Kernel 2.4.17 Out

ThatComputerGuy writes "Linux kernel 2.4.17 is final, with a lot of fixes/updates. Check out the huge changelog. If you're on a desktop machine, you should try using RML's preempt patch, it definitely helps response times."

Comments

Instability with the preempt patch?

[Dimensio]

I used the preempt patch back when 2.4.14 was released and I kept getting consistent kernel panics. Mind you, I'd also applied an -ac patch, so I can't say for certain that preemption was the cause, but it was troubling and the panics went away once I disabled preemption.

Re:Instability with the preempt patch?

[donutz]

I've applied the the preempt patch along with win4lin patch to a 2.4.14 kernel and experienced no kernel panics, no stability problems. I am not sure, however, if i have seen any positive effect from having this patch applied. it hasn't hurt, at least...

Re:Instability with the preempt patch?

[KnightStalker]

I had similar problems with 2.4.14 and the kernel preemption patch, but I'm running it against 2.4.16 right now with spectacular results. I can play a DVD full screen when the system load (i/o bound) is greater than 3 on my single-Athlon 900. No skips, pops, jumps, anything. Smooth as silk.

Re:Instability with the preempt patch?

[indigo78]

I had some problems, too. I ran RML patch on 2.4.14 on an k6-III and I've experienced strange locks of the system (Slackware 8.0 with some updates). Suddenly everything freezed and remained blocked for about a minute, then everything got back to normal. It could happen around one time every two hours. After I updated to 2.4.16 (and didn't install the RML patch), I've never experienced this kind of problems (but my gateway's Davicom ethernet card seem not to like the new kernel very much).

Re:Instability with the preempt patch?

[Spoing]

Make sure that any kernel modules you load are SMP-safe. While the preempt patch does not magically make a uni-processor system into a multi-processor, it does create an environment where modules that aren't SMP-safe tend to fail or act unusually.

Rule of thumb: If it's in the base kernel sources, you're OK. If it's a patch to the kernel sources, be careful but not overly concerned. If it's a pre-compiled binary (even if only in part), be very cautious. Remember: Google is your friend.

Example: The Alcatel external USB DSL modem, for example, has a binary part that isn't fully SMP-safe. Because of that, it can't be used with the preempt patch even on a system with a single main CPU.

Re:Instability with the preempt patch?

[Darren Winsper]

May I recommend Benoit's Alcatel Speedtouch driver at http://benoit.papillault.free.fr/speedtouch/index. en.php3 ? It's a user mode driver, but it does need the official drivers for the modem bios object which it uploads to the modem.

There are many reports that these are far more reliable than the official drivers and since it's entirely usermode it doesn't suffer from the SMP-safeness issues the official drivers have.

Re:Instability with the preempt patch?

[BeeazleBub]

Preempt patch hmmm.

I haven't ever seen any use to this patch. The only thing that I've ever seen it do once applied is make system performance choppy on the user end. I have better response times without it on a workstation than with.

Re:Instability with the preempt patch?

[pivo]

Yes this is off topic, but I wonder what you're using to play DVDs. I haven't had great luck with any DVD player.

Re:Instability with the preempt patch?

[KnightStalker]

I'm using Xine to play, with the d4d and dvdnav plugins. Menus don't always work correctly, but I usually can play the movie decently, with subtitles and alternate soundtracks. I also have a Matrox G400 and XFree86 4.1.0.

It turns out the preemption patch didn't do much. I rebooted into a kernel without the preemption patch and I could play movies with the same reliability as with the patch. When I played through the Xshm driver (no acceleration) I got perhaps half a frame per second as opposed to one frame per second with the preemption driver enabled.

The tests I was doing to load the system were:

while (true); do tar -czv . > /dev/null; done (two of these on two different partitions)
gzip < /dev/urandom > /dev/null
cat /dev/urandom > /dev/null

top reported the load as being around 4.0 with 60% user and 40% system.

Re:Instability with the preempt patch?

[Spoing]

You're right. Thanks for the correction!

Kernel Tree

[CaffeineAddict2001]

Ive got dibs on putting the star on top

Who's gonna do via?

[Kernel Panic]

From the changlog:

Jeff Garzik is not the via82cxxx driver
maintainer anymore: "No time, no hardware".

So who's gonna do it? Am I an idiot or is that a fairly common chipset among linux users?

Re:Who's gonna do via?

[WoofLu]

I'm using 2.4.16, compiled the OSS sound module (CONFIG_SOUND_VIA82CXXX)
And it's working quite well.

But the sound hangs when it's being used with aRts, then only message i get is this one:

mybox:$ dmesg
[snip]
Assertion failed! chan->is_active == sg_active(chan->iobase),via82cxxx_audio.c,via_c han_maybe_start,line=1201
usb.c: USB disconnect on device 2
mybox:$

Anyone experiencing this problem/knows how to fix it ? It's pretty annoying, sind I have to reboot everytime it hangs to keep listening to music ..

Re:Who's gonna do via?

[jbailey999]

I have a Via VT82C686 and hand all sorts of randomness with this driver. The output was always corrupt unless I used esd, and the microsoft didn't work. I went out and bought an SBLive! card so that I could use gnomemeeting with it (The savings on my phone bill made it pay for itself in about 3 days).

It's not obvious from the ALSA web pages if this chip is supported there or not. Once the kernel updates slow down, I'm going to try ALSA - It would be cool to have a sound card just for telephony, and the SBLive! for games...

Wait for Final Preempt patch

[pbur]

You will want to wait until RML releases the finale preempt patch. It will just be the kernel version (2.4.17) without the rc on the end. His patches are very version specific.

Pbur

Re:Wait for Final Preempt patch

[signe]

Actually, there is a final 2.4.17 patch, but it's not on the page linked to. Go to
http://www.kernel.org/pub/linux/kernel/people/rm l/ preempt-kernel/v2.4/ and grab it from there (or use a mirror).

-Todd

Re:Wait for Final Preempt patch

[devphil]

Uhhhhh... on the page linked to in the /. article, there is no "rc" on the end. The link there for "2.4.17-1" goes to the same place (kernel.org) as your link.

Re:Wait for Final Preempt patch

[pbur]

You're right, he seemed to be updating the page when I posted because at the time an "rc" link existed but didn't go anywhere. RML is good like that. :)

Changelog...?

[Hiro Antagonist]

I haven't looked at the changelog yet, but I'm sure that there's a line reading something like

"This time we /promise/ not to corrupt filesystems when you 'umount /mnt/tmp/lifes_work'."

All the same, many kudos to the kernel guys for giving me something new to play with for the holiday!

It appears to actually be fixes

[Anthony Boyd]

It looks like we're actually seeing 99% bug fixes this time around, rather than new features being added. Yay for having a 2.5 branch, it seems to be getting the experimental code now. This may be the first 2.4 kernel I compile for my system (I'm not saying I'm still stuck in 2.2, just that I've kept the default 2.4 kernels from my Mandrake and SuSE installs). I also see a couple ext3 fixes, which means I'm pretty comfortable having this replace the patched-to-use-ext3 2.4.10 kernel in my SuSE 7.3 box.

Re:It appears to actually be fixes

[ansible]

Well, it's gonna be a while before I try 2.4.anything on a production machine.

On the one hand, I'm glad to be seeing a lot of bugfixes in the changelog. On the other hand, I'm worried about seeing lots of bugs in the changelog, because I wonder how many more serious ones are lurking in the codebase.

With 2.2.whatever, the bugfixes there have been recently have mostly been related to specific drivers, which I usually don't care about (unless I'm using that driver). But with 2.4.x I'm still seeing lots of fixes in the main kernel code, because so much fundamental structure has changed since 2.2.

I guess we'll have to see. Maybe if some decent RAID card support starts showing up in OpenBSD, I'll think about switching my servers to that.

Re:It appears to actually be fixes

[mbrod]

This is good news for Debian which is probably not to far from releasing Woody (3.0) onto the world. Last time I checked they were teetering on not putting 2.4 in there as default (I know easily apt-get'd but it would be nice as default).

