Jason Thorpe writes:

>I don't think asking for concrete justification is unreasonable.  And,
>until you provide it, I'm not really willing to consider changing the
>interface in any way that you suggest.

Jason,

one of the things Justin added was a driver-specific filter function.
Justin says that's a better way to detect mappings that oculd tickle
driver bugs.  VLB buslogic boards were given as a specific example; it
sounds plausible to me.  (I'd assume that if NetBSD added this, the
filter argument would be NULL in almost all cases).

Does that count as concrete justification -- at least if the
filter-function works without switching to a callback-based model?