"H. Peter Anvin" wrote:
>
> Followup to: <20000530211154.A1909@yahoo.com>
> By author: "Kenneth C. Arnold" <kcarnold@yahoo.com>
> In newsgroup: linux.dev.kernel
> >
> > With that said, here's the very preliminary list (the names are the people
> > who submitted the itmes to me):
> >
> >
> > Linux 2.5 / 2.6 TODO (preliminary)
> > N Documentation
> > W Merge ext3
> > W Merge ReiserFS
> > N VFS changes
> > I Get rid of SCSI host template
> > I Handle replugging
> >
>
> N!!!! dev_t resizing. This is absolutely vital for a whole bunch of reasons.
>
> > Replace the cache of filenames in the VFS with a B+ tree
> > algorithm.
>
> B-trees and derivatives are good on-disk structures. They aren't
> particularly optimal in-memory structures.
>
> -hpa
> --
> <hpa@transmeta.com> at work, <hpa@zytor.com> in private!
> "Unix gives you enough rope to shoot yourself in the foot."
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.rutgers.edu
> Please read the FAQ at http://www.tux.org/lkml/
To explain what hpa says:
hashing destroys correlations in access (locality of reference), B-trees can
preserve it, RAM is less sensitive to correlations in access than disk is, thus
there is less motivation for balanced trees for in RAM accesses. The more you
employ technologies like RAMBUS (somewhat) and NUMA (much more so), the more
motivation to use trees.
What we will try to do:
It will be interesting to see if we can make the reiserfs trees sufficiently
general purpose that people will find it convenient to use them --- that is one
of our design objectives for 2.5/2.6. People tend to prefer hashing because the
code is simple: we need to make trees an easy toolkit to employ. I don't think
people should be asked to employ trees until we have successfully put out an
easy to use toolkit, and ReiserFS is not such a toolkit yet. If we do succeed
in making distributed ReiserFS an easy to use multi-host tree toolkit then it is
my hope lots of people will use it for lots of things never imagined by me.
Hans
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed Jun 07 2000 - 21:00:11 EST