--BOKacYhQ+x31HxR3
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Mon, Nov 12, 2007 at 11:18:56AM -0800, Bill Stouder-Studenmund wrote:
> That's the idea. The loading code still has to slog through the whole=20
> file, but we shouldn't need kernel memory for the debug stuff.

This is one of the things that bugs me about just using one file...
I have spent a lot of time working with systems that boot from flash
or from the network.  I don't want to load a full netbsd.gdb that way.
I'd rather have two files around.  I like what I think Andrew suggested
of having a file with some minimal debug information that you could boot
=66rom and which could be used for more extensive in-kernel debugging.
Having a full netbsd.gdb and better two-machine debugging support would
be really nice, too--I did a little two-machine debugging of a the OS X
equivalent of an LKM in my previous job, and it was quite handy.

-allen

--=20
Allen Briggs  |  http://www.ninthwonder.com/~briggs/  |  briggs@ninthwonder=
=2Ecom

--BOKacYhQ+x31HxR3
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (NetBSD)

iQEVAwUBRzio5mIPSzajGzeGAQLkFAgAnIVJ2ShWrYGCrJ3hXrydPQ7K5qckN0/7
1R3BjqZCWKSdT6JuUzwHWt7ck/8pv/1y27+sH1pI2yheh3cQdQnbxtZywxXkrqJo
nVaIk6wxkGfkRNigHjmRvwgyrSjx/NvcTHLllYUhmN5X9+LHFdasLylvlfh0Fqlf
ZBj8HwsoMDW4NzMUOcQUB3jhNvZDAdIHgN2lpHcUjMl4r5saFtvFPfEjO0CvzN2u
/mg4mtXd/fl2owe3bJzVP+CicOL3L9+jjiAR4Phuf5kRmK4OXB8uEclJjelVTH6u
hJON/lMgWOX9B7/5ibRC02sz1ddHhuvxwCn1Ie9iaODaE5NAEtOZEw==
=GpHs
-----END PGP SIGNATURE-----

--BOKacYhQ+x31HxR3--