Kernel 2.4.14 is out

MrSnivvel writes: "The new 2.4.14 kernel is out for the taking." It's in all the usual mirrors. Check out the ChangeLog and revel in the newness..

Comments

Last Linus 2.4 kernel

[reynaert]

Isn't this supposed to be the last Linus kernel in the 2.4 series?

Re:Last Linus 2.4 kernel

[worldwideweber]

From: Linus Torvalds
CC: Kernel Mailing List

On 31 Oct 2001, Michael Peddemors wrote:
>
> Lets' let this testing cycle go a little longer before making any
> changes.. Let developers catch up..

My not-so-cunning plan is actually to try to figure out the big problems
now, then release a reasonable 2.4.14, and then just stop for a while,
refusing to take new features.

Then, 2.4.15 would be the point where I start 2.5.x, and where Alan gets
to do whatever he wants to do with 2.4.x. Including, of course, just
reverting all my and Andrea's VM changes ;)

I'm personally convinced that my tree does the right thing VM-wise, but
Alan _will_ be the maintainer, and I'm not going to butt in on his
decisions. The last thing I want to be is a micromanaging pointy-haired
boss.

(2.5.x will obviously use the new VM regardless, and I actually believe
that the new VM simply is better. I think that Alan will see the light
eventually, but at the same time I clearly admit that Alan was right on a
stability front for the last month or two ;)

> My own kernel patches I had to stop because I couldn't keep up .... Can
> we go a full month with you just hitting us over the head with a bat
> yelling 'test, dammit, test', until this is tested fully before
> releasing another production release?

I think we're really close.

[ I'd actually like to thank Gary Sandine from laclinux.com who made the
"Ultimate Linux Box" for an article by Eric Raymond for Linux Journal.
They sent me one too, and the 2GB box made it easier to test some real
highmem loads. This has given me additional load environments to test,
and made me able to see some of the problems people reported.. ]

But I do want to make a real 2.4.14, not just another "final" pre-kernel,
and let that be the base for a reasonably orderly switch-over at 2.4.15
(ie I'd still release 2.4.15, everything from then on is Alan).

Linus

Re:Last Linus 2.4 kernel

[FredGray]

"but Alan _will_ be the maintainer,"

What happened to the report (see this Slashdot story from Nov. 2) that Alan Cox would be replaced by Marcelo Tosatti as the stable release coordinator?

Re:Last Linus 2.4 kernel

[worldwideweber]

Linus posted this note to the linux kernel mailing list before the story of Nov. 2, so this should not be seen as a reversal of that story.

Alan is free to do what he wants with the kernel including pass on its maintenance to someone else :).

Re:Last Linus 2.4 kernel

[KidSock]

The latest Kernel Traffic suggests that Alan is considering using the AA VM anyway. Looks like Rik's stuff is getting dumped entirely. That is unless they can figure out how to parameterize VM's and make it a compile time option as they've been talking about.

Re:Last Linus 2.4 kernel

[42forty-two42]

Why not have it patch and unpatch the VM during the configure script?

Re:Changes/Improvements

[Eslyjah]

More readable version.

final:
- David Miller: sparc/scsi scatterlist fixes
- Martin Mares: PCI ids, email address update
- David Miller: revert TCP hash optimizations that need more checking
- Ivan Kokshaysky/Richard Henderson: alpha update (atomic_dec_and_lock etc)
- Peter Anvin: cramfs/zisofs missing pieces

pre8:
- Andrea: fix races in do_wp_page, free_swap_and_cache
- me: clena up page dirty handling
- Tim Waugh: parport IRQ probing and documentation fixes
- Greg KH: USB updates
- Michael Warfield: computone driver update
- Randy Dunlap: add knowledge about some new io-apics
- Richard Henderson: alpha updates
- Trond Myklebust: make readdir xdr verify the reply packet
- Paul Mackerras: PPC update
- Jens Axboe: make cpqarray and cciss play nice with the request layer
- Massimo Dal Zotto: SMM driver for Dell Inspiron 8000
- Richard Gooch: devfs symlink deadlock fix
- Anton Altaparmakov: make NTFS compile on sparc

pre7:
- me: reinstate "delete swap cache on low swap" code
- David Miller: ksoftirqd startup race fix
- Hugh Dickins: make tmpfs free swap cache entries proactively

