FreeBSD 5.2.1 Released

← Back to Stories (view on slashdot.org)

Posted by michael on Thursday February 26, 2004 @02:57AM from the now-in-technicolor dept.

Kalev writes "The FreeBSD Release Engineering Team has announced FreeBSD 5.2.1-RELEASE. This is intended to address several bugs and vulnerabilities discovered in the FreeBSD 5.2 release. See the Release Notes. The release is now available for downloading. If you are currently running FreeBSD 5.x, you can easily cvsup to it or use binary upgrade feature of sysinstall."

20 of 110 comments (clear)

Min score:

Reason:

Sort:

Point point releases? by Anonymous Coward · 2004-02-26 03:05 · Score: 4, Insightful

Is it just me or are point point releases of FBSD pretty rare? Almost seems like 5.2 was a bit of a rush job.
1. Re:Point point releases? by Anonymous Coward · 2004-02-26 03:28 · Score: 5, Informative
  
  No, it wasn't a rush, but people didn't test it enough.
  That is usually the problem, 5.1 ran so well that people didn't want to test the RC's, thus some bugs didn't get ironed out for _their_ hardware.
  The thing is, if these people had downloaded the livecd of RC2 and sendpr'ed this release wouldn't be needed.
  
  You should blame people for their lack of will to test but strong will to always complain.
2. Re:Point point releases? by cperciva · 2004-02-26 03:33 · Score: 5, Informative
  
  Is it just me or are point point releases of FBSD pretty rare?
  
  You're right. The only other one was 4.6.2-RELEASE. (I'm not counting the 2.2.x releases -- 2.2 was a major version number :) ).
  
  Almost seems like 5.2 was a bit of a rush job.
  
  5.2 was right on the boundary between "experimental" and "stable". As such, lots of people started using it once it was released, but few people actually participated in testing it. I believe that 5.2 had one of the longest ever periods between code freeze and release.
  
  --
  Tarsnap: Online backups for the truly paranoid
3. Re:Point point releases? by shlong · 2004-02-26 04:24 · Score: 5, Informative
  
  Almost seems like 5.2 was a bit of a rush job.
  
  As Colin pointed out in a peer post here, 5.2 had quite a long release cycle. If you look at the 5.2 release schedule you'll note that we spent almost 2 months on it. Add in that 5.1 was released in June of 2003, and you have quite a long dev cycle. We did the best that we could to manage risks in the 5.2 cycle, but shortly afterward it became apparent that there were some significant bugs in certain modules that didn't gain much attention until after the release was made.
  
  --
  Cat, the other, tastier white meat.
4. Re:Point point releases? by cperciva · 2004-02-26 04:28 · Score: 4, Informative
  
  4.6.1 never existed. It was going to exist, but some security issues appeared, so it was aborted. But you're right, I forgot about 3.5.1-RELEASE, mostly because it was never tagged in the CVS repository.
  
  --
  Tarsnap: Online backups for the truly paranoid
5. Re:Point point releases? by shlong · 2004-02-26 04:41 · Score: 5, Informative
  
  4.6.1 never existed. It was going to exist, but some security issues appeared, so it was aborted. But you're right, I forgot about 3.5.1-RELEASE, mostly because it was never tagged in the CVS repository.
  
  FreeBSD 4.1.1 existed also, and was tagged. However, it was a branch off of RELENG_4 instead of RELENG_4_1 and turned into a disaster. But yes, ever since 3.0, we've had few point releases.
  
  --
  Cat, the other, tastier white meat.
6. Re:Point point releases? by Anonymous Coward · 2004-02-26 05:53 · Score: 5, Informative
  
  actually if you had looked at http:www.freebsd.org/releases/index.html
  
  You would have noticed that 4.1.1 was the first point point release in almost two years after the decision that they were unneccesary extra work.
  For 4.1.1 it was decided it was worth it because of the expiring of the RSA patents, it allowed the security pieces to be more easily merged in for US users.
7. Re:Point point releases? by dohcvtec · 2004-02-26 08:59 · Score: 4, Interesting
  
  Funny, 4.1.1 was the first version of FreeBSD I used, and I was hooked right from then. Heh, even a "bad" release of FreeBSD is still pretty good compared, oh, say, a bad RedHat release (anyone remember the whole GCC 2.96 fiasco?)
  
