Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 22 May 2000 10:52:51 -0600
From:      Brett Glass <brett@lariat.org>
To:        Matthew Dillon <dillon@apollo.backplane.com>, Doug Barton <DougB@gorean.org>
Cc:        Warner Losh <imp@village.org>, cjclark@home.com, freebsd-security@FreeBSD.ORG
Subject:   Re: The procfs Hole in 2.2.8-STABLE?
Message-ID:  <4.3.1.2.20000522103818.04569960@localhost>
In-Reply-To: <200005220657.XAA56693@apollo.backplane.com>
References:  <Your message of "Sun, 21 May 2000 14:08:47 EDT." <20000521140847.G96573@cc942873-a.ewndsr1.nj.home.com> <20000521140847.G96573@cc942873-a.ewndsr1.nj.home.com> <4.3.1.2.20000521225733.048a0c40@localhost> <3928D2D5.DCA07289@gorean.org>

next in thread | previous in thread | raw e-mail | index | archive | help
At 12:57 AM 5/22/2000, Matthew Dillon wrote:
   
>     I think 3.4 was our 'golden release' for the 3.x series.  

I've seen some really important fixes (not enhancements, but
fixes) added after 3.4-RELEASE. Enough that I'd want to see a 3.5.

>     3.5 is only
>     going to have small cleanups, sort of like 2.2.8 had only small cleanups
>     over 2.2.7.  People have been MFCing bugs fixes reasonably well, but 
>     that's as far as it's going to go.  Many of the new features in 4.x 
>     would simply be too difficult to backport into 3.x, and a lot of the
>     really new stuff is being built on the older new stuff in 4.0-release. 
>     There is no chance of any of that being backported.

I understand that the backporting will be limited to important features.

>    My personal opinion is that the 4.0 release *already* exceeded 3.4 in
>     regards to stability, and 4.x in general is far, far superior in 
>     virtually all regards.  SMP, VM, NFS (my babies) are direct examples. 

I hope you don't think that I'm putting down your work in these areas; I'm
not. In fact, from what I've seen, I think it is quite good. It's simply
a matter of policy. We never use a .0 version of ANYTHING, from ANYONE,
in a production environment. That's true whether it's FreeBSD or (shudder)
Microsoft. We've noted that FreeBSD releases usually become as stable as
the last release on the previous branch at about the .2 release. But
not always. There were problems in 3.2 that caused us to delay full
deployment until 3.3.

I might add that we do put later releases on workstations -- just not on 
key servers.

>    4.0 is the first release where you can actually *TRUST* all the memory
>     manipulation and mapping syscalls (madvise, msync, mmap) to work
>     properly!

That's good. There's been some VM strangeness in earlier versions
that has hopefully been removed now.

>    I expect 4.1 will be the banner distribution for us if 4.0 hasn't 
>     already stolen the thunder!  3.5 will be an afterthought at best.

It will be nice if a .1 release is truly that stable! However, I think 
you'll agree that it would be the first time that this was so. So, we
will wait and see.... We can probably make that judgment about 6 weeks
after 4.1 comes out.

There really is a great benefit in running software that's a bit behind
the curve but whose quirks are known. The DoJ is demanding that Microsoft
allow customers and OEMs to buy earlier versions of Windows rather than being
forced to accept a new version.... The reason is that this is what many of
them not only want but need. Forcing an upgrade is ungood.

--Brett





To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-security" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4.3.1.2.20000522103818.04569960>