pre6:
- me: remember to bump the version number ;)
- Hugh Dickins: export "free_lru_page()" for modules
- Jeff Garzik: don't change nopage arguments, just make the last a dummy one
- David Miller: sparc and net updates (netfilter, VLAN etc)
- Nikita Danilov: reiserfs cleanups
- Jan Kara: quota initialization race
- Tigran Aivazian: make the x86 microcode update driver happy about
hyperthreaded P4's
- me: shrink dcache/icache more aggressively
- me: fix up oom-killer so that it actually works

pre5:
- Andrew Morton: remove stale UnlockPage
- me: swap cache page locking update

pre4:
- Mikael Pettersson: fix P4 boot with APIC enabled
- me: fix device queuing thinko, clean up VM locking

pre3:
- René Scharfe: random bugfix
- me: block device queuing low-water-marks, VM mapped tweaking.

pre2:
- Alan Cox: more merging
- Alexander Viro: block device module race fixes
- Richard Henderson: mmap for 32-bit alpha personality
- Jeff Garzik: 8139 and natsemi update

pre1:
- Michael Warfield: computone serial driver update
- Alexander Viro: cdrom module race fixes
- David Miller: Acenic driver fix
- Andrew Grover: ACPI update
- Kai Germaschewski: ISDN update
- Tim Waugh: parport update
- David Woodhouse: JFFS garbage collect sleep

Changelog?

[atrowe]

Does anyone have a link to the changelog? I'm interested to see if they've fixed any of the bugs present in 2.4.11a and previous.

I'd also be interested to know if they've fixed the Integer floating poing RAND bug that can occur on some older MMX Athlons. I've got a friend who has one of the affected processors, and he's been dying to upgrade, but can't because his tarball segfaults on compile.

Also, what's the status of that TCP/IP stack optimization that Theo De Raadt has been working on? Has this been incorporated into the new kernel? I'm getting sick of compiling my winipcfg.dll file every time I install a new kernel. Hopefully the new winmodem streaming support will take care of this problem. It shouldn't have taken as long as it already has, considering DeRaadt switched from C to VisualBasic for his new code. Any info would be helpful...

Re:Changelog?

[ConsumedByTV]

Here is the change log
Ironicly for a mensa memember you could have at least read the summary that included the link. What is your offical IQ anyway?

Re:Changelog?

[Agent Drek]

wtf is up with the moderation? /me remembers a fine old day when slashdot articles weren't taken so seriously and this kind of stuff was good for a quick chuckle. And as for being marked off-topic, I think that these conversations would get quite stale without any tangents at all. my 2 cents. I'm relurking.

Re:Changelog?

[OblongPlatypus]

The parent isn't a Troll - if anything, it's Funny. A real troll should at least TRY to appear legitimate...

Re:Changelog?

[Darkstar9969]

Is the word sarcasm lost on /. readers?

I found it hillarious!!

My $.02

Is Fatal VM Bugs Fixed?

[robbyjo]

The last pre seems to be OK, but Linus is still "working" on that. Did anyone know that the VM bugs are fixed? I don't see that in the change logs after pre8...

That's nice, but its not really news...

[Nailer]

There's so many interesting userspace apps slashdot could write about. Unless the kernel has some new feature or fixes a major secutiy hole, I personally don't see how interesting each minor release is. Slashdot isn't freshmeat.

If /. is going to write about apps, why not focus on the new and clever ones - like
* MPlayer allowing us to play WMVs under out OS of choice
* Xine, finally maturing into a solid high quality DVD player
* Partimage providng a useful and open source disk imaging system
* Ximian's setup tools beta making an X config tool that doesn't suck

OI don't have anything against the kernel, but we all know there's always goiung to be a new kernel every couple of weeks. There's so many interesting userspace Open Source projects we could be hearing about.

After all, isn't the point of an OS to run *apps*?

Re:That's nice, but its not really news...

[Mr.Ned]

As I see in my upper-left hand corner, Slashdot bills itself as 'News for Nerds. Stuff that matters.'

->linux kernel = news for nerds of the linux type

->to nerds of the linux type, linux kernel = stuff that matters

->nerds of the linux type = majority of slashdot

I personally would like to hear about stuff like Xine and MPlayer and how that enables the world to break free of all things proprietary. Not saying I don't.

