On Wed, 2006-04-12 at 00:23 +0200, Nicolas Boullis wrote: > Can you think of a nice way to handle this?
Well, "Plan A" would seem to be to get the driver into the kernel.org source tree so that it ends up with well-defined device numbers that are in devices.txt and no longer ambiguous or in conflict. I've generally made a habit of pushing back on supporting device information in makedev for drivers that aren't in the kernel.org source tree, though I'm not pedantic about it. Short of that, I'd suggest getting a major number assigned that won't need to change again, and just require existing users to go through an upgrade event. Bdale -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
