev 1.10/1.10, addr 2> on uhub4
>   ums0: 3 buttons and Z dir.
>
> and run moused in debugging mode I can see mouse clicks being handled,
> but mouse movement `dies' after a bit of time.  Pressing one of the
> mouse buttons generates moused events like:
>
>   moused: received char 0x83
>   moused: received char 0x0
>   moused: received char 0x0
>   moused: received char 0x0
>   moused: received char 0x0
>   moused: received char 0x0
>   moused: received char 0x0
>   moused: received char 0x7f
>   moused: assembled full packet (len 8) 83,0,0,0,0,0,0,7f
>   moused: ts:  3708 752982137
>   moused:   :  3649 133538726
>   moused: flags:00000001 buttons:00000001 obuttons:00000000
>   moused: activity : buttons 0x00000001  dx 0  dy 0  dz 0
>   moused: mstate[0]->count:1
>   moused: button 1  count 1
>   moused: received char 0x87
>   moused: received char 0x0
>   moused: received char 0x0
>   moused: received char 0x0
>   moused: received char 0x0
>   moused: received char 0x0
>   moused: received char 0x0
>   moused: received char 0x7f
>   moused: assembled full packet (len 8) 87,0,0,0,0,0,0,7f
>   moused: ts:  3708 859980149
>   moused: flags:00000001 buttons:00000000 obuttons:00000001
>   moused: activity : buttons 0x00000000  dx 0  dy 0  dz 0
>   moused: mstate[0]->count:1
>   moused: button 1  count 0
>
> but moving the mouse doesn't show anything in the log of moused, as if
> the movement event were never delivered by the USB stack to `/dev/ums0'.
>
> Some times moused starts receiving movement events for 1-2 seconds and
> then they are gone again.
>
> One of the ways I can reliably get the mouse in a `dead' state is by
> typing in an xterm window.  When xterm hides the mouse pointer of X11,
> mouse movement is gone for good.
>
> A second mouse -- a wireless Microsoft mouse -- attaches as ums0 but
> never delivers anything to moused:
>
>   root: Unknown USB device: vendor 0x045e product 0x00e1 bus uhub4
>   kernel: ums0: <Microsoft Microsoft Wireless Optical Mouse 1.00, class 0/0, rev 2.00/0.07, addr 2> on uhub4
>   kernel: ums0: 5 buttons and Z dir.
>
> Both of these used to work a couple of weeks ago, and I'm looking back
> through history to find out when this started.
>
> Is anyone else seeing something like this?