Linux Kernel 2.4.21 Released

An anonymous reader writes "After > 6 months of waiting, 2.4.21 is here. Lots of cleanups, and a patch which gives a MAJOR boost to the 'feel' of the system under heavy disk IO, especially on IDE systems. As usual, available from your local kernel.org mirror or ftp.COUNTRYCODE.kernel.org/pub/linux/kernel/v2.4/! Tidbit: 'Current bandwidth utilization 131.72 Mbit/s '." See the Changelog for new stuff.

Comments

Looks good..

[NightWulf]

Now to attempt to get it before it's slashdotted.

Re:Looks good..

[Anonymous Coward]

Is this version of the kernal free of the Unix code SCO owns? I don't want to get in trouble.

Re:Looks good..

[Metasquares]

Perhaps if SCO would actually reveal the code to the public, a version could be made that is clean of "SCO's" code. Of course, they might just get laughed at if they reveal the code.

Re:Looks good..

[42forty-two42]

ftp.us.kernel.org is a DNS round-robin for 27 hosts, some with multiple OC3's. Try slashdotting *that*!

Re:Looks good..

[Kyeo]

I'll give it a shot, but I'm not making you any promises.

Re:Just compiled it

[Anonymous Coward]

You must be compiling on Gentoo...

Started to think 2.4.x was dead

[phathead296]

I was seriously starting to think the 2.4 series was dead in preparation for 2.6.0. The ChangeLog is impressive though.

Phathead

Re:Started to think 2.4.x was dead

[mahdi13]

Far from dead...hell, the 2.2 kernel is still being maintained and patched (mostly by Alan Cox, but still...it's active)

Re:Started to think 2.4.x was dead

[$calar]

Yeah, especially since all the hype on 2.6. But, this comes so late it's at a state of irrelevance. For instance, Mandrake has already put the "2.4.21" kernel in their 9.1 release, which was 3 months ago! I just hope that 2.6 would pick up some momentum and be out by the next series of distro releases. Probably not though . . .

Re:Started to think 2.4.x was dead

[axxackall]

Well, untill 2.6.4 or something I am going to download new 2.4.x patches anyway. I guess I am not alone in it.

Re:Started to think 2.4.x was dead

[gunpowder]

I can't believe some still don't get this:

As long as it's not available from ftp://ftp.kernel.org or http://www.kernel.org, there is no newer version of the linux kernel.

The same applies to the ftp://gcc.gnu.org and GCC (not the website, they are always a little bit lame updating it).

So if you say

'Mandrake has already put the "2.4.21" kernel in their 9.1 release'

your are wrong! They didn't. They lied to you. Or you are mistaken. Or they used a prerelease and renamed it 2.4.21.

And no, 2.4.21 it's not 'too late'. Look at the Changelog and what huge amount of bugfixing has been done. And all those updates to the drivers!

True, 2.6 will feature a log of nice extra stuff, but I guess 99.9% of all linux users are happy with just the features 2.4 has. They simply don't need support for NUMA, 64bit dev_t or Zero-copy NFS.

Re:Started to think 2.4.x was dead

[destiney]

Mandrake has already put the "2.4.21" kernel in their 9.1 release

This would not really be possible since 2.4.21 was only released today. Unless Mandrake has a time machine or something..

They may have released their distro with a heavily patched 2.4.20 or possibly a 2.4.21-rc*, but not 2.4.21.

Re:Started to think 2.4.x was dead

[Anonymous Coward]

Chill dude. He probably already knows this. If you look, he put 2.4.21 in quotes. This is a very strange thing to have a knee jerk reaction over.

Re:Started to think 2.4.x was dead

[Jayanef]

Mandrake named the pre-release kernel with suffix 0.XXmdk
I use the source from mandrake and recompile for some tunes

$ uname -a
Linux lavender.mania.or.id 2.4.21-0.13mdk-lavender #19 Sen Apr 14 11:56:48 WIT 2003 i686 unknown unknown GNU/Linux

Re:Started to think 2.4.x was dead

[presroi]

Far from dead...hell, the 2.2 kernel is still being maintained and patched (mostly by Alan Cox, but still...it's active)

Even 2.0.X is still maintained. It currently stands at 2.0.40-rc6 (almost one year old).

Re:Started to think 2.4.x was dead

[gunpowder]

Chill dude
Thanks. I needed that!
It's Friday, 13th, you know ...

Re:Started to think 2.4.x was dead

[sflory]

What he may have meant is they put 2.4.21pre or 2.4.21rc in their kernel. All of the linux vendors do this. If you extract a Red Hat kernel src rpm. You generally find that there is a pre or ac patch applied.

Re:Started to think 2.4.x was dead

[$calar]

The Mandrake kernels are always heavily modified versions of the official kernels.

The one packaged with Mandrake 9.1 is a pre-release of the stable 2.4.21 kernel, hence the quotes I used.

According to Mandrake, they make so many modifications to the kernel that the difference in waiting for the final .21 isn't worth the time, and probably wouldn't make much of a difference.

I'm sure there are many bug fixes in waiting around, but they have to get this product to market.

My point in saying that it is irrelevant (to me, anyways), is that I like using the modified Mandrake kernel. I trust that they have done what's best for my system and when I recompile the kernel, I use that one.

I am really looking forward to 2.6, though. I am hoping that the incorporation of the ALSA drivers will eliminate all of the sound server mumbo jumbo.

Re:Started to think 2.4.x was dead

[Anonymous Coward]

It's Friday, 13th, you know ...

At least it's a full moon, so I have some chance at playing nethack.

Re:Started to think 2.4.x was dead

[mateub]

>> The ChangeLog is impressive though. Sure, just removing all those SCO copyrights took days! :-) Matt

Re:Started to think 2.4.x was dead

[sglines]

I looked through the changelog and didn't find the SCO code that was added. Does anyone know where it went?

Re:ot - google

[Anonymous Coward]

no

BitTorrent

[Dreadlord]

I wonder if they are planning on an official BitTorrent.

Re:BitTorrent

[aliens]

Bit torrent works best on large files. The source is only like 20megs. They could do it, but people should be only downloading the patch anyway.

Re:BitTorrent

[p3d0]

Huh? 20MB is certainly large enough for BitTorrent to help.

And the patch is only suitable for people who already have 2.4.20. Everyone else either must download and apply multiple patches, or just get the full source and be done with it.

Re:BitTorrent

[Tumbleweed]

> people should be only downloading the patch anyway

Yeah, right, and people should be writing in Esperanto on their Dvorak layout keyboards, too, but that's about the same likelihood.

If they _really_ want people to actually use the patches, they should release the patches first, wait about a week, THEN release the full kernels. Evil, I know. :)

Re:BitTorrent

[narfbot]

If they _really_ want people to actually use the patches, they should release the patches first, wait about a week, THEN release the full kernels. Evil, I know. :)

Even better: On releases, they should first post the patches and a torrent for the full thing, then like a day later, post the regular link for the full thing.

Re:BitTorrent

[Tumbleweed]

I don't think a mere day is long enough - many people don't even know about it within a day. If the bittorrent is up immediately, then a week for the regular full kernel would work fine.

Re:BitTorrent

[Xerithane]

Yeah, right, and people should be writing in Esperanto on their Dvorak layout keyboards, too, but that's about the same likelihood.

Est tajp jam nun, kun Dvorak... sed mi.

Re:BitTorrent

[Anonymous Coward]

Dvorak is useful.
Esperanto sucks.

Re:BitTorrent

[Jetson]

And the patch is only suitable for people who already have 2.4.20.

It's been six months! The only people who don't have 2.4.20 sources are people who don't compile kernels, and I doubt they'll be wanting 2.4.21 in that case....

Re:BitTorrent

[Wakkow]

I've never done this before, but I thought I'd give it a shot:

http://66.227.104.34/linux-2.4.21.tar.bz2.torren t

Re:BitTorrent

[Anonymous Coward]

Not official, but try this one:

torrent

Re:BitTorrent

[aliens]

Didn't 2.4.19 and before have a local exploit? And I guess you're right 20mb would be good for BT

Re:BitTorrent

[BrookHarty]

but people should be only downloading the patch anyway.


And if someone is using a binaryed based Distro, should they download all the patches? Much easier to grab the single bziped tarball. Bit torrent would still save the ftp servers bandwidth on the first couple of days, no reason not to use one.

Re:BitTorrent

[Anonymous Coward]

Or everyone can stop being nazis about it and just quit complaining. I always download the entire package even when a patch is available that is less than 10% of the size because I like having the latest "full" version available.

Re:BitTorrent

[Anonymous Coward]

this doesn't work

Re:BitTorrent

[greck]

It's been six months! The only people who don't have 2.4.20 sources are people who don't compile kernels...

...or have uptimes >= 6 months.

Re:BitTorrent

[Lussarn]

If they _really_ want people to actually use the patches, they should release the patches first, wait about a week, THEN release the full kernels. Evil, I know. :)


Well. Then we had to download both linux-2.4.20 and patch-2.4.21 hogging 2 sockets.

Re:BitTorrent

[clacke]

Cool! I reached 280 kB/s.
The Sunet mirror runs at over 600 kB/s, though, so not really useful. But still, cool!

Re:BitTorrent

[Kalak]

Since BT is for large files, and many don't believe in patching, it sould make sense to use the kernel.org's rsync server. Patching w/o the patch. *However*, rsync.kernel.org::pub/ is only good for mirroring the ftp site, not for mirroring the released kernels.

Putting an un-tarred version of the kernel up on a site would be like CVS read only. It seem seems like this would make sense for release kernel distribution. The PPC kernel tree is distrubuted this way, and it's really handy. Why isn't / where is the main tree distributed this way? You get a fresh, updated kernel by only updating the changes, and you don't have to do a make mrproper as rsync -avz --delete would give you a pristine version of the tree.

Re:BitTorrent

[bpmcdermott]

hmmm.. though i'm not positive, you might have started a trend.
possibly one tantamount to posting nytimes.com urls with fake referers.
currently i'm getting 50kB/s and that's not too shabby.

thanks.

Re:BitTorrent

[Tumbleweed]

> Dvorak is useful.
> Esperanto sucks.

Okay, then, Interlingua. There _are_ no elegant-looking or -sounding artificial languages because none are based on French. Erm, I mean 'Freedom.' :)

Re:BitTorrent

[AvitarX]

So then everybody downloads 2.4.19 and the patch.

That accomplishes a lot.

If they really had a huge bandwidth concern they would only post patches, let other people patch and host the whole thing.

Re:BitTorrent

[damiam]

What's your point? People with >= 6 months uptimes aren't very likely to want to install a new kernel on its release day, putting them in the category of people who don't compile new kernels.

Re:BitTorrent

[Anonymous Coward]

Thanks! I'm getting much better results from your BT stream than from my local mirror.

kernel.org should really consider putting up some torrents sometime. Especially for heavily anticipated releases such as this one.

Ok, I'm ready to recompile my brand new kernel. I'm hoping to get rid of Red Hat's slow IDE performance, whatever it is.

Re:BitTorrent

[damiam]

If you're using a binary distro, you should use your vendor's update system to update the kernel. If you're using a locally-compiled kernel on the binary distro, then you should still have the source tree around from last time, ready to be patched.

Re:BitTorrent

[liverbugg]

Maxing out my cable at 250k/s...torrent finished downloading before the kernel.org web page loaded.

Re:BitTorrent

[Anonymous Coward]

the French cock suckers strike again!

Re:BitTorrent

[greck]

sorry! you're technically correct of course, I was just reading with my /. goggles on...

Re:BitTorrent

[kasperd]

Didn't 2.4.19 and before have a local exploit?

IIRC 2.4.20 has one as well. Vendor kernels are (should be) patched against it. But if you wanted a nonexploitable kernel from kernel.org you'd have to use a prerelease. Nobody wanted to rush 2.4.21 through because of that particular bug.

Re:BitTorrent

[Hugonz]

Cxu vi vere parolas Esperanton? Ne sxajnas....

Sed...bona sxerco!

Amike,

Mi

Re:BitTorrent

[Wakkow]

Currently 35 people are leaving their completed downloads open for others to leech, and 5 are currently downloading.

The log file shows 2336 people loaded the torrent file, but I don't know how many actually completed it.

Re:BitTorrent

[Anonymous Coward]

The fact is I tried all 2.4 kernel versions so far and none are compiling successfully. It either won't compile or when it does compile then it kernel panic on next boot.

I have yet to come across some kind of hat trick from usenet allowing me to compile 2.4 kernels. I'm starting to beleive that Linux zealots are actually just pretending to use 2.4 kernels.

Re:BitTorrent

[patbernier]

Wow! I got 2400 kB/s! I don't think I've ever downloaded kernel source that fast...

I was really hoping that someone would start a torrent -- actually, I was thinking of starting one myself if I didn't find one here...

<paranoia>
Now, I still have to get the PGP signature from kernel.org since you didn't redistribute it, you know, just to make sure you didn't add anything funny to the kernel ;->
</paranoia>

Still, thanks ^_^

Re:BitTorrent

[ChrisCampbell47]

2400 kBytes per second? That's 19.2 megabits per second. Really? If not, please be more careful with your b's vs. B's next time.

Re:BitTorrent

[mnmn]

Yes. Releasing the patches first should alleviate the bandwidth problem. At least there should be a script in the kernel sources that downloads the latest patches with wget and applies them, also changes the directory name to reflect it. Such ease should convince people not to use 20 times the bandwidth they really need.

To be fair anyway, I admit I'm downloading the full sources :)

Re:BitTorrent

[Wolfrider]

--Try 27 Megs, and that's .bz2. Gzip'd is over 30Megs.

Re:BitTorrent

[factotum]

Versxane tute ne. Sed li akiris lian 5+ Funny... kia Slashdote!

Martin

Re:BitTorrent

[patbernier]

2400 kBytes per second? That's 19.2 megabits per second. Really? If not, please be more careful with your b's vs. B's next time.

Yes, I really mean it! 2400 kb/s -- close to 20 megabits per second.

The system I was using for this transfer has 100 megabits per second of Internet transit connectivity (12.5MB/s), so it's not impossible. I often get around 800 kilobytes per second (6.4Mb/s) when using Bittorrent to download well-seeded, popular anime fansubs. My previous torrent speed record was 1300 kilobytes per second (10.4Mb/s).

Perhaps one or a few of the users that were seeding the kernel were topologically close to my system, or had similarly great connectivity...

[And, being lazy, I'm using kilobytes and megabytes here, not "kibibytes" or "mibibytes" -- or whatever it is we're now supposed to call 1024 bytes and 1048576 bytes..]

Re:BitTorrent

[ChrisCampbell47]

You did it AGAIN. Please use little 'b' to mean bits, and use big 'B' to mean bytes. I couldn't care less about the 1024/1000 distinction (2.4% error), but the B/b distinction is rather big (800% error).

So that's 2400 kB/s, not 2400 kb/s.

Aaanyway, I'll still assume this was on a corporate/campus network, not at home.

Re:BitTorrent

[patbernier]

I'll still assume this was on a corporate/campus network, not at home.

I /wish/ I had that much bandwidth at home. Although I'm pretty comfortable with my home SDSL link, I'd rather run my bittorent downloads on a well-connected system... In the case of well-seeded torrents, the end result is actually a bit worse since I still have to download it home after, but OTOH, I usually get to contribute significantly more to the torrent.

You did it AGAIN. Please use little 'b' to mean bits, and use big 'B' to mean bytes.

"Again"? While it is ironic that I apparently _did_ make that mistake in my second comment, I believe that the first one was actually quite correct.

Just to get these out of the way...

[Drakonian]

Oh man! My 286 just finished compiling 2.4.20!

or

Oh man! I was downloading at 4000 K/sec before this story showed up. Thanks a lot Slashdot!

Re:Just to get these out of the way...

[s20451]

Nice, but unfortunately, Linux does not support the 286, and never has.

Re:Just to get these out of the way...

[Neon Spiral Injector]

Forgot, "this isn't Freshmeat."

Re:Just to get these out of the way...

[Spock the Vulcan]

Theoretically, he could be using a cross-compiler that runs on DOS, and generates Linux binaries :-)

Re:Just to get these out of the way...

[reynaert]

Sorry, no.

Re:Just to get these out of the way...

[shepd]

Actually, it does.

Sorry to burst your bubble... :)

Re:Just to get these out of the way...

[Prince_Ali]

Sorry to burst your bubble, but ELKS is not Linux. The top of the first page makes that very clear.

Re:Just to get these out of the way...

[Anonymous Coward]

Also true, but you don't need to be running Linux to compile the kernel...

Re:Just to get these out of the way...

[DShard]

Is GCC not supported on a 286? I think as long as you could run almost ANY operating system you can run GCC. You could then cross compile... you just couldn't run the resulting binary on the compiling system. But that wasn't what the grandparent said.

Re:Just to get these out of the way...

[Anonymous Coward]

It (WWW at least) was actually slowly BEFORE it was posted on /. already. It was at 50% of it's usage at that moment.

Re:Just to get these out of the way...

[Shwag]

>All models are wrong. Some are useful.

What is your tagline based upon?

Re:Just to get these out of the way...

[Anonymous Coward]

this is ridiculous...

Re:Just to get these out of the way...

[lacrymology.com]

Technically, he said that his 286 was compiling the Kernel... not running it.

Re:Just to get these out of the way...

[Drakonian]

Damn, good point. I did forget that.

Re:Just to get these out of the way...

[Drakonian]

It's attributed to George Box, a "famous" statistician. I like it because it's a reminder that we shouldn't get too caught up in any one set of beliefs because it's likely to all chanage at some point in the future.

Re:Just to get these out of the way...

[shepd]

>Sorry to burst your bubble, but ELKS is not Linux. The top of the first page makes that very clear.

Then why call it Embedded Linux Kernel Subset?

Because it isn't?

