On Mon, 2 Dec 2002, David Brownell wrote:
> Oliver Neukum wrote:
> > Am Samstag, 30. November 2002 22:48 schrieb David Brownell:
> >
> >>In short, you're saying that request does exactly what
> >>it's intended to do.
> >
> >
> > Upon further thought, it doesn't. You see, simulating
> > a disconnect will not ensure that the device is in a usable
> > state. In fact, in a disconnect handler trying to quieten
> > a device would be wrong, but strictly speaking it's needed
> > here.
>
> There's no "simulating" involved: the driver is really
> getting disconnected from the device. Drivers that don't
> stop using the device at that point ("quiescing") are
> seriously (oopsably) buggy -- always have been. You're
> wrong here.
Nope. If it's physically gone, you kill all urbs and
in terms of USB you are done. Doing anything else
with a device that is gone is indeed a bug.
But a physical disconnect is not the same thing as
disconnecting the driver from the device. You do that
to use the device afterwards. To do so it must be
in a useable, known state. Simply calling the disconnect
handler will _not_ ensure that. In fact as the disconnect
handler must assume a physical disconnect, it cannot
put the device into a known, useable state, as that
would require communication with the device.
Regards
Oliver
-------------------------------------------------------
This SF.net email is sponsored by: Get the new Palm Tungsten T
handheld. Power & Color in a compact size!
http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0002en
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
