Slashdot Mirror
Kernel 2.4.2 Released
Three weeks after 2.4.1 hit the streets, 2.4.2 is now available. Here's the Changelog, and here's the download link that I know you're looking for.
← Back to Stories (view on slashdot.org)
← Back to Stories (view on slashdot.org)
all your this is slashdot not freshmeat are belong to us.
This just in...
Oddly enough, I agree. I'm from the UK and I use UK spelling, but I've worked on a project where I had to correct my "honour" to "honor" because the standard was US spelling. This was a company founded in the UK with offices in the US, and I thought they made the right decision to standardise on US spelling. Er, I mean standardize.
--
Xenu loves you!
Torvalds and Tannenbaum hashed out this argument long ago.
Tannenbaum lost . . .
hawk
This form of governmnet can only last until the hippo decides to sit on the horses . . .
hippocracy, n. The use of one hippo-like status (i.e., fat, bloated, and heavy) in the market to squish competitors. see also "Microsoft"
:)
hawk, who knows better than to flame spilling, but this was just to juicy
It's good for Linus to want to keep the 2.4.x kernels stable, but at the same time, I don't want to "downgrade" to a later kernel. :)
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
Our (6-way) router has been experiencing stability problems for some time now. I upgraded to 2.4.2-pre4 after one of the ChangeLog entries caught my eye, and the stability problem went away. (I believe the problem was in the 3com drivers.) Finally I've found a stable kernel for our router. No more down time! :)
Jason.
But have you tried cramfs? It works pretty well for systems that are relatively static - I don't think you can use it on
...well, as soon as the loopback bug gets fixed.
---
http://www.google.com/search?q=cramfs
---
Yah, but turning off javascript means that a large number of sites will not work (Internet banking, webmail, auto-cursor focusing) All kinds of creature-features that I'm not willing to do without. I run Javascript under Win98, WinNT4, and Win2K with no problems. Now, if I could just get javascript to work reliably with Netscape 6 or Konqueror under Linux, talk about flaky. I hope changing to Mozilla .8 will help... I am sick of dual-booting and would like to just go for the Linux uptime record instead of rebooting every time I need to do something that I cannot do with Linux browsers... :(
The text login screen displays the kernel version. If you have your machine set to auto-start X, then hit CTRL-ALT-F1 to switch to text mode, look at the screen, then hit CTRL-ALT-F7 to switch back to X.
Exactly how long did it take you to figure that out? I'm sure it wasn't so intuitive that you just knew. First you have to know to right-click, rather than left-click, to get the properties menu. That alone is nothing I would call more intuitive than looking at the login screen (You don't even have to log in...) and about equal with the uname -r as far as intuitiveness.
Open you mind a bit, remember that there was a day that you were just as lost with Windows as you seem to be currently with Linux. You'll learn, it'll get easier.
They may not be able to make it fool-proof, but they can already use it to show proof of fools.
Hi.
I'm not too sure what's going on, some can someone help me: I downloaded 2.4.2 and was hapily compiling it until at the end of "make bzImage" I got an error about ld not being able to find "binary." The line in the makefile was something along the lines of:
ld -elf [..] -s -oformat binary bootsect.s bootsect.o
(forgive me if I'm a little off, I'm away from the machine at the moment..)
I quickly flipped through the ld manpage and saw that -oformat is a valid option. I then tried recompiling 2.4.1 which I had installed cleanly when it came out and got the same problem. I looked at the Makefile in 2.2.18 and there is no -oformat for the ld call there.. at last I can still compile it..
Does anyone have any ideas about what my problem is? I don't know too much about the Gnu linker, but it looks like maybe the version in sid doesn't jove with the Linux kernel..
Help?
Ben
My girlfriend has informed me that Sid was the bad kid in the first movie. I haven't seen it though, so don't quote me on it..
Ben
Don't forget web pages! Those are generally stored uncompressed, and a large web site can literally have gigs of HTML files.
Boy, you've got a slow system if it takes you two months to compile three releases. Maybe if you invest in some more RAM or faster hard disks, you won't have this problem.
The other alternative is to stick with 2.2.19 or 2.0.39.
I have discovered a truly marvelous sig, unfortunately the sig limit is too small to contain i
It's not the size, it's how you use it.