If it says that it isn't linux, then I think I'll call OSX "Embedded Microsoft Windows Kernel Subset" because titles don't mean jack, ever.

Also note this:

"Q1.2. How does ELKS compare with standard Linux?
ELKS is intended to be a subset of true Linux, and ought to be small enough to be understood by one person, so it should be invaluable as a learning tool. ELKS recently entered the beta development state and most of it's functionality is there."

"Q2.2. How do I make an ELKS kernel? ...The development environment will be created in /usr/src/linux-86"

"Q2.3. I get an error saying that /usr/include/linux/vm86.h does not exist."

"Yes - by loading a kernel module you can run ELKS binaries directly [in Linux]."

"This a list of answers to Frequently Asked Questions about ELKS, also known as Linux-8086."

And, to put the cherry on top:

shepd@crashy:~/elks-0.1.1$ grep -r -i linux * | wc
1377 5848 90626

WTF? And they're not all references saying "this is not linux" either...

I think the ELKS team should decide if they're using Linux (even if it is a big hack of it) or not. It seems they don't know themselves! They keep saying it's Linux, then they say it isn't, then all the files are in linux named directories, and contain a hell of a lot of references to the word linux.

I'm VERY confused at this point.

Re:Just to get these out of the way...

[kasperd]

Is GCC not supported on a 286? I think as long as you could run almost ANY operating system you can run GCC.

I have used gcc under DOS 6.20 on a 486. It was not particularily fast and did use protected mode. I know the 286 has protected mode, but I think gcc uses 32 bit code, which is not an option on the 286. It came with some runtime environment to setup protected mode and even included virtual memory. I later found the same virtual memory could help when running MCAFEE on a computer with only 4MB of RAM. It was implementing some standard whose name I don't recall.

Re:Just to get these out of the way...

[kasperd]

Then why call it Embedded Linux Kernel Subset?

I think the word subset is an important one here.

Re:Just to get these out of the way...

[DShard]

Having just checked out DJGPP's site.. and since I have already followed this rathole to the bottom, It appears that this is indeed the case as it REQUIRES at least 32bits to compile 32bits. It appears that you cannot compile from a 16bit platform targeting a 32bit one. Sorry for doing this frightening display of utter compulsiveness in public.

Re:Just to get these out of the way...

[Shwag]

Thats good. I like.

I also like, "Substantiality is inversely proportional to ponderability." - Franklin Merrell-Wolff

Re:Just to get these out of the way...

[Jonner]

What compulsiveness? One certainly could write a compiler hosted on a 16-bit architecture and targetting a 32-bit one. I don't know if such a beast exists, though. Come to think of it, there probably were such things in the early days of 32-bit machines.

Re:Just to get these out of the way...

[Sabalon]

First time Caldera was at Networld, the rep was standing there telling people that Linux would run on a 286 and up. I clarified this to him, and he argued with me saying that a 286 would run it (before Elks.)

Some guy was sent away with a Caldera CD to try and install on his 286.

Re:Just to get these out of the way...

[DShard]

My compulsiveness was to actually verifying that with GCC you could not infact due this. You COULD do it but that would be a project onto itself with very little benefit. though spending some time yesterday looking at Lunix I geuss I was amazed that there DJGPP couldn't do that already. Considering it is made for DOS makes me wonder why it isn't 16bit, since the OS is at heart that way.

Re:Just to get these out of the way...

[Jonner]

Yes, Lunix does sound pretty impressive. I believe the machines Unix originally ran on had a similar amount of memory, though they may have been 16-bit.

As for DJGPP, it's a port of GCC to protected mode DOS. It's still a 386 program all the way. I'm fairly certain GCC has always run on at least 32-bit hardware, as it started life on a 68020 or 68030 or something.

Be gentle to the mirrors

[Spock the Vulcan]

Instead of downloading the entire kernel, download just the patch file if you are running the previous version. Then patch your source tree using:

cd /usr/src/linux
bzcat /blah/patch-2.4.21.bz2|patch -p1
make oldconfig

Re:Be gentle to the mirrors

[narfbot]

kernel.org was already slow before slashdot posted the story. And even then I was only trying to download patch-2.4.21-rc8-final.gz ! What good that did...

Re:Be gentle to the mirrors

[Spock the Vulcan]

When I can download the Windows kernel source code and compile it in less than 5 commands, give me a call.

Re:Be gentle to the mirrors

[narfbot]

cd /usr/src/linux
bzcat /blah/patch-2.4.21.bz2|patch -p1
make oldconfig
blah blah blah

And you people still wonder why so many people stick to Windows instead of trying Linux? When I can double-click an icon and click OK and have it install the updates give me a call.

What do you expect? Vanilla kernel compiles/updates aren't for the average joe. They can wait for their easy distro update.

Re:Be gentle to the mirrors

[hendridm]

> When I can double-click an icon and click OK and have it install the updates give me a call.

Should we let you know when it gets oppressive EULAs and DRM controlled by a single company too? I wouldn't hold your breath.

re: be gentle to the mirrors

[ed.han]

"should we let you know when it gets oppressive EULAs and DRM controlled by a single company too?"

wouldn't he know when he begins the download process?

ed

Re: be gentle to the mirrors

[Anonymous Coward]

Are you a moron? He asked to be made aware of when Linux was as "easy" as Windows. It is likely he wouldn't download it until being made aware that it has "caught up".

Re:Be gentle to the mirrors

[Jetson]

Instead of downloading the entire kernel, download just the patch file if you are running the previous version.

They should seriously consider moving the mirrors list off of the main website. I can't get to www.kernel.org/mirrors right now, presumably because it's buried in requests for 25Mb downloads.

Re:Be gentle to the mirrors

[koreth]

I bet you can get the Windows source code with one command:
mail -s "My resume" jobs@microsoft.com

Re:Be gentle to the mirrors

[volkerdi]

Instead of downloading the entire kernel, download just the patch file if you are running the previous version.

But, is anyone actually running vanilla 2.4.20, or keeping it in /usr/src/linux? Odds are that most people are running vendor kernels (even in Slackware the usually virgin kernel has been slightly violated this time ;), or at least applied the ptrace hole fix. As that fix is implemented differently in 2.4.21 it would be enough to keep the patch from applying cleanly.

Of course, you might still have the official linux-2.4.20.tar.bz2 tarball sitting around somewhere, and if you do you can use that and the patch-2.4.21.

Re:Be gentle to the mirrors

[Shagg]

And you people still wonder why so many people stick to Windows instead of trying Linux? When I can double-click an icon and click OK and have it install the updates give me a call.

Wow, what icon can I click on in Windows to compile a new version of the OS from source code?

Re:Be gentle to the mirrors

[42forty-two42]

Use ftp.countrycode.kernel.org - e.g. ftp.us.kernel.org. Most won't be swamped with requests. Also, the .us one has 27 DNS round-robin mirrors, so that will probably always work...

Re:Be gentle to the mirrors

[42forty-two42]

You can back out patches, you know :)

Re:Be gentle to the mirrors

[Leebert]

cd /usr/src/linux
bzcat /blah/patch-2.4.21.bz2|patch -p1
make oldconfig

Or, if you are patching up multiple levels (and you're lazy like me), put all the intermediate patches in /usr/local/src, then in /usr/local/src/linux run: ./scripts/patch-kernel . ..

and it will apply the patches automatically.

Re:Be gentle to the mirrors

[Angry Pixie]

I think the fact that I can't download anything at greater than 5.40K/s in the age of broadband means that I'm being very gentle to the mirrors :)

Re:Be gentle to the mirrors

[kasperd]

I can't get to www.kernel.org/mirrors right now

I recognize that problem. I recall being unable to get to kernel.org for multiple days. At that time I decided to mirror the list of mirrors for my country.

Re:Be gentle to the mirrors

[kasperd]

When I can double-click an icon and click OK and have it install the updates give me a call.

Sorry I cannot help you there. RedHat has something very similar, but you only have to click once on the icon not doubleclick.

Torrent

[angryLNX]

Is there a torrent out there with the kernel in it? Mod the reply with torrent link up.

Re:Torrent

[Anonymous Coward]

Why doesn't somebody just patch wget so that it first checks to see if anyone else has wget-torrent'ed a link, and then if not, it creates a .torrent and exports its presence to some central location? That way we could all use wget and there'd always be a torrent-downloadable version.

Re:Torrent

[zoloto]

this may seem a bit redundant, but I haven't seen anyone post about it yet. Release the patch on the kernel.org site a week in advance. Post the gpg signature for the patch AND the full kernel source. Then post the full kernel on kazaa, eDonkey, torrent etc. they download the kernel and then can verify the archive with the gpg sig and md5sum.

though those who provide the spaces can handle the demand, i figure we can ease up on costs for them by using some of our own precious bandwidth even if we set a cap on it. 100,000 users at 5k is faster then 1 server at 225k. even if it it takes a large community effort :D

My Thanks goes to

[Lolaine]

to the kernel team. I hope this release makes my system faster, since it uses a lot of disk I/O. Will *try* to download it now.

The question is: Will slashdot be slashdotting the kernel mirrors more than they are slashdotted after mail is sent to the kernel list?

Good luck mirror's sysadmins :D

Whoo hoo

[Anonymous Coward]

1) Cowboy Neal back in the Polls.
2) Linux Kernel update articles.
3) Anonymous Cowards getting voted Trollbait
or whatever.

Slashdot is back !!!

Admin Question

[peripatetic_bum]

Im in charge of keeping production servers up to date. Im using a 2.4.20-rc7.
should I go up tp 2.4.21

How often should I build the new kernels for production servers (should I even be using 2.4? ie, stick with 2.2)

Thanks for the Help!

Re:Admin Question

[Malc]

Why are you running a release candidate on a production server?

If the server(s) is/are performing correctly, why bother upgrading? What will it buy you? Just apply security patches and don't tinker with anything else.

Re:Admin Question

[xanadu-xtroot.com]

Im using a 2.4.20-rc7.

Well, the 2.4.21 kernel was (in reality) the RC-8. Look at the changelog and see if any of that applies to you. If so then yes, it's would be wise to upgrade. If not, then it's your call. This is why the changlog exists...

Re:Admin Question

[Anonymous Coward]

You are either an incompetent administrator or a troll. Either way I don't give a fuck.

Re:Admin Question

[oohp]

What release candidate are you talking about? This is an *official* kernel release. It's not like 2.4.15 or anything. Go figgure.

Re:Admin Question

[Malc]

Doesn't the "rc" in "2.4.20-rc7" indicate it's a release candidate? If not, then I take back that part of my comment.

Re:Admin Question

[stefanlasiewski]

Im in charge of keeping production servers up to date. Im using a 2.4.20-rc7.
should I go up tp 2.4.21

If stability is important to you, you should only use proven, stable kernels on a production server.

Unless there is some new feature that you absolutely need RIGHT NOW and cannot wait, it is very bad to use 2.4.20-rc7 on a production server. The "rc" stands for "release candidate", which means that the kernel is almost ready to be used by the public, but needs people to test it first.

If you care about system stability, you should not be testing the kernel on a production machine. If you do want to test the kernel, do so on a test machine that is not a mission critical machine.

On several occasions in the past, a release-candidate kernel introduced new code which would crash or corrupt systems that used the kernel.

However, if 2.4.20-rc7 is not crashing on you, you don't need to upgrade to 2.4.21 right away. Review the kernel changelog, and see if any of the changes apply to you. Wait a few days (or weeks), and upgrade to 2.4.21 when it's convenient to you.

Personally, unless there is some urgent fix that I need in the new kernel, I always wait a few weeks or months before upgrading the kernel, just in case some wierd bug was introduced into the new kernel version. During those weeks or months, I usually test the new kernel on a test machine and see if anything wierd happens.

Re:Admin Question

[Anonymous Coward]

Yes -rc means 'release candidate'. And -pre means 'prepatch'.

Re:Admin Question

[DShard]

I don't think there is anything wrong with using release candidates on ANY system. I do think you should avoid the development branch unless you have done enough testing to be sure it is stable enough... and even then only after a feature freeze.

Re:Admin Question

[mr.nicholas]

Im in charge of keeping production servers up to date. Im using a 2.4.20-rc7. should I go up to 2.4.21

If you are in charge of administering a number of production-level servers, you should know the answer to this question already.

(And people wonder why there are 100,000 open proxies/mail servers on the net today. Geez!)

Re:Admin Question

[Necrobruiser]

What I would do if I were you is tell your boss to hire someone new, if you aren't just a troll. You don't seem to have the skills yet to maintain production servers.

Good Answer! I'm pretty sure that he was looking for a sarcastic and elitist response to his question rather than, say, useful information. Assclown.

Re:Admin Question

[ChadN]

Obviously "oohp" didn't read the post you responded to. That said, it is not inappropriate to use a release candidate on production servers, especially when the rate of rc releases was a slow as it was. If the RC fixes a specific problem you are having, or offers a specific improvement you could really use, it is worth investigating. There are also issues of security to be concerned about; the iperm and networking bugs that exist in 2.4.20.

The main point is that one should do a small test deployment, and some heavy testing, before a wider deployment. There a probably few, if any, user mode level compatibility problems between 2.4.20 and 2.4.21, so reverting back to 2.4.20 should be fairly easy if there are problems. It all depends on the situation.

Re:Admin Question

[Some Dumbass...]

Why are you running a release candidate on a production server?

Uh, because there were several vulnerabilities found in the 2.4.20 kernels which were only fixed in the -rc's? See this summary.

Just apply security patches and don't tinker with anything else.

Sometimes the kernel needs security patches too.

Re:Admin Question

[kelleher]

Quiet! Leave the guy alone!

I get paid good money to come in and clean shops up after sloppy Admins have created unstable messes... :)

Re:Admin Question

[Jack Auf]

Im in charge of keeping production servers up to date. Im using a 2.4.20-rc7.

Dude you are *so* fired. What in the world possessed you to use an RC kernel in prod? Are you that foolish or that bored?

I have public servers running 2.2.20 that have been up for months and months and months. There's no *reason* to upgrade kernel.

You really should be fired. Or quit and find a new line of work. Really.

Re:Admin Question

[sjames]

For starters, don't run pre or release cantidates on a production machine, run them on a test machine.

Second, is it working now? If yes, it's low priority except for the sopecial case of the PTRACE security hole. Next, look at the changelog. Does anything there significantly improve the performance of your apps/hardware. If no, don't upgrade.

Re:Admin Question

[morgajel]

I could see upgrading a production server if the kernel enhancements improve disk IO.
If there's some benefit, then try it- however, use a test system first. Also, wait a while before you install it- make sure there aren't any nasty bugs that are found a week or two down the road.

I still don't understand why he's running a RC on a production server. that's a Bad Idea.
Since you are running an RC, you'd probably be better off with the 2.4.21 release.

Re:Admin Question

[Ringlord]

Should'nt the parent to this be modded 'Funny'?

Re:Admin Question

[ichimunki]

Do you call requiring doctors to have gone through medical school elitism? I don't. I call it proper training. This guy needs to get some proper training or he should let one of the many unemployed Slashdotters he expected to do the difficult work of reading the Changelog to him take over for him.

Re:Admin Question

[real_b0fh]

pfft I used 2.4.0 in production servers, as well as some late 2.3.x kernels, ZERO problems.

the "how dare you to run this unstable shit on production" is mostly exageration.

Re:Admin Question

[peripatetic_bum]

Ha!

What's funny is I am a doctor and have passed all the tests.

So yes, You are right, I should have known better than to put a 2.4 kernel on a production server.
Unfortunately, i have no proper traning for server admin, just learning as I go.

The non-profit i *volunteer* for is happy with the results, but I shall endeavor to do better.

THanks!

G

unstable 2.5

[jmays]

2.5 has a release today as well. 2.5.70

Re:unstable 2.5

[Jayr]

Actually, that was released May 27th.

Re:unstable 2.5

[jmays]

Yeah, you are right. I suppose I should have hit kernel.org. Stupid freshmeat lag.

Broadcom support

[DeathPenguin]

Any news on supporting those damned Broadcom 802.11 chipsets for wireless? Prism seems to have been phased out already :(

Re:Broadcom support

[mahdi13]

Most likely not. That would of been a whole Slashdot article in itself!

Anyway...I don't think Hell has frozen over quite yet

Re:Broadcom support

[wazlaf]

Just apply the Host AP patch and your Prism cards will work just fine.

I take it...

[Albanach]

SCO won't be offering this new kernel for existing Caldera Linux users then?

Re:I take it...

[WindBourne]

Perhaps, not for caldera users, but I would not be surprised to hear that unixWare users suddenly get a big patch.
No your honor, we that was our code that Linux put out there. We simply put our stuff out later.

Re:I take it...

[Anonymous Coward]

SCO won't be offering this new kernel for existing Caldera Linux users then?

I doubt it, but they're still distributing 2.4.13.

Since they've stopped distributing the infringing code, I presume it must have show up in 2.4.14 or later.

why not posting the http links instead of the ftp?

[Anonymous Coward]

http://www.de.kernel.org/pub/linux/kernel/v2.4/Cha ngeLog-2.4.21

for example.

ftp needs much more time and authentication stuff for login, commandos and so forth.

fr

jp

and so on...

Another version!

[Anonymous Coward]

Oh man! My 286 just finished compiling 2.4.20!

Get it before it's SCOne ( SCOre : -1 )

[Anonymous Coward]

You just know those SCrOtums will be SCOrching penguins tonight (It's Friday the 13th remember).

At last

[SquadBoy]

explicit Nforce support. This is a happy day. :)

Re:At last

[jericho4.0]

For those of you with nforce chipsets who have non-nvidia graphics cards, this link will give you agpart support. It's the new release of the linux drivers. It's not linked to from the nvidia.com main page, for some reason, but it works.

Enjoy, losers!

[Anonymous Coward]

Ha, I'm glad I have a life, unlike you losers! Knock yourselves out compiling the newest kernel; I'll be out having a blast with my girlfriend in my Porsche.

