Linux Kernel 2.4.5 Released

John Jasen writes: "Join the kernel of the month club! Order yours now!" See the Changelog, I would link to the mirrors but I doubt they're updated yet, so just head to kernel.org.

Comments

Re:Cool Patches!

[stripes]

The Real Time Scheduler does not really make Linux an RTOS because in and of itself it does not provide kernel pre-emption - the ability for the kernel to interrupt kernel-space code to deal with incoming events that _must_ be processed. This is a requirement of a 'proper' hard-RTOS because such an OS must be able to guarantee a response time, and if it cannot interrupt kernel code the OS scheduler may be stuck waiting for kernel code to return before it can go on to deal with the input.

As a minor nit real time kernels do not require kernel pre-emption. The require a bounded maximum interrupt latency time. In theory the bound can even be high (100+ms, or hours even). In practice the bound has to be low just like you said. In practice kernel pre-emption is the simplest way to do it (one could also use a true micro kernel that only passes messages, and does that really fast, making all of the "real work" done in premptable user level code).

There is also the difference between hard and soft real time. Soft real time like a video game can't handle going above the stated latency very much or the animation will stutter and the user will become displeased and play a different game, but it can handle once in a while blowing the stated latency. Hard real time can't handle missing the promised latency, a computerized fuel injector might be a good example of this. If it misses, even just once it could inject fuel at the wrong time, and might blow out a delicate gasket and cause $1000 of damage to your engine (this may also be a bad example, I'm not sure how tight the timings really are for CFJ).

Re:VERITAS?

[otis wildflower]

(2) Disksuite is a free Volume Manager that does various levels of RAID.

However, it requires a lot more legwork.

You pretty much have to slice and dice your HDDs identically (and have identical HDDs in the normal case). Thus, you are still limited to 7 partitions (minus some for metadbs of course) within a "volume".

Also, you only get concatenation when you want to increase filesystem sizes, and fairly dumb concatenation at that. And IIRC if you want to concat you have to take the filesystem offline.

Disksuite is nice for small systems and root/boot/swap mirrors. Much nicer IMHO than setting up similar service (converting a single disk R/B/S system into a mirrored one) in Linux using md. I just did both on separate boxes in the last week, and I am still cringing from the md mirror "procedure" (though it did remind me that I actually don't suck ;)

A true LVM beats DS up and down the square. Many flavors of Unix come standard with LVM for "free" (though you usually have to license the OS, and Sun now beers it away up to 8 CPUs) and IMHO it's about time for Sun to give it away as well, whether they license Veritas or port/write another solution.

ps: when you've got your kernel installed into the boot sector and you've gotten your / to start mirroring (by in my case booting from rescue cdrom, copying over the /dev/md instances into the ramdisk /dev, insmodding md, raid1 and reiser, mounting the partitions to mirror, chrooting to that mntpoint, editing lilo.conf (btw that's the LATEST lilo with md bootsector support) and /sbin/lilo), BE SURE to specify your md=X,/dev/hdeX,/dev/hdgX for your root drive in your kernel append for your mdX mirror label.

It's days like that you don't feel overpaid. ;)

Your Working Boy,
- Otis (GAIM: OtisWild)

The mirrors ARE updated !

[Taco Cowboy]

FYI -

The mirrors are updated !

I have tried the mirrors at .no (norway), and
version 2.4.5 is now available at
ftp://ftp.no.kernel.org/pub/linux/kernel/v2.4/

Cool Patches!

[Ex Machina]

Here are some patches for 2.4.X I find essential

• Alan Cox's Patches - Nice!
• Real Time Scheduler - Aside from making Linux a RTOS, it improves app performance!
• GetRewted - Similar to the Openwall pacthes for 2.2.X - NonExec stack, improved filesystem security, stealth networking, Trusted Path Execution

My personal box runs all but GetRewted. My server will run them all very soon. Enjoy!

Re:Cool Patches!

[marm]

The Real Time Scheduler does not really make Linux an RTOS because in and of itself it does not provide kernel pre-emption - the ability for the kernel to interrupt kernel-space code to deal with incoming events that _must_ be processed. This is a requirement of a 'proper' hard-RTOS because such an OS must be able to guarantee a response time, and if it cannot interrupt kernel code the OS scheduler may be stuck waiting for kernel code to return before it can go on to deal with the input. The rtsched patches do appear to integrate with MontaVista's kernel pre-emption patches however, and together they would indeed form a proper hard real-time OS.

Kernel pre-emption does not come without a price though - it can make a significant dent in overall performance, and it is tricky to implement in a clean way, and this is why kernel pre-emption will probably stay out of the mainstream kernel for the forseeable future. It also isn't necessary for 99.9% of people, who, as long as the latency, the time to respond, is on average less than a few ms, are happy. This is called 'soft' real-time and is more than adequate for any video or audio work.

Linux is actually pretty bad at soft real-time as standard, with typical latencies around the 100ms mark, which is rather worse than any version of Windows 9x or NT, and a lot lot worse than BeOS, which has latencies in the sub-5ms realm. Andrew Morton's Low-Latency patches deal with this quite nicely, taking typical latencies down to the 1.5ms mark by improving various kernel algorithms and adding a few points where the kernel can reschedule itself during long periods in kernel-space code. This represents the best latency in just about any OS that does not do hard real-time with kernel pre-emption (QNX, vxWorks etc.) and does not hit performance in the way that pre-emption patches do.
What would be very interesting is to combine the low-latency patch with the improved scheduler in the rtsched patches...