So if they stabilize it out good the Debian guys might go with that (I hope).

On another note mostly bug fixes is a beutiful thing... ANYONE AT MOZILLA LISTENING????

Re:It appears to actually be fixes

[josepha48]

2.4.10, 2.4.13, 2.4.16 have all been fairly stable on my desktop workstation.

The fact that 2.4.17 was ALL bug fixes is a good sign. It means that they are fixing the issues that they know about, like parport IEEE1394 seems to work better.

I see no reason not to use 2.4 at this point on a produciton machine. It is not until people do start doing so that we are going to find out all the kernel issues. I'd recommend trying this release if no other.

VIA KT133 chipset?

[blogan]

Does it work with the KT133a Chipset and Athlons? I looked and google and there were reports of the problem, but no report of a fix anywhere that I could find.

Re:VIA KT133 chipset?

[MikeO]

A bios upgrade fixed the problem for me. Look for an update on your motherboard vendor's website.

Re:VIA KT133 chipset?

[Otter]

What is the problem? I looked at Google as you mentioned but didn't see what you were talking about.

Re:VIA KT133 chipset?

[KidSock]

Does it work with the KT133a Chipset and Athlons?

This was a long standing motherboard/BIOS issue. I think VIA actually found and acknowledged that some register was getting scrambled inadvertantly but I could be thinking of another peice of hardware. It's so hard to find these messages in the lkml archives even after only a few days. At any rate, I have this board. At one time I had problems but it was difficult for me to tell if it was the VIA board or the infamous IBM Deskstar 30GB drive I had that was failing. So I RMA'd the drive, bought another SCSI drive and put my 2940UW on there with the latest BIOS update. I have not had any problems since and I tortured it pretty hard. Of coure all of this is off topic because I'm using Red Hats 2.4.7 but again, this is/was a hardware/BIOS issue.

Re:VIA KT133 chipset?

[josepha48]

how do you upgrade a bios in Linux??

Also does anyone know if they fixed the parport problems?

To answer your question about KT133, there is a bug on the VIA chipset that causes some people problems.

In my case and a few others there are IDE bugs. IDE IOMEGA zip 100 and VIA686B don't go togeather to well.

Also some people have had problems with using the ECC/ECP part of the parport. Again, in my case it was my webcam did not work. It stopped working in 2.4.13 or 14 not sure. I had a fix for it and explained on the lkml what the problem was, but not sure if that was ever acknowledged.

Well now I have a very good testing mechanism for testing kernels like this. I have a foo.test kernel and a foo.custom kernel. The custom kernel is the current kernel and the test is, well a test kernel. Nice thing about lilo is you can have multiple kernels. I can then boot the test kernel and make sure all works on my system. I'll be doing that tonight...

Re:VIA KT133 chipset?

[SaDan]

I've got 2.4.17 running on a Sony VAIO FXA36, which uses the KT133 chipset, and it's working just fine.

Re:VIA KT133 chipset?

[jmv]

I have a board (ASUS A7V133) with a KT133A+T-Bird 1.33 and haven't experienced problems with any kernel I tried so far (2.4.5, 2.4.7, 2.4.8, 2.4.10 and 1.4.16). The only problem I've heard (never experienced it) with my bord was actually the VIA (866B?) southbridge.

Re:VIA KT133 chipset?

[josepha48]

Thanks, I do have dos disks still. I was hoping I would not have to do that again though,but I was able to upgrade my BIOS. It had fixes for several issues.

Also it seems that between my biios update and 2.4.17 the parport issues has gone away. I think it was that they finally fixed the IEEE1394_ops.c code. Yes there were changes in this file that now makes my web cam work again ;-)

Re:VIA KT133 chipset?

[Karpe]

The fix was included somewhere before 2.4.16. It disables a single bit in a undocumented PCI register, fixing most problems.

How stable is the preempt patch?

[cornice]

I want to try the preempt patch but I finally got a kernel from my favorite distribution that works for me right out of the box. To get a raw kernel with all the stuff that I want plus the preempt patch I'll have to spend more time than I have patching and compiling. I hear how great this patch is about once a month. Why isn't it included in the major distros? Would it "harm" in any way my desktop/server running X, KDE, FreesWan, Apache, Samba, etc?

Re:How stable is the preempt patch?

[nusuth]

I used it against 2.4.12-ac? on a single athlon machine, it seemed to be stable and useless. I didn't run FreesWan or Apache, so I can comment on them, but X, KDE and Samba was OK. Infact, none of the programs I tried was affected in a negative way. It just didn't help.

Re:How stable is the preempt patch?

[MindStalker]

Well, why don't you try the source provided for you by your distro, it should have all the distro changes, in it. Then see if you can get the patch to work. If you can't, find a linux guru who can do it for you, and publish the patch, many people will love you :)

Re:Wow

[sam@caveman.org]

you don't HAVE to upgrade your linux box, either. but at least you have the OPTION of upgrading for FREE, instead of paying year after year if you want to upgrade.

as you said, you are happy with win98, more power to you. but many people are not happy with windows and have to shell out big bucks every couple years to upgrade.

yeah, i know, the post was flamebait. but hey, i'm a sucker for anything this obvious.

-sam

New Maintainer

[krackbebe]

Looks like the new kernel maintainer is really working out. I enjoy seeing these kind of detailed changelogs, to determine whether there is anything critical enough to upgrade my system.

Seems like Alan and Linux lately haven't been all that hot about doing the drudge detail work. This arrangement seems to be the best solution for everyone.

Re:New Maintainer

[kill-1]

Alan's changelogs of the -ac kernel tree have always been quite detailed. But Linus' changelogs were really bad. A lot of code that came from Alan's tree only had changelog entries like "merge with Alan".

Re:New Maintainer

[krmt]

Yeah, and after all the flack everyone gave him here for being terse with his interview, it's great to see his changelog speaking for him instead. Much kudos.

Re:New Maintainer

[MicroBerto]

Not only that, but consider how many fixes were done at the RC level! Can never be too patient, those bugs could have been in 2.4.17.

Re:New Maintainer

[Per Wigren]

Anyone saying Alan and sexual in the same sentence must be REALLY perverted.. :-P

Oh wait, I just did! Damn! Well, nevermind...

What's the point?

[Anonymous Coward]

I'm all for stable vs. unstable forks, but at several points the changes to the 'unstable' one just start rolling back anyways. So, why keep maintaining 2.4? or 2.2? or 0.1? Another thing is even vs. odd naming, why not just put a b beside the version number of an unstable release? How does it make sense that odd is unstable? Lastly, who the hell decided to do the 0.10 > 0.9 thing? Would it kill people to make a small bugfix update a .01 or .001 incriment if it meant you wouldn't have to fuck over the decimal system?

Re:What's the point?

[lamj]

Because it is a way of version management and control.

2.4 and 2.5 are maintained by different group of people. 2.5 is basically a play ground for the kernel playground to test big changes and it might take serveral revision to get such big changes right.

On the other hand, in 2.4 changes are much smaller, mostly fixes and very important proven improvments.

Please understand that changes in 2.5 does not sync with 2.4, things that should up in 2.5 does not mean it will be in 2.4, this gives developers (or kernel hackers) more flexibility to play around with things.

To learn more about this, play around with some source version control software like CVS and start branching off different version of your code then you might understand the need for a totally testing branch.

Re:What's the point?

[prog-guru]

Another thing is even vs. odd naming, why not just put a b beside the version number of an unstable release?

They have pre releases, those are more like beta versions than the development branch is.

Would it kill people to make a small bugfix update a .01 or .001 incriment

Well you also can't have 2 decimal points in a real decimal number, and you would have to decide how many subdivisions there are when you release (if they made 2.4.1, then needed something higher than 2.4.9 they would need to break the system).

I think it's pretty easy to tell where 2.4.17 is, it is after 2.4.2, and 4 is even so it is stable. Plus it doesn't suffer from version inflation like many distros do.

