> But what
> if the device snapped a new picture while it was mounted? How would the
> host filesystem learn about it without unmounting/remounting?
>
> With a "real" mass storage device like a disc drive the host f/s never
> has to worry about somebody else changing the media (I think!). All
> changes to the media pass through the host f/s.
>
> But with a camera, the camera firmware can spontaneously change the
> media.
The "Still Image Capture Device" USB class spec is basically the PTP
spec, which sends an event from camera to device when that happens.
There are also the media change events that Pete mentioned, for when
media (say, CF cards) get, um, hotplugged.
PTP basically provides a semi-smart filesystem model. However, the
Linux support for it is entirely user mode (http://jphoto.sourceforge.net)
and doesn't currently implement the event support. (Seems to need a
bit of usbdevfs rework: one I/O per device, not per endpoint, is the
current limit for synchronous calls.)
The other standards-oriented model for cameras, mass storage, would
have that problem (as Matt agreed). But that doesn't support things like
"tethered mode" operation, useful for remote camera control: Take this
picture now, using F/2.8 at 1/125 second, and so on. PTP does.
Proprietary camera protocols (mostly legacy) tend to handle remote camera
control more than they handle these media/filesystem change events, at
least to the limited extent I've looked.
- Dave
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
http://lists.sourceforge.net/lists/listinfo/linux-usb-devel
