On Sat, 14 Jan 2012, Greg KH wrote:

> On Sat, Jan 14, 2012 at 06:05:24PM -0800, Tony Luck wrote:
> > On Sat, Jan 14, 2012 at 8:30 AM, Alan Stern <stern@rowland.harvard.edu> wrote:
> > > The driver core isn't designed to handle device structures that get
> > > unregistered and then spring back to life; callers are supposed to
> > > allocate a fresh new structure instead.
> > 
> > Does every caller use "kzalloc()" for this fresh new structure (or even
> > for the original registration) rather than "kmalloc()"?
> 
> They better, as that's what they are required to do :)
> 
> The kernel would be crashing with slab debugging enabled, if it wasn't
> so today.

These are both important restrictions on the use of the driver core,
and they don't seem to be explained anywhere in the kernel source.  
Would you like to receive a documentation patch?

Alan Stern