Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 28 Feb 2005 10:08:48 -0800
From:      Brooks Davis <brooks@one-eyed-alien.net>
To:        Hans Petter Selasky <hselasky@c2i.net>
Cc:        freebsd-isdn@freebsd.org
Subject:   Re: mechanical i4b cleanups
Message-ID:  <20050228180848.GB438@odin.ac.hmc.edu>
In-Reply-To: <200502262014.36481.hselasky@c2i.net>
References:  <20050226004635.GA2922@odin.ac.hmc.edu> <200502262014.36481.hselasky@c2i.net>

next in thread | previous in thread | raw e-mail | index | archive | help

--uZ3hkaAS1mZxFaxD
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Sat, Feb 26, 2005 at 08:14:33PM +0100, Hans Petter Selasky wrote:
> On Saturday 26 February 2005 01:46, Brooks Davis wrote:
> > Below are some fairly mechanical patches I did 6-9 months ago to drag
> > some of i4b's drivers in the direction of a modern implementation.  I
> > don't have the ability to test them and I don't have time to work on
> > them.  Perhaps someone here does.
> >
> > The patches are also at:
> >
> > http://people.freebsd.org/~brooks/patches/i4b.diff
>=20
> If you want to modulize the devices in i4b/drivers you need some kind of=
=20
> i4b_driver_attach() routine, to perform linking.
>=20
> What mechanisms exist to initialize variables in a module or kernel durin=
g=20
> load, hence you want to have NI4BXXX stored in a variable?
>=20
> The manual pages in /usr/src/usr.sbin/i4b/man/ still say that you have to=
=20
> enter a count after "device i4bxxx", instead of using the new "option=20
> NI4BXXX=3Dcount" to set the number of device units. I think one should ke=
ep the=20
> "device i4bxxx unit_number" syntax, because then the user is made aware t=
hat=20
> the device requires a unit number. When the number of device units is=20
> unknown, and not directly depending on the number of hardware units prese=
nt, =20
> the kernel should keep a table of settable unit numbers in a special=20
> data-section, initialized with default values, reachable by  the kernel- =
and=20
> module- loaders ??

The man page is actually wrong in 6 and has been for quite some time.
Since neither peter or I are in a position to do all the work required
to modernize I4B, the NI4B* values are now options instead of counts.
As phk said, these should die.

Most of these devices should probably be created when a piece of
hardware is attached.  A framework to notify all loaded i4b drivers that
hardware has been attached is probably needed.  Hopefully someone who
cares about ISDN will find time to work on this.

-- Brooks

--=20
Any statement of the form "X is the one, true Y" is FALSE.
PGP fingerprint 655D 519C 26A7 82E7 2529  9BF0 5D8E 8BE9 F238 1AD4

--uZ3hkaAS1mZxFaxD
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQFCI14vXY6L6fI4GtQRAggZAJwJwWiZ/4Lzbaq7+0zPd66OJNgVsgCeINNV
vEGPv3OHcwg3dyJ/jxa/Nhs=
=dg0p
-----END PGP SIGNATURE-----

--uZ3hkaAS1mZxFaxD--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050228180848.GB438>