As for GetRewted patches... well, I'm not entirely convinced about the value of a non-executable stack. The problem is whether they actually do any useful good - they give a warm fuzzy feeling of security while only actually preventing a limited subset of attacks. In addition, it's in the wrong place. It's a kernel-space fix for what is really a user-space problem - and certainly I think it's better to fix problems at source than patch them up elsewhere - otherwise you end up with code spaghetti.

My own personal favourite anti-stack-smashing add-on is libsafe, originally a Bell Labs project, which overrides dangerous libc functions with its own, safe functions, either by using the LD_PRELOAD feature of ELF shared objects to protect existing binaries, or by being linked in to a binary at compile time, preferentially to the existing libc functions. In addition, version 2 of libsafe now includes protection against format-string attacks that appear to be the new scourge of unix. Of course, the best place for this protection is in libc itself, and glibc 2.2 does include some protection like this, but it is a compile-time option only, and further, is primarily designed to help developers fix overflows during program testing rather than helping sysadmins in the wild - it causes more of a performance hit than libsafe does.

Anyway - as for 2.4.5, nice to see the VM is sorting itself out - I was that close to turning my desktop machine's ext2 partitions into UFS. I think I might convert them to ReiserFS now. :)

Re:NOT a service pack

[throx]

Once the dev kernel gets forked off, the kernel releases become much more like service packs. I know if you are running anything less than 2.2.16, most people will suggest you upgrade.

In the early numbers however, it is probably worthwhile upgrading now and again to get rid of those bugs that surface up in the major version change.

Re:great...

[treke]

get the patch for the 2.4.4 source? It's only about 900k

Merging Alan

[FattMattP]

Alan Cox: further merging

I'm glad to see that we're still working hard to merge Alan Cox directly into the kernel. And news of this right after the AI post, too!

Re:Merging Alan

[V50]

Dude! That's my post!! Now if I were a big company I'd probably sue you... But seeing as I'm not I'm glad more people get to see my post, which I came up with at 2.4.4-pre6 or something and had to wait until 2.4.4...

But anyway glad you got modded upto 3 so I don't have to look like moron "recycling" my old post...

--Volrath50

Call it what it is

[Mr. Sketch]

Linux 2.4 Service Pack 5. I'm running Service Pack 2 just fine and I haven't really seen a reason to apply the latest Service Pack as soon as it comes out, unless the changelog mentioned a significant security fix. Otherwise, if it's not broke, don't fix it.

FreeBSD 5.0-CURRENT-CVS-20010525 *just released!*

[Carlos Laviola]

You can't miss this! Go check it out right now!

*Sigh*. How long are we going to have to read kernel = kernel + 0.0.1 just released stories? What is the relevance of this, truly? This thing shouldn't even be at freshmeat, for christ's sake.

If not, then I want daily CVS announcements. Please, either completely bore me, or do not bore me at all.

Annoucing this kernel version is VERY IMPORTANT !

[Rosco P. Coltrane]

According to David L. Kitts, well-known specialist of God's Divine Number, the phrase "Lord of Hosts" appears 245 times in the King James version of the Bible.

Furthermore, 245 is divisible by 7, and everybody knows 7 is God's Divine Number.

Don't you understand ? 245 times == kernel version 2.4.5 ??? It's OBVIOUS : God has decided that this version of Linux will be Lord of Hosts, therefore making NT and Solaris server looking like toys for pagans !

Many thanks to Michael and Slashdot for reporting such a CRUCIAL event in the history of Humanity !

"A door is what a dog is perpetually on the wrong side of" - Ogden Nash

University of Wisconsin mirror

[patco15]

Enjoy...

http://mirror.sit.wisc.edu/pub/linux/kernel/v2.4/
ftp://mirror.sit.wisc.edu/mirrors/linux/kernel/v2. 4/

Virtual Memory System

[Cardhore]

According to Alan Cox, the VM system seems (finally) sane now (since 2.4.4-ac10). Check out Alan's full changelog for extreme details of changes at http://www.uwsg.indiana.edu/hypermail/linux/kernel /0105.2/1618.html.

linux-2.4.4-ac10 : Anonymous coward kernel ?

[stud9920]

linux-2.4.4-ac10 ? Now an anonymous coward can add code to the kernel ? I guess Cowboy Torvalds and Linus Malda did that to allow people to protect their anonymity in coding, e.g. for Microserfs who would want to participate.

But then again, don't expect them not to add boot code to display an ascii goatse.cx picture. Or to call all identifiers nathalie_portman. An also expect the karma whores to add empty for-loops to gain karma.

Business Critical Servers and Upgrades

[einhverfr]

I too would hesitate to go out and upgrade a server farm in any way without first testing the software in a controlled environment. Too many things an go wrong, and business critical systems are not the place to be experimenting.

If it ain't broke, do some testing before you fix it.

That being said, I am continually looking at migrating over to later kernels as the performance boosts I have seen have been pretty incredible particulalry for my Athlon boxes. One of them is a PDC/File/Print server (SAMBA), Apache+PhP web server, MySQL and PostgreSQL database servers, etc. mostly for development work, and the memory usage is way down from the 2.2.x kernels.