And how come FreeBSD is at 4.4 when Linux started first and 2.4.17 has lots of stuff not in 4.4 yet?

Re:What's the point?

[ahaning]

I think what the previous poster means is that, in basic math, the 0 in .10 would be truncated, and thus (.1 < .9).

The answer to their question is that, yes, (.10 < .9). However, in software releases, the period is just a separator, and not a decimal point. Thus, (10 > 9) is true.

Re:What's the point?

[geekster]

I think it's probaly confusing when you use only one point:
0.9, 0.10
because we're used to reading that like a decimal number.

But when you have two points:
0.9.0 0.10.0
I think it's more obvious.

Re:What's the point?

[be-fan]

Because they have a sane versioning system and don't try to be all 'leet and artificially keep their version numbers down! What is it about OSS developers that makes them dispise version numbers above 1.0?

Re:I haven't finished downloading the last release

[Ranger Rick]

get the patch

Re:I haven't finished downloading the last release

[Black Parrot]

> yet on my 56k modem, and a new one is already out? I'm hopeless...

If you're modem bound, you're supposed to start downloading the next kernel, before they finish working on it.

I take that back

[xercist]

150.208.72.11 has it right here!

patch mirror

[noodlez84]

I have mirrored the patch and signature:

patch-2.4.17.bz2 (388KB): http://home.earthlink.net/~noodlez84/patch-2.4.17. bz2
patch-2.4.17.bz2.sign (1KB): http://home.earthlink.net/~noodlez84/patch-2.4.17. bz2.sign

Re:patch mirror

[Anonymous Coward]

Does it really make sense to mirror the signature? You'd have to be pretty dumb to download the sign from an untrusted source. That's what the sign is _for_.

Re:patch mirror

[zdzichu]

Mirroring signature is useless.
Signature should be downloaded form kernel.org to check if patch mirrored by you has been altered.
If you mirror signature, you could have altered both patch and .sign.

Re:patch mirror

[hpa]

If so, he would have to possess the kernel.org secret keys. If he does, we have some really nasty security problems that we didn't know about :-/

This isn't just an md5sum, it's a gpg signature. You can verify it as long as you have the public key, which is widely available.

Re:Wow

[sam@caveman.org]

try using windowsupdate to get the latest directx (8.1) for windows 95 (see older slashdot story).

-sam

2.5.0

[havardi]

Im running the 2.5 kernel and it seems as snappy as the 2.4.blah with the preemptpbile kernel patch-- winamp doesnt skip even when burning cd's and starting mozilla... is there something different here? besides that i cant find the preemtpaible kernel patch for 2.5

Re:2.5.0

[aeil]

if you are running after a 2.5.1-pre3 then the bio changes are what's making most of the difference, better io scheduling and consistant buffer handling..

Re:2.5.0

[Phork]

winamp? is that an error, or do you realy do something like run winamp in wine?

Re:2.5.0

[zdzichu]

There is Winamp for Linux from some time.

Re:YAY !

[Publicus]

Read The Cathedral and the Bazaar

Release early and release often! How else are those bugs going to get squashed?

Preempt Patch?

[jasno]

Has anyone actually tried this and noticed a difference? I was under the impression that a lot of people thought this was useless.

Re:Preempt Patch?

[itarget]

The patch makes a big difference for me when using latency-sensitive software (xmms) while I'm really pummeling my system (big compile).

xmms usually skips a bit while I'm compiling something large, but it hasn't even once after applying the preempt patch.

I haven't noticed performance degredation from any effects on throughput, so it's all good here.

Re:Preempt Patch?

[Anonymous Coward]

It definitely makes a difference on my (400Mhz) system. I suspect, however, the faster your system, the less noticeable the effects will be, since it'll take less time inside the kernel regardless of whether the kernel can be pre-empted or not.

That said, the patch is, architecturally, a good thing, as making it work better automatically makes SMP work better, and vice versa. I think it should go into the offical kernel in the 2.5 cycle, personally - even on servers, it often matters that clients are responded to quickly, to keep people at a website, say, rather than sending the absolute max amount of data down the line - so even if the preempt patch has a slight negative impact on total throughput, it's a win for responsiveness.

Re:Preempt Patch?

[Legion303]

Preempt made my system (around 2.4.9 or so) noticeably slower.

-Legion

Re:Preempt Patch?

[PD]

I installed it on 2.4.14 and I didn't notice a single bit of difference.

BUT, I'm also not doing anything that can really take advantage of it. I don't play music or movies. The most I will ever do is run vi and gcc at the same time.

So, it depends on what you do with your machine. The patch seems to work for people who want to play video on a machine with a high load.

Re:Preempt Patch?

[X-ViRGE]

I definitely have noticed a huge difference in the load times and responsiveness of applications. X and my desktop start up in like 2/3 the time they used to take.

Re:Preempt Patch?

[LinuxGeek8]

Good question.

I never used the preempt patch, because I'm on a smp system, and it would make it rather unstable.

But from what I hear, it helps with latency problems, but has a decrease on throughput.
So I'm wondering, just increasing the latency does do exactly the same, doesn't it?
I mean, I increased the latency on the pci bus, and my mp3's are almost never skipping anymore. Which is the same claim made by lovers of the preempt patch.
I never use video stuff though, so I can't comment on that.

For changing the latency, you can find a good read here:
Latency

Re:Preempt Patch?

[jbailey999]

Most systems I see this kind of problem on haven't enabled DMA on their drives.

I use:
/sbin/hdparm -d1 -c1 -m16 -u1 /dev/hda

That's DMA, 32bit IO, Multiple sector count: 16, Unmask IRQs

YMMV

2.4.x bigger than 2.2.y

[valentyn]

Ha! 2.2.18 on december 11, 2000. 2.4.0 in january, 2001. That means, roughly, 1.5 versions of 2.4 per month, while we have only 1 version of 2.2 in 6 months.
Somewhere in april we'll have 2.4.21 and 2.2.21 and one month later, 2.4.22 will be out. Hooray!

Re:2.4.x bigger than 2.2.y

[ocelotbob]

Kernel development is a funny beast. Sometimes, things just work, and kernels only have to be released when there's a significant amount of hardware. Other times, things don't work and you have to put out a lot of fixes - the early days of the 2.4 VM system sucked; it only started to really rock about 2.4.10 or so. What you didn't take into account with your guestimates is that kernel 2.2.x also had some growing pains in the early stages; it just took a bit longer for 2.4.x to get stable. Just take into consideration the 2.0.x series, they're getting close to a .40 for that line. Trust me, 2.0.x is still being used; there are some vendors who supplied binary-only patches for the 2.0.x series which don't work with the newer lines.

Re:2.4.x bigger than 2.2.y

[GigsVT]

That's the way it is supposed to work. As a program matures, there will be less updates, and things to fix. I thought this was a pretty basic concept...

Obligitory karma whoring

[damiam]

Don't forget to use the mirrors.

Re:Preempt Patch? - have not noticed a difference

[donutz]

I've tried the preempt patch and not noticed a difference either...

maybe what we need is someone to give us some concrete, real world examples of where the preempt patch would do some good? Then we'd know if it would make much of a difference for what we use our systems for...

Misguided versioning?

[kurtras]

After several of the last few kernels being released with major bugs, I thought the consensus on LKML was to use -rc versions for bugfixes, and then release a 'final' without making any changes in it. Yet, when I read this changelog, I see that changes were made in the final version. A lot of people will only download a 'final' kernel, because they think that it contains only tested, stable code. That is what the -rc system was to ensure, but releasing a 'final' with changes means that a partially untested kernel is being released to the unsuspecting public. Now, I will admit that there's a very good solution that any user can implement - just don't upgrade. However, these recent quality control problems have given Linux something of a black eye in the public's mind. Therefore, it just seems common sense to not release a kernel with code that hasn't been in for at least one -pre or -rc revision. So, if I were a kernel maintainer, about to release kernel 2.4.18, and I received a 'critical' patch from a project maintainer, I'd make one last -rc release to ensure that the code gets tested before I release it. However, I'm not a kernel maintainer, so take this as you will. I don't mean it as a flame, and I think that Linus and Marcelo have done a wonderful job so far with Linux 2.4.