(Okay, okay; I'm sitting at home in my underpants reloading slashdot to see if new stories are posted. I'll go download it now.)

Re:Enjoy, losers!

[Zeriel]

(retraction at bottom of comment ignored *g*)

>Knock yourselves out compiling the newest kernel; I'll be
>out having a blast with my girlfriend in my Porsche.

*grins* Knock yourself our in your tiny-ass car, I'll be compiling my kernel from the comfort of my bed with my geek fiancee. =P

Re:Enjoy, losers!

[Anonymous Coward]

my girlfriend in my Porsche.

Imagine a Beowulf cluster of those!

Re:Enjoy, losers!

[Synn]

[quote]
I'll be compiling my kernel from the comfort of my bed with my geek fiancee. =P
[/quote]

We hate you.

She have any sisters? :)

Re:Enjoy, losers!

[Anonymous Coward]

She have any sisters? :)

His right hand.

Kidding, kidding.. 0: )

Re:Enjoy, losers!

[Bronster]

Hehe, mine isn't a geek - well, not about computers anyway, but I'm cuddled up in bed with my 2 week old daughter - maybe she will be a geek.

Re:Enjoy, losers!

[Zeriel]

Good news: Yes, she does.
Bad news: Said sister is 11. ...stop thinking that. I WILL remove parts of your anatomy. =P

Re:Enjoy, losers!

[Zeriel]

Pardon me, sir, but your jealousy is showing.

Ywah!

[Anonymous Coward]

ah finally better ide performance!! thanks kernel-crew ;)

excellent, we can switch to this

[mikeee]

when our AIX licenses expire today.

2.6?

[Damek]

I thought 2.6 was supposed to be out by now?

Quick Question

[dracocat]

I have a server running 2.4.20 and it took forever to get the kernel compiled with all the correct modules.

When is it worth upgrading kernel versions?

Is there a way I can easily use the old configuration? Any HOW-TO on this?

I must say I am very reluctant to upgrade the kernel.. especially when I don't have physical access to the machine. But I would of course love to the the fastest and most secure server as possible. Just curious what rules and procedures others use.

Re:Quick Question

[Neon Spiral Injector]

$ cp linux-2.4.20/.config linux-2.4.21/
$ cd linux-2.4.21
$ make oldconfig

Re:Quick Question

[Anonymous Coward]

Use the old config?

Get the patch from 2.4.20 to 2.4.21 instead of getting the full source.
Then just go through and check for anything new you want/don't want and compile away.

Re:Quick Question

[Isaac-Lew]

Did you try backing up your /usr/src/linux-2.4*/.config file and then run:

make oldconfig

That will only prompt you for new stuff, rather than go back through every single option.

Re:Quick Question

[Osty]

When is it worth upgrading kernel versions?

When there's a compelling reason to upgrade. Those fall into two categories:

• The kernel fixes a previous security problem, or
• The kernel provides new features that you require for your product (not "want", but "need").

Any other reason is superfluous, especially for a server machine.

Is there a way I can easily use the old configuration?

The kernel config writes a .config file in the source root. Use that. If you patch rather than grabbing completely new sources, you won't even need to worry about copying that file around (unless you do a make mrproper, which you probably don't need to do unless stuff starts breaking during compile).

I must say I am very reluctant to upgrade the kernel.. especially when I don't have physical access to the machine. But I would of course love to the the fastest and most secure server as possible. Just curious what rules and procedures others use.

Years ago, back when the kernel was being updated nearly every other week rather than once every few months (2.0/2.2 time frame), I would always download the very latest kernel and compile that. Coincidentally, I was also learning Linux at the time, so I didn't mind spending time on stuff like that, and I was in school which meant a lot more free time. These days, my only linux box is a server, so unless there's a security fix I'm inclined to just leave the box alone. It's certainly easer not to upgrade than it is to upgrade.

Re:Quick Question

[tka]

When is it worth upgrading kernel versions?

This seems to be frequent question. I'd say that you don't need to compile new kernel until the old one doesn't have the xyz feature that you need and/or you feel that the new kernel is far more reliable and faster.

In other words: some still use 2.2 series because there is no reason (for them) to upgrade.

Remember that you can use modules to get that xyz feature..

Re:Quick Question

[bourne]

Is there a way I can easily use the old configuration? Any HOW-TO on this?

Copy the old arch/$(ARCH)/config.in file to the new tree and 'make oldconfig'

Re:Quick Question

[Anonymous Coward]

Personally when I've done that I've had nothig but problems. I'm sure it must work for everyone else, but I find uless I do it from scratch everytime ther are always module problems.

Re:Quick Question

[praedor]

Years ago, back when the kernel was being updated nearly every other week rather than once every few months (2.0/2.2 time frame), I would always download the very latest kernel and compile that. Coincidentally, I was also learning Linux at the time, so I didn't mind spending time on stuff like that...

Same here. I used to download and build kernels up the yingyang as each new update/version/subversion came out. That was when it took a bit longer (or at least the same amount of time) to build a kernel as compared to KDE. Now the kernel builds fast and KDE takes for-freakin-ever...if you manage to meet all the dependencies... In any case, I only upgrade kernels on my personal system if security is improved, a feature that I really want (almost even need) has been added, or as in this case with 2.4.21, something really painful has been addressed...slow IDE i/o. I'll upgrade to this kernel and build it as usual (with all the nice grsecurity patches).

Re:Quick Question

[Anonymous Coward]

These days, my only linux box is a server, so unless there's a security fix I'm inclined to just leave the box alone. It's certainly easer not to upgrade than it is to upgrade.

[looks over changelog]
Um, what's your IP? I wanna check out your, um, content! Yea, that's the ticket.

Re:Quick Question

[Alioth]

I found it worth upgrading because:

1. I was running 2.4.19, which has a local root hole (I had band-aided it, but I really wanted a fixed kernel).
2. I wanted the skas3 patch for UMLs.

I also don't have physical access to the machine - it's 4800 miles away. I *hate* remotely upgrading kernels. However, this can be mitigated by doing local testing with the newly built kernel image (or in my case, since the machine at home is P4, and the remote is Athlon, a build with the same config options apart from the different processor being selected).

Re:Quick Question

[/dev/zero]

What I do is do a "save configuration as" (from make xconfig) into a file of the form $host-$date.config, and then copy it into my configuration tree.

When I build a new kernel (like right now), I copy this saved config into /usr/src/linux-2.4.21 (for example), and then from make xconfig I do a "load configuration from". Then I do a quick once-over to make sure all is well. If I make changes, it's "save configuration as" again. Then "save and exit" an on to the build.

This way I can reproduce old kernels if need be (like if something breaks with the new one, and I decide to patch the old one and keep using it).

Gordon.

Re:Quick Question

[JDWTopGuy]

Now the kernel builds fast and KDE takes for-freakin-ever...if you manage to meet all the dependencies...

Ironically, that's what I've been doing lately. I just grabbed the patch update to 2.4.21 and will be compiling it as soon as I finish recompiling QT (again).

Adding to the irony, I'm compiling this stuff on my mac (under VPC... too scared to try making a GCC cross-compiler) so I can use it on my 90Mhz Pentium that I only use on weekends.

And I'm replying to a foe of mine.

I think I'll go choke on my pocket guide to the heimlich (sp?) maneuver.

heres how to compile the kernel

[net_bh]

$ cd /path/to/kernel
$ make menuconfig # Replace menuconfig with xconfig if you want
-Select the options you need in your kernel and save it
$ make dep bzImage
-Look in arch//boot/ for the bzImage file
-Install it for your favourite bootloader (grub/lilo) and reboot machine
-gloat :)

Re:heres how to compile the kernel

[Anonymous Coward]

some people might have modules

try this:
# make menuconfig
# make dep && make bzImage modules modules_install

Re:heres how to compile the kernel

[Cable_Monkey]

Whatever happened to "make modules" and "make modules_install"? ;)

Re:heres how to compile the kernel

[mahdi13]

You forgot some...

make modules
make modules_install

Unless you are a masochistic...er, monolithic kernel builder =)

Re:heres how to compile the kernel

[xanadu-xtroot.com]

# cd /usr/src/linux
# make menuconfig
# make dep clean modules bzImage
(assuming all's good)
make modules_install install


The scripts have been able to install the kernel itself by itself for quite a while and also detect lilo or grub or whatever. I have no idea why people keep telling the n00bs to cp ./arch/what/ever/

Re:heres how to compile the kernel

[Osty]

# make dep && make bzImage modules modules_install

Why run make twice when you could run it once? Also, it's a good idea to run the "clean" target, especially if you used a patch.

# make dep clean bzImage modules modules_install

Re:heres how to compile the kernel

[tka]

You forgot modules and lot's of pain & sweat and also reboots..;)

Re:heres how to compile the kernel

[Anonymous Coward]

and if you are using lilo, don't forget to run "lilo" :)

Re:heres how to compile the kernel

[Laplace]

That's me! Seriously, if I know that I will be using certain modules without fail, why not add them to the main kernel executable? My kernel may be big, but it kicks ass.

Re:heres how to compile the kernel

[Anonymous Coward]

Everyone says this is easy, then gives your list. Step #2 is deadly complex. I wasted most of a weekend trying to upgrade a RedHat 9.0 kernel.

After much work, I found how to do it:
$ cd /path/to/kernel
$ cp /boot/YourOldConfig .config
$ make oldconfig
$ make menuconfig
ETC

Re:heres how to compile the kernel

[FyRE666]

The scripts have been able to install the kernel itself by itself for quite a while and also detect lilo or grub or whatever. I have no idea why people keep telling the n00bs to cp ./arch/what/ever/

Because it doesn't work very well on all systems. I've tried the "make install" last stage and always ended up having to edit lilo.conf anyway to be sure the machines could boot afterward. Far easier to just mv the old bzImage and System.map (optional), then copy the new ones into the /boot partition and run lilo... Never failed me in years.

Grub, well, that's failed me several times, and I certainly wouldn't ever use it or recommend it again (unless you need a pretty boot screen and don't mind the risk).

Re:heres how to compile the kernel

[Minna Kirai]

In the past, putting all the targets on one make invocation could fail mysteriously. I never tracked the problem to it's source (the workaround is easy enough), although presumably it involved the makefiles created in the "dep" stage not being read in time for the other targets to see them. (This might be dependent on the version of "make" you have)

Re:heres how to compile the kernel

[pnatural]

Seriously, if I know that I will be using certain modules without fail, why not add them to the main kernel executable?

Because some modules won't work properly compiled into the kernel!

Re:heres how to compile the kernel

[42forty-two42]

Don't forget:
make modules modules_install

Re:heres how to compile the kernel

[jooon]

I suggest people who use Debian do it the Debian way.

apt-get install kernel-package

This will install make-kpkg which compiles the kernel into a debian package. Go to the source directory, configure it with make menuconfig or something of your choice. Then:

make-kpkg --revision custom1 kernel_image

then after it is finished:

dpkg -i kernel-image-2.4.21_custom1_i386.deb

Re:heres how to compile the kernel

[parkanoid]

Because you usually want *control* of where and how the kernel goes. install just puts it into the default path and (AFAIR, I haven't used it in a bit) updates lilo accordingly. I name my kernels according to version and specific features I compiled in. A naming scheme allows me to roll back to a specific older kernel that I know to be working if something goes wrong. It's possible to modify the makefile to do that for me, to an extent, but I just don't compile kernels often enough to bother.

Re:heres how to compile the kernel

[Anonymous Coward]

For the Gentoo crowd, you can get the kernel sources using:
# emerge rsync
# emerge -u vanilla-sources
before the compile steps.

Re:heres how to compile the kernel

[damiam]

In that case, there shouldn't be the option to compile them in. If there is, it's a bug.

Re:heres how to compile the kernel

[skookum]

So, can I have some karma for copying and pasting a README file into a post? Seriously. Informative my ass.

Re:heres how to compile the kernel

[Laplace]

Which I am aware of, and consider to be a bug.

Re:heres how to compile the kernel

[anonymous cupboard]

Personally, I save my config outside the tree as linux-yymmdd.config so:

# cd /usr/src/linux
# make mrproper

The 'mrproper' seems extreme but I really prefer to clean up properly between releases. # cp ../linux-yymmdd.config ./.oldconfig # make oldconfig # make dep bzImage modules # make modules_install install # reboot RH automatically connects the kernel to the grub or lilo boot loaders. The

2.4.21 final is the same as RC8

[Anonymous Coward]

just for information according to the changelogs:

final:

- 2.4.21-rc8 was released as 2.4.21 with no changes.

dont upgrade if u have RC8 running already

Re:2.4.21 final is the same as RC8

[Neon Spiral Injector]

That how they will always do it now, and have done it for a few releases. There were problems with the final rc working okay, but one little patch that shouldn't have hurt anything, hurting something.

2.4.21-rc8 will not patch cleanly to 2.4.22-rc1, because there is one tiny change, the version number.

2.4.21?

[bazik]

bazik@synack bazik $ uname -r
2.5.70
bazik@synack bazik $


Runs perfect on my system =) Looking forward for 2.6.x!

Re:2.4.21?

[joe_bruin]

[joe@circe ~]$ uname -r
86.9.14
[joe@circe ~]$

mine's better, i win.

Re:2.4.21?

[FroMan]

I've played with 2.5 on a couple less used machines, but I always seem to have module problems. Is this still the case?

Re:2.4.21?

[bazik]

About the module problems:

You have to install "module-init-tools" if you use the 2.5.x series: *click*

Re:2.4.21?

[ldspartan]

$ uname -r
1.3.22(0.78/3/2)

mine's harder to read, i win.

Re:2.4.21?

[Kourino]

The module-init-tools link is valid, but you really should read this if you want to try 2.5 and haven't been following the development.

Re:2.4.21?

[Fembot]

I'll switch to 2.5.x or 2.6.x when the nvidia module works with it.

Re:2.4.21?

[Tony Hoyle]

I'll use 2.5.x when the damn framebuffer works... I like my console to be nice and wide.

2.5.70 just locks hard if you enable the framebuffer... doesn't even start to boot properly.

Also, my network card isn't supported yet (they have a driver but it's been broken since 2.5. and nobody seems to care about fixing it).

Re:2.4.21?

[bazik]

I have framebuffer AND nvidia drivers working.

Gentoo has patches for the nvidia drivers in portage :)

Re:2.4.21?

[Fembot]

I thought that the NVidia drivers were closed? How can that work?

Re:2.4.21?

[Genyin]

I thought that the NVidia drivers were closed? How can that work?

They're basically a wrapper around a binary driver... patching the wrapper is sufficient.

Re:2.4.21?

[JDWTopGuy]

% uname -a
Darwin JDWTopGuyPBG4.local. 6.6 Darwin Kernel Version 6.6: Thu May 1 21:48:54 PDT 2003; root:xnu/xnu-344.34.obj~1/RELEASE_PPC Power Macintosh powerpc

Oh wait, was this about Linux? I have to build 2.4.21 for my VPC installation and my 90Mhz Pentium... "Behemoth"

Re:2.4.21?

[MS_is_the_best]

The framebuffer works for me fine in 2.5.69? I have patched the source a bit for a proper default resolution/refresh rate (nasty...), but before that it already worked in 1024x768 (NVIDIA).

However notice that the fbset is not updated for the new kernel, so don't use it.

Re:2.4.21?

[Fembot]

so is there a way for us mortals to download the patch without having gentoo?

Re:2.4.21?

[bazik]

Yep! Nvidia driver patch for Kernel 2.5.x:

http://www.minion.de/files/NVIDIA_kernel-1.0-4363- 2.5.diff
See http://www.minion.de/nvidia.html for more info.

Disc IO Related Lockups?

[Jerk City Troll]

I have two systems that receive heavy use. Both of them are often used for ripping and the dual processor system is used for encoding. Whenever either of these systems is under heavy load, and I rip a DVD or image a CD, weird things happen. I get IO timeouts and sometimes even lock ups. Under normal load, there is absolutely no trouble at all, except with the dual processor system. That machine does filesystem crypto and thus, it's processors are quite stressed by cryptoloop processes whenver the disks are active. Dumping a disc to a filesystem on that box sometimes produces annoying problems. I've had solid lock ups, inability to unmount and eject discs because processes won't release them, and sometimes even X just stops responding.

Both systems are running 2.4.20. Now, question: are problems like these resolved in 2.4.21 with these IO fixes? Remember, the drives doing the reading are probably fine. The one machine has two, a DVD-ROM and a CD-RW, and the other has a DVD+-RW. All three drives cannot possibly be faulty, nor can both IDE controllers. The problem has got to be with software. I cannot think back to when this began, but it may have been for the life of 2.4.20.

So is there been something screwy with the IDE-CD subsystem in Linux lately?

Re:Disc IO Related Lockups?

[bulletman]

Do you overclock? It (ahem ) has been known to cause problems.

Re:Disc IO Related Lockups?

[Cnik70]

Did you mount your drives with defualt in the /etc/fstab file? If so, use noatime, notail instead. it will prevent those lockups.

Re:Disc IO Related Lockups?

[grigori]

That's one of the problems with Linux as it is today - it doesn't have a preemptable kernel, you can get priority inversion, etc. Plus, the IDE code that's in the process of being ripped out has lots of nastyisms in it. So, you can get race conditions, lockups, and DEFINITELY get very jerky, nonsmooth response for interactive apps while doing heavy IO. Some of this supposed to be lots better in 2.6. We'll see. What, didnt you know that Linux wasnt the perfect OS of all time? (Ssssh, dont tell anybody, it might burst there bubble, and distract them from sliming other OSes that solved this junk years ago)

Re:Disc IO Related Lockups?

[Anonymous Coward]

How is one supposed to know that? Is it a problem that will be fixed in some future kernel, so that "default" will set the correct options for problematic disks?

Re:Disc IO Related Lockups?

[arjun]

Instead of whining here, why don't you test the rcX candidates and report your findings on lkml, like the rest of us. You have better chances of getting things fixed.