--
--
"Outlook not so good." That magic 8-ball knows everything! I'll ask about Exchange Server next.
cpeterso
I've heard the Debian's next release is codenamed Sid, but which Toy Story character was Sid? I've never seen the movies.
cpeterso
Whistler Beta 1?
cpeterso
If you'd prefer that they slow down the release cycle and include more changes in each release, then just install every other kernel.
Staying on the bleeding edge does sting, that's why they call it that...
Kevin Fox
--
Kevin Fox
> o Fix 48 misspellings of interrupt (André Dahlqvist)...
Maybe they need to add a make spellcheck step to kernel compilation.
--
Sheesh, evil *and* a jerk. -- Jade
Maybe you should read the the announcements that come with every kernel release before proclaiming that:
Don't just go out and grab 2.4.2...
From Linus's announcement:
The IDE driver bug that Russell found has, to my knowledge, never been shown to happen on anything but his ARM machine, but for all we know it could be quite bad even on x86. Similarly, the elevator bug could cause corruption, but probably has not actually bit people in practice. But both are definitely deadly...
I don't know about you, but I think this may be significant and worth the upgrade if you use IDE in your systems.
Remembering your name in the morning is already a good start...
<duck>
I'm a leaf on the wind. Watch how I soar.
Although it's only just gone into the kernel in 2.4, reiserfs has been stable and working for a while, and is already included in a number of distros.
I use Mandrake 7.1 (current is 7.2) which has reiserfs - I use it for all my non-root filesystems to avoid the fsck checks.
Glad to hear that you are going to do the rewrite. The source is all there. Are you setting up your efforts on sourceforge or will you just announce it view usenet ala Linus?
You might be surprised. When I was playing with a local news spool I had to reformat the partition because I ran out of inodes. At the same time I dropped the blocksize from the default 4k to 2k, and recovered nearly 1 GB out of a 4 GB partition.
Further research showed that the average file size was around 5k, so it required 8k of disk space (3k unused). A 2k blocksize required 6k of disk space (1k unused). A 1k blocksize freed up even more space.
If you have a lot of small files you can eat up a surprising amount of disk space in the tail of your files.
For every complex problem there is an answer that is clear, simple, and wrong. -- H L Mencken
I spent almost a week downloading Red Hat's 2.4.0 beta and getting it installed (all I really wanted was iptable support). And, after getting it installed and customizing it so it would actually work (Hey! It's a beta distro...) I was amazed at the overall speed and performance. But, iptables would not work (xinetd required a backpedel to an earlier version) and I was forced to use IPChains. As it is, the box is still sitting behind my firewall rather that on the front line.
Downloaded 2.4.1 and tried to compile it. It broke things in the RD distro. Downloaded the most current iptables and recompile the 2.4.0-99..whatever RH kernel. iptables still would not work.
Today, out of a whim, I downloaded and recompiled 2.4.2. Not only did it compile without any issues, but iptables works as well (imagine that).
Can't wait to see tomorrow if, when I reboot, that it tells me nfsstatd didn't start like 2.4.0 and 2.4.1 did when I recompiled. Well, I gotta see, recompiled the kernel from home and it just sounds too good to be true.
RD
The reason, which you and the people who replied to you seem to have missed, is because Microsoft said themselves that Windows 2000 would never required fixes. They lied to the corporate managers and other people who choose what to run on the servers in order to get more money.
That is what people don't like about MS Win2k. Linus never claimed that 2.4.0 would be bugfree (or if he did, he did it tongue-in-cheek). If MS had more truth in their advertising, I know I'd be happier.
--
--
Internet Explorer (n): Another bug -- that is, a feature that can't be turned off -- in Windows.
Give us a break. The whole point of linux is "do what works" not "do what's propper." Not to mention, why don't you quit gabbing about it and do it yourself. That's the beauty here. If you think Linus is doing an ass job, he actively encourages you to try and do better. You don't have to overrule him to make some changes, you just do it.
Someone you trust is one of us.
.technomancer
.technomancer
.technomancer
.technomancer
Obviously you never read the flame war between Tannenbaum and Linus that took place in 1991. Theoritical design is one thing, but writing something that works and works well is another.
But, if you don't like how things are going with the linux kernel, nobody is stopping you from starting your own fork of the linux kernel. Import the whole source tree into your own CVS repository, get some developers and get some work done.
All those things don't well, if at all, so a compressed filesystem would just be redundant. The exegz thing might help some, but stripping your binaries is probably just about as good without the runtime performance hit. And I'd think that even if you compressed your whole root partition with a scheme like this, the savings would be negligible but everything you did would require packing/unpacking stuff so the whole system would be generally slower.
