USB Mouse Problems in -BETA5 [Was: logitech wireless mouse works for a while but then disconnects]

[Chris Gilbert]

Just to narrow things down a bit, I switched the mouse onto my onboard USB controller. 

Before, it was on a PCI USB 2.0 card. 
('uPD720100A/101 USB 2.0 Enhanced Host Controller')

Now it is on the onboard USB 1.1 controller. 
('AMD-768 USB Controller')

However, the problem still persists, which leads me to rule out a faulty controller/hub, and leads me to believe that the problem most likely lies somewhere in the uhid/ums parts of the USB stack.

To give an idea about the severity of the issue, just in the last 12 hours, the mouse has "attached and detached" over 25 times.

Today I will switch my mouse for an identical model on a FreeBSD 5.3 machine. 

This way I will be able to confirm that it's not a problem with the mouse itself, and a real problem with the USB stack.

To summersize, here is a quick run-down of what I know so far:

1) The problem does not lie in moused, it happens both with moused active and without.

2) The problem is not X related, it happens both with and without the X server running.

3) The problem is irritated by lots of frequent mouse activity. It has never happened while the mouse is sitting still.

4) The problem is not specific to any particular USB controller, or ochi/uhci/echi. It happens with both ochi and ehci, and both with AMD and NEC USB controllers.

5) For me, the problem was not present in FreeBSD 6-CURRENT or prior. It started after updating to BETA2, and continues while using BETA5.

Also, for the record, ACPI is disabled at boot via device.hints. I thought it may have been some sort of strange power management issue with the system thinking it should power down the USB ports to save energy... but this doesn't seem to be the case.

I will send another mail once I have made sure it's not an issue with my particular mouse. But judging by the fact that the mouse has worked perfectly in the past, and others are having the same error, I am confident that this is not the issue.

Once I have affirmed this, I will start looking into the USB stack myself. 

However if anyone else is interested in looking into this problem, I would be more than happy to assist them in any way I can.

Thanks again,
Chris Gilbert

On Sunday 02 October 2005 02:25, Chris Gilbert wrote:
> I just updated to 6-BETA5 and have the same results.
>
> I confirmed that it happens even when I'm not in X, and just at the
> console.
>
> However, it only happens when I am moving the mouse around...
>
> It's almost like it's getting a bytestream from the mouse that causes it to
> go bonkers internally... but I'm not sure.
>
> In the mean time, I have worked around it by having X use moused/sysmouse
> instead of opening the ums0 device directly. This doesn't fix the problem,
> but at least I don't have to cycle my VTs every time it dies.
>
> Any ideas on this guys? Since it's not just me, this could be a potentially
> annoying issue in the release unless it gets some solving...
>
> Thanks!
>
> --
> Regards
> Chris Gilbert
>