Re:Disc IO Related Lockups?

[bogie]

"Ssssh, dont tell anybody, it might burst there bubble, and distract them from sliming other OSes that solved this junk years ago"

Then why does windows lock up tight whenever you need to delete a large file? Windows falls to pieces whenever your doing anything with high IO. Feel free to lower the priority of the high IO process because that's the only way you'll even be able to open Notepad once a high IO task starts in Windows 2k/XP. You really think explorer isn't a crippled piece of shit? Windows doesn't do anything that could be called multitaking once you have a process hogger. At least with linux there are patches which can make it useable under the most severe loads.

Nice try though.

Re:Disc IO Related Lockups?

[orcrist]

I had symptoms like that on my laptop with my dvd drive. Turning off dma for that drive solved it. Don't know if that applies to your situation.

-chris

Re:Disc IO Related Lockups?

[rabtech]

Yes, his was a nice try as opposed to your troll which is a very very poor try.

First, you are intermixing processor hogging and I/O hogging, both of which Windows 2k/XP handle just fine. Deleting a large file is the same thing as deleting a small file: marking it as deleted in the journal & the MFT. Perhaps you mean that you have explorer's previewing feature turned on, and when you click on a large movie file, it cannot be deleted until explorer stops loading it to show you the preview? If so, turn off the previewing function. Besides, if you only have a single IDE hard disk, it is trivial to lock up the disk with lots of I/O, and this works for any OS.... unless the OS is willing to preempt and greatly slow the high I/O process to make room for other stuff. Of course you could just give that process a priority of "Low" and that would help.

However, there are some SCSI situations for XP that were fixed in SP1a, as well as some driver bugs (OMG!) that could cause systems under heavy load to start crawling. Why, sounds just like some Linux drivers I know. For shame!

Re:Disc IO Related Lockups?

[Cnik70]

I don't see it exactly as a problem, more like a result of a setting that allows the system to constantly update the drive info with file access time and tail info (which can require loads of extra io's when loading a file). when you set the system to not update access time and tail info on a drive, it removes this extra io taskload.

Re:Disc IO Related Lockups?

[Anonymous Coward]

Slashdot moderation in action:

"What, didnt you know that Linux wasnt the perfect OS of all time" = Troll
Offtopic "slime" against Windows = Insightful.

They should just change the labels to "-1 Microsoft Sucks" and "+1 Linux is great".

Re:Disc IO Related Lockups?

[oscarcar]

My guess is that it's Direct Memory Access (DMA) issues.

A lot of DMA and IDE issues have been fixed in the 2.4.21prerelease series. the fixes have been in the Alan Cox (ac) series for awhile, but only very recently in Linus' kernel.

My setup works awesome since I've had these fixes. I use the VIA KT400 chipset with two CD-Writers and a DVD-ROM.

I can have 2 burning at the same time and read from the DVD without a hitch. Can't even notice anything from the user perspective.

Getting the latest "cdrecord" can also help but I think the main issues were the IDE/DMA ones.

will tell you a lot of things about your IDE device: /sbin/hdparm /dev/hdx

will tell you specifically if DMA is being used: /sbin/hdparm -d /dev/hdx

Oscar

Re:Disc IO Related Lockups?

[lmfr]

I had a similar problem. Under heavy I/O, my system locked up for some seconds and complained about lost interrupts.

The new kernel (well, 2.4.21-rc8) fixed that.

One workaround mentioned in the LKML is to disable IO-APIC.

Re:Disc IO Related Lockups?

[Anonymous Coward]

You don't see machine lockups as a problem? That's the silliest thing I've ever heard.

Re:Disc IO Related Lockups?

[sludg-o]

Are you sure you have DMA enabled? Several distros (including RH 7.3 and 8, not sure about others)
don't have DMA enabled by default. Log in as root and type "hdparm /dev/hd(a,b,c,d) and see if the using_dma flag is set to 0 (off) or 1 (on). If it's turned off, the -d flag toggles on/off for dma.

Ripping and encoding do not really cause heavy disk IO. It's all sequential. High CPU time no doubt, but not disk. I run a squid proxy for 5000 pr0n hungry college kids. That has high disk IO.

Re:Disc IO Related Lockups?

[grigori]

Peace. I agree 100% I didnt mean Winblows. Relax, dude

Re:Disc IO Related Lockups?

[grigori]

too bad this post got modded down by oafs who thought it was a troll for MSFT. Makes the same points others making about current Linux, but noooo, people have to get defensive if you have the nerve to point out anythings wrong. read the kernel mailing list if you think this stuff is madeup, and have a grownup exaplain what the big words about OSes mean

Re:Disc IO Related Lockups?

[Cnik70]

no, i see lockups as a problem, but it's not the kernel's fault. in cases such as this it's usually the filesystems fault. two completely different entities.

Re:Disc IO Related Lockups?

[Stardate]

Actually there are plenty of conditions in XP (even the latest fully patched version) that can cause a process to suddenly start hogging 100% of the CPU, such that even if you press CTRL+ALT+DEL, XP will take an absurdly long time to bring up TaskMan. Even if it does come up, it's iffy as to whether you'll be able to find the process that's causing the problem (either explorer, csrss.exe, or a lovely 'System' process that you can't kill).

A lot of the new 'chrome' features in XP (as well as the horribly annoying preview features that don't allow you to delete files) can be the cause of it. But even under heavy load, linux always allows you to kill processes.

Re:Disc IO Related Lockups?

[ralphclark]

Well, except that what he said about the IDE code is true, at least in 2.4.20. Systems under high load can freeze up badly, this is particularly noticeable when ripping a CD to disk but that's not the only way it can happen.

Also once the task responsible has gone you'd think the system would return to normal. It doesn't. God help you if you had a lot of web pages open in Mozilla; it's the first thing to freeze up and once it does that it doesn't come back even if CPU & memory utilization elsewhere is cut back to the minimum. Parts of KDE can go into a permanent sleep like that too.

I don't know why some applications should be so utterly incapable of recovery after a resource shortage, I'm just reporting an observation.

This is the first IDE system I've run Linux on. it's never happened to me on any of my SCSI systems running the same 2.4.x kernel series.

Re:Disc IO Related Lockups?

[pkplex]

Yep, FreeBSD owns for smooth multitaskng :)

Its also a lot cleaner then Linux, and as you mention, the licence is much more free and proper.

The GNU licence its self is bloated.

2.6 kernel

[haydenth]

Does anybody know when the 2.6 kernel releases are coming out? I can't wait.

Re:2.6 kernel

[Neon Spiral Injector]

Last week I saw on l-k, 1 year, but it was later ammended to 6 months.

So 6 months to a year from last week.

The late 2.5.x series is shaping up nicely though. I run it on my home machine and my workstation in the office.

Re:2.6 kernel

[haydenth]

Last October Linus expected it to be released this month... I hope it doesn't get delayed anymore.

Re:2.6 kernel

[FroMan]

Better to hope that the delays will make a better kernel for you. This is software development. Nothing is ever on time. And almost as rare is that software is done right.

I'd prefer late software done right to on time software done wrong.

Re:2.6 kernel

[Cramer]

2.5 is nowhere near ready to be called 2.6. There's just too much stuff that's been changed that hasn't made it to all corners of the kernel. And the kernel maintainers have started taking a Microsoft-like stance of publicly stating they'll remove some stuff that doesn't work ("how many people still need support for Adaptec 1542CF's anyway?")

Ah

[HughJampton]

But does it run linux?

RedHat kernels

[DreadSpoon]

Hoping RH pushes updated kernels for RH9. Piss-poor IDE disk performance is my one big gripe with my Linux boxen at the moment; whole machine feels like shit when something heavy is running the disk in the background. :(

Re:RedHat kernels

[Anonymous Coward]

Have you run hdparm and turned on a bunch of flags to speed up the disk?

Re: RedHat kernels

[Black Parrot]

> Hoping RH pushes updated kernels for RH9. Piss-poor IDE disk performance is my one big gripe with my Linux boxen at the moment; whole machine feels like shit when something heavy is running the disk in the background. :(

\AOL{meetoo}. Actually, even if I just had lots of windows open and not much CPU or disk traffic my UI felt like Windows 95, repeatedly coming to a screeching halt for several seconds at a time, usually when switching from one window or desktop to another.

I finally failed back to an older kernel I still had around, and the problem went away. I don't know whether the problem was with the 2.4.20 series kernels (I tried three) or the rumored Red Hack kernel hack that they purportedly distribute for RH9 (all three I tried were from RH RPMs). I'm just glad I was able to make it go away.

Re:RedHat kernels

[mahdi13]

Red Hat 9 back-ported a disk scheduler which is a dog. either download this 2.4.21 vanilla kernel and compile it or recompile the default RH9 kerenl and remove the scheduler (I believe it is under General, been a while since I looked at this...the option right after the CPU info)

Re: RedHat kernels

[IamTheRealMike]

Ditto.

In fact, I switched to a stock Linus 2.4.20 kernel just last night. So far, no problems. Before, my desktop felt like treacle.

I think I might upgrade to 2.4.21, maybe even apply to pre-empt/low-latency patches. I'm feeling adventurous.

My old SuSE 8.0 box, that used 2.4.18, worked just fine, so I think it's something that red hat screwed up. There's a bug in their bugzilla about some people experiencing swap storms so bad it takes 10 minutes for a newline in a shell!

Re:RedHat kernels

[SuperQ]

Redhat 9 2.4.20 is patched with 2.4.21pre3. so the major fixes are already in there.

Re:RedHat kernels

[bogie]

So your saying the alternative Windows doesn't run like shit once a heavy disk task starts? That's sure not my experience. I just posted this in another thread, but once you start any heavy disk task under Windows you an forget about doing anything else with the OS as you get to watch everything slow to a crawl.

Try out this newest kernel or the preempt patches. ALso like someone else said make sure you have the right hard drive flags set. After making sure your hdparm setting are correct tune ext3 as well.

to change ext3

add data=writeback to the mount points you want to be writeback not data=ordered in fstab
mkinitrd /boot/initrd-kernalversion.img kernelversion (yes this is odd but its the way until you hard code ext3 into the kernel)
update grub if you used a different name to point to correct .img, usually you don't have to do this though

My experience with RH 8.0 after upgrading to 2.4.20, making sure hdparm was correct, and change the ext3 journal type improved dramitically.

Re: RedHat kernels

[Anonymous Coward]

> I switched to a stock Linus 2.4.20 kernel

Geez, get with the times already! Linus 9 has been out for months now, and that Linux Torvagae guy has been really working hard. Least you could do is run the most recent version!

Re: RedHat kernels

[Anonymous Coward]

For those not in the know, treacle is British for 'molasses.' Welcome to America.

Re:RedHat kernels

[Kashif Shaikh]

You have to watch out with Redhat8(don't know about 9). It disabled DMA on my hd causing massive performance problems /w any disk access.

use hdparm and hdparm defaults setting in /etc/sysconfig.

My computer was running flawlessly with DMA enabled on Redhat 7.2/7.3. So I don't know why they disabled it.

Re: RedHat kernels

[mj01nir]

Same here. On my desktop the stock RH9 kernel sucked in a major way. Rolled a new bare 2.4.20 + preemptive kernel and it's flying. Wonder what magic RH was attempting that gummed up the works?

Re: RedHat kernels

[checkyoulater]

I think I might upgrade to 2.4.21, maybe even apply to pre-empt/low-latency patches. I'm feeling adventurous.

If your're feeling really adventurous, try and use Con Kolivas' low-latency patches. I haven't tried them yet, but lots of people on the Gentoo Forums talk about how responsive they make your system.

Only problem I see is that Con admits that his patches are focused more on responsiveness and less on stability.

Re: RedHat kernels

[xtrucial]

For what it's worth, I'm running Gentoo 1.4rc2, kernel 2.4.20, and I get the disk swapping/slowness thing too. So I suspect it's the kernel itself that causes the problem.

Re: RedHat kernels

[Anthony Boyd]

My old SuSE 8.0 box, that used 2.4.18, worked just fine, so I think it's something that red hat screwed up. There's a bug in their bugzilla about some people experiencing swap storms so bad it takes 10 minutes for a newline in a shell!

That would be me. My SuSE 8.1 laptop will take about 5+ minutes to alt-tab to an already open shell window, and type "kill -9" whatever. In my experience, it's Kicker. When I kill Kicker, everything goes right back to snappy.

Of course, it takes a while to get this way. My laptop right now has about 35 days of uptime, with Mozilla & Quanta & Konq & a shell window open the entire 35 days. I'm sure there must be some leaky memory in there somewhere....

Re: RedHat kernels

[Anonymous Coward]

If I open too many windows, then the bugs come in.

ps: I life in Florida...

Re: RedHat kernels

[Bombcar]

Rolled a new bare 2.4.20 + preemptive kernel

And here I read that as:

Rolled a new pre-ban 2.4.20

and though SCO won?!?!?!

Re: RedHat kernels

[raboofje]

Are you sure you have 'hdparm -c1 -d1 /dev/yourdisk'? It makes a huge difference, and seems to be off by default quite often still...

Re: RedHat kernels

[slacker775]

I've been running with Con's patches for a few months now and they work great. I'm on the 2.4.20-ck7 right now and things fly. More importantly, my VMWare works without a hitch which is pretty critical. Stability has not been a problem at all.

Re: RedHat kernels

[jbayes]

I noticed similar problems. My RH 2.4.18 works fine, but 2.4.20 was dog slow. After some investigation, I found out that I can't turn on dma on a 2.4.20 kernel. (I forget the error message, but I remembed hdparm -d1 /dev/hda barfed.) Anyone else notice this?

Impressive!

[SuperBanana]

Oh man! My 286 just finished compiling 2.4.20!

That's especially impressive, since Linux won't run on a 286 ;-)

Re:Impressive!

[protek]

Actually, Linux will run on 8086 and 80286 CPU's. Check out the ELKS project.

Re:Impressive!

[KewlPC]

He could be running a 386 emulator on his 286.

Or maybe he's cross-compiling.

Or...

Shut up ;)

Re:Impressive!

[Anonymous Coward]

s/Shut up ;)/I've run out of ideas/;

Re:Impressive!

[Anonymous Coward]

s/Your Face/My Foot/g

Uhh....what timing

[Cable_Monkey]

All joking aside, I was in the middle of compiling 2.4.20 when I saw this article...

Re:Uhh....what timing

[tim_m]

You're telling me:

Linux athens 2.4.20 #1 Fri Jun 13 09:59:22 EDT 2003 i686 GNU/Linux

I always thought that was just something people said for humor value....

Re:Uhh....what timing

[Lennie]

Same here, well, downloaded the patch...
Control-C, lftp ftp.xx.kernel.org/pub/linux/kernel/v2.4/
cp -va .config ../
bzcat ../patch-2.4.21.bz2 | patch -p1
cp -va ../.config .
make oldconfig
make menuconfig
make...

Re:Uhh....what timing

[phorm]

Using Debian 3: 2.4.18 is the newest available kernel from the standard (stable) apt sources. Never needed more though, why upgrade it if there's nothing you actually would need/use?

Re:Uhh....what timing

[Zerbey]

...and do you have any idea how long it takes to compile a kernel on a P75 with 40Mb of RAM.

(ok, ok it's 83 minutes but, still...)

Re:Uhh....what timing

[Cable_Monkey]

Probably no longer than it takes to recompile the kernel on a P2-400 system that hosts a game that is tieing up 80% of the CPU due to a bug within the game itself. ;-)

Re:Uhh....what timing

[samhalliday]

security updates?? remember the good 'ol kernel/kmod.c thread race condition which gives anyone root access? ;-)

Re:Uhh....what timing

[bfree]

Debian have a tendency to backport security fixes rather than updating a system. The security team released this update to the kernel this month. So with Debian there's no need to update the kernel version for security fixes, instead you can just get the same kernel with the security fix(s).

bad timing

[bumby]

JUST as I was building my new lfs-system, and now kernel.org is slashdotted! GAH!!!

Re:why not posting the http links instead of the f

[Isaac-Lew]

wget will work just as well for http and anonymous ftp transfers. even a little faster for ftp, since it's less letters :).

whatever

[Anonymous Coward]

I don't really know what the commando "blah blah blah" does, but... If you wanna pay big bucks for klicking do so.. I am content with typing three lines of text.. ;P

Arr Laddy!

[jellomizer]

Tell this to me Captin Jello! I got the Hook to replace my hand loss from Compiling Linux Kernel 2.4.21. And I lost an Eye from it too. Arr!

Re:Arr Laddy!

[jellomizer]

After it compiled I got better.

Patch

[Anonymous Coward]

MAJOR boost to the 'feel' of the system under heavy disk IO, especially on IDE systems

Is nVidia optimizing IDE drivers now?

Re:Patch

[FroMan]

Yeah, the driver writes to /dev/null instead of the disk. They get nearly 30% better performace over their normal drivers.

ATI's drivers reorder the writes so all your data is munged, but they only get 2-5% increase in performace. :-)

Re:why not posting the http links instead of the f

[SuperBanana]

ftp needs much more time and authentication stuff for login, commandos and so forth.

Man, can't a guy download a kernel via FTP without a bunch of armed dudes storming the place? Those SCO guys just don't know when to give it up.

Behind the curve already-

[IWantMoreSpamPlease]

Dang!

I just installed Rehhat 9 on my tecra 8100 and am in the process of understanding linux and now this comes along! I'm already behind the curve and I feel my knowledge slipping away from my grasp, furthe... ..wait a minute, what is this clicky thing here???

(with apologies to Dilbert)

Read between the lines.

[Prince_Ali]

He is obviously running a 286 emulating a 386. That would be hell...

Super!

[Anonymous Coward]

I hope they ripped out all that UNIX code too, so I can start using Linux again.

GrSecurity update.