However, linux is one thing among others that binds this community together. It's stuff that matters. So what if a new kernel comes out every couple of weeks? If nothing else, it adds another news item that provides banner impressions that keeps /. afloat, and I'm for Jon Katz stories if they do that.

Re:That's nice, but its not really news...

[Jburkholder]

>There's so many interesting userspace Open Source projects we could be hearing about

almost afraid to ask, but...

have you tried submitting something? It's not like there is a finite limit to the number of articles that can be posted each day and an article on a new kernel version means there is one less article that can be posted on something else, no?

Kernel releases are news. Not earth-shattering news, but events that generate interest around here (for the most part, YMMV obviously).

>why not focus on the new and clever ones

Someone has to write something up and post it somewhere that can be slashdotted. It's not like /. developes its own content in the form of essays, editorials or the like. Think you need to go somewhere like linux journal for stuff more like that.

Re:That's nice, but its not really news...

[geekoid]

Wouldn't it be nice if the included a link to somewhere that gets a newbie the opportunity to learn where "the usual" places are?

Re:That's nice, but its not really news...

[garcia]

I don't read freshmeat. I search for things there. It isn't something I really care to see everyday.

I don't subscribe to the kernel announce list. I don't care to read a bunch of garbage everday.

A new kernel may not be news for everyone but it is certainly news for nerds.

If they tell us when the release of XP is, or when the release of a new variant of Code Red or the like is out, why not the kernel.

Sorry but I feel that the latest kernel is quite acceptable to be shown on /.

If you don't, tough. It obviously is going to stay no matter how many of you comment on it (remember, we have seen these posts before).

I like it.

Re:That's nice, but its not really news...

[Angst Badger]

Unless the kernel has some new feature or fixes a major secutiy hole, I personally don't see how interesting each minor release is.

Well, the 2.4.x series has been marred by some serious and not-yet-entirely resolved virtual memory flaws that can cripple some mission-critical server applications. I'm not sure that qualifies as security hole, unless you consider that it makes DOS attacks easier, but it is a major performance problem. Until a final fix is in, every release in this branch is news.

Re:That's nice, but its not really news...

[Anonymous Coward]

wouldn't it have been nice if you did? or I did? :)

Re:That's nice, but its not really news...

[drinkypoo]

If they tell us when the release of XP is, or when the release of a new variant of Code Red or the like is out, why not the kernel.

Those are not sufficiently similar things. Code Red is big news for the 'net at large. XP is a major rev; Linux kernel 3.x coming out is big news, but a teeny revision is not.

If this rev of the kernel had something new - PS2 support, perhaps :) - Then it would be front page news. If it's just bugfixes, which is mostly how it appears (looks like there's something about hyperlinked P4s, that's pretty big and important I guess, but I'm getting to that) then it's not worth the front page. If there's something worth mentioning in the article (and not just in the changelog) then it's frontpage news. Otherwise, why bother?

Re:That's nice, but its not really news...

[cvincent]

I don't read freshmeat. I search for things there. It isn't something I really care to see everyday.

you dont have to, you can subscribe to the linux kernel entry on freshmeat and recieve an email when a new version is out.

I don't subscribe to the kernel announce list. I don't care to read a bunch of garbage everday.

so release information is not garbage when its on slashdot but it is garbage when it comes from the actual source?

If they tell us when the release of XP is, or when the release of a new variant of Code Red or the like is out, why not the kernel.

Then by this very logic, CNN should be running stories on the latest Linux kernel release? XP is a bit different from 2.4.13 to 2.4.14 and Code Red made quite an impact on the Internet.

If you don't, tough. It obviously is going to stay no matter how many of you comment on it

That may be so, but dont expect the opposite and rely on slashdot to only run the stories you want. I bet if slashdot wasnt running stories on kernel updates that stories would still be submitted every release.

Re:That's nice, but its not really news...

[_fuzz_]

there should be a fourm on Freshmeat

There is a forum on Freshmeat. You can attach comments to releases and projects.

Re:That's nice, but its not really news...

[lupercalia]

I think you're overstating the case a bit. Linux accepted the new VM but so far reports have been pretty good. Yes, there have been some problems, but weren't we having problems with Rik's VM as well?

