On Tue, 6 Jun 2000, Hans Reiser wrote:
> Alan Cox wrote:
> >
> > > fs
> > > --
> > > W Merge ext3
> > > W Merge ReiserFS
> > W Merge XFS
> > W Merge NWFS
> >
> > (All four depend in part on putting the right core journal layer there for
> > them all to use)
>
> This is just a RedHat bureaucratic excuse to delay reiserfs. Tweedie has long
> suggested that we should use his journaling code, and that his code should form
> some sort of standard journaling layer. The only problem with that is that he
> showed no interest in making it work for reiserfs, or in tuning it for reiserfs,
> or doing anything else for any FS other than ext2. We have working journaling
> code, and yours is bad enough that you tell us it is horribly unfair to
> benchmark against it! So you suggest that we wait until you have something that
> will allow us to use your code!? If you want us to use your code, show the
> code, stop promising the code, and stop telling us to wait for it rather than
> using our benchmarkably better code. Prove that your code can work for reiserfs
> as fast as what we use today. The fact is, Chris Mason kicked your asses
> performance wise and now you want to employ some journaling standardization
> process to stall your competitors that you can't program as well as.
>
Hans, take a deep breath. Alan's opinions about reiserfs haven't changed
in a long time. The fact that he disagrees with us does not make a redhat
conspiracy.
I'm not sure all 4 filesystems can use the same core journal layer...and
if they could, I'm not sure they want to, as each one will have its own
performance needs. But, if we can find a single layer for 2.5 that
supports the different needs of each FS (GFS too while we are at it), I'd
love to see the reiserfs code use it.
It seems that Stephen started out with the goal of making a generic, easy
to use layer for any FS, with the best interests of the linux kernel in
mind. I wrote something just for us, because I wanted a challenge. Look
through the ext3 code and the reiserfs journal.c, and it will be clear
Stephen knew what he was doing from the start, and I learned as I went
along.
In other words, quit saying that I did a better job than ext3. We had
different goals, and it will be a long time I before understand things as
well as Stephen does.
> Why don't you help us with ReiserFS instead of standing in the way of people
> trying to program?
>
You're right hans, at least once a week, alan comes over to my house and
breaks one of my fingers. I've just been afraid all this time to talk
about it.
[ ... ]
> We will be doing wandering logs,
> and journaling in ReiserFS has another 18 months of DEEP innovations planned for
> it. Standardization bureaucrats should stay out of the way, we need to code.
>
What Hans forgot to mention is the 3.6.x format (includes 3.5.x) will be
around for a *long* time, regardless of if we get into the 2.4 kernel. My
primary job will be supporting the existing code.
-chris
-
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 - 21:00:21 EST