[cyt0plas]

Grsecurity has released a version for 2.4.21. I highly recommend it, especially for production environments with multiple users. I have found the ability to limit outgoing sockets by group invaluable.

n00b3

[Anonymous Coward]

:P ... :)

Got it!

[Isaac-Lew]

Hehe...downloaded it at 166 KB/s.

Hint...UM doesn't only stand for University of Michigan :).

Speed

[Anonymous Coward]

Wow, a MAJOR boost in speed and all they had to do was remove the SCO code..

Question

[CaptainZapp]

make oldconfig doesn't seem to capture the stuff, which is compiled in as module. This means going through all options anyway.

Is there a good way to avoid that, or did I miss something in the RTFM?

Re:Question

[GigsVT]

Well, "compiled in" and "as a module" are two different things, so your question is kinda ambiguous.

But in any case, make oldconfig will carry your entire config forward, what was a module will stay a module, what was compiled in will stay compiled in.

The only thing you need to go through are any new options that were added since the last release.

Re:Question

[minaguib]

make menuconfig, don't change anything, exit. copy the .config file from your old kernel source directory to the new one. make menuconfig again, doublecheck the options as well as look at any ne ones.

Re:Question

[cbcbcb]

you don't need the first make menuconfig, and you should do a make oldconfig before the second. The second make menuconfig then becomes optional.

13th?

[jmays]

Wait, recompile my kernel on Friday the 13th? I think not ....

Re:13th?

[mahdi13]

Sure, what could go wrong?
Just don't delete the old one until you tested the new one...

This is of course if you machine doesn't burst into flames from the HDs spinning so fast with the new optimizations

Re:13th?

[Cruel Angel]

Heh, full moon too. I'll wait until tomorrow, thank you.

Re:13th?

[cperciva]

Friday the 13th means nothing to Linux. The day you want to avoid upgrading to a new kernel is thanksgiving.

Re:13th?

[Anonymous Coward]

Oh come on, where's your sense of adventure?
Did you leave it in your lucky pants, and forget to wear them today?

Re:13th?

[Roofus]

What happened on thanksgiving?

Re:13th?

[NanoGator]

"Friday the 13th means nothing to Linux. The day you want to avoid upgrading to a new kernel is thanksgiving. "

I'm not familiar enough with Linux to understand this joke. Somebody please explain it?

(sorry to be a spoilsport!)

Re:13th?

[cperciva]

What happened on thanksgiving?

Thanksgiving 2002 saw the 2.4.20 data corruption on umount kernel released; thanksgiving 2001 saw the famous 2.4.15 "greased turkey" data corruption on umount kernel released.

There have been other kernels with problems, but it seems that the data corruption bugs tend to arrive with thanksgiving.

Re:13th?

[cperciva]

The past two thanksgivings have seen kernels released (2.4.15 and 2.4.20) with serious data-corruption-on-umount bugs.

Re:13th?

[NanoGator]

"The past two thanksgivings have seen kernels released (2.4.15 and 2.4.20) with serious data-corruption-on-umount bugs."

Ah!! Thank you. :)

Re:13th?

[Anonymous Coward]

He's referring to that you should be with family that day. The other day to avoid is halloween.

Make mine unixfree, please

[yukster]

Um, can I get mine without the SCO code, please? That Darrell McBride is a scary man:

The Linux business model was bound to change, and some people are having a hard time accepting this, McBride says. "The whole concept of getting something for nothing just doesn't hold up," he says. "The notion that you're going to run a Fortune 1,000 company on something that in the end could be more like Napster than an enterprise software system, it's a big question mark."
InformationWeek, today

Re:Make mine unixfree, please

[John Hasler]

> That Darrell McBride is a scary man:

No. He's merely another fool.

Re:Make mine unixfree, please

[Anonymous Coward]

"Scary" in that there are people out there this stupid.

Re:Make mine unixfree, please

[Anonymous Coward]

Maybe I'm too emotional, but personally I just don't understand those people - do they want to be the most hated after Bill Gates? It's ok if you want to sue somebody who has stolen something from you, ok, proove your justice, but blame everyone else?
It's feels more like a something is rotten here in this system, which we call 'capitalism', but what I call simply 'capitalistical anarchy'.

Just can't get it...

Re:Make mine unixfree, please

[tupshin]

What's that old saying??? Oh yeah. "Suffer not a fool to live."

-Tupshin

Re:Make mine unixfree, please

[Anonymous Coward]

It's just people going after money and power. It is what drives everyone, whether they are capitalists or communists. Capitalists are just generally a lot better at doing whatever it takes to get ahead.

Re:Make mine unixfree, please

[spitzak]

As probably 1000 or more lines were deleted for this version, it is almost certain that the 80 lines of SCO code were. Or at least you can probably tell your phb that.

Agree, noticed odd CPU load when burning

[SuperBanana]

So is there been something screwy with the IDE-CD subsystem in Linux lately?

I'll second that. I noticed after going to 2.4.20 that whenever I burn a CD(and this is on an 8x writer, not exactly 'fast'), the CPU(Athlon 1.5ghz) goes to 30%, all of it system time-and the system slows to a crawl, cursor jumping and everything-which is new- and I've got 32bit IO and unmasked IRQs set on all my IDE devices....

Re:Agree, noticed odd CPU load when burning

[Spoke]

Do you have DMA on as well?

Re:Agree, noticed odd CPU load when burning

[spencerogden]

It must be something specfic to your system. I usually burn at 42X with no problem on an IDE CD Writer. I am usually listening to MP3s at the same time and system is completely usable.

Re:Agree, noticed odd CPU load when burning

[bahamat]

I noticed after going to 2.4.20 that whenever I burn a CD(and this is on an 8x writer, not exactly 'fast'), the CPU(Athlon 1.5ghz) goes to 30%, all of it system time-and the system slows to a crawl, cursor jumping and everything-which is new- and I've got 32bit IO and unmasked IRQs set on all my IDE devices....

That's because converting mp3's to wav so you can burn them takes some overhead.

Re:Agree, noticed odd CPU load when burning

[Anonymous Coward]

Not 30% of a 1.5GHz Athlon. Maybe 30% of a 60 MHz Pentium.

huh?..

[Anonymous Coward]

what fsckin scripts?.. lol!.. really.. I still do it the make menuconfig, make bzImage/modules/blah, cp arch/yaddayadda, vi lilo.conf, lilo way... what's the name of the scipt and where is it at.. :?

Re:huh?..

[Anonymous Coward]

what's the name of the scipt and where is it at.. :?

Just do as I typed. make install. It'll do all the copying and shit like that for you.

-Xanadu

Re:huh?..

[KewlPC]

cd /path/to/kernel
make xconfig
make dep # also make clean if you've upgraded via a patch
make modules modules_install bzLilo

Re:huh?..

[KewlPC]

I should add that you don't even need to run lilo after this. bzLilo builds the kernel, installs everything where it needs to go, and then runs lilo.

You still might want to edit /etc/lilo.conf so that you can boot into the old kernel if the new one crashes.

Re:huh?..

[Sloppy]

You use make? What a pussy. I just gcc whatever looks like it's dated later than the .o. Then I boot my new kernel by running a.out.

Linux Tech. Support

[TrollBridge]

The parent post is a perfect example of what Linux tech support usually looks like.

Sure, running an RC on a production box isn't really bright, but instead of explaining that (and yes, it really shouldn't have to be explained), the typical response to any such Linux-related questions usually is something to the effect of "You're a fucking moron, RTFM".

Not a good way to establish the Linux community as a helpful technical resource, my friends.

Re:Linux Tech. Support

[ichimunki]

Sounded more like laziness than cluelessness/newbie-ism to me. If the person was smart enough to get a release candidate kernel onto his server, he's either smart enough to figure out if he wants/needs to update to the actual release or he's using a distro and should ask them for an updated package that is known to work with their other packages.

Re:Linux Tech. Support

[Anonymous Coward]

The grandparent post is more of an example of the condescending attitude that seems to be encouraged by this website.

If I should conclude the whole Linux community is exemplified by that person's post, then I should also conclude that most atheists, Christians, conservatives, libertarians, Americans, and Europeans are assholes too, based upon the posts frequently seen here.

Re:Linux Tech. Support

[cbreaker]

There were also some informative posts.

At least it's not "uhh.. reboot." or "re-install windows" or "update your drivers." Like you would get on a Windows forum.

There's asses everywhere, and in my experience the asses don't rule the Linux support community. Most people are pretty nice and good about giving out suggestions.

Mailing lists are usually a great source of good help. On Slashdot, you get a lot of trolls.

Re:Linux Tech. Support

[Anonymous Coward]

Oops, sorry, I forgot that there are some asshole liberals too.

Re:Linux Tech. Support

[Angry Pixie]

If you're on Linux and someone tells you that, then you should consider yourself lucky. You have decent manuals. On Windows we have that stupid Windows help that responds using the fucking Socratic Method! And all our product documentation are really just post-purchase marketing material.

Besides, it's not just a Linux community thing. Pretty much all techies do that. A lot of you remind me of that Saturday Night Live sketch about the company's computer repair guy. ;)

Re:Linux Tech. Support

[Anonymous Coward]

You're a fucking moron.

Re:Linux Tech. Support

[kasperd]

At least it's not "uhh.. reboot." or "re-install windows" or "update your drivers." Like you would get on a Windows forum.

When a Windows user would have rebooted, I log out and log in again. When a Windows user would have reinstalled Windows, I just reboot the computer. If I still have problems after rebooting, I use df to find out which partition is full.

Re:Linux Tech. Support

[cbreaker]

Good for you! But that wasn't the point of my post...

Re:why not posting the http links instead of the f

[tomhudson]

1. because of the extra overhead in http transfers (the server's overloaded already - why put extra load on it)
2. because of the ability to resume ftp transfers (w/o browser add-ins)
3. because the people who are actually going to comile it are at ease with ftp (prevents wannabes from uselessly sucking up bandwidth and/or hosing their systems doing a fucked-up compile/install :-)

I'm a big stupid idiot

[Saint Stephen]

But does anybody know how to get the fglrx ATI driver working on the 2.5.xx kernels? Or what are those other drivers?

Re:I'm a big stupid idiot

[FroMan]

Have you gotten the drivers working on anything other than the 2.4.18 kernel? If so, might sharing how?

I can only get the ATI drivers working with 2.4.18.

Re:I'm a big stupid idiot

[Saint Stephen]

Yeah, that was no big deal -- as a matter of fact, I've only ever used 2.4.20. Using Debian's kernel-source-2.4.20 and even the vanilla kernel. I followed these directions: http://space.virgilio.it/flavio.stanchina@tin.it/d ebian/fglrx-installer.html

Excelent

[zulux]

Linux has always felt sluggish to me compaired to FreeBSD as a workstation. There are hese little lags where the system diden't apear to be doing anything when you do taks that require disk IO.

Re:Excelent

[PipoDeClown]

Your writing on a Linux machine now?

Re:Excelent

[Anonymous Coward]

Your writing on a Linux machine now?

His writing on a Linux machine now what? Or did you mean "Your're"?

Re:Excelent

[termos]

His writing on a Linux machine now what? Or did you mean "Your're"?

Yes, he meant that and "Excelent" and "compaired" and "hese" and "apear" and "taks". ;-)

Re:Excelent

[DWIM]

His writing on a Linux machine now what? Or did you mean "Your're"?

Actually, I think he meant "You're".

Re:Excelent

[zulux]

Yes, he meant that and "Excelent" and "compaired" and "hese" and "apear" and "taks". ;-)

Thank you for coming back. I was wondering where my spellcheck went. ;)

Bit Torrent

[isorox]

Why dont they give out bittorrent links?

Re:Bit Torrent

[Anonymous Coward]

cos u should be using the local country mirrors

insert cuntrycode between ftp. .kernel.org or www. .kernel.org

and you are all set

i dont understand why slashdot always does the very same mistakes over and over.
posting ftp links where there is http ones, and posting the link to the main site, rather then a few more links but to the mirrors.

jesus, slashdot staff grow up at last.

Re:Bit Torrent

[jelle]

Saw a torrent link here

Re:Bit Torrent

[jelle]

And here...

Re:Bit Torrent

[Anonymous Coward]

Last I checked (5 minutes ago), it wasn't on the mirrors yet, but the torrent worked.

1kps, and you're complaining??

[TrollBridge]

I'm on a 9600 baud modem, you insensitive clod!

Now with DRM....

[Anonymous Coward]

Has anybody started work on a patch to strip this garbage yet? (No offense intended to the quality of Mr Cox's code).

how ironic...

[Imoen1337]

That a new kernel comes out on Friday the 13th.

Re:excellent, we can switch to this (SCO)

[Anonymous Coward]

From the Release notes:

> o [Bluetooth] Use very short disconnect timeout for SCO connections.
> o [Bluetooth] Kill incoming SCO connection when SCO socket is closed.
> o [Bluetooth] Support for SCO (voice) over HCI USB

Are these the lines SCO's bitching about?

But...

[blair1q]

...does it include any code from CP/M?

Re:But...

[Sloppy]

There's some new Z-80 code in there, but it isn't from CP/M. It was copied from something called .. MSX? Never heard of it, myself. Don't worry, I'm sure the company that originally wrote it, isn't litigous.

Misread?

[DamienMcKenna]

You must have misread what he said. 2.4.20-rc7 was a release candidate for the 2.4.20 release, not the 2.4.21 release.

Re:Misread?

[stefanlasiewski]

I think it was a typo anyways, since there is no 2.4.20-rc7.

some interesting updates...

[Xpilot]

...Besides the ptrace vulnarability fix and the ext3 bug fix, the new kernel also includes interesting netfilter updates, such as ARP filtering and MAC address filtering. Which is real useful in my institution where stupid XP boxes crapflood the whole LAN with ARPS.

What? No torrent link?

[Dr. Smeegee]

And to think I was all prepped to see how well bit torrent worked on something as slashdotted as this! Alas!

Re:why not posting the http links instead of the f

[QuMa]

1. HTTP has no more overhead than FTP, and in most cases less.
2. So get a better browser (opera comes to mind) or don't use your browser. Just because it's http doesn't mean you have to use your browser.
3. which they can do anyway

Obsolete

[masouds]

Didn't someone from Netherlands say that Linux is obsolete? So, why being happy about releasing something that is obsolete?
Just kidding!

oops! My bad....

[Penguin Follower]

... I almost snorted my pepsi straight out my nose when I saw 2.4.21 was released, because this morning I decided I was going to try Debian, and had chsen ftp://mirrors.kernel.org/debian-cd/i386 as my download location. I'm on the last CD now... that's #7... getting 240K/sec at the moment :0

Re:oops! My bad....

[Minna Kirai]

Why would anyone download multiple Debian CDs before installing? That just doesn't make sense. The most you should download is a 150 meg bootable CDROM image, then let the rest of the packages you want come over HTTP when you choose to install them.

Debian prides itself on an enormous amount of packages... nearly twice as many, counting bytes, as RedHat provides. To attempt to download "a copy of Debian" is wrong and wasteful. The fun of Debian comes in when you decide, on the spur of the moment, to try some exotic free software program and can apt-get it in a much less time that it would take to even figure out the name of the RPM you'd need to install on a "normal" Linux system.

Even if the desired install computer doesn't have fast internet access, burning 7 CDs is excessive. There probabably won't even be 2 CDs worth of packages you really want to install. Of the top ten largest packages in Debian, six of them are only desirable for hardcore software developers.

Re:oops! My bad....

[Anonymous Coward]

CD of everything is nice for multiple installs of machines on same (slow) subnet. Or for copying discs to give to friends on dial up accounts. Or because he/she likes shiny discs ;-)

Re:oops! My bad....

[joeljkp]

It's extremely useful if you have an unreliable or slow internet connection.

Re:oops! My bad....

[Penguin Follower]

Why would anyone download multiple Debian CDs before installing? That just doesn't make sense. The most you should download is a 150 meg bootable CDROM image, then let the rest of the packages you want come over HTTP when you choose to install them.

1. Because I've never used Debian before, I may fsck it up and wanna reinstall (I'm not a guru by any stretch of the imagination -- the Unix shell scripting course I just finished last quarter still gives me nightmares...)
2. I like hard copies. I have a shelf full of downloaded and paid-for CD's of Linux software... probably about 100 discs of Linux software altogether.
3. And last but not least, because I can!

Debian prides itself on an enormous amount of packages... nearly twice as many, counting bytes, as RedHat provides. To attempt to download "a copy of Debian" is wrong and wasteful.

Lots of packages = Choice. I like choice ;) Wrong and Wasteful? In the sense of bandwidth (for the server), yes. Although, I am spending $50/mo for cablemodem, so I got my money's worth this month! Can't say so for kernel.org :( However, if I really like Debian, I will probably buy a copy from now on. As I said, I like hard copies.

The fun of Debian comes in when you decide, on the spur of the moment, to try some exotic free software program and can apt-get it in a much less time that it would take to even figure out the name of the RPM you'd need to install on a "normal" Linux system.

Yes, I have heard of the great "apt-get" and will definitely exercise it a bit. I don't know why you have a problem with the naming of RPMs. I find that it is usually the same as the program or package name. Then again, I've been dealing with RPMs almost entirely since '97, when I started out on RH 5.2 Matter of fact, RPMs are a nice idea, just implemented somewhat wrong, IMHO.

Even if the desired install computer doesn't have fast internet access, burning 7 CDs is excessive. There probabably won't even be 2 CDs worth of packages you really want to install. Of the top ten largest packages in Debian, six of them are only desirable for hardcore software developers.

I'm used to swapping 3 discs around during an install... I started out with RedHat & Mandrake...

Anywho, it's pointless, as the downloads are already finished. I've already started burning the images to disc.

Re:oops! My bad....

[hal200]

I've not done this myself, since I've only got a couple Debian boxen and a zippy connection to the 'Net, but I believe it's possible to set one of your boxes up as an apt source and cache the packages locally on your subnet. So, you download them once, then have them available to all your machines.