Linus' philosophy is usually to stick with a simple system a programmer can fully understand. Neat tricks sometimes give better performance, but simple systems are robust and tunable. Selecting right now to add the patch rather than forking off 2.5 and then backporting it doesn't seem very wise to me, but even Linus can make a mistake.

Re:That's nice, but its not really news...

[Dahan]

It's not like there is a finite limit to the number of articles that can be posted each day

Yes, there is.

an article on a new kernel version means there is one less article that can be posted on something else, no

Yes, it does. The editors go through the submissions and pick the ones they want to accept.

Re:That's nice, but its not really news...

[cowens]

But they don't have to post those stories. All they have to do is post a kernel release story and they know that some whiner will complain that they aren't focusing on relevent software like X and Y. This person will then get modded up to level four or five and everybody can get what they want.

Re:That's nice, but its not really news...

[GdoL]

One way to put the announce, IMHO, could be to put also links to how people can install the kernels, why they should do with, etc... And do the same when posting news about a new version of other projects.

Re:That's nice, but its not really news...

[koekepeer]

teeny revision?

Re:That's nice, but its not really news...

[zmooc]

A new kernel may not be news for everyone but it is certainly news for nerds.

Actually I think the things currently going on @ mplayer are much more important to nerds. Especially since for many nerds playing movies is still an important reason to run windows because they don't know about mplayer. Since /. is a very open-source-minded site, this would be much more important to cover. It's not even news but the impact of posting an article about mplayer would be much larger than posting about every new kernel.

Re:That's nice, but its not really news...

[Denny]

It says "the usual mirrors" with a link on mirrors...

Did they post-edit the story, or did you post without reading it?

Regards, Denny

Re:That's nice, but its not really news...

[mr3038]

Unless the kernel has some new feature or fixes a major secutiy hole, I personally don't see how interesting each minor release is. Slashdot isn't freshmeat.

How about publishing all the news covering software releases (including kernel releases) under new topic called "software releases" and people like you can opt-out in preferences. Others, including me, can read about these from /. because we don't read both slashdot and freshmeat.

Re:That's nice, but its not really news...

[naasking]

I personally don't see how interesting each minor release is.

The REAL question is: does anybody really care what you think Slashdot should be like?

The answer is: not really! But thanks anyway. Next time, just try ignoring the article. That's usually what you do in real newspapers when you come across articles that don't interest you.

Re:That's nice, but its not really news...

[Eil]

That may be so, but dont expect the opposite and rely on slashdot to only run the stories you want. I bet if slashdot wasnt running stories on kernel updates that stories would still be submitted every release.


CmdrTaco has stated on many occasions and in many interviews that the news that gets posted to slashdot is only that which interests him. Nobody else. Except maybe the other staff at /.. And it's not going to change no matter how much bitching goes on.

Re:That's nice, but its not really news...

[drinkypoo]

Er, I should have said "small" revision. Major.minor.small.teeny. Or at least, that's how I learned it.

Re:That's nice, but its not really news...

[koekepeer]

tiny? teeny is open for ambiguous interpretation ;-) hehe

meneer de koekepeer

incorrect mirrors link

[fonebone]

I think they meant to link to this url.

Re:incorrect mirrors link

[fonebone]

yeah, i think someone actually fixed it.

nevermind, then.

Version niceness

[Anonymous Coward]

The 2.4 kernel will be complete only when it reaches the "Version Niceness" factor, that is version 2.4.24.

Re:Version niceness

[grammar fascist]

No, I think you're missing the point. "Version Niceness" can't have seemingly arbitrary numbers in it - that's why Microsoft ditched their scheme after Windows 3.1.

I think that 2.4.XP is a much nicer version number.

Re:Version niceness

[Bradee-oh!]

I think that 2.4.XP is a much nicer version number.

And to all those disgusted about this apparent relation between a good version number for the kernel and M$'s new o/s, what he really meant was to relate it to the new Athlon's which obliterate their Intel counterparts :)

Re:Version niceness

[GC]

Actually WinXP is just another name for Windows version 5.1.2600

5.1.2600 just doesn't have the same ring to it as XP though... I think the 2600 is the build number...

I wonder what build number Linux would be if it incremented everytime Linux make a kernel change.

