Re: USB devices changing UIDs
Re: USB devices changing UIDs
- Subject: Re: USB devices changing UIDs
- From: Mark Cookson <email@hidden>
- Date: Wed, 12 Dec 2001 14:59:37 -0800
On 12/12/01 12:03 PM, "Jeff Moore" <email@hidden> wrote:
>
on 12/12/01 12:40 AM, Sven Duwenhorst <email@hidden> wrote:
>
>
> the Roland UA-30 changes its uniqued ID string, when switching thr sampling
>
> frequency from 44.1 to 48. This is a strange behaviour, because unique IDs
>
> of devices shouldn't change at all, especially when changing a property of
>
> the device.
>
>
>
> The unique ID string should NEVER contain any setup information of the
>
> device as sampling frequency, number of channels, bit depth, clock source,
>
> input ...
>
>
Agreed! Please write a bug report about this.
>
>
That said, the UID for a device can change since it is derived in part from
>
its position in the USB network topology, assuming the device doesn't have
>
it's own UID as described in the USB spec.
So I looked at my UA-30 to see what it does when you throw that switch...
It disconnects from the USB, then reconnects with a new config descriptor
with a new product name.
It changes its product name from:
UA-30 (48000Hz, Analog REC)
to:
UA-30 (44100Hz, Analog REC)
so its UID changes from:
AppleUSBAudioEngine:Roland:UA-30 (48000Hz, Analog REC):19160000:2
to:
AppleUSBAudioEngine:Roland:UA-30 (44100Hz, Analog REC):19160000:2
The driver makes the UID using human readable strings so that a person might
be able to read this and interpret it if the driver ever needed to present
the string to the user. I could have it use the hex device and manufacturer
IDs, but that would defeat that purpose.
How does this cause problems for you?
--
Mark Cookson
Engineering Droid
Apple Computer, Inc.
Core Audio CPU Software
6 Infinite Loop MS 306-2CW
Cupertino, CA 95014