Re:oops! My bad....

[Dachannien]

The fun of Debian comes in when you decide, on the spur of the moment, to try some exotic free software program and can apt-get it in a much less time that it would take to even figure out the name of the RPM you'd need to install on a "normal" Linux system.

As in, "Tank, I need a pilot program for a V-212 helicopter."

I always forget to install wget until I type it in and realize I haven't installed it. Five seconds and one apt-get later, I can just hit up twice and enter once. ;)

Re:oops! My bad....

[Minna Kirai]

Yes, I have heard of the great "apt-get" and will definitely exercise it a bit.

Even if you install packages from CD-Rom, you'll use the same interface as if you were getting it live from HTTP. The only difference is that it'll prompt you to insert the right disc first... and since most packages are small, it'll often take more time for you to find the disc than to just get it from the server.

You didn't mention if you had downloaded the stable or testing Debian... testing is generally prefered, because it's not as painfully obselete. If you value stability, "stable" is good of course. But if you want to have fun and experiment, then newer is better. And if you're using "testing", then you'll probably want to keep up with changes made after the CDs were burnt. Debian "testing" CD-Roms go obselete really fast.

I don't know why you have a problem with the naming of RPMs. I find that it is usually the same as the program or package name.

RPM names also contain at least the version string, and often an indication of which architecture the software will run on. Sometimes supported OS versions are mixed in too. For example, when I tried to install a package on a Red Hat system, I had to download that RPM. Then go to install it, and find out I needed multiple other RPMs first, which need even more RPMs to work.

The point of apt-get is you, the installing user, never even see the *.deb file that the package actually comes in. The hunt for dependencies is completely hidden from you.

Of course, RedHat users can optionally run apt-get themselves, but that's not formally supported by the distribution developer.

I won't go into the whole problem of not getting *.deb files for new, bleeding edge software. It's an accepted fact that Debian users who wish to try something brand-new will be compiling it themselves.

Re:oops! My bad....

[fearlessrogue]

Installing the base system from a netinstall is a bear. It is much easier to install the base sytem of debian disk number one.

Re:oops! My bad....

[boomka]

I agree in general, 7 CDs is just way too much, and shouldn't be downloaded.
However, having a CD or two with the .deb packages that are most likely to be used would be very nice - because connection isn't something you can always rely on, because it's less trouble to download stuff once than at 5 different points, and because you may always need these .debs later for isntalling on the other machine or just reintstalling the system (what if you break something or fail to install at a later stage and want a clean start over?)

And that is one of the big weaknesses of Debian - you don't have that option of downloading a CD with a bunch of most often used packages. You have to go with either full network install, or buy all 7 CDs. I imagine that's a big deterrent to many people out there who may want to try out this great distribution otherwise. And by the way, if you want to know how much software can fit on one CD, look at the list packages in Knoppix 3.2. It's really amazing - it has practically everything any normal user would ever need on a Linux box. With a CD filled with .deb packages like that, no user will find himself downloading more than a handful of files by http apt-get. One good step in that direction I saw recently is Bonzai Linux - it's a 180M CD image of Debian-based distribution that has KDE 3.1.2 and some other usefull stuff - all in 180m iso image! And it is great at hardware autoconfiguration. However, Bonzai isn't Debian per se, whereas I would like to see similar effort for the official Debian...

Re:oops! My bad....

[Rinikusu]

Uh, I'll give you a reason:
I have an old desktop. It doesn't have a NIC, a modem or anything, which stands to reason as it's a standalone workstation. Ever hear of one of those? No network connectivity. Anything I need to transfer from it goes on a floppy. The computer was free and it's honestly not WORTH the $14 to put a NIC in there, as that $14 would go a long ways towards picking up another machine (say, a $299 Walmart PC) later. The machine doesn't NEED network access (it's used for Gimp and Blender work), and hey, I don't have to worry about security patches because it's not ON A NETWORK, which makes network installations a bit of an impossibility.
What's wasteful is network installing debian on multiple machines when you could save bandwidth (and some of us are bandwidth limited or metered) and download a few .iso's and be done with it. Quit telling me what I *should* do and mind your own damn business.

Re:oops! My bad....

[Minna Kirai]

What's wasteful is network installing debian on multiple machines when you could save bandwidth (and some of us are bandwidth limited or metered) and download a few .iso's and be done with it.

If you want to save bandwidth, you can download a single ISO of Debian's desktop packages and stick with that. The highly popular Knoppix CD, for example.

Anyone willing to download seven ISOs of stuff he'll probably never even look at obviously has little concern for conservation of bandwidth.

Most of the benefits that make Debian unique and valuable come from its network installability. If you don't intend to take advantage of that feature, you've got little reason to use Debian at all. (Unless you were already very familiar with Debian, but then you'd know how to assemble a custom install CD without grabbing a ton of ISOs)

Re:oops! My bad....

[Tony Hoyle]

Yup, it's called apt-proxy.

It works quite well... I keep 5 machines fed from Unstable through it.

Re:oops! My bad....

[Anonymous Coward]

And that is one of the big weaknesses of Debian - you don't have that option of downloading a CD with a bunch of most often used packages. You have to go with either full network install, or buy all 7 CDs.

You can install Debian with the first ISO (preferably the non-US ISO). That's it. One issue with 'popular packages' is the target audience. The cds are filled with packages in order of importance. Other distros also fill cds in order of package importance. The difference in order reflects the difference in target audiences of distros. Debian order is oriented torwards server setups, then popular desktops, then programming environments, and then infrequently-used languages and programs. Most user friendly distros place the desktop first.

Also, it is possible to create your own apt repositories. If you are installing Debian on multiple computers, it it worth the time to create your own repository cd. I even had a CD with OpenOffice.org and KDE3.x repositories before they became official Debian packages.

Re:oops! My bad....

[Crazy Eight]

For the record, you could have used jigdo to spread those downloads over multiple mirrors.

Re:oops! My bad....

[BrokenHalo]

It's an accepted fact that Debian users who wish to try something brand-new will be compiling it themselves.

So why don't they just use Slackware?

Re:oops! My bad....

[SmittyTheBold]

Some of us don't have broadband, and a modem's not exactly desireable when downloading ~500 MB of software. What's even worse is when your connection is limited to four hours at a time.

By the time I got testing installed by that method, half of the packages would require updating! =P

Re:oops! My bad....

[Minna Kirai]

Yes, but the guy I responded to does have broadband, and he was downloading ~4550 MB of software.

If you want Debian, and have broadband, then a netinst is almost always best. If you don't have broadband, then ask (or pay) someone for a copy of the CDs. Downloading the CDs gives you the worst of both worlds.

(People with a packrat tendency, and many blank CDRs, are free to expand their collections if that's what makes them happy. Just don't expect to ever use half those discs.)

NO! Cause BitTorrent SUCKS

[Anonymous Coward]

The only thing I've seen, more stupid than an A+ "tech", is you weenies that think BitTorrent is the answer to everything. BitTorrent SUCKS ASS!!!!

Re:NO! Cause BitTorrent SUCKS

[Zork the Almighty]

How about BitTorrent a catalyst for world peace!

This is a known problem, if:

[WD]

If the following are true, then the high CPU usage is expected:
-You have an IDE CD burner
-You are using IDE-SCSI emulation
-You are burning a CD with a blocksize other than 2048 (such as redbook audio, or (S)VCD, etc..)

If the above are all true, the ide-scsi emulation reverts to PIO mode. Supposedly this will be fixed for 2.6

Re:This is a known problem, if:

[Anonymous Coward]

I had some difficulty parsing your post, but anyway: SCSI-burners can't go to PIO mode because PIO (programmed input output) is an IDE transfer mode like DMA.

I assume that ide-scsi was mentioned since this problem is caused by the ide-scsi module. Since there are no programs for burning without SCSI emulation that I know of, this does happen for every IDE burner.

Re:This is a known problem, if:

[staed]

actually, i think you can burn atapi with cdrecord now... dev=ATAPI:?:?:? or something...

Re:This is a known problem, if:

[Drathos]

the most recent version of cdrecord can burn without using ide-scsi emulation. now all we need is for the front ends to know that and support dev=ATAPI.

Re:This is a known problem, if:

[Anonymous Coward]

SCSI-burners can't go to PIO mode because PIO (programmed input output) is an IDE transfer mode like DMA.

Oh yeah? You obviously never used an Adaptec 1510 ISA card. ;)

Re: This is a known problem, if:

[Omniscient Ferret]

Er, hang on, what?

This is news to me. My Pentium-level machine had no problems burning VCDs or audio CDs on an IDE CD burner using SCSI emulation. This is without burn-proof to avoid buffer underruns, too.

Re:excellent, we can switch to this (SCO)

[nchip]

From the Release notes:

> o [Bluetooth] Use very short disconnect timeout for SCO connections.
> o [Bluetooth] Kill incoming SCO connection when SCO socket is closed.
> o [Bluetooth] Support for SCO (voice) over HCI USB

Are these the lines SCO's bitching about?

Nope. SCO in the bluetooth world means Synchronous Connection Oriented link.
Used mostly with bluetooth wireless handsfree devices.

Amen...

[YellowSubRoutine]

Same here, but not jut with cd drives.

apt-get upgrade (during configure scripts, they tend to do a lot of disk access) locks up my system for almost half a minute.

I think it started with 2.4.20, but it meight be earlier.
As soon as the acpi patch is ready, I'll try 2.4.21

Re:Amen...

[dsplat]

That sounds pretty familiar. I've been seeing symptoms like that under 2.4.18 with Debian. I thought I was tripping over a bug in the nVidia driver, but I'm not sure anymore. The video certainly went to "plaid mode". I can't reliably reproduce it now because I tripped over what appears to be an ext3 bug that is fixed in 2.4.20. The ext3 driver didn't handle corrupted inodes at all well. I got a kernel oops while running fsck after a hard reboot because the thing hung. If you want to see an unusable file system, try crashing part way through an fsck. /var is one of the file systems that got hit. I've been intending to reformat and reload that box when I finally diagnose the root cause of the problem. This sounds like it might be it, although I welcome other insights.

2.4.21

[JWSmythe]

Yippie! 2.4.21 is finally here! :) (I got it about an hour before the /. story showed up.)

They now have Opteron support in there.. I knew it was in the pre, I was just wondering if the new kernel or the rest of my hardware would show up first. I have everything for a dual Opteron system, except the processors and case. I'm so anxious, I'm going to burst. :)

( ) 386
( ) 486
( ) 586/K5/5x86/6x86/6x86MX
( ) Pentium-Classic
( ) Pentium-MMX
( ) Pentium-Pro/Celeron/Pentium-II
( ) Pentium-III/Celeron(Coppermine)
( ) Pentium-4
( ) K6/K6-II/K6-III
( ) Athlon/Duron/K7
(X) Opteron/Athlon64/Hammer/K8
( ) Elan
( ) Crusoe
( ) Winchip-C6
( ) Winchip-2
( ) Winchip-2A/Winchip-3
( ) CyrixIII/VIA-C3
( ) VIA-C3-2

I'm going to be a compiling fool when the rest of the parts show up. :) I'm going to try to get Slackware running on it. :)

Re:2.4.21

[njchick]

You are compiling for 32-bit Intel architecture. If you want to run a 64-bit kernel, you need a cross-compiler, and don't forget to set ARCH and CROSS_COMPILE on the command line for make.

Re:2.4.21

[Anonymous Coward]

Not that I want to smash on Linux, but NetBSD already has fully functional single and dual-CPU support for Opterons.

Why not give NetBSD a try too?

ftp://ftp.netbsd.org/pub/NetBSD/arch/amd64/iso/

Re:2.4.21

[JWSmythe]

I had read something to the effect that with GCC, if you compile on an x86-64, it'll make 64-bit binaries.. But, I will play, experement, and abuse it until it does what I want..

In the back of my mind, I keep hearing the echo of "you'll need a cross compiler", so we'll see how it goes once I get started. I have nothing against building a cross-compiler, if that's what it takes.. I already did it for my ARM machine (ipaq with Familiar), so I can't imagine it being much harder for the x86-64..

Re:2.4.21

[randombit]

I had read something to the effect that with GCC, if you compile on an x86-64, it'll make 64-bit binaries..

You can compile an x86 GCC to support x86-64 as well (by passing -32/-64 to pick which one you want to compile for). IIRC by default GCC for x86 won't support x86-64, there is an extra flag to enable it at configure time. So your cross compiler will be an easy job.

Which model are you going for? The 1.8 Ghz ones look _nice_, but at $850 a pop they're a bit pricey still...

Re:2.4.21

[JWSmythe]

That sounds pretty easy. :)

The machine we've ordered goes like this:

MSI K8D Master
Two Opteron 240
two 512Mb PC2100 (1Gb total)
Enlight EN-8950
3 180Gb Western Digital Special Edition IDE

The price wasn't bad for the whole thing. Getting most of the parts from wholesalers, it came to about $2300.

The motherboard, memory, and drives have already arrived. I'm just waiting on the CPU's and case before I can start playing.

The machine it's replacing doesn't use a lot of memory, so 1Gb will be fine to start. The plan is, we'll swap out CPU's later for faster ones when they come down in price, and add memory as necessary.

I didn't want to go really nuts with it on the first attempt. I needed something the boss wouldn't freak out about the price, but still be impressed with the performance.. If this one works out well, there's a database server that needs to be upgraded, and we'll blow lots of money on that. :)

The guy who handles purchasing (like, hunting down vendors, and placing the orders) found a place selling the 240's for just under $300/ea. The 242's are just over $700/ea from the same place. He didn't get prices on 244's yet. Since they didn't mention it to him, I'd have to assume they aren't carrying them yet.

Re:2.4.21

[randombit]

found a place selling the 240's for just under $300/ea. The 242's are just over $700/ea from the same place.

That's consistent with what's on pricewatch, too. The price jump between 240 and 242 is pretty amazing - I guess they're having fab problems still (which is hardly a surpise this early in the game).

I'm really wanting a nice x86-64 box, but with my limited budget (read: blew it all on booze), I'll probably wait until sometime this fall.

Have fun with new box! :)

Obligatory SCO comment...

[sahonen]

Hopefully, by 2.6 SCO will actually TELL US what code got copied, so we can change the comments and variable names so it doesn't match anymore...

Re:why not posting the http links instead of the f

[mrpull]

Site News

April 2, 2003: We have gotten reports that www.kernel.org is occationally slow in responding. We are trying to resolve the problem, however, in the meantime, the FTP server does not have this problem.

If you look at the bottom of kernel.org, you'll see the quote above. (I know it is cliche, but I _really_ was there yesterday downloading 2.4.20 and just finished compliling when I saw this post :). Between their existing problem and the Slashdoting, the FTP is probably a better option anyway.

Oh well.
mr.

Re:excellent, we can switch to this (SCO)

[gunpowder]

No this was done after SCO was bitching about Linux. Here an explanation what it actually does:

> o [Bluetooth] Use very short disconnect timeout for SCO connections.

The timeout is actually 0 seconds. Once a connection with a SCO machine is already established, kill it immeadiately! We don't like SCO!

> o [Bluetooth] Kill incoming SCO connection when SCO socket is closed.

The socket is closed by default. So this actually means, never accept new connections from SCO machines!

> o [Bluetooth] Support for SCO (voice) over HCI USB

This allows using your USB Modem to call and harass the SCO support center.

(Note for SCO employees: don't take this post too seriously)

Is it worth upgrading for old Red Hat Linux 7.x?

[antdude]

Is it worth getting this version for my old Red Hat Linux v7.1 and v7.2 boxes (currently running Kernel 2.4.20)? Will I see any performance improvement?

I mainly use them as a workstation (7.1 as a private server for personal usage) for gaming (e.g., RTCW, RTCW:ET, Q3A and its mods, surfing, IRC'ing, watching videos (DVDs, AVI, etc.). I haven't really noticed any IO problems.

I did noticed dragging video windows (e.g., MPlayer) and my SB Live! (emu10k1 driver from opensource.creative.com) stop playing audio until I let go of the mouse button. Does this Kernel version fix this problem?

You can see my system setups. Thank you in advance. :)

webserver slow

[Anonymous Coward]

Other reasons were posted already. Also important:

"April 2, 2003: We have gotten reports that www.kernel.org is occationally slow in responding. We are trying to resolve the problem, however, in the meantime, the FTP server does not have this problem."

Source: scroll down @ http://www.kernel.org

/cry

[FroMan]

Its nice to see kernel develoment moving forward. Too bad ATI doesn't update their drivers with each kernel release. There's a couple cool features floating around since 2.4.18, but I don't get to use them.

I wish video card manufactures would keep uptodate with their drivers. This is probably my biggest bummer of linux. Oh-well. I can always use the open source drivers, just no decent 3d acceleration.

Re:/cry

[Anonymous Coward]

Google does wonderful things

Now we wait

[Uteck]

for the offical Debian version in a few months.

Re:Now we wait

[Anonymous Coward]

Yeah sucks for them. I've been running the Alan Cox patched 2.4.21 kernel for quite some time now....

Intel SATA Support

[commander salamander]

Snipped from the changelog:

Summary of changes from v2.4.21-rc2 to v2.4.21-rc3
Alan Cox :
o add Intel ICH5 Serial ATA

w00h00! I can connect some nice Seagate SATA drives to my ASUS i875 board, and toss out these old parallel junkers!

--

Re:Intel SATA Support

[Alan Cox]

If you do then you should also get Jeff Garzik's new sata as scsi stack patch and use that.

BitTorrent Link

[redhat421]

linux-2.4.21.tar.gz.torrent

Yay!

[Anonymous Coward]

So in a year maybe Debian will include this kernel with it's standard install!!