Re:Misguided versioning?

[aeil]

I dissagree, 2.2.x series was handled in a similar fashion as the 2.4.x is now. so one could really say that the lack of a development kernel to let the "kids" play with their new patches is what really hurt 2.4 so far. cause if you hadn't noticed, most of the changelog is BUGFIXES.

I just remembered... RTFM (or RTFL in this case) archives can be found on the web.

Re:Misguided versioning?

[Sits]

I don't quite understand your point. Unless releases are introducing new features they must be fixing bugs otherwise what's the point of changing the version number?

Surely all that happens is that a nominated pre is renamed to an rc and if the rc proves to be stable it is renamed to final. So all final releases will have bugfixes in them because they were either a pre or they fixed something in a previous rc.

Re:Misguided versioning?

[be-fan]

Basically, you have for each kernel version you have several -preX kernels. Once you get the thing stable, you rename the line to -rcX kernels. From here, you accept only bug-fixes. When the -rcX kernel gets really stable, you rename it to -final, without changing anything else.

Re:Misguided versioning?

[Anonymous Coward]

> After several of the last few kernels being released with major bugs, I thought the consensus on LKML was to use -rc versions for bugfixes, and then release a 'final' without making any changes in it. Yet, when I read this changelog, I see that changes were made in the final version.

You're correct that this is the general principle that will be used. BUT, Marcelo reserves the right to make judgements case-by-case, and in this case felt that the fixes were small enough to go in the final build.

My favourite change:

[schon]

From pre1:

- Speeling fix for rd.c

Gotta love that sense of humor (at least I HOPE it was intentional :o)

Re:My favourite change:

[sluggie]

this was intentional... you know chargon file compatibility... ;)

Cache/Buffer Fixes

[Mish]

I don't know if anyone else is as happy as I am right now to see these fixes in the changelog...

- Make kernel try a bit harder to shrink caches
instead swapping out
- Fix VM problems where cache/buffers didn't get
freed

The 2.4 series has been plagued by these problems, thank god that they might finally be over...

NTFS bug fixes?

[Mr. McGibby]

Any word on what the NTFS bug fixes involved? Any closer to a usable readwrite mode?

Re:NTFS bug fixes?

[bucky0]

I've been wondering about that too. Why is linux support for NTFS lagging?

Re:NTFS bug fixes?

[Mr. McGibby]

Lack of solid documentation and licensing issues as I remember.

Re:NTFS bug fixes?

[Cheesy Fool]

Well, if you have checked the linux-ntfs website, especially the mailing list archives you would see this question comes up a lot.

Here's one reply to why it hasn't been done yet.

Re:NTFS bug fixes?

[Anonymous DWord]

You remember right. Lack of *any* documentation, more like it. Licensing I don't think is a problem, as everything is done blind. MS changes NTFS every time they put out a new OS though, so that kind of complicates things.

Re:NTFS bug fixes?

[rfredell]

It doesn't seem to bother the guys over at sysinternals too much.

Re:NTFS bug fixes?

[afidel]

Yeah because they write a wrapper around the Microsoft libraries. It's kind of easy to support it if you can use the native code =)

Re:Wow

[cscx]

You're comparing apples with oranges.

You're saying that a kernel upgrade is an OS upgrade... WRONG! Someone can go to Windows Update and get new drivers, updates, etc.

If I want to upgrade the OS, sure I'll have to get the upgrade... That's like saying you have to get the upgrade if you want to go from RH 5.2 to 7....

Of course Windows users don't have to face the task of recompiling their kernel every so often. From what I hear, many linux distros are taking the Windows route and encouraging the use of loadable modules (just like Windows) instead of compiling them in.

AMD K7 SSE

[BrookHarty]

in the changelong I noticed...
pre5 - Enable K7 SSE (John Clemens)

So we now have SSE for the K7 cpu? Does any programs on linux even take the extra speed of SSE/MMX/3D NOW? I have always wondered since these type of optimizations are only visible when the software application lists it, and most software is for windows.

Re:AMD K7 SSE

[Algorithm wrangler]

SSE requires some extra effort on the OS on context-switch, since the SSE registeres are not normally saved/restored (they are only available on SSE-enabled CPUs). So SSE has been unusable on K7 up until now since you might have your programs screwed up by a context switch.

Re:AMD K7 SSE

[Anonymous Coward]

What are you talking about?

The message is about SSE being enabled on AthlonXP's.

It doesn't work on 'normal' Athlons!

Re:AMD K7 SSE

[fr2asbury]

Actually XMMS has the ability to make use of 3D-NOW!
It's not what you might think though.
It's go tnothing to do with 3D, but it does make it write to disk MUCH faster.

On my site K6-rpms ou can find XMMS rpms that take advantage of this.

Cheers,
Jonathan

Re:AMD K7 SSE

[fr2asbury]

Oops
Don't worry, if you don't see XMMS in the list of files, it's there. Just click on any of the other projects.
I'll take care of that as well as get the latest pre-release of XMMS up, over the weekend.

Cheers,
Jonathan

Re:AMD K7 SSE

[Hoser McMoose]

Just a little FWIW for anyone who's wondering, the following current AMD chips support SSE:

AthlonXP
AthlonMP
Athlon 4 Mobile
Duron chips of 1GHz and faster (and new 900 and 950MHz chips)
Mobile Duron chips at 800MHz and greater

Presumably any future AMD chips will also support SSE. To date, none support SSE2 (as found in the P4), though rumor has it that this will show up in the Hammer series of AMD chips to be released in about a year.

Re:So what happens if 2.4.99 is released?

[donutz]

ummmmm, 2.4.99.2

Nah, probably 2.4.100; the kernels start at x.x.0, not x.x.00, so it seems we're using the convention of just tacking the incremental version on without any worry about leading zeros...

Re:Wow

[alen]

I guess then they will be just like Apple charging for OSX upgrades.

So what kernel does slashdot use?

[Dante]

I wonder what kernel slashdot uses on it's servers.
are up to 2.4 yet? Or still running 2.2?
If they are still running 2.2 when will slashdot upgrade?

What is "preemtive" for ?

[hemabe]

I don't understand the difference between preemtive and the normal way (btw: which?). Could someone explain this to me? What are the advantages? Thanks, Hermi

Re:What is "preemtive" for ?

[itarget]

Having a preemptive kernel just means that the kernel will allow itself to be interrupted by other programs and give them some cpu time.

This improves response time for your programs as now they won't get stuck waiting for the kernel to finish doing something time-consuming (like disk I/O) before they get some cpu cycles.
In most cases this isn't a big deal, but you'd definitely notice when your mp3 player skips because it's stuck in line behind the kernel.

Re:What is "preemtive" for ?

[agrapa]

When a process starts a kernel call, it will not be context-swapped until it finishes executing the call. That is, the kernel will not change the executing process as long as it is kernel code that is in execution. There are many reasons for this. The patch makes it OK to to context swapping in the middle of kernel execution, helping all processes get an equal amount of CPU time.

Re:What is "preemtive" for ?

[hemabe]

Thanks you for your answer, i don't understand how it is possible, that the kernel right now is not preemptive? Is it possible, with the old kernel, that one process will consume all cpu time. Is it possible, that the process will not "give back" to other processes ? ciao, Hermi

Re:What is "preemtive" for ?

[jquirke]

No, a process when in userspace cannot hog the CPU. Even if a process does not sleep on a wait-queue, say the program consists of:

for (;;);

It will be forced off the current CPU when its counter expires. The period of time depends on its nice value. The counter is decremented each clocktick, and a re schedule() happens when the counter is found to be =0.

The only way a program can hog the CPU if it is running real-time (SCHED_FIFO), but only superuser processes can do this.

