Hi Dave, if, during probe() for an interface, a driver does usb_driver_claim_interface on the interface being probed, what will happen? I think the answer is: bad things.
Which wouldn't be a new behavior. That's basically a "let's clobber the driver model state" action; Bad Idea.
Look in drivers/usb/image/mdc800.c for an example of a driver doing this. Maybe the driver core should protect itself against this...
I'm unclear why "mdc800.c" still exists, I though that was suppose to go away during 2.5 development ... time maybe to mark it as obsolete? Or at least, fix that particular bug.
The driver core shouldn't be trying to protect itself against every possible mis-use; there are too many of them. As for this particular one ... I'd guess not, but it could be argued either way.
Hi Dave, do you think this bit is really needed? Of course it doesn't hurt and is safer, but I was wondering if you saw a real need for it? [ set up all interfaces before probing any]
I didn't see a need for it, but that was just a bugfixed version of something from your patch. As a precaution it seemed safe.
Hi Dave, why this comment change? Isn't this obvious? [ document that claim() applies only to the active config ]
Maybe it's obvious to folk who understand how all the linux-usb API calls relate to the driver model ... likely not to people just starting to use them.
- Dave
------------------------------------------------------- The SF.Net email is sponsored by EclipseCon 2004 Premiere Conference on Open Tools Development and Integration See the breadth of Eclipse activity. February 3-5 in Anaheim, CA. http://www.eclipsecon.org/osdn _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
