--RXc6EO4W1yUvSQ0X
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Tue, Nov 06, 2007 at 02:29:54PM -0500, Matthew Mondor wrote:
> On Tuesday, 6 Nov 2007 9:18:35
> Adam Hamsik <haaaad@gmail.com> wrote:
>=20
> > IMHO something like boot -c debug/lockdebug... doesn't hurt and=20
> > should be useable for everyone. if you need debug you can turn it
> > on  at boot.
>=20
> > is it possible to do this or we have to change to much code for
> > this?
>=20
> A possible issue might be performance, considering that as a compile
> option macros may be used which generate no code at all when compiled
> without debugging support.
>=20
> A boot option might need for empty function stubs, checks against
> NULL function pointer or boolean option variable in
> performance-critical paths, which could no longer heavily use
> assert/diagnostic features unless the functionality could still
> completely be disabled at compile-time.

You're basically describing the reasons Sun came up with dtrace. :-)

Sounds like we should port dtrace and add boot support. :-)

Take care,

Bill

--RXc6EO4W1yUvSQ0X
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (NetBSD)

iD8DBQFHMgmcWz+3JHUci9cRAgpUAJ4516xF/IKrV2J/EWtxryS3gxAswwCeMa6c
W/uCkvdnK1ivfB04HIIUPEk=
=/8xZ
-----END PGP SIGNATURE-----

--RXc6EO4W1yUvSQ0X--