  --
  -- Never hit a man with glasses. Hit him with a baseball bat.
8. Re:Point point releases? by cant_get_a_good_nick · 2004-02-29 17:55 · Score: 4, Informative
  
  When RedHat decided to release the 7.0 distro, they decided that 2.95 was too broken to serve as their system compiler. Too much C++ stuff was broken, not ANSI compliant, etc. So they grabbed a dev branch of 3.0 (at that time tagged 2.96) polished it, and called it gcc 2.96. Cool in some ways, because it was the first time the rubber hit the road for the 3.x branch. A lot of bugs got fixed. But it was a dev branch, and the code changed on the way to 3.0. So now you have a compiler that is C++ binary incompatible with anything before it, or anything agasint it. RedHat 7.x users (we support it at work) find it real hard to get 3rd party binaries compiled by 7.3. Either they use the more common 2.95, or they're finally trickling in to 3.2 or above (which are finally C++ binary compatible). Just a real PITA getting any C++ code using 2.96.
Cool! by Anonymous Coward · 2004-02-26 03:07 · Score: 5, Interesting

Despite the claim that 5.x isn't yet the
production branch, we've been running it on
all our development machines and servers for
6+ months now. Apparently the FreeBSD
release engineering team has pretty high
standards! We're really looking forward to
FreeBSD 5.3, which has M:N threading and
the new O(1) scheduler as the default.
Thread creation in our application is
blindingly fast *and* runs on many CPUs at
once. After getting off the poor Linux 2.2
and 2.4 threading, there was no turning back.
1. Re:Cool! by Anonymous Coward · 2004-02-26 06:27 · Score: 5, Informative
  
  For the record, FreeBSD's scheduler was already O(1), it just didn't handle SMP extremely well. The new ULE scheduler handles the SMP case much better, along with other nice improvements. See Jeff's paper at http://www.chesapeake.net/~jroberson/ULE.pdf
  
  The release engineering team certainly does have high standards. Trying to live up to the stability reputation. But keep in mind that 5.x still is considered in testing and major changes can still be afoot that can cause instabilities. So please still keep in mind what -current means, http://www.freebsd.org/doc/en_US.ISO8859-1/books/h andbook/current-stable.html, and read the early adopters guide.
  
  5.x will get better and better as it approaches 5.3R, so while some of the problems running a -current release are lessened, one should still be aware of all this and the higher standard for fixing one's own problems when running 5.x. RTFM is not an insult when running 5.x, its simply a price of entry to a great OS.
Updating from 5.2-RELEASE to 5.2.1-RELEASE. by cperciva · 2004-02-26 03:37 · Score: 5, Informative

[I posted the message below to -current and -security, providing an easier upgrade path from 5.2-RELEASE to 5.2.1-RELEASE] In order to provide an easy update path for i386 systems from FreeBSD 5.2 to FreeBSD 5.2.1, FreeBSD Update will now update systems running FreeBSD 5.2-RELEASE to 5.2.1-RELEASE. To take advantage of these updates, install and run FreeBSD Update, and reboot into the new kernel: # cd /usr/ports/security/freebsd-update && make install clean # cp /usr/local/etc/freebsd-update.conf.sample /usr/local/etc/freebsd-update.conf # /usr/local/sbin/freebsd-update fetch # /usr/local/sbin/freebsd-update install # shutdown -r now If you have recompiled any files locally, FreeBSD Update may not be able to update them automatically (it will complain). With the latest version of FreeBSD Update (version 1.5), you can use one of the following commands: # /usr/local/sbin/freebsd-update --branch crypto fetch or # /usr/local/sbin/freebsd-update --branch nocrypto fetch depending upon whether you installed the "crypto" distribution, to force files to be updated. (If you're not sure if you installed the "crypto" distribution, you almost certainly did). FreeBSD Update will update a 5.2-RELEASE system to the exact binaries distributed with 5.2.1-RELEASE, with the following exceptions: 1. Files under the following directories will not be updated: /usr/ports /usr/share/doc /usr/share/man/cat* /usr/src The ports and src trees can be updated using cvsup; the files in /usr/share/man/cat* are rebuilt from (updated) man pages automatically. 2. FreeBSD binaries include, in their headers, the value of __FreeBSD_version on the machine where they were compiled. This value was bumped from 502000 to 502010 as part of the release engineering process; binaries for which this is the ONLY change will not be updated. As always, this is something I'm providing personally; it is in no way endorsed by the Security Officer, Release Engineering team, or the project as a whole. Colin Percival

--
Tarsnap: Online backups for the truly paranoid
1. Re:Updating from 5.2-RELEASE to 5.2.1-RELEASE. by cperciva · 2004-02-26 13:01 · Score: 4, Interesting
  
  Do you have any idea how many people actually use your service?
  
