--ZmUaFz6apKcXQszQ
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Tue, Jul 19, 2005 at 05:30:31PM -0400, Matthew Orgass wrote:
> On 2005-07-19 matt@3am-software.com wrote:
>=20
> > One thing you may want to consider is that eventually that the tmpfs
> > maybe replace the current md-root mechanism but using an external tool
> > to write a tmpfs into a kernel.
>=20
>   Consider also the possibility of designing tmpfs such that it can
> survive reboots (either battery backed RAM, or soft reset on a system that
> doesn't clear memory).  This could make it more complex than desired
> (particularly the abililty to survive interrupted activity), but would at
> least be a point of comparison.

Let's save that idea for next summer's SoC project. :-)

My concern is that if we add the ability to survive reboots, then we=20
really aren't talking about a "temporary" file system. While someone can=20
look into this later, I think that'd be a lot of work. Among other things,=
=20
one of my hopes was that we would use kernel pageable memory, so some of=20
tempfs's data will be in the swap partitions. Making swap survive reboot=20
(and still be usable as swap, as opposed to being used to store a kernel=20
core dump for a little bit) strikes me as a sizable task in its own right.

Take care,

Bill

--ZmUaFz6apKcXQszQ
Content-Type: application/pgp-signature
Content-Disposition: inline

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

iD8DBQFC3Zt4Wz+3JHUci9cRAvYNAJ0QF2PhE4J1jRIXm5cxyGSCBrkG1ACfbQVO
GS/z/IC5O5TaVHYCWlfp0CY=
=ix8p
-----END PGP SIGNATURE-----

--ZmUaFz6apKcXQszQ--