Re:What is "preemtive" for ?

[be-fan]

There are two things here. Preemptive in userspace, and preemptive in kernel space. Linux is preemptive in user space, meaning that when a process is running application code, it can be preempted. When it is running kernel code, however (ie. during a system call) the scheduler will not preempt the thread, even if a higher priority one becomes available. Essentially (on a single processor) the kernel code is cooperatively multitasked. Kernel code runs until schedule() is called to invoke the scheduler. Some kernel code paths are very long, which leads to long periods where the current process cannot be preempted, which kills latency.

Also, to expand on the original question, there are a couple objections to the patch: It has the potential total throughput, because more locking must be used since the kernel can be preempted at any time, not just at specified points. However, in practice, the effect on throughput seems to be negligible. It also increases complexity, due to additional locking, but most of the complexity is there anyway, in the form of the SMP locks.

Re:What is "preemtive" for ?

[UnknownSoldier]

> I don't understand the difference between preemtive and the normal way (btw: which?).

The normal (old) way is "cooperative" -- meaning you don't yield a task until you're ready.

Pre-emptive means you can be forced to give up your task.

Re:What is "preemtive" for ?

[ASM]

That's not *quite* right. Linux has been pre-emptive since the beginning, but in userspace, not kernelspace. IE, system calls, driver code, and other kernel stuff couldn't be preempted, but user code could.

Windows, on the other hand (9x, I don't know about NT) is fully cooperative, meaning that userland isn't preempted either. That, and poor memory protection, is why buggy windows programs can bring the system down, Whereas in linux, only kernel space stuff can lock up the system.

The preempt patch, then just makes the kernel preemptable, so that Linux is fully preemptive, instead of just in userland.

Re:What is "preemtive" for ?

[be-fan]

Ah! Partially right! But sorry, no dice. The only cooperative multitasking that is done in Windows 9x is with 16 bit applications, since all 16-bit apps are run in the same virtual machine. All 32-bit apps are fully preemptively (userspace multitasked. As for memory protection, that's only partially true. Application memory is indeed protected from each other. However, there is a big 1GB region of shared memory that is unprotected. Apps that use this region and asking to hose the system. Also, some bits of kernel memory are unprotected because DOS apps need access to them.

As for NT, it is a fully preemptible kernel, both in userspace and kernel space. Like all preemptive kernels, of course, it is not preemptible when interrupts are disabled (since the clock interrupt can't happen). The main reason why NT has always been preemptive is because its always been SMP. The locking requirements on SMP are similar to to locking requirements for a preemptible kernel, so you can get both together for the price of one. Indeed, the preemptive patch for Linux is very small because it uses the existing SMP locking mechanism.

Re:What is "preemtive" for ?

[be-fan]

Hey, I love BeOS as much as the next fanatic, but (as much as I would love to) I can't use it because its compiler is old and there is no support for the rear-channel on my SB Live!

I'm going to have to create a "LinBoys" fangroup..

[cybrthng]

I have upgraded from Windows 2000, to W2k SP1, to W2k SP2, IE 5.5 SP2, DirectX 8.0-8.1, IE 6.0 and lots of other hardware upgrades, driver updates and crap and have paid

$.00

Infact i only paid 149.00 to upgrade to Windows XP after running 2k for a while that wasn't a bad deal either.

Now if i have any problems i can submit the crash dump to Microsoft and within 3 days they have analyzed the dump and written back if it is hardware, software, os bug or program bug.

I can't get that from Linux being an endusers, i don't have time to peruse the code to fix all the bugs that i may come across and well, features WIndows has em all, i'm not waiting for one thing or another then what the latest and greatest will be.

But waite, i also Have a Suse 7.1 server i bought for a few hundred running an Portal 9iAS server, and i have a Professional RedHat 7.2 i have purchased from redhat. Both of which cost me just as much as my windows solution with the same amnount of support. Linux happily runs my Resin/JSP/XML/XSL development server and makes a great workstation for coding..

But believe me, your underestimating the value of having upgrades and at the same time not having to download new kernels, compile, tweak, reboot, and crap like that. That is the luxury of Windows. If you need an update, point your browser to windowsupdate and voila.

Sure Ximian has it, but upgrade your OS to something else and wait... wait.. wait.. and maybe the support will come, or quite possibly after you have already purchased ximian and paid the 9.95 a month for preferred access your already paying out the wazooo for more then what windows costs..

anywhooo, your not right or wrong, but your point is moot without you actually giving any reasons linux would be cheaper. Time is money, and if it takes a while day for me to download an iso, install, patch, secure, tweak and fix a linux box then that has cost me alot more then a 149.00 cost of XP that has installed out of the box on more hardware then i have ever seen..

oh well, to each there own.. just thought i'd poke you for a bit

Re:Already? x = 36

[pheph]

I think 2.0.36 was the hugely popular RedHat 5.2 based kernel.

loopback deadlocks

[Anonymous Coward]

final:

- Fix more loopback deadlocks (Andrea Arcangeli)

the very first line of the changelog is scaring my ass of. this sounds like there are some / an unknown number of loopback deadlocks still lurking and nobody knows where, until it jumps out to rip your head off.

Your out of your league...

[didyaseethat]

Case in point, go to windows update on your NT kernel 5.0 (win2k) pc, and try to update to NT kernel 5.1. Surprised your upgrade is not there. For the upgrade you will have to buy Windows XP. Of course Windows bug fixes are there, but they could not get away with blatently charging to fix their mistakes. You can also download a new 2.4.x kernel onto your Redhat 5.2 PC, compile the kernel, and Bang, for free, you have all of the kernel improvements since Redhat 5.2 was released, without the overhead of running new window managers, or whatever. Granted most people will just acquire the newest release, but it is not the only forced solution if you want the improvements. All distro's I've ever used/seen have used all modules rather than compiled in drivers. It would be near imposible to support even a few different machines with hard coded kernels. Expert users may compile their own kernels to acheive even beter performance than offered in a stock kernel, but no one ever has to. Every distro I have ever used releases RPM packages of the new kernels pre-compiled and ready to go. Its as easy as just clicking the icon in your file manager and your ready to go with the new features. It is clear that you have very little, if any exposure to linux. For the sake of those that may be interested in trying something different, please keep your mis-information to yourself.

Re:Your out of your league...

[Wumpus]

You can also download a new 2.4.x kernel onto your Redhat 5.2 PC, compile the kernel, and Bang

Congratulations on your choice of words. Bang, indeed.

Re:Your out of your league...

[cscx]

It is clear that you have very little, if any exposure to linux.

It is clear that you have very little, if any exposure to WinNT.

Ever heard of service packs? They have new builds of all the OS files bundled. Yes, that means a kernel upgrade. Just because the version number didn't increment doesn't mean the build number didn't.

Suck it, Trebek.

Re:Your out of your league...

[didyaseethat]

Service packs....actually, I have installed two of them on my Windows 2000 pc (not to mention the millions of updates on 9x). Lots of bug fixes, but unfortunately no new features. Take your 2.0.X kernel, upgrade it to a 2.4.X and the worm has turned for you my friend. The capabilities grow exponentially. The title "Service Pack" says it all. Service what you got. If you want better (not that XP is better) pay up.

Build number...heh heh

By the way, I keep 2000 around for the TV out hacks. I can't yet get my tnt2 overscanned on my TV. But the day I can....

Patching the kernel

[Corporate Troll]

I know you are supposed to download the patch. It would be ideal since I only have ISDN. Unfortunatley I just downloaded it, and now I don't know what to do.
Downloading the kernel is easy, it's a tar.bz2 and with tar -Ixvf I get the whole tree, unfortunately I just notice a patch is only "patch-2.4.17.bz2", I can unzip it, but then? What do I do?

