Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!nntp.coast.net!howland.reston.ans.net!agate!theos.com!deraadt
From: deraadt@theos.com (Theo de Raadt)
Newsgroups: comp.unix.bsd.netbsd.misc
Subject: Re: How do NetBSD and FreeBSD differ?
Date: 17 Jun 1996 13:59:30 GMT
Organization: Theo Ports Kernels For Fun And Profit, Inc.
Lines: 158
Message-ID: <DERAADT.96Jun17075930@zeus.theos.com>
References: <4pvg1u$3aj@news.umbc.edu> <SOUVA.96Jun16153706@aibn58.astro.uni-bonn.de>
<4q2kma$qru@voodoo.pdx.oneworld.com> <4q30j4$7gf@panix2.panix.com>
NNTP-Posting-Host: zeus.theos.com
In-reply-to: tls@panix.com's message of 17 Jun 1996 03:13:40 -0400
In article <4q2kma$qru@voodoo.pdx.oneworld.com> downsj@threadway.teeny.org
(Jason Downs) writes:
Of course, there's also OpenBSD, which is not only multi-platform (with
different architectures and better support on others), but also supports
things like the FreeBSD package system. OpenBSD is not only superiour
technology to both of the other systems, but also easier to use by regular
people, and easier to do development with then NetBSD.
In article <4q30j4$7gf@panix2.panix.com> tls@panix.com (Thor Lancelot Simon)
replied (but I see someone has cancelled his article already, oh well, I will
reply to it anyways):
Of course, some people are of the general opinion that OpenBSD is the freak
product of a ridiculous pissing fight between various people in the NetBSD
camp on the one hand and Theo DeRaadt on the other, and that aside from the
issue of who's in control there really isn't enough difference to pay a whole
lot of attention to. I mean, the whole thing came to life as
"s/NetBSD/OpenBSD/g" writ large...
Using your notation, I'm sure a lot of people recall "s/386BSD/NetBSD/g" ?
I remember when you phoned me and got me out of bed, begging me not to
start OpenBSD; you asked me to hold off for a few more days while you
talked to Charles in person. You said you understood my position, you
said the whole affair made you sad. That was also the first time I
found out you and another NetBSD fellow were going to produce a NetBSD
CD. Actually, I don't think I ever heard back from you.
That was you, right? How'd that CD go, anyways?
Anyways.
Actually, OpenBSD is quite a bit different from NetBSD.
Anyone can see that if they check the pages at http://www.OpenBSD.org,
read the general mailing lists, and *in particular* peek at the
source-changes mailing list to see what kind of work is being done.
I presume you have done all the above, right? -- that you are
qualified to make an accurate statement when you imply nothing much
has happened except a name change and `leadership'?
Basically, OpenBSD is way way OPEN. This is largely due to the
influence Chuck Cranor had on me; he wanted a BSD where everyone could
get more involved. I was afraid after what had happened with NetBSD,
but Chuck pushed and I've learned that it is a really good way of
operating a large project. One major part of being `open' is that
people who want to get involved can get involved... at the level they
want.. very easily. They don't just get to file PR's and mail patches
around. Using CVS access to send detailed diffs is easy. CVS commit
access is easy to get. I'm giving commit access to someone about every
2 days.
Anyways, for those who have not checked the WEB page or the mailing
lists yet, I'll paste a few things that OpenBSD has added or fixed..
ie. these are things which are not in NetBSD.
First: OpenBSD tracks NetBSD changes, plus we do our own additions and
fixes. As of today OpenBSD is merged up to all the changes NetBSD had
on June 7th; in a few more days (tomorrow, if it keeps raining!) we
will sit down and merge the NetBSD changes over the last 10 days.
Even though we don't have access to the real NetBSD source tree
(ie. CVS) we've been managing to merge their changes using sup. I see
a lot of conflicts because OpenBSD typically fixes NetBSD PR's before
NetBSD gets around to fixing them.
These are the major changes to the machine-independent parts of the
kernel. Many machines benefit from these changes. (I could make a
longer list, but most would not appreciate anything longer :-)
New curses library, including libform, libpanel and libmenu.
a termlib library which understands termcap.db, needed for new curses.
(yes, nvi is using new curses, and yes, it works fine)
The FreeBSD ports `framework' was integrated and is usable by you! This
permits you to download the newest revision of FreeBSD's ports
tree and build application from it by typing "make"
ipfilter for filtering dangerous packets
better ELF support
nlist() that understands ELF, ECOFF, and a.out, allowing non-a.out ports
to use kvm utilies
Verbatim integration of the GNU tools (using a wrapper Makefile)
(new: cvs, gcc, new binutils, libg++, texinfo, in fact on my sunos
machine I can "cd /usr/src/gnu/usr.bin/gcc; ./configure; make" and
it WILL build correctly)
All the pieces needed for cross compilation are in the source tree.
Some LKM support in the tree.
ATAPI support (should work on all ISA busses, soon)
new scsi, md5, pkg_* commands (from FreeBSD)
Numerous security related fixes
Kerberos and other crypto in the source tree that is fully exportable
(OpenBSD is from Canada, eh)
Solid YP master, server, and client capabilities.
/dev/rnd -- source of random data
In-kernel update(8) with an adaptive algorithm
Some ddb improvements and extensions
(try "hangman" -- the interactive kernel symbol table teacher, <1K)
Numerous scsi fixes
(NetBSD incorporated some of these ELF support changes just a few days ago).
Now, let's see what's happened on the i386 port, which is what a lot
of you are interested in. Again, this is just a minimal list -- think
of it as a list that people often ask about, "Did you fix this yet?"
the addition of ISA bounce buffer support so that ISA scsi cards
(ie. Adaptec 1542) can be used in machines with more than
16MB of memory.
The same one GENERIC kernel works on all machines, whether it
has BusLogic or Adaptec scsi cards.
some APM and PCMCIA support (3c589, NE2000-clone, and COM ports)
(i can sit on the couch upstairs with my laptop and lap-cat :)
some DDB improvements
OPTI mcd support
ATAPI cdrom support
P5/P6 improved NTP support
pccons has userland replaceable keymaps, and screen blanking, and a
bunch of other fixes and improvements
boot "-a" support like other ports (asks for root partition)
The mvme68k port is completely different. I've been working with a
few people (Hi Chuck! Hi Dale!) and it solidly and reliably supports
the 68040 models of Motorola's cards. This has no relationship to
the code in the NetBSD tree.
Niklas has been hard at work on the Amiga port, and it supports the
ISA bus on the Golden Gate II card quite well now. A few days ago he
read the ID off a IDE drive hooked up to an regular PC IDE card
plugged into his machine (uhm, it said "oCnnre ePirhprela s" :-).
he's hoping to make the ATAPI stuff work too... All this work is
applicable towards making the ISA bus stuff work on other kinds of
machines.
Oh, and the sparc. Well, there's some news coming about that very
soon.
The Acer PICA is sitting beside me running X11R6 using MIPS shared
libraries. Per Fogelstrom did that stuff, and the new gnu binutils
stuff (ie. mips shared libraries) which he worked with cygnus on will
be available for use by all the mips ports (including i suspect Dave
Miller's new SGI linux when it comes out). It is now time to make
this port work with the other ARC mips consortium machines.
Ah heck, I give up -- the list is too long. Like people always tell
the newbies -- if you wanna know what it has, give it a try.
Certainly it's been around 3/4 of a year.. we haven't been idle, and
we're certainly not tired of hacking on this stuff.
time to deal with netbsd pr#2555.
--
This space not left unintentionally unblank. deraadt@theos.com