Comment removed

[account_deleted]

Comment removed based on user account deletion

BitTorrent Download Link

[Brian Ristuccia]

http://osiris.978.org/~brianr/torrent/linux-2.4.21 .torrent

Re:BitTorrent Download Link

[jelle]

Wow! As fast as my cable modem lets data get through! And it needs only 0.5kb/s up... I guess a lot of people are seeding this (translation: keeping their download windows open). Thanks guys!

It wasn't even on the mirrors yet.

Re:Is it worth upgrading for old Red Hat Linux 7.x

[Anonymous Coward]

Sounds like you need a real system that's optimized for your hardware. Try Gentoo or LFS.

ccache for the compulsive kernel compiler

[Corporate Gadfly]

For the compulsive kernel compilers amongst us, ccache is a lot of help.

From their page:

ccache is a compiler cache. It acts as a caching pre-processor to C/C++ compilers, using the -E compiler switch and a hash to detect when a compilation can be satisfied from cache. This often results in a 5 to 10 times speedup in common compilations.

So, if your normal kernel build command is:

make dep clean && make bzImage modules modules_install

then, your ccache version of that command would be:

make dep clean && make CC="ccache gcc" bzImage modules modules_install

Trust me, it will save you a lot of time, especially if you are constantly tweaking your config settings and recompiling all the time.

If you compile as root, usually the cache directory will be /root/.cache
You can tell ccache to only reserve so much disk space for itself by issuing a

ccache -M 100M

command to reserve 100M.

Re:ccache for the compulsive kernel compiler

[caluml]

Something better than that is a distributed C compiler.

distcc.samba.org. It's another fine piece of coding from the Samba team. It uses your standard gcc, and does parallel builds on other machines. Really speeds things up.

Re:ccache for the compulsive kernel compiler

[Corporate Gadfly]

Good point.

You can always use distcc and ccache in combination. From distcc website:

You might like to also look at Andrew Tridgell's ccache compiler cache, which can avoid running the compiler altogether if the object file is cached from a previous identical run. This is great if you work in several similar source trees. Of course it works well with distcc.

This way if you DO have multiple machines at your disposal, use distcc (even distcc/ccache combo). Otherwise, if you only have one machine, then use ccache.

Free time?

[MikeFM]

You had free time in school? I'm assuming you mean college and not like highschool. Jeez I worked a full time job, a part time job, and went to school full time. Any extra time I had was spent sleeping. Compared to school just being a working stiff is easy. NOW I have the time to play around with compiling things pointlessly. Especially since the dot bomb.. lots and lots of free time. School really pays off when the job market is dry by the time you get a degree. :)

I know, I know...

[TheQuantumShift]

This has probably been adressed but, is there a non-programmer version of the changelog? I mean I like to know whats going on, but I only keep track of releases like this one. I'd like to see maybe at the top of the changelog, a simple 2.4.20 to 2.4.21 rundown. I guess I just don't have the patience for starting at the bottom and working my way up...

Omitted From Changelog

[Anonymous Coward]

- Removed offending SCO source code.
- Replaced offending SCO source code with ASCII art of a middle finger.

Re:Omitted From Changelog

[Wild Bill Hickock]

Since we are at it, we should include the goatse guy too :) excuse me while i puke...

Re:Is it worth upgrading for old Red Hat Linux 7.x

[bazmonkey]

Is it worth getting this version for... Will I see any performance improvement?

That's more of a personal question than anything else. Seeing as how these aren't mission-critical systems, there's nothing *wrong* with trying a new kernel. I think the question you're needing to ask is "Is a recompile of my kernel worth the effort of a new kernel?" Assuming you're slick enough with Linux to do it, it's no big deal to recompile (on a high-end Athlon no less), edit a couple files (or bust out some nifty RedHat tool), and reboot.

If you have a backup computer as well (which I see you do), I would personally let the newer computer follow the 2.5.x branch of the kernel. You'll be able to try out all kinds of new features, see serious performance increase, and help find bugs (which you will... but that's why you have a backup). Besides, the 2.5.x kernels of late are *useable*, just not stable.

So, basically, yes, upgrade away. Your question would be a lot harder if you were my boss, managing systems that, if down, cause some 50 people or so to get very disgruntled. IOW, it would be a different situation because in your case no one really cares if your computer goes down. We need a damn good reason to bring one of our boxen down... but you can play with yours all you want.

Re:Shouldn'thave to turn off DMA

[McQuaid]

I installed a live based distro to my harddrive and DMA wasn't working by default, recompiled the kernel and it's much smoother now. Before I was experiencing slow downs and lockups. I tested with hdparm and before I was getting 3.04MB, now I'm getting 40MB.

Also, I didn't have to force dma with hdparm, once I compiled for my chipset DMA was detected. I only used hdparm for testing, as I've read that invoking dma and 32bit with hdparm can add overhead as well.

I've never done this before

[Anonymous Coward]

Eye of the Tiger, baby.

Braindump

[Anonymous Coward]

I'm trying to learn more about the operation of the Linux kernal. What is this ring 0 people keep talking about? If there a poster I can print from the Internet that details the architecture of the linux kernal?

Thanks

Re:Is it worth upgrading for old Red Hat Linux 7.x

[antdude]

What do you think about the video window issue with audio? Do you think this is fixed with 2.4.21? I am not sure if that is related to IO stuff.

The mirrors have awoken

[serenemy]

ftp.{countrycode}.kernel.org are starting to get copies now.

The Onion

[Anonymous Coward]

Why does your comment remind me of the Taurus horoscope from The Onion's Vol 39 number 21 issue:

"You'll admit that the errors MIT found in your coordinate system for a nine-dimensional plenum check out, but it still seems like they did it just to be dicks."

Some little-known 2.4.21 change log entries:

[lacrymology.com]

: o Removed SCO/Unix code o Obfuscated SCO/Unix comment blocks

Overview of 2.4.21 vs 2.4.20 please

[McQuaid]

I've been playing more with linux, compiled 2.4.20 to my system, but overlooking the changelog, I can't assess if there is anything I might find interesting.

Looking over the change I don't have a clue what these address.

Ex.
Backout erroneous kiobuf dcache flush changes Cset exclude:
Uh, yeah.

Some of them you can kind of get an understanding of what it's probably for, but a lot of them, forget it.

It would be nice to know the big points of 21 over 20. Like this IO overhead thing, if it only applies to people with problems is one thing, but say 21 had a 10% gain in performance in that area across the board, more people get interested.

Here's a patch mirror

[simetra]

Please slashdot my ISP. They're wankers.
Get the patch here: www.squirm.net

Wow, lotsa changes!

[mj01nir]

Just finished reviewing the ChangeLog. Did anything not get touched this time? Good lord, there are arcnet patches in here.

Also, when does Alan Cox sleep? Prehaps he's a new form of undead with an affinity for coding? Maybe he has cloned himself a few times? Alan, however you're doing it, we appreciate all of your hard work. Now go take a nice, long nap.

Re:Wow, lotsa changes!

[rtscts]

I think Slashdot reported that Alan had been merged directly into the kernel a few years back. He no longer has any need for his physical body.

I believe there was even an X-File about it.. some shack in the middle of nowhere with a T1.

Re:Wow, lotsa changes!

[Space cowboy]

there was even an X-File about it.. some shack in the middle of nowhere with a T1.

Hmm. Maybe Alan's gone over to the dark side.... Sounds dangerously like an i-loo, or whatever that crap is called that Microsoft were pushing. Yeah, pun intended :-)

Simon

Re:Wow, lotsa changes!

[/dev/trash]

I didn't think Alan Cox was in charge of 2.4 stuff?

Re:Wow, lotsa changes!

[Anonymous Coward]

Thank Marcelo Tosatti instead. He's actually been getting alot of flack for being slow on the release cycle.

I think he's doing a good job, and, right now he's the captain of 2.4 - that means, he decides. I don't see any of the complainers on LKML jumping to take on this big responsability and loads of work.

Re:Is it worth upgrading for old Red Hat Linux 7.x

[Minna Kirai]

I did noticed dragging video windows (e.g., MPlayer) and my SB Live! (emu10k1 driver from opensource.creative.com) stop playing audio until I let go of the mouse button.

That seems like something wrong with MPlayer itself, not the sound driver or kernel. Assuming the audio is coming from the videofile, MPlayer might stop updating the output stream until the window drag is complete.

(Or, do you mean sounds from other applications stop as well when MPlayer is dragged?)

Re:Is it worth upgrading for old Red Hat Linux 7.x

[bazmonkey]

Well, I didn't look too hard into it (sorry, I'm at work and it's not exactly priority), but the Changelog doesn't mention emu10k1 or anything about soundblasters. Do you know if you can make the problem happen on other video players, xine, aviplay, xanim, etc.?

Also, you might want to give the ALSA drivers a chance. The new kernels are pushing it as the new sound architecture for Linux, and sometimes they make a big difference. Besides, having everything set up for ALSA and running properly will make it easier to move on to 2.6 kernels.

Like I said in my first post, compiling a kernel for a computer of your computer's stature is a matter of copying the old config, checking to see if there's anything else you need to change, and then waiting for a couple minutes. Work a little Grub/Lilo magic, and shzaam!, new kernel. You could be telling me if the new kernel helps within a half hour if you started soon.

If nothing else, IDE I/O is always a good thing to work on. My lowly laptop is definitely getting a 2.4.21, if not a 2.5. It takes me a tad longer; I use the cryptoapi modules and that always seems to take extra time.

COX is a MAN!!!!!

[Anonymous Coward]

Thank God that Alan is part of this world.

What a fucking MAN!!!!! His change entries triples everyone else.... Not that anyone else name in the ChangeLog is a slouch by ANY means, but Alan just KICKS ASS!!!!

Rock on and live long Alan..

Thank You!!!!!!

Re:Is it worth upgrading for old Red Hat Linux 7.x

[antdude]

It is not just MPlayer. It also includes Xine, Ogle, etc.

XMMS is OK if I drag it around. I think it is only with video files and DVDs.

It's no big deal, but it can be annoying. Other than that, my boxes are happy.

Re:Shouldn'thave to turn off DMA

[orcrist]

I'm talking specifically about cd/dvd drives, not hard drives. I got the DMA tip after a google search which turned up the relevant comments from Alan Cox; I'm too lazy to do the same search again, but he basically said some cd/dvd drives caused these lockups in Linux 2.4.x kernels and turning off DMA might help. In my case the system was locking up occasionally when firing up the SuSE package installation module, which does a lot of heavy I/O with the dvd when reading the package info. After turning off dma I never experienced the problem again. I still have dma enabled for my harddrive though.

-chris

What DRM??

[ShortSpecialBus]

I know I shouldn't repond to this, but what DRM are you referring to?

Nobody ever said linux would have DRM in it...they simply said they wouldn't stop others from putting DRM stuff in the linux OS because it DOES have some benefits from time to time.

Don't be an idiot.

XFS patches

[caluml]

Finally, I can upgrade, and remove "/some/shit/file" from /proc/sys/kernel/modprobe to prevent the ptrace local root from working.
As soon as there are XFS patches for it....

Re:XFS patches

[boomka]

I remember a story on slashdot a while ago about XFS support
in the development version of the kernel. I thought it would make it into 2.4.21.
Does anyone know why this isn't in yet?

What sort of a slashdotting is this?

[meringuoid]

Downloading from kernel.org at 57.12 K/s... which is just about as fast as my ADSL link can gulp it down. People, you're not doing your duty! Melt that server!... as soon as my download finishes, that is.

as a side note....for LP Patch users

[Lumpy]

The Linux Progress Patch for 2.4.20 still works with this IF you do the following....

Use 2.4.20 source
patch for LPP
patch to 2.4.21

Voila... that nice Purdy linux boot screen is still there for your relatives that ball up on the floor crying when the boot messages start flying...

Vendor Kernels

[Bilbo]

> Odds are that most people are running vendor kernels ...

Of course, anyone running a stock, vendor kernel is unlikely to be interested in downloading and building a brand new 2.4.21 kernel.

Having said that, I agree that it's highly unlikely that the patch will cleanly install against a vendor kernel, since they usually have lots of little mods in them.

Re:Vendor Kernels

[kasperd]

Having said that, I agree that it's highly unlikely that the patch will cleanly install against a vendor kernel, since they usually have lots of little mods in them.

If you install the .src.rpm for the RedHat kernel, you will actually get the original .tar.bz2 file and all the individual patches. So it would be no problem to patch against that.

Uh oh

[kosamae]

Guys, I think we just DOSed kernel.org, I can't connect. Good job.
Maybe we should do this to microsoft.com next time they release a new patch, which'll probably be tommorrow.

Re:why not posting the http links instead of the f

[gearheadsmp]

Step 1: Print source code to Mac OS X's Mach 3.0 microkernel
Step 2: Position crate of code behind front door
Step 3: Watch from street as a bunch of people dressed in black try to ram open a door barricaded with >1,000lbs of paper in broad daylight.

Re:why not posting the http links instead of the f

[Anonymous Coward]

What horseshit.

1. HTTP has more overhead than FTP???

2. Modern browsers can resume, as well as many download tools.

3. Yeah, people who are versed with GET/PUT/... are obviously Linux experts.

Re:Is it worth upgrading for old Red Hat Linux 7.x

[Minna Kirai]

But, if you drag MPlayer around, does the audio stream from XMMS get interrupted too? Or does it continue?

Re:Is it worth upgrading for old Red Hat Linux 7.x

[BobaFett]

Depends on what you've got and how you use it. Slowdown during intensive IDE output was a major issue for me, so I'll be trying 2.4.21 out tonight. Also, usb-storage over USB2 (ehci) is not very stable in 2.4.20 (although fixed in RedHat's version of it), and 2.4.21 changelog says that ehci is updated to what's not in 2.5 kernels, so I want to try that too. Finally, I have Hipoint onboard IDE controller which was not supported in 2.4.20 but again changelog mentions it (and RedHat's 2.4.20 supported it).
So, think of what problems and issues you have now, read the changelog to see if they were addressed, if a subsystem which did not work well for you is touched build the kernel and see if your issues are solved.

Re:Is it worth upgrading for old Red Hat Linux 7.x

[antdude]

I am not sure. I never tried it. :) Let's say, it does. What do you think the problem is?

How about a nice piece of STFU! :)

[Anonymous Coward]

- Who wants a Porsche if you can own an entire SUN?
- My Athlon is faster than a Harley

expensive bandwidth?

[7-Vodka]

Tidbit: 'Current bandwidth utilization 131.72 Mbit/s

Hey is someone actually *paying* for all this bandwidth?
Wouldn't it be nice if kernel.org ran a bittorrent for the latest kernels just so they save some bandwidth?

Re:Shouldn'thave to turn off DMA

[reboot_imminent]

I have to second this... I was experiencing all sorts of lockups with my CDRW under 2.4.x until I turned off DMA. Can't wait for 2.6 :)

Re:why not posting the http links instead of the f

[Anonymous Coward]

hmm stupid noob question i guess... but what are some ways to download through http sans a browser?

Re:why not posting the http links instead of the f

[Anonymous Coward]

• hmm stupid noob question i guess... but what are some ways to download through http sans a browser?

These two come to mind: wget and curl

Attention Linux Developers!

[Anonymous Coward]

I am offering upwards of $5 for an enterprising young Linux entrepreneur to fuck my fat face until his tiny, pathetic wanglet puts the "creme" in my "creme de mouth".

The March of Progress

[colmmacc]

net/network.o: In function `irda_device_init':
net/network.o(.text.init+0x2c 70): undefined reference to `toshoboe_init'
make[1]: *** [vmlinux] Error 1
make[1]: Leaving directory `/usr/src/kernel-source-2.4.21'
make: *** [stamp-build] Error 2

Oh well, it's going to be a fun weekend.

Final Nail in Bill's Coffin (for me, anyway)

[jhylkema]

This kernel release includes support for the Palm Tungsten T. Finally, I can wipe my Windows partition. WOOHOO!

On another note, kernel.org does seem to be slashdotted.

usb mass storage corruption

[Anonymous Coward]

Anyone know if this release fixes the usb mass storage devices? My camera's compact flash was hosed not long ago. I saw some fixes in the changelog but I'm not sure if the updates have been used or tested by someone?

OT: but here we go...

[Penguin Follower]

Even if you install packages from CD-Rom, you'll use the same interface as if you were getting it live from HTTP. The only difference is that it'll prompt you to insert the right disc first... and since most packages are small, it'll often take more time for you to find the disc than to just get it from the server.

You're making me think I'm not going to like the installer... related packages should be on the same CD. Then again, I go by categories instead of individual package selection... takes too long, and hard drives are cheap... use that space!

You didn't mention if you had downloaded the stable or testing Debian... testing is generally prefered, because it's not as painfully obselete. If you value stability, "stable" is good of course. But if you want to have fun and experiment, then newer is better. And if you're using "testing", then you'll probably want to keep up with changes made after the CDs were burnt. Debian "testing" CD-Roms go obselete really fast.

I downloaded the 3.0r1 Stable. It's not going on a very fast machine, anyway. Part of the reason why I am even investigating other distros is that Red Hat is making some choices in their package selection (or deselection as it may be) that I don't agree with. I just happened to have a need for dosemu recently, only to find out it hasn't been shipped in RedHat for at least the last couple releases. Upon downloading the source I find there are other packages not installed that was needed to compile dosemu (an assembler -- bin86 i think it was). I even fetched that package... more problems. I would imagine that Debian still includes dosemu, although I may be wrong on that. Either way, the machine will be a server and so stable appeals to me...

Of course, RedHat users can optionally run apt-get themselves, but that's not formally supported by the distribution developer.

We can also run "up2date packagename" and it will grab it... as long as it was something shipped with the distro, that is. And up2date does dependencies for you, which is nice. Unfortunately, up2date doesn't grab programs that weren't in the distrobution. :(

Re:OT: but here we go...

[Zarquon]

http://packages.debian.org/stable/otherosfs/dosemu .html

Stable is at 1.0.2.1-7.
Unstable and testing are at 1.0.2.1-13.

Re:OT: but here we go...

