hi,

the patch looks fine, but as far as I can tell, map->size isn't
used for anything other than that assert.  how about we just
get rid of it instead?

-Chuck


On Mon, Feb 09, 2004 at 03:58:18PM +0200, Antti Kantee wrote:
> DIAGNOSTIC kernels panic at KASSERT(map->size == 0) in uvmspace_free()
> when a child process exits after a fork() with a MAP_INHERIT_NONE
> entry.  This is because size bookkeeping is not done for those
> entries in uvmspace_fork().
> 
> Is the attached patch okay, or are there more far-reaching consequences
> related to the problem?