I'm only using Linux since about a year seriousely and downloaded and compiled nearly all kernels since then but always the complete package. It would be great if I knew how exactly to patch the kernel.
Yes, I know this is a newbie question. Forgive me for this :-(

Re:Patching the kernel

[mboedick]

I used to always download the whole tarball too, until I discovered /usr/src/linux/scripts/patch-kernel. Now all I do is grab the patches, verify the signatures, drop them in /usr/src/linux, run that script, and compile.

It's a fairly intelligent script and makes things painless.

Re:Patching the kernel

[Corporate Troll]

Thanx....I just read that in the readme, and I'll try it tomorrow :-)

Re:Patching the kernel

[Corporate Troll]

The kernel script worked fine, and my machine is running 2.4.17 happily now. Thank you for the extra information, and I'll try that at the next kernel upgrade. (I printed out your explanation :-))

good thing

[7-Vodka]

Good thing it was intentional! Being brazilian I recognized it as a common spelling mistake for brazilian english and thought the worst haha :)

I'm stupid!

[Corporate Troll]

Aaaagh! Now I feel extremely stupid, I just did 'man patch' and it's all there! I just should have RTFM and shut up :-(

Re:I'm stupid!

[Anonymous Coward]

at least you realized you asked something stupid... :)

the trickiest part is figuring out that you need to use "patch", there's also the README file in the root of the linux source that I believe gives you instructions

-- gid

Re:I'm stupid!

[Corporate Troll]

Gid,
It's not that obvious if you're new to all of this. Of course I am in fault not to read the REAMDE, but then I got used to README's in download directories that explain server policies and such. Perhaps my post will point users that don't know about the 'patch' command to using patches instead of using the source tree. I know it sounds silly, but this is the kind of help newbies need.

Thanks for your reply :-) I do realise I asked something stupid, but that doesn't make me a bad person, does it?

Re:I'm stupid!

[greenrd]

If you didn't have that username, someone might actually believe you.

Re:I'm stupid!

[Corporate Troll]

You're lucky you can. At work we're an Microsoft Shop mainly. I'm one of the few Java people, and I'm the only one using Linux regularly. People mostly ask me questions, but one cannot know everything :-)

Re:Wow

[__past__]

In fact, *only* Linux users have to upgrade their kernel every so often.

Users of proprietary OSes don't have a chance to, and users of about
any other Free OS - well most other free OSes kernels just aren't
broken every second week.

Re:Preempt Patch? - have not noticed a difference

[Darren Winsper]

My MP3s stopped skipping under heavy(ish) load. For example, when checking out the Mozilla tree, there is a lot of hard disk access for up to a minute. During this time, XMMS would often skip and crackle, whereas with the preempt patches this no longer occurs.

Xine also seems to like the patches. I can often have two compiles going on in the background with a fair bit of swapping and DVD playback is still smooth. This was not the case without the preempt patches.

In general, the preempt patches help if you use your system as a desktop/workstation, it could actually harm system performance if it's primarily a server.

Why are you upgrading?

[nbahi15]

I use my office and home server on 2.4 kernels without trouble. I got over upgrading to the latest kernel a long time ago. I only upgrade if I need functionality or encounter a serious problem. I moved to 2.4 for USB support. I've only had one kernel panic on my machine in the 2.4 series. And as far as the preemptive patch I have never found it necessary. I do add the microcode support and load the microcode fixes from Intel and my array uses reiserfs which I patched.

Re:Why are you upgrading?

[chill]

Depends on your hardware. Firewire support has been improved/stabilized in successive kernels; there have been filesystem fixes to Reiser and Ext3 (both of which I use).

As far as preemptive: you see it when running something like an MP3 player; DVD/CD/Video player or anything else that is very timing sensitive. Try doing a big compile or massive file copy with XMMS running in the backgroung -- skips, stutters and glitches. The preemptive patch addresses this a lot.

Re:slave ATAPI CD on soundblaster

[ghack]

that is (a) off topic and (b) the sb16 scsi cdrom uses the aha152x driver. look it up somewhere...and dont post OT questions

Re:Wow

[rseuhs]

Of course Windows users don't have to face the task of recompiling their kernel every so often.

Neither have Linux users.

Did I miss something? Did Bill Gates redefine the english language so that "be able to" and "have to" mean the same thing? Why don't Windroids understand the difference then?

Re:slave ATAPI CD on soundblaster

[Craig Davison]

Or (c) Some SB16s use a Sony, Panasonic or Mitsumi proprietary inferface, shitstain.

Seriously, this guy is probably looking for help with one of the "non-ATAPI/SCSI" cd-rom interfaces that are so rare now.

Re:Wow

[nbahi15]

If all you want from computer is slick looking marketing. I encourage you to buy Windows. Buy two instead. Aside - I bet everyone who swears by Windows violates their EULAs.

What is the better operating system? Stupid question. Dumber argument.

If you want to create the tools that canned apps don't provide you choose Linux. First because programming is powerful, easy and free in Linux. Second you can manipulate the damn files since they aren't in a proprietary binary format.
Linux provides something that Microsoft never can. Multiplatform support. Linux will run on a toaster (provided you really wanted it to). It can be customized however you like and it is a solid platform to build any truly efficient appliance. Tivo is a good example, so easy, so solid, sooo Linux. Also infinitely hackable.

No nasty licensing issues. I would never want to compete in a Microsoft world because you can only play by their rules (read EULA).

Finally be concerned about the politics of your OS. Digital Rights Management is coming, and it will not be negotiated in your favor. You like Fair Use, the massive corporate concerns want us to take only what they provide and buy only on their terms. I'm not a big fan of restrictive EULAs. I'm also not a fan of copyrights. And before you ask, no I see no need for inventors to receive anything other than the glory of their own innovation. And before you say, I disagree that the world would come to an end and the economy would collapse.

cameron.

Re:bios flashing, not just from DOS anymore

[thing12]

You should try out /dev/bios. It's a very cool little kernel module. I recently used it to flash the bios on one of my linux boxes that didn't have a floppy installed. Just cat the file out to /dev/bios and reboot. It's a beautiful thing, though obviously you probably wouldn't want to have it installed all the time :-)

KT266A

[green pizza]

I noticed someone else asking about VIA KT133 support, so I thought I'd inquire about the KT266A...
We have two new office-brew systems, one mobo from Asus and one from Abit, both are based on the KT266A and neither will boot any flavor of kernel 2.4.x that we've thrown at it. I've done the normal google and usenet searches, but haven't found much other than a few "works for me" posts. Anyone have some pointers or patches?

Re:KT266A

[theMissingLink]

We have 3 soyo dragon plus motherboards at work which I believe use this chipset. I have a soyo dragon at home. All running SuSE 7.3 2.4.10 kernel with no problem. Well no problem booting anyway. We have some flakiness we haven't figured out with the dragon plus boards.

Re:I haven't finished downloading the last release

[vrws]

damn, man! Why is this rated 'funny' with _every_ kernel release? Or am I the only one noticing/nagging?

Asking /. staff...

[TeknoHog]

Many people keep asking if the /. staff actually read their site - when commenting on multiple instances of the same article. Now, I wonder whether the staff ever read/reply the comments, because I've never encountered any of them replying. Of course, they would probably reply under some obscure aliases, but you should still be able to tell...

Re:Asking /. staff...

[Ziviyr]

Commander Taco has spoken. :-)

BZ2 vs GZ

[jquirke]

Something I feel like asking as 2.4.17 (bz2) trickles down the connection at 0.2K/sec from Australia's Planetmirror...

The kernel's are posted in both GZ and BZ2 formats. What do you guys mostly use? I can't see much point these days with having the Gzip format, I mean is there still a point to downloading it? Or even having them available in that format?

From what I can see, removing the Gzipped versions

*reduces network congestion
*saves space on the mirrors
*saves space on local storage (yeah only a couple megs)

Of course, it requires more processing time to extract, but that seems to be no big deal these days. I'm pretty sure everyone has bzip2 installed , and those who don't can easily get it, so that can't be a problem.

So is it really just traditional reasons it's posted in Gzipped format? Tell me if I've missed something. It would be interesting to know what everyone thinks about this.

