lkml.org 
[lkml]   [1997]   [Oct]   [30]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    Date
    From
    SubjectRe: Linux 2.0 and 2.1 stability..

    Okay.. I'm curious, what major feature of 2.2 are we going to tell the
    users about? I doubt they care about the new exception handling or the
    dcache.. Saying Faster seems to take care of that...

    Also, is the release going to somewhat conencide with the libc6ifying of
    most distributions?

    Does this also mean there will be a new smbmount util? :)

    At least, I'll have a chance to try out all the jucy patches that are only
    being maintained for the stable kernel (like raid[1,5] and QNX (the 2.1
    ver dies with SMP) and others)..


    On Thu, 30 Oct 1997, Linus Torvalds wrote:
    > >
    > > Ok,
    > > as people have noticed, 2.0.31 isn't the greatest kernel ever: it works
    > > ok most of the time, but a failed try to fix a memory leak actually
    > > results in much (MUCH!) worse problems than the leak itself under certain
    > > circumstances.
    > >
    > > I'm not personally too interested in the 2.0.x series (everybody together
    > > now: "No, really?"), but I hope we can make a 2.0.32 with the help of the
    > > LMP group which contains a fix for the 31 mmap ring problem (and probably
    > > also the updated fs/locks.c). I further hope that this will be a question
    > > of days rather than weeks (much less months, like 2.0.31).
    > >
    > > However, at the same time I actually want to start getting 2.2 ready too.
    > > I've mentioned this a few times before, that I hoped for a code-freeze
    > > before the end of this year so that we could have a 2.2 early next year.
    > > Now, I know this is going to take some people by surprise, but the end of
    > > this year is actually coming pretty close.
    > >
    > > So people, I would ask for some stability work be done to get a stable
    > > 2.0.32, but as of next week I'd prefer if developers started thinking "Oh,
    > > sh*t, Linus thinks he can freeze us, let's panic".
    > >
    > > Yes, I'm serious. There are lots of things that people want to get done,
    > > but we have to start freezing some day, and I _have_ warned people (even
    > > if nobody ever seemed to actually notice), so I'm announcing a code-slush,
    > > that will eventually turn into a feature-freeze and then a code-freeze.
    > >
    > > The code-slush means that there are a few outstanding issues: some
    > > patches accumulated by David (mainly sound, architecture and networking),
    > > and the NFS issues. The kernel NFS server is broken and is outside the
    > > code slush - I assume it will get a major overhaul still. The NFS client
    > > is being debugged.
    > >
    > > I just made a 2.1.60 which contains some VFS changes to support
    > > pread/pwrite (not too interesting in themselves, althought "Single UNIX"
    > > does want them, but the async IO people might be able to use them to good
    > > advantage with threads). It also contains the Joliet fs etc. That's going
    > > to make some people happy feature-wise.
    > >
    > > There are still some obvious problems in 2.1 (like ncpfs still doesn't
    > > work because it hasn't been updated to the new VFS interface etc), but I'd
    > > like for people to really start thinking about "Ok, what can I do to make
    > > this _stable_" instead of "Hey, what else do I want?".
    > >
    > > I'll be away for a few days to be at LISA, but I hope that by the time I
    > > get back I'll have a 2.0.32 to bless, and that the worst panic over an
    > > eventual 2.2 code-freeze will have settled and people are constructive
    > > again. Comments?
    > >
    > > Linus
    > >
    > >
    > >
    >


    \
     
     \ /
      Last update: 2005-03-22 13:40    [W:0.074 / U:1.060 seconds]
    ©2003-2014 Jasper Spaans. Advertise on this site