Re: modutils, the next generation
H. Peter Anvin (hpa@transmeta.com)
14 Jan 1997 20:57:10 GMT
Followup to: <Pine.LNX.3.94.970113121433.16950A-100000@Cavern.NMSU.Edu>
By author: Derrik Pates <dpates@Cavern.NMSU.Edu>
In newsgroup: linux.dev.kernel
>
> On Mon, 13 Jan 1997, Martin Buck wrote:
>
> > Sorry, no answer to your question, but I've got another question concerning
> > modules and their recognition:
> > Why do modules use .o as extension? Of course, they're object files, but
> > very special ones, so an extension like .mod would make much more sense,
> > IMHO. If I remeber correctly, the old insmod/depmod/modprobe (never looked
> > at TNG) even checked for .mod in some places, but it didn't work because
> > the check for .mod was missing at some other places.
>
> I'd say .mod is not a good extension to use - it's already used by a music
> file format. But, if we want to give loadable modules their own extension,
> how about .m or .k or maybe .km ?? Just food for thought.
>
There is another problem with kernel modules: they are not
self-contained, but are only part of what a device driver needs to
be. I have been meaning to write up a proposal for a "supermodule"
system to solve this, which would include additional components than
the kernel module proper.
-hpa
--
This space intentionally has nothing but text explaining why this
space has nothing but text explaining that this space would otherwise
have been left blank, and would otherwise have been left blank.