Re:BZ2 vs GZ

[damiam]

Tradition, some people don't have bunzip2, and BZ2 takes longer to decompress (a real issue on slower systems).

Re:BZ2 vs GZ

[TaliesinWI]

I'm sure I'm going to be showing my age when I say that I remember the Big Switch from .Z (Compress format) to .gz on most FTP sites. Is .bz2 really becoming that common? The only place I'm really exposed to it are the kernel sites, most other source code repositories are either just .gz or still have legacy .Z stuff lying around...

Re:BZ2 vs GZ

[Gregg Alan]

I only use bz2 for the kernel. Seems to me that processing power is cheaper than bandwidth (lucky for me I have plenty of both.)

If the extra time to decompress a bz2 over a gz is that great a factor, why would you even want to compile a kernel on that particular machine? Compile it on your fast box and just copy it over. That's what I do.

Re:BZ2 vs GZ

[spuk]

Bzip2 is *way* slower and memory hungry than Gzip, and the gain is not necessarily great ...

Re:BZ2 vs GZ

[|_uke]

the difference between downloading a full bz2 compressed kernel and a full gz compressed kernel IS a fairly big deal on my little dialup

not to mention that if you remove whatever percentage of users that still downloads the .gz version, you save the difference in size times the amount of users in bandwidth!

a few megs per user times just even as little as 25 users makes a difference

the bz2 is not there so much for your sake, but for the servers :)

(although again for little modem users like me, it DOES make a difference... for me that is :P)

Re:BZ2 vs GZ

[jbailey999]

I use my favourite format: .deb =)

The stock Debian kernels have been just fine for me. I used to enjoy going through and selecting all of the options in versions 1.2.x. Now with all of the options, it's become an all-day project to try and remember which options to pick. The debian kernels nicely have (almost) everything available as a module, and has all the scripts to setup the initrd.

Re:BZ2 vs GZ

[Ed Avis]

If you're worried about download time, just get the patches!

(Okay that only works if you've got the previous version - you can download a whole sequence of patches for bigger jumps, but after a while it gets bigger than the kernel itself. I'd like to see a general patch generator where you type in what version you currently have and it generates the smallest patch file just for you. Alternatively some way of using rsync would save on bandwidth.)

Re:BZ2 vs GZ

[bconway]

You realize you can re-use your old .config, right? There's even an option to load or save old non-.config files in the various menuconfig/xconfig/whatever menus.

Re:BZ2 vs GZ

[spuk]

Sorry, I misread the original post. What I meant was that Bzip2 may not be worth it to be used everywhere. For the specific case of Linux, I think it could be reasonable to use only Bzip2.

Re:BZ2 vs GZ

[Anonymous Coward]

I use Debian source kernels; the trick is to copy .config from the old tree to the new one, then connect to the new tree and make oldconfig. Voila', you only answer the new questions; all your existing settings get setup automatically. Build as usual from this point.

Re:3DNow!

[TeknoHog]

MPlayer makes use of MMX and 3Dnow! if they are available. Makes my K6-III+, 400 MHz, play DivX quite well :-)

can anyone compile it?

[seandarcy]

I've got it. I'm trying to compile it. It fails at make bzImage compiling network.o. See the snippet.

Anyone know how to fix it?

ld -m elf_i386 -T /opt/kernel/linux-2.4.17/arch/i386/vmlinux.lds -e stext arch/i386/kernel/head.o arch/i386/kernel/init_tas
k.o init/main.o init/version.o \
--start-group \
arch/i386/kernel/kernel.o arch/i386/mm/mm.o kernel/kernel.o mm/mm.o fs/fs.o ipc/ipc.o \
drivers/char/char.o drivers/block/block.o drivers/misc/misc.o drivers/net/net.o drivers/media/media.o drivers/char
/agp/agp.o drivers/ide/idedriver.o drivers/scsi/scsidrv.o drivers/cdrom/driver.o drivers/sound/sounddrivers.o drivers/pci/d
river.o drivers/pnp/pnp.o drivers/video/video.o drivers/md/mddev.o \
net/network.o \
/opt/kernel/linux-2.4.17/arch/i386/lib/lib.a /opt/kernel/linux-2.4.17/lib/lib.a /opt/kernel/linux-2.4.17/arch/i386/
lib/lib.a \
--end-group \
-o vmlinux
net/network.o: In function `__rpc_schedule':
net/network.o(.text+0x49a0d): undefined reference to `rpciod_tcp_dispatcher'
make: *** [vmlinux] Error 1

Re:can anyone compile it?

[jdg]

I built it without problem. net/network.o build
okay. I haven't installed it yet. I'd be happy to email you my .config file if you want.

jdg at eznet dot net

Re:can anyone compile it?

[skriefal]

Appears to be a known problem when compiling with RedHat's prerelease of the GCC 3.1 compiler.

Here is some more information. In general, I recommend that you do not attempt to compile the Linux kernel using any version of GCC newer than 2.96.

Re:can anyone compile it?

[jrockway]

> In general, I recommend that you do not attempt to compile the Linux kernel using any version of GCC newer than 2.95.

I'll second that. gcc3 really fouls up the kernel. Everything compiles okay, but on boot the system goes right back into OpenFirmware with a "DEFAULT CATCH!". gcc2.95 with the same config fixes this.

Woo Hoo!!!!

[_01001001]

I've been waiting about a week or two for this kernel to come out. There is quite a few updates that I looked at in the changelogs a few weeks ago, that I wanna get. Mostly, the update of the sym53c8xx_2 driver.

I would have used one of the beta versions, but I have something against using a beta kernel on my workstation. So sue me ;)

--Frank

Re:So what happens if 2.4.99 is released?

[floodle]

Nah, probably 2.4.100; the kernels start at x.x.0, not x.x.00, so it seems we're using the convention of just tacking the incremental version on without any worry about leading zeros...

This is correct. The 2.1 devel series went to well over 100, 2.3 should have but 2.3.99+ became "2.4.0-testX" for some reason, although it went for a number of months and some major changes were made.

Another RML?

[RML]

"RML"? Robert M. Love stole my initials! Now I have to worry about people confusing me with someone who knows what he's doing.

This is fixed

[free!arrow]

A patch specific to this was applied a while back - I think in 2.4.10. It fixed the problem in my case. You will see "Trying to stomp on Athlon bug" in one of the very first boot messages - this is the fix.

Re:This is fixed

[Sits]

This guy is right - the workaround for the problem has been about for a while. For more information on the problem take a look at the Via Hardware FAQ. The whole via problem has been known about for sometime (a search on Kernel Traffic for KT133 turned up a few references. The most recent reference was 2.4 Kernel freezes on VIA KT133.

As mentioned in other comments, motherboard makers were encouraged to workaround this at bios level.

2.4

[Arjuna Theban]

I've been using linux since the very early 2.0.* days and for the most part I keep up with every kernel released. Since I've moved to 2.4.*, I've notices an incredible slowdown on my machine, even in post-2.4.13 kernels which supposedly did something to improve performance.

Personally I'm about ready to go back to good old fast&stable&reliable 2.2 tree. I wonder if we really need to make the kernel this sluggish for the sake of introducing new stuff in the kernel level though. I know I'm not the only one who noticed the performance drop with 2.4.*.

--

Re:2.4

[Arjuna Theban]

Well, for one, the basic program startup time. Your average bloatware (xemacs, staroffice, mozilla, etc) takes longer to start than they did on the 2.2 tree. I didn't time the processes with the two different kernels or anything so I guess I can't back that with hard evidence but it feels that way. The second one (which I've also heard from many other people) is the slowdown with disk i/o. When I am copying large files from one drive to another or there is simply a large amount of drive activity, my computer becomes virtually unusable. I mean, it takes 2-3 seconds to switch desktops in X. I've heard from a friend that 2.4.13 seemed to improve this situation for him, but it didn't for me.

