Dan Streetman wrote:
Actually it's khubd that notices physical disconnects and disconnectsThe reason _why_ a driver releases an interface doesn't change the semantics of releasing it. Releasing involves your point #1 (which implies #2).
It just seems strange to me to be aborting URBs for a device that no longer exists...seems like the HCD would just clean them out during the next frame. Especially since it would be easier (more centralized) to do it in the HCD, while doing it in the drivers is more work...
drivers from the device. And "rmmod" (or other calls) do them for
other reasons. So the HCD really isn't involved until much later; it
eventually gets told "device is gone, recycle its memory".
One of my "worth finishing" patches adds a new interface between
usbcore and the HCDs, so usbcore can tell the HCDs to clear out the
state for any given endpoint. That'd be mostly useful for handling
configuration events correctly (set_configuration, set_interface),
but there's a similarity: in the same way those config events can
make endpoints vanish, so does physical disconnect. In all of
those cases, it'd seem good to make usbcore nuke pending urbs.
- Dave
-------------------------------------------------------
This SF.net email is sponsored by: Microsoft Visual Studio.NET comprehensive development tool, built to increase your productivity. Try a free online hosted session at:
http://ads.sourceforge.net/cgi-bin/redirect.pl?micr0003en
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
