>> 4) Having the kernel know what boot kernel image was used.

This comes up periodically.  The problem is not one of what to do with
the kernel image; the problem is getting hold of the kernel image at
all.  Some ports, it seems, get the kernel into memory in ways
unrelated to any NetBSD-accessible filesystem, and thus not only does
the kernel not know where its on-disk copy is (assuming it's on disk at
all), it may not be accessible even if we knew.

Various people have made various suggestions on this subject in the
past; the full discussion is in the list archives for some one of the
NetBSD lists, though I can't recall offhand which one.

>> 5) How about a kernel that could recognize certain errors at not
>> being misson-critical panic attack problems, and just disable
>> certain parts of the kernel on error?  Like debugging an lkm
>> filesystem: if the author screws up, no need to panic the system,
>> just kill that lkm.

Except that you don't know what the lkm has hooked itself into.  You
also don't know what else the lkm may have damaged before going wonky
enough to be noticed.

> Or a memory failure in a non-kernel area of memory (just map it out
> after killing the process using it).

This one is entirely reasonable, conceptually.  If it wouldn't be too
hard to do (whatever _that_ means), it'd be nice to have.  (But I don't
want it badly enough to put any effort into it myself....)

					der Mouse

			    mouse@collatz.mcrcim.mcgill.edu