I guess my problem is that I used to be able to run 2.2.* with X on a 233mhz machine just fine, and 2.4.* feels just like that on a 500. Maybe not *that* bad but 2.4 is a steep performance drop compared to 2.2 for me.

--

lame question.

[jon_c]

I just put RH 7.2 on my laptop (compaq 1720US). the eepro100 drivers that come with it do a timeout deal constantly. I am hoping that this new kernel includes some fixes for it.

Previesly i appempted to get the latest official kernel (2.4.16) and make it work with redhat. I got as far as using kgcc and getting it to compile but i was unable to get xwindows working after that. (i think it might have had something to do with gartagp not working, not sure)

does anyone know of some simple steps to make RH work properly with the latest kernels?

-Jon

Re:lame question.

[Anonymous Coward]

The thing about RedHat with newer kernels make sure that you sym link include/linux in source dist to /usr/include/linux to make sure you have the proper headers when you compile modules.

Also sym link asm, asm-i386 or anything else you may need form the include directory in the linux kernel source dist.

I usualy link these to /usr/src/linux which is itself a link to the current linux source dist.

I don't know if there's an easier way of doing this, but it seems that RedHat just copies header files from the kernel that it comes with into /usr/include which I think is nasty.

Re:lame question.

[Anonymous Coward]

Download and compile eepro100-diag.c from the Scyld website then follow Donald Becker's instructions here.

alcatel driver

[Renaud]

you're talking about the initial kernel-based driver from alcatel, which is now obsoleted by the open source driver by Benoit Papillault (http://speedtouch.sf.net).

This one works better and only requires the standard kernel option "HDLC line discipline" to be compiled in.

2.4.16 works great with the Gigabyte 7VTXH

[sunset]

I've built and tested a couple of these for clients. No problems at all. Seems about 15-20% faster than KT133A systems.

Re:DAMN! Am compiling 2.4.16 on a 486 as I write t

[rythriynn]

eeek! this is off-topic, sort of, but you really need to start rethinking your firewall... over here I do the same thing but DO NOT have gcc or anything remotely "code-powerful" (would kill mount if I could) installed on my firewall box... as such, yes I HAVE to take advantage of my protected computers increased performance for compiling the kernel and the modules I see fit (VERY few), then ship them over via scp (just for practice) to restart the new "test" config with... ...has anyone heard of patches btw?!!? also, why would you "upload" the source instead of the .tgz?!?!? also, having the kernel source on your firewall box is probably not the best thing to do.... especially since gcc appears to be there as well (and I didn't even have to crack your box to find out!) I appreciate your "pain" but please consider that this is a large community and "somebody's already thought of that"... instead of suffering, please do a little checking and spare yourself... the kernel will help you :) if you do insist on keeping the source on your firewall, then patches will certainly benefit you! but DON'T... kernel source+ gcc+ floppy disk=serious fun! kernel source+ gcc+ root+ lilo=serious problem... of course, the reasoning behind this is that my "protected computer(s?)" all run firewalls and "suspect" my firewall... hehe.... just a thought...

Re:although you have to wonder...

[ndogg]

They're being left out because they haven't been extensively tested yet. That's the point of stable. You work on what you have, don't add new features (see my first point), and make sure that all the bugs are out.

Re:Wow

[didyaseethat]

Depends on how you program. It can be a sticky situation in some cases. I mean, if you steal someone else's code. If you respect other peoples licenses you will have no problem.

You probably don't need preempt patch

[ndogg]

The preempt patch is mainly for embedded devices. This is definitely not something for production level machines. Most applications that don't target embedded devices aren't ready for preemption. To use the preempt patch is to introduce instability.

Some Vacation

[matusa]

Alan Cox seems to be the source of some 25% of the fixes. If that's taking it easy, than I'm a 17 tongued goat.

Yeah yeah, parts of it are merges of old fixes, but even with that in consideration, he is definitely superhuman

my thanks

The new kernel maintainer is doing a great job too. I can't believe he stepped in and had so much momentum immediately! Rather amazing. He's definitely living up to the very high expectations

thanks.

Marcelo's true merge plans

[matusa]

the pic says it all:

Here's a picture of marcelo wearing FreeBSD horns

I'm fireproof so don't bother with flames. I'm just having a good time, you know.

Preempt patch and 3rd party modules

[loftwyr]

I installed the patch, recompiled and both ALSA 5.12a and nvidia's kernel module were broken. I got unresolved symbols.

It would be nice if there was some way to exempt these two from the optimization or there was a doc explaining what I would need to change

Re:Preempt patch and 3rd party modules

[dossen]

Have you tried to rebuild the modules, this used to fix similar problems for me. I know the modules are binary, but there is still some linking to be done before they load on a given kernel. I usually just make modules modules_install and rebuild any seperate modules (nvidia!) every time I mess with the kernel (not that often)...

Re:Wow

[jbailey999]

This hasn't been my experience at all.

First that I've only run 3 of the 2.4 kernels over the last year, but second that I've run Windows Update at least every 2 weeks for updates that are marked ``critical''.

Half of these updates aren't critical: They're not even for software that I use - But I don't trust Microsoft to actually list all of the fixes in their release notes. I've had Internet Explorer updates fix overall system crash problems. Incredibly sad.

Could be pnpbios related

[Sits]

You don't mention which distribution you are using... I noticed that on one of the Mandrake forums someone had found that setting nobiospnp fixed the problem on their KT266A system.

Re:Does freebsd kernel have this

[athlon02]

Recompiling the FreeBSD kernel is pretty easy:

http://www.freebsd.org/doc/en_US.ISO8859-1/books /h andbook/kernelconfig-building.html

however, the FreeBSD authors don't go around releasing new kernels everytime you blink, they do it right the first time (=

My username

[Corporate Troll]

Don't always judge people by theire names. This is indeed my troll account, you are right. I stopped trolling by the way because it's not as fun as I thought it would be. Why did I use this account to ask such a question? Right, just because I didn't want to spoil my karma on the regular account. I actually thought I'd be modded as a troll or offtopic, but instead nice people tried to help. Now, isn't that what one should call "a community"? Of course it could just be the Christmas spirit :-)

2.4.17 with pre-emptive patch works great

[supaphinn]

looks good, my first time with the preemptive patch, as far as responsiveness i dont feel a difference, maybe slight but that might be psychologic.

anyway, it seems to work great. Looks like its worth using if it does what it says

not to mention the l33tn3ss of using it ;)

Numbering scheme - is it just me?

[simetra]

At first glance, it seems like 2.4.9 is bigger than 2.4.17. Obviously, 17 is greater than 9, but maybe the numbers should start with a zero for the 1 digit numbers... like instead of 2.4.9, 2.4.09
My 2 cents.

Re:3DNow!

[jrockway]

Eh... my 233MHz iMac plays DiVX fine (MPlayer screws up the sound, but Xine is great)

Re:Wow

[DavidTC]

There's not actually anything restrictive about the GPL at all. If the software you receives wasn't licensed under the GPL, and was instead just giving out straight copyright law, you couldn't do anything more with it, and a great deal less.

Re:Linux Kernel is lagging behind

[00Monkey]

LOL, "-1, Troll" Slashdot admins's heads are too big.

I posted that comment just to see how many flames and what score I would get.

I'm pro-Linux and the comment was a JOKE. I've never met more of an a$$ crowd than the pro-linux, Slashdot crowd.

Security

[Sloppy]

I use my office and home server on 2.4 kernels without trouble. I got over upgrading to the latest kernel a long time ago. I only upgrade if I need functionality or encounter a serious problem.

While this does seem wise at first, be careful. You can have a serious problem even if you don't encounter one -- What I'm getting at is security. You gotta keep an eye on the bugfixes. It wasn't that long ago that a local root exploit was found, in both the 2.2 (!) and 2.4 trees. Your 2.4.9 (or whatever old one you have) might not be safe, even if seems to be running fine.

Re:Wow

[sam@caveman.org]

lately try using win 95 as a firewall/dns/router on a 386?

have fun.

-sam