[Minna Kirai]

You're making me think I'm not going to like the installer

That's a common opinion, although it somewhat depends on what you mean by "installer". The installer that initially prepares your hard drive is quite servicable (in terms of partitioning, bootloaders, etc), although not as pretty as a recent RedHat or Mandrake offering.

However, once that setup is over, you're left with some bad options to add further packages. There is dselect, which is full-featured and informative but painfully convoluted to use. And there is apt-get, whose minimal CLI is adequate for those who can handle it, but unhelpful for newbies. An assortment of "Graphical APT frontends" have come and gone over the years, but they never manage to survive very long ("aptitude", etc).

Unfortunately, up2date doesn't grab programs that weren't in the distrobution. :(

In general, Debian's apt system doesn't support that either. But their larger number of packages in "the distribution" makes that less problematic. Since most users never attempt to burn a full set of discs, they have little pressure to save space by dropping packages.

If a 3rd party developer supports it, she can provide an apt.config line to add HTTP access to auxilliary deb files via your normal apt-get routine.

Re:OT: but here we go...

[Styx]

We can also run "up2date packagename" and it will grab it... as long as it was something shipped with the distro, that is. And up2date does dependencies for you, which is nice. Unfortunately, up2date doesn't grab programs that weren't in the distrobution. :(

You can easily set up your own up2date server, using current. Apart from letting you add your own packages, it will also save you a bunch of bandwidth, if you run multiple RedHat systems.

Three cheers for Marcelo!

[nightfire-unique]

Thanks for all your hard work Marcelo! I think I speak for the community when I say we all you a great debt of gratitude. And of course, thanks Linus for choosing him! :) And last, but certainly not least, kudos to all kernel contributors. You guys make it happen. Ok, now back to my compile. ;)

Re:BitTorrent - good for anything over 30 seoonds

[Splork]

bittorrent is useful for any download that would take you over 30 seconds at your max thruput rate.

i just used the .torrent link posted & modded up below to get the kernel at 160k/sec no problem.

obviously not many others using it though as my uplink hasn't gone over 25% of full capacity.

http://66.227.104.34/linux-2.4.21.tar.bz2.torren t

Arrgh! Why didn't I wait?

[Nucleon500]

So last night I decided to try Gentoo for the heck of it. So here I sit, reading Slashdot with Lynx while my system builds on the other console, and I read about the new kernel! I should have waited another day or so. I suppose it is Friday 13.

(Yes, I know how easy it will be to upgrade.)

Re:why not posting the http links instead of the f

[tomhudson]

1. the overhead for http as opposed to ftp varies between 10% and 20%
2. many browsers cannot resume w/o browser add-ins
3. I said that people who are not comfortable w. ftp are less likely to actually compile the kernel.

Great work!

[plutoid]

Everyones' efforts are much appreciated. Thanks, Jake Johnson http://www.plutoid.com

Re:Great work!

[plutoid]

plutoid

Re:why not posting the http links instead of the f

[tomhudson]

1. an ftp command-line client has much less resource requirements than a browser, can be started and stopped by cron, etc. So the actual user overhead is much smaller, and the system overhead is also smaller. Or you could do a perl script to get it w. http, but why bother.
2. If you're not going to use a browser, why not just ftp or bittorrent it?
3. you'd be surprised how many people don't have a clue about how to ftp a file, never mind setting up a cron job to do it during off-peak hours.

Re:why not posting the http links instead of the f

[RustyTaco]

the overhead for http as opposed to ftp varies between 10% and 20%

10% is VERY high estimation for "GET /pub/linux/kernel/v2.4/linux-2.4.21.tar.bz2 HTTP/1.1" then 30M of binary goo

With FTP that would be something like:
USER anonymous
PASS noneof@yourbusiness.net
CWD /pub/linux/kernel/v2.4/
BIN
PASV
RETR linux-2.4.20.tar.bz2
...binary good over a different socket
Where is this imaginary overhead coming from again?

- RustyTaco

From the changelog, it looks like it roxorz

[CaptainPhong]

Man, look at all the USB fixes in there! There must be 10 lockup bugfixes in USB-storage alone. I wonder what "fix usb-storage horkage" means...

Oh, and for those of you wanting a preview of some of the stuff in 2.6 but are afraid to try 2.5, I highly recommend the jiffies, preempt and low latency patches - the performance increase is tasty (don't forget to set your jiffies to something like 500 or 1000).

Re:why not posting the http links instead of the f

[tomhudson]

Over a large file, the difference isn't much, if any. Unfortunately, have you evere d/l'ed an iso w. your browser only to find that its' md5 is crap? So, the overhead of ftp'ing the file after a failed http transfer makes it 100% more in some cases :-(

I admit to my prejudice in favour of ftp (or bittorrent) over http in such cases, as well as cases where the browser gets the type wrong and either doesn't do a binary transfer (shouldn't matter under linux, but it does) or tries to open it as a document instead of downloading it.

Sure, a 20 meg file might not seem like much nowadays, but when you get /.ed, evrry bit hurts (oh - bad pun, sorry)

Anyway, it's the end of the week, and I'm going home. Have a nice w/e.

NNTP

[Anonymous Coward]

why don't things like this get posted to USENET, so people can grab them off local news servers? Even better than bittorrent because you're not going to tick off your ISP...

aureal

[Anonymous Coward]

Does this mean I'll finally be able to use sound w/ Linux and Aureal.

matroxfb

[oohp]

Wtf is wrong with matroxfb? It stopped working in 2.4.21!

Re:matroxfb

[Replicant7]

Good question.. There are some patches for matroxfb around, but they no longer apply cleanly...

Re:matroxfb

[Replicant7]

I take it all back. The patches seem to have been merged, I just forgot to compile in matroxfb statically. Good work!

Re:why not posting the http links instead of the f

[Angry Pixie]

So am I an evil bitch for using 'wget -c' from my Windows console no less ;) ?

Kernel update

[Anonymous Coward]

Wow, looks like the linux kernel is updated more often than the Italians rebuild their government. Wait, that's a good thing right? the updated kernel, not the change of government.

Good things in 2.4.21 = AMD64 + NFS for 2.4.22-pre

[Anonymous Coward]

We've been waiting on this kernel for a long time. I guess good things DO come to those who wait.

1. AMD64 support!

Yup. SuSE is far, far ahead of RH in AMD64 space. We've been running no problems since the rc-6 release. ...and for 2.4.22-preX series, we'll have...

2. NFS (fixes to tcp/udp nfs)

(http://www.fys.uio.no/%7Etrondmy/src/2.4.21-rc6 )
+ linux-2.4.21-14-xprt_fixes.dif
+ linux-2.4.21-15-fix_tcprace.dif
+ linux-2.4.21-16-fix_tcprace2.dif
+ linux-2.4.21-17-fix_tcprace3.dif
+ linux-2.4.21-18-fix_lockdmon.dif

Under heavy tcp-nfs load, NetApp filers will set the window size to 0. The linux client will continue to poll the filer (as expected). After about 5 minutes the filer RSTs the connection.

Older kernels won't pick up on this, and hang. We owe Trond a lot of thanks for NFS improvements that can be patched cleanly against 2.4.21 (see thread at nfs.sourceforge.net). Most of your questions as to the environment, test load, versioning, and whatnot can have been answered already.

More testing of these will probably get them fed into the main tree for 2.4.22-preX series, improving everybody's life.

Re:why not posting the http links instead of the f

[tomhudson]

So am I an evil bitch for using 'wget -c' from my Windows console no less ;) ?</quote>

Well, Microsoft might think you are, after all, they want you to use their gui for everything :-) Also, wget is one of those /evil pacman-like open-source license/threat to microsoft/viral license/ programs that are "unpatriotic" or a threat to the economy for some reason.

CHEER! The Highpoint RAID Drivers Made it in!

[Proudrooster]

Alan, Thanks for the including hpt37x driver! I've been waiting for this one for awhile! No more boot time loadable drivers on my ABIT motherboards!

Alan Cox alan@lxorguk.ukuu.org.uk:
o hpt37x

kernel.org is inappropriate

[rossz]

When I tried to read the changelog, I got this:

You have attempted to access an Internet site that has been judged inappropriate. This is a violation of the companyâ(TM)s Internet policy. This attempt has been logged. Repeated attempts will be reported to your manager and may result in disciplinary action.

Not doubt because the changelog is terrorist related.

Re:why not posting the http links instead of the f

[makapuf]

1. use curl or wget
2. why not use a browser ?
3. tell them to grab it via kazaa

any chance its 80 lines of code shorter?

[The Lynxpro]

You know, so each of us aren't sued by the SCO litigation machine (of no substance)?

Re:any chance its 80 lines of code shorter?

[Anonymous Coward]

Someone should prefix all the comments in the code with "FUCK SCO: ".

Re:why not posting the http links instead of the f

[tomhudson]

1. on using curl or wget, agreed - did that to grab the whole history a few months ago, though for one file, it seems a bit overblown
2. on using a browser - kernel.org is slow right now. Can't tell my browser to go there at 2 am tomorrow unless I stay at the office
3. grab it via kazaa? sure, but why not a .torrent file, to show that .torrents are for more than movies? After all, the MPAA has been harping on Bittorrent not having any legit uses.

Re:why not posting the http links instead of the f

[Anonymous Coward]

Then why deoes MS ship over 1000 command line tools for Windows?

Oh right, you're full of crap, I forgot.

see that changelog?

[brauwerman]

see that changelog?

big ups to hugh dickins!

Re:why not posting the http links instead of the f

[rutledjw]

I dunno, maybe. I've never had an HTTP download run near as fast as pure FTP. I thought it had something to do with the protocol.

It's too the point I'll actively avoid HTTP d/l if at all possible. Maybe if you're talking over dial-up there's no difference, but over DSL (and any greater high-speed for that matter) that's what I've seen.

Re:Is it worth upgrading for old Red Hat Linux 7.x

[mj01nir]

Are you running 2.4.20 with the preemptive patch? If not, you may want to try it. On my previous system, Ogg file playback would be choppy and interactive events like this would be choppy as well. The preemptive patch fixed 'em all. I've not had any trouble with it and it's mature (it's actually part of the 2.5 kernel now).

It doesn't look like there is a preemptive patch for 2.4.21 yet, but check back in a couple of days.

Linux kernel?

[CoolVibe]

A true FreeBSD Jedi warrior craves not such things.

(yes, a joke, get over it already)

Got it to compile with gcc 3.3 debian (sid)

[peope]

Just a note. Successfully compiled the kernel 2.4.21 with gcc 3.3 debian (sid). Works nicely. 2.4.20 barfed though.

Re:Got it to compile with gcc 3.3 debian (sid)

[getoblstr]

worked fine for me on debian (sid) as well. Even patched in pre-empt!

OT: 2.5.x and nvidia drivers

[havardi]

Slightly OT; I was using 2.4 kernel mostly because of nvidias drivers not compiling for 2.5.x. Then I found this site that provides up-to-date nvidia driver patches for 2.5.x including 2.5.70. (I had found patches before but they were always not-so-fresh)

One can download the nvidia driver from ftp://download.nvidia.com (the website only seems to link an "installer" version, which was irritating)

Now I'm happily runny mozilla-xft without those buggy artifacts the nv driver has-- and GL is always nice...

Caps

[theTerribleRobbo]

Some of us have download caps (1-gig) at home, insensitive clod.

I usually use the Uni connection to download anything above 40MB, hence I'd need all the CDs if I was going to install a CD-worth of programs, regardless whether each program is tincy-wincy.

Konspire p2p network is broadcasting it

[Anonymous Coward]

It seems it is being broadcasted using Konspire (http://konspire.sf.net). I got mine there, hell fast.

Re:why not posting the http links instead of the f

[JDWTopGuy]

1. just get over it already and download the stupid patch version
2. ????????
3. PROFIT!!!!!

Dang, I need a beowulf cluster to compile this new kernel! I oughta get a soviet russian to do it for me. All your base are belong to slashdotted kernel.org!

Ahhh, feel the karma burnin'...

High Latency

[KidSock]

Has anyone using RHL 7.3 noticed that occasionally X will freeze for 5-10 seconds while the disk grinds? This happends to me about once a day. It's very annoying. Also, when typing into an xterm I frequently type more quickly than the system is willing to accept. This is very annoying and more importantly it didn't use to happen prior to RHL 7.2. I recently upgraded the kernel to 2.4.20-13 and X 4.3 and there has been no change. How can this not be a high prority problem for people? It's very annoying!

Re:High Latency

[deaddeng]

using RH7.3, and don't have that particular problem, but the behavior reminds of an earlier install where I had manually partitioned and set up physically separate /boot, /home, /usr, /etc. Turned out my /home was almost full (all those kernel downloads). Do a df -h and see what your space looks like.

Re:why not posting the http links instead of the f

[Angry Pixie]

I love wget. I heard about a browser that uses wget as its system for saving files. I wish Internet Explorer could be retro-fitted to use it also. That way, large incomplete downloads wouldn't simply disappear after I cancel or get disconnected.

No, I don't want to use Netscape, so shut up about it already ;)

Re:why not posting the http links instead of the f

[RustyTaco]

Unfortunately, have you evere d/l'ed an iso w. your browser only to find that its' md5 is crap?

No. Also, the only way besides not downloading all of it the only way you should be able to ruin the md5sum over either FTP or HTTP(Both just shoving the raw data over a TCP socket) is by not downloading all of it.

If browsers are causing you so much trouble maybe you should consider wget

- RustyTaco

Re:why not posting the http links instead of the f

[Combuchan]

Um,

As I understand it, http will generally be faster than ftp. Unlike FTP, HTTP is a stateless protocol with no control connection. HTTP doesn't usually have username/password combinations. Nor does it have ugly passive and active transfer modes.

Always use HTTP for anonymous downloading from servers.

--sean

Re:why not posting the http links instead of the f

[QuMa]

Purely a misconfiguration or software issue then (and don't discont the placebo effect based on one or a few faster ftp downloads), once the initial protocol chatter is done, ftp and http are exactly equal.

With Suse yoy can!!!

[alonso]

On my suse 8.2 you can click the green icon near the digital clock. Than you'll have the system updated....
What's your number?

New Kernel, No Reboot. Nya nya ne nya nya

[rimu guy]

Downloaded the new kernel, got the latest UML patch. make oldconfig ARCH=um etc, and am now trying it out on a couple of test Virtual Private Servers. No (physical server) reboot required. Oh yes. The joy of userspace kernels.

BTW: So what was the patch that improves the 'feel' of the system when under heavy IO load?

And what is happening with the O(1) scheduler. Is that going to ever be in 2.4 (or just 2.5/6). Anyone running the O(1) patch against a 2.4 kernel? (How does it go?)

I don't think so

[Bilbo]

> you will actually get the original .tar.bz2 file file and all the individual patches.

I don't think so. Red Hat at least puts in a lot of small modifications to the actual source code, sometimes back-patching fixes for security updates, or making changes to integrate better with certain applications. (They've got changes to the NFS client support that I find particularly irksome.) They usually do a pretty good job of keeping up with new kernels when they come up, rolling their changes forward, so they aren't really "forking" the code, but they are definitely NOT the same as the tar files you download from kernel.org.

Re:I don't think so

[kasperd]

I don't think so.
It doesn't matter what you think. I tell you the .src.rpm contains the original .tar.bz2 file, and that is a fact. I even verified that file together with the .sign file from kernel.org. Anybody who doesn't believe me can install the .src.rpm and see for themselves that I'm right.

Red Hat at least puts in a lot of small modifications to the actual source code, sometimes back-patching fixes for security updates, or making changes to integrate better with certain applications.
Yes, that is why there are also patches in the .src.rpm file. The latest kernel 2.4.20-18.9 contains 152 patches. The .src.rpm contains 189 files in total:

• 1 spec file
• 1 .tar.bz2 file
• 152 patch files
• 25 .config files
• 10 other files

Re:Is it worth upgrading for old Red Hat Linux 7.x

[antdude]

Minna Kirai: Other applications using audio don't get interrupted. Only whatever program has video.

Mirror

[ddent]

http://www.omegasphere.net/oss/linux-2.4.21.tar.bz 2

I'll take it down if it gets to be a problem, but I don't think it will. It should be nice and fast :).

Re:Mirror

[ddent]

Oh, and don't take my word for it being the real thing... check the MD5s for yourself!

Re:why not posting the http links instead of the f

[tomhudson]

the TCP/IP protocol doesn't catch all errors, for either HTTP or FTP. So, yes, it's possible to get bad md5 sums on a complete download. The larger the download, the more likely it is to happen, and it only takes one.

I happen to prefer ftp, 'cause it does what I want, when I want, except for those times when I prefer wget, 'cause it too does what I want, when I want. I just use my browser for browsing ...

Re:why not posting the http links instead of the f

[tomhudson]

Now to reply on-tiopic - maybe the reason they're only posting an ftp link is because:

1. anyone who wants to download it w. a browser is free to do so - most browsers support ftp
2. they might not want to run both an ftp and http server serving the same files - the http server would then just be for web pages, and ftp for downloads. Simpler setup == easier to maintain, more secure
3. it auin't broke, so don't fix it - it's working fine as is.

Re:Is it worth upgrading for old Red Hat Linux 7.x

[Minna Kirai]

That makes it sound really like a problem in the application (MPlayer or xine) itself, then. Not something related to your kernel or sound drivers.

It's as if, when the window is being dragged, the program doesn't bother updating the audio output buffers.

However, I tried Mplayer myself, and it continues to play sound during a move or resize. You might try adjusting your window manager properties (in KDE or Gnome control center, for example) to change the effect used when moving a window. Try toggling things like "display content in moving windows" on, and stuff involving "translucency" off.

Re:Is it worth upgrading for old Red Hat Linux 7.x

[antdude]

Thanks. I will try that. I am using Ximian's GNOME v1.4. I usually turn off show contents in moving windows to speed things up.