Daily use of Neo1973

Tilman Baumann tilman at baumann.name
Wed Feb 20 19:48:25 CET 2008

Joe Pfeiffer wrote:
> Tilman Baumann writes:
>> Hm. I never tried Windows. But this does not make much sense to me.
>> The Os just has to select a device configuration to allow it to draw 
>> power. This does not mean it has to have a driver for it.
>> I'm pretty much sure my linux server which i sometimes use for charging 
>> does not have any driver whatsoever for this device. But it charges nicely.
>> Strange. But however, it is windows...
> I don't know anything about Windows, but -- the problem with not
> charging on dumb cords is that the NEO is polite about needing to
> handshake, and being able to request 500mA, before it will draw
> 500mA.  If the device on the other end doesn't even do that, the NEO
> will only draw 100mA which isn't enough (the fast_cccv parameter
> somebody else mentioned will, apparently, force it to draw lots of
> power anyway).

I know. But selecting a device configuration (allowing the device to go 
into a status that consumes more than 10mA) has nothing to do with 
drivers or anything like that.
At least not necessarily.

Any USB device just offers the host one or many possible device 
configurations (descriptors) with they respective power consumtion 
profiles when it is plugged in.
The next step for the host (operating system) is to select one of these 
configurations to allow the device to go in this mode.

The os does not need to now what a device does, in order to allow it to 
do anything.
I'm really surprised and not entirely convinced that windows does not 
select a profile on any unknown device.
A mobile phone would be not the only situation where this behaviour 
seems like a bad idea.


More information about the device-owners mailing list