Re:Version niceness

[Nau.dk]

"Actually WinXP is just another name for Windows version 5.1.2600"

Here's the proof, WinXP is unstable. ;-)

/Nau

VM finally there...

[RadioheadKid]

It looks like there were many fixes to the VM, which is good news. Hopefully this is the one we can all be happy with (well that will never happen, but at least be content for a bit) and let the Linux team move on to 2.5. The VM talk has seemed to calm down a bit on the LKML.

Re:Stop it.

[mabinogi]

nightlies?....

aren't you confusing this with Mozilla...

There's no (official) public CVS or nightly builds of the Linux kernel, so these releases (and the pre-xx ones) are all there is.
No one's forcing you to upgrade, and you really shouldn't upgraded as often as possible, you should only upgrade if you know that the new version has an imrovement / bugfix / new feature that you need or want.
And even then, in theory you should test it thoroughly on a test machine before putting it on a production box. (Though of course for desktop users this is usualy not much of an option).

release often

[archen]

you know, at the rate these updates are comming, they're going to run out of numbers, and they'll have to start calling them 'service packs'...

Re:release often

[C0vardeAn0nim0]

Well, if you think this is fast wait for 2.5.x to start. it's not unusual for a development kernel to go above x.x.100.

Ohhh, and 2.0.x was in 36 revisions when i shifted to 2.2.0

I still don't know why 2.2.x had so few releases. only 17 before 2.4.0... weird...

Re:release often

[Ctrl-Z]

Uh, maybe because it was good and didn't need that many releases? I never had any problems with any of the 2.2.x kernels I used.

Re:release often

[prog-guru]

Ohhh, and 2.0.x was in 36 revisions when i shifted to 2.2.0

I was starting to worry about having so many in 2.4.x, now I don't feel so bad.

Re:release often

[Cryptnotic]

It's impossible that they would run out of numbers, since the set of version numbers has infinite cardinality (i.e., is infinite).

Cryptnotic

Re:release often

[ozbird]

... they'll have to start calling them 'service packs'...

Nah, call them "feature releases" - it's the same thing, but you get to charge money for them.
(Citrix *really* sucks.)

Re:release often

[jsse]

No, they will increment the middle no., just like what they did when it went beyond 2.3.99

Just kidding. :)

Re:release often

[Tony-A]

2.4.infinity+1 (we're talking ordinals here)
2.4.infinity+2
...
2.4.inifinity+infinity+1
...
2.5.0
3.0.0

Re:release often

[death_denied]

it's not unusual for a development kernel to go above x.x.100.

I would not call it "not unusual" if something only happens once.

Re:release often

[Tony-A]

You are thinking cardinals, where an infinite set has exactly as many elements if you add any finite number of elements to it. I.e. there are exactly as many primes as there are fractions.
Problem is, the is no last element added that made it infinity.
If you are counting (ordinals), +1 makes a different number, so if you have infinity, infinity+1 is a different number.

The line looks like this:
1,2,3,4,5,...
...an infinite number of them here...
infinity,infinity+1,infinity+2,...
...another infinite number of 'em...
2infinity,2infinity+1,...
...and so on....

freshmeat

[Anonymous Coward]

Is VA Linux / Research / Software / Panhandler trying to save domain registration fees by combining slashdot and freshmeat or something?

Mirrors?

[EvlPenguin]

Maybe chrisd meant here. The page he posted was the mirrored sites AT kernel.org.

Re:call it what it is

[mabinogi]

Actually, A Service pack is a collection of patches and updates to a wide variety of services and applications that make up the complete system, and have been tested (in theory) to work together, and can be installed in one go.

A new kernel version would be just one small part of a service pack.

So no, it's not a service pack, and even if it were.....would that be a bad thing?

Personally, I'd rather be seeing lots of updates, which indicates that development is still being done, than to see nothing, and live under the uncertaintly of not knowing whether this meant that the system I had was perfect, or if the maintainters just didnt care.

Though you're definitely right about "if it's not broke, don't fix it" part.

New rules when announcing Kernels.

[SirStanley]

Since when is the Linux Kernel "The Kernel" "Kernel 2.4.14 is Out" is being highly biased towards Linux users... I thought Affirmitive Action was Illegal.. Isn't showing bias towards a Lesser Operating system in order to get it more accepted Affirmitive Action? Dammit. I run FreeBSD. My Kernel isn't at 2.4.14.... So I'm alienated.... Where's the ACLU?