  Yes. Around 100 systems per day; in total, somewhere around 2500 systems have been patched.
  
  Are there any plans to make it an "official" offering of the FreeBSD project?
  
  Yes. I recently joined the FreeBSD security team, and now that 5.2.1 is out of the way, I'll be pushing to get the integration underway.
  
  --
  Tarsnap: Online backups for the truly paranoid
Re:Unfortunately by anthonyrcalgary · 2004-02-26 06:34 · Score: 4, Funny

Even Theo deRadt is talking about a combined *BSD userland utility team.
oh god I hope not. They'd kill each other over philosophical issues.

--
When someone might yell at me, it has to be OpenBSD.
Re:How's FBSD on AMD64? by eht · 2004-02-26 06:57 · Score: 5, Informative

The AMD64 platform is currently a Tier 1 FreeBSD platform.

Current Tier 1 platforms are i386, Sparc64, AMD64, PC98, and Alpha.

Current Tier 2 platforms are PowerPC and ia64.

Current Tier 3 platforms are S/390(R).

All systems not otherwise classified into a support tier are Tier 4 systems.

All information lifted verbatim from the FreeBSD website most of it from Section 10 of the Committer's Guide, Support for Multiple Architectures

So expect as much support for AMD64 as you would for the standard PC version, the only thing keeping AMD64 back is it's not a widely distributed and therefore not as well tested.
Re:How's FBSD on AMD64? by mi · 2004-02-26 09:04 · Score: 4, Informative

It works, except for the kernel modules. Currently, you need to compile everything you need into the kernel. kldload-ing does not work yet.
The 32-bit emulation is supported and turned on by default, although some 32-bit binaries, may have problems controlling some hardware with ioctl-s, because the sizes of structures are often different.
I wouldn't recommend it as a workstation, because too much stuff out there (open source and not) is poorly written and thus unportable and will break during compile time (at best) or at run-time (at worst). Think about all the foolish assumptions, that sizeof(int) == sizeof(void *) and shudder.... I don't think NVidia offers their drivers for amd64 either, and so on.
Makes a (very) nice server, though...

--
In Soviet Washington the swamp drains you.
What is it about by coughski · 2004-02-26 12:41 · Score: 4, Interesting

Operating Systems that invokes such spite between proponets? IMHO FreeBSD works very well I run 4.8 for my server and have never had a problem never crashes and it does what I need it do. For my Desktop I run SuSe 8.1 and it works very well for most everything I need. I suspect I could use both as a server or a desktop with few problems. I am not a developer or sys admin but I do enjoy tinkering with networks and computers and *nix or *bsd based systems. I just wonder why I see posts like "FreeBSD is dead" from a supposed linux user when clearly both operating systems are actively growing. cough

--
two cans and a string, now that's innovation
IPSEC still broken by dokebi · 2004-02-27 03:41 · Score: 4, Informative

It looks like the KAME implementation got borked between 5.1 and 5.2. ISAKMP packets get filtered even when they're not supposed to.
see here.

--
In Soviet Russia, articles before post read *you*!
Re:I'm confused... by Anonymous Coward · 2004-02-27 08:07 · Score: 4, Informative

Then be enlightened.
Re:Fuck Apple. by cant_get_a_good_nick · 2004-02-27 08:19 · Score: 4, Informative

If it weren't for the Mach micokernel from Apple
Mach is from Carnegie Mellon, by way of NeXT.

Windows NT / 2000 / XP runs on a variant of the Mach kernel.
XP does not run on Mach. It was a microkernel, made with a lot of input from DEC VAX guys. Over the years it has shed a lot of Microkernel attributes and become more of a macro style kernel.

Mach is a fairly standard, well documented design principal.
Microkernels are a fairly standard, well documented, design principal. Mach is an instance of them.

I actually agree with some of your other statements, your parent poster was an uninformed fanboy. Apple has contributed to BSD though, check out the BSD project list and see where.