GigaByte GA-MA69VM USB+mouse problem [WAS: AMD690G/V issues with 7-current (sata, usb)]

[韓家標 Bill Hacker]

Michal Varga wrote:
> On Sun, 2007-10-14 at 11:04 -0400, 韓家標 Bill Hacker wrote:
> 
>> The rarity of other similar reports HERE on the list suggest that whatever it 
>> is, it is probably not FreeBSD code.
>>
> 
> I'm thinking opposite - it definitely is FreeBSD code (in combination
> with whatever AMD and/or GigaByte internally do with their USB
> controller and BIOS).
> 

Keep on 'thinking opposite', then.

I'll kick back and have a Bushmills, neat.

Testing of 4 different MB, 3 USB and one PS2 mice with 6 flavors of FreeBSD, and 
one of DragonFlyBSD turned up:

Option USB mouse support ON in BIOS, waste two days.

Option USB mouse support OFF in BIOS, mouse responds as it should.

> Two facts:
> 1. This is a really generic low budget A4Tech mouse, I buy this single
> model for about 5 years, because I love their design (they don't have
> any) and their functions (they don't have any). I've had them attached
> everywhere, from Pentium II, PIII, P4, Athlon, AM2 boards and this is
> the first time the mouse doesn't work. If there was something really
> special about this particular model, I'd be aware of that for some years
> already..
>

...but as you've just enumerated - the MB, its 'bridge' and USB chipset has 
changed, and the *BIOS* has changed....

> 2. Windows 2000 can initialize the mouse without any problems (and with
> no specific drivers). FreeBSD can do that too, *after* the mouse is
> plugged in when the OS is already running. I wouldn't argue if this is
> GigaByte BIOS's fault or not, or if FreeBSD should do something more (or
> less) during the initialization, this is not my field of expertise. But

Clearly so.

Windows is an entirely different environment, and not 'of interest'.

DragonFly was close enough to be of interest.

> if this is the case that FreeBSD just refuses to take away control of
> the device previously initialized by the BIOS and it can be safely done
> (see Windows 2000), then maybe a systcl for such cases, or at least a
> quirk would be a nice thing to have.
> 
> m.
> 

The ums and usb are kldload/unload 'able modules.
I've been banging on that very approach..

Didn't help.

BIOS setting USB kbd DISABLE, USB mouse DISABLE, all other USB ENABLE did.

To be fair, your 'thinking opposite' has generated a lot less of a 
global-warming carbon footprint than all these machines I have running and the 
A/C to cool them.

So keep at it...

OTOH, my mouse JF works....

Bill