Re:New rules when announcing Kernels.

[NonSequor]

Maybe someone should make her a heart donor.

Useless point...

[LSD-OBS]

"from the 14-and-counting dept."

We're actually at 15 (not 14) and counting - kernel version numbers are 0-based, and sequential.

Sorry moderators - I just had to :)

Re:Useless point...

[LSD-OBS]

You can start counting at 0, but when you reach 14 you've counted 15 numbers, and I do believe "14-and-counting" alludes to the number of point-releases counted.

"from the point-release-14-and-counting-dept." would be more accurate, but I'll try not to lose any sleep over it. *yawn*

Re:Useless point...

[larien]

Well, there was the "sorry excuse for a kernel" 2.4.11 which was withdrawn due to the symlink bug. So you could argue that we are at number 14...:)

Re:Useless point...

[WWWWolf]

We're actually at 15 (not 14) and counting - kernel version numbers are 0-based, and sequential.

Actually 14. Unless, of course, you count 2.4.11 as a real "release". =)

Re:Useless point...

[shanek]

And how do you know the "14-and-counting Dept." didn't start counting at 0 as well? :^)

Re:Changes/Improvements

[Saint Aardvark]

Hm...these are always interesting, and I do like keeping up w/the Jones' -- but I'm nto a kernel hacker, and I don't know, for the most part, what they're talking about. Can anyone give some insight about what these changes mean for ordinary users? Not something like "Better, Stronger, Faster", but "this kernel call is used by GnomeFoo for reading your mail and sending it to Linus. It's now 2.5 times faster and encrypted."

Or is that not a good question -- is there, maybe, no way to dump this stuff down w/o getting to cliches like "Better, Stronger, Faster"?

Re:why is this news?

[jfmiller]

Normally I just lurk but I feel the need to disagree with Mr. Grimes.

I like many who read Slashdot am a linux enthusiast. I appriciate having a place where my OS of choice is reported on with the zeal and timeliness that Microsoft gets in the main stream press. Slashdot is one of the few sights I read on a daily basis, and I for one appriciate hearing about the new kernels. along with the rest of the pertinante Tech news.

My local paper still lists The High School football team's results on the front page every Saturday during the fall, even though most of it's readers dont have kids on the HS football team. It builds a sence of community pride wich is greater than the information that is contained or it's relevance to peoples lives.

Kernel anouncements on slashdot are appriciated by me and I hope a great many other readers. And even if there are only a few of us Linux is one of the things that form Slashdot's Community, and therefore in the interest of Community pride I encourage CT to continue to post all the kernel updates.

JFMILLER

Link error

[killme]

ld -m elf_i386 -T /usr/src/kernel/2.4.14/arch/i386/vmlinux.lds -e stext arch/i386/kernel/head.o arch/i386/kernel/init_task.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/parport/driver.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/char/drm/drm.o drivers/ide/idedriver.o drivers/scsi/scsidrv.o drivers/cdrom/driver.o drivers/sound/sounddrivers.o drivers/pci/driver.o drivers/pcmcia/pcmcia.o drivers/net/pcmcia/pcmcia_net.o drivers/pnp/pnp.o drivers/video/video.o drivers/usb/usbdrv.o drivers/input/inputdrv.o \
net/network.o \
/usr/src/kernel/2.4.14/arch/i386/lib/lib.a /usr/src/kernel/2.4.14/lib/lib.a /usr/src/kernel/2.4.14/arch/i386/lib/lib.a \
--end-group \
-o vmlinux
drivers/block/block.o: In function `lo_send':
drivers/block/block.o(.text+0x854f): undefined reference to `deactivate_page'
drivers/block/block.o(.text+0x8599): undefined reference to `deactivate_page'
make: *** [vmlinux] Error 1

Re:Link error

[jrockway]

make clean dep, please

Re:Link error

[killme]

make dep clean bmImage
was what I did. Looks good to you?

Re:Link error

[baptiste]

make dep clean bmImage
was what I did. Looks good to you?

Um not unless this is a typo - you cleaned after you dep'ed - ooops! make clean dep bzImage would be more like it!

Re:Link error

[baptiste]

Its a bug. See further down

Re