Oliver Neukum wrote:
> 
> in the long and celebrated tradition of sending in patches which happen
> to be completely untested here it is for discussion.

At a first glance it seems the right kind of thing, though it needs a few
more updates.  What did you have in mind for testing this?


> This patch disconnects all interfaces of a device before changing
> a configuration and introduces proper locking for that.

I don't think I noticed it creating driverfs nodes for the interfaces
on the new config.  And if the SET_INTERFACE request (over the bus)
fails, I think you don't release the dev->serialize semaphore.
But that was just from scanning the patch.

I'd like to move _all_ configuration handling logic into "config.c";
right now it only has some of it.  From this patch, that'd include
usb_set_configuration() and your new reap_interfaces() helper.

- Dave








-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to