--JYK4vJDZwFMowpUq
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Sat, Feb 05, 2005 at 04:41:16PM +0100, Manuel Bouyer wrote:
> On Sat, Feb 05, 2005 at 01:18:46AM +0100, Reinoud Zandijk wrote:
> > > [...]
> > > Well.. you'd need an explicit kernel/userspace flag as far as I can s=
ee.
> > > Like uio does.
> >=20
> > isn't there allready one? i.e the VOP_IOCTL() call does call with `stru=
ct=20
> > proc *' being zero isn't it? i.e. its not afiliated with a process. Thi=
s=20
> > fact enough points that its kernel-called isn't it? A copyout_proc() li=
ke=20
> > call with the proc pointer passed would then know what to do.
>=20
> Would proc still be NULL if the IOCTL is called from a kernel thread ?

We're making the semantics of such a call up as we go along. However I=20
think that passing proc =3D=3D NULL if the secondary addresses are in kerne=
l=20
space is reasonable.

Take care,

Bill

--JYK4vJDZwFMowpUq
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (NetBSD)

iD8DBQFCB6zqWz+3JHUci9cRAqYFAJsHG5AxHqIeGN4KNIqL+5cJ7kSTlQCfQicR
hUTBMx7FBuvU12rw9SpRHmE=
=RfvP
-----END PGP SIGNATURE-----

--JYK4vJDZwFMowpUq--