With the ever expanding size of hard drives, I think this is a pretty small issue.
Yes, I realize that maybe single floppy distros and embedded devices may find this kind of thing useful, but I'm talking about the other 95% of the linux community here.
Free Online Woodworking Resources Directory
Don't just go out and grab 2.4.2. Odds are you don't even need it. It's merely another stable release that spreads what Linux can run on.
You haven't read the fscking Changelog have you?! 2.4.2 fixes a serious IDE multimode write bug. If anyone has even somewhat-modern IDE hard drives in their system, it is certainly worth their while to get it.
I think SCSI-only boxen can wait, though -- but Linux was (and still is, to a degree) all about Unix on low-cost x86 hardware, so methinks there be plenty of IDE-based Linux systems around...
--
From the changelog:
-pre2:
- Russell King: fix serious IDE multimode write bug!
If you have IDE hard drives, I recommend you pop 2.4.2 into place purdy quick. Write bug == bad.
--
Buy windows 2000 and you have to download 1 service pack and thats means its a terrible OS that needs constant patching. Download linux 2.4.0 then 2.4.1 and then 2.4.2 in less than half that timespan and rather than meaning that 2.4.0 was a terrible OS that needs constant patching it is an innovative OS with a rapid turnaround. Am I the only one who spots the hipocracy here?
I see no hypocrisy.
First, when I make a judgement like that about a Microsoft product, it's not because of the number of service packs. I realize that all software of that complexity has bugs. It's instead that their software doesn't work well for me, even after applying all the service packs. Their service packs just don't seem to fix all the important problems for me, no matter how many I apply. For example, I've patched my Windows 98 system to the latest Windows Update stuff, but I'm still having some weird problems with Internet Explorer.
Second, Linux x.y.z releases are not just bugfix releases. In this case, it probably is...x.y.[1-5] typically are. But there are many new features introduced in point releases. 2.2.18 (or was it 2.2.19?), for example, backported USB support to the 2.2 series. I see a lot of important new features introduced in new Linux point releases, which I don't see in Windows service packs. Having a specific x.y just means the basic architecture is constant, not that the feature set is.
Third, as someone else mentioned, you don't need to upgrade to a new kernel revision to fix a bug. You have the source code, and you have the full patches. If you just want to fix one bug, you can do that. You don't have that option with Microsoft code, since it's not open-source.
Looking at some of these other posts you'd think it were a bad thing to have bug fixes and updates in a timely fashion.
until 2.5.2 (or 3.0.3)
Je t'aime Stéphanie
André Dahlqvist is fusing the line between English major and CS major.
"My mother never saw the irony in calling me a son-of-a-bitch." - Jack Nicholson
If you are truly being serious, the reason is because a kernel number makes it MUCH easier to understand the sort of kernel you are using.
The first number is a major code change, fundamental in nature. After about 7 years we are now up to 2.x.x. The second number also shows major revisions, but of a less fundamental variety than the first number. An odd second number denotes a development series, not intended to be used for production computers. This is why most users went from 2.0.x to 2.2.x to 2.4.x, because 2.1.x and 2.3.x were development versions. When the development version is deemed stable enough to be used in some production platforms, it moves to an even second number, like the recent 2.4.0. However, the 2.2.x kernel series is still being maintained for use as an ultra-stable kernel, while the 2.4.x is more cutting-edge for the latest hardware support and performance.
The third number indicates a small change, usually bugfixes but some small amounts of new features supported. When going to purchase new hardware it is easy to tell if you have a "2.2.0 or later kernel".
Finally, a service pack generally implies a large set of bugfixes, as Microsoft had somewhere around 7 (maybe 8) for NT 4. The Linux Kernel version system allows for a few small changes to be made at every release, decreasing the waiting time for users to wait for a desired bugfix or feature (instead of months for a new service pack).
If my overgeneralization of the Linux kernel was incorrect, my apologies, but I think an overall understading of how the Linux Kernel numbering system works is important for those who don't know yet.
"The universe seems neither benign nor hostile, merely indifferent." --Carl Sagan
If it ain't broke, don't fix it. Currently, 2.4.1 is the latest "something for everyone" kernel, and only because of a change in memory handling. Read the changelogs before downloading the kernels and see if there is anything you'll actually USE!
The problem with capped Karma is it only goes down...
SIG: HUP
Aparently this has been known about for a couple weeks and a patch has been made but for some reason didn't make it in 2.4.2
--
Garett
"A few atoms won't even light a match" - Dr Jones, 1933
Because the linux advocates constantly criticise microsofts service packs means of upgrading. Buy windows 2000 and you have to download 1 service pack and thats means its a terrible OS that needs constant patching. Download linux 2.4.0 then 2.4.1 and then 2.4.2 in less than half that timespan and rather than meaning that 2.4.0 was a terrible OS that needs constant patching it is an innovative OS with a rapid turnaround.
Am I the only one who spots the hipocracy here?
Slashdot: Proof that a million monkeys at a million typewriters can create a masterpiece
Ahhhh, but when the HURD is heard from, will the herd keep HURD from being heard? Word.
--
-- Geof F. Morris
I don't know if that makes much difference. Let's say SP 1 == kernel x.y.1 through 5. How many people will be running x.y.5 without the changes introduced in the 2->3 transition? Yes, I know you can do it and yes, I know how to do it. But realistically there's almost no one doing that. For most users, using kernel x.y.z means using all the patches introduced between 0 and z.
How does the fact that an NT box somewhere has a 3 day uptime means the guy you're responding to doesn't have a 342 day uptime?
And why are you assuming that whatever web server Netcraft is tracking is part of the Datacenter Server setup that's supposed to be providing the 99.999% uptime?
Come on, my Linux box at home isn't even on now - does that mean Linux doesn't run at all?
Alan Cox is doing that voodoo he do do: LWN report on 2.4.2ac1. Also, the ftp link.
More on 2.4.2 from the LWN is here.
It's technically called 'clustering', but think of it like this: take the raw bits of your hard drive and divide them into 8k blocks, or whatever. Now compress those blocks and store them on another filesystem. The filesystem-under-the-compressed-filesystem takes care of where to put these physically compressed blocks. In the worst case scenario, it says 'out of space' and the write of the new bytes fail.
Does my bum look big in this?
Interestingly, disk caches store data uncompressed. Therefore, the pages that are actually read will most likely be available from memory or swap uncompressed.
Does my bum look big in this?
Not when you have an extremely large website. Consider the fact that a site that has lots of stuff compressed gets /. all the sudden. How about that?
/.ing, decompression only happens on the first page hit, and the following 10000 hits are served direct from RAM. Unless the /.'ed content is larger than you have RAM for, there's no issue.
You're missing the point. Any file that gets read more than once is likely to be in the cache. Just depends on how much RAM you have. As for
Does my bum look big in this?
I speak for people with small drives everywhere when I say: When will the kernel support default compression of filesystems.
Use gzexe - which needs no special kernal magic, or apply the ext2compr patch to the kernel, which isn't that great.
Does my bum look big in this?
"...the kernel can effectively take on modern kernel implementations, such as BeOS and QNX, and Plan 9's kernel."
1 346238&mode=thread
Maybe they are too busy taking on operating systems that people actually use to worry about conforming to an academic's idea of how software should be architected..
That sounds like I'm advocating market share over correctness, but I'm not. I'm saying that "correctness" is in the eye of the beholder. And the beholders who agree with you and Tannenbaum haven't made much headway in the Real World whereas Linus has.
--
http://www.geekizoid.com/article.pl?sid=01/03/03/
Non-meta-modded "Overrated" mods are killing Slashdot
(Hey Ryan! Here's your proof!)
On http://www.microsoft.com/windows2000/guide/server/ solutions/overview/reliable/default.asp Microsoft explains how Windows 2000 is "Five nines 99.999% uptime", and on the right is a link to a Starbucks Windows 2000 study.
r bucks.com indeed says "The site www.starbucks.com is running Microsoft-IIS/5.0 on Windows 2000." not a big surprise.
It goes on to say how much better the coffee is and how much happier customers are because of Windows 2000.
http://uptime.netcraft.com/up/graph/?host=www.sta
However, this is one of the sites that netcraft tracks for uptime.
Note: Uptime - the time since last reboot is explained in the FAQ
Plotted Value : Windows 2000
No. samples : 250
Max : 215.28
Latest : 3.18
3 days uptime is not something to brag about, Microsoft. Do some research before you stick your foot in your mouth, but only after shooting it.
Why can't Linux just name it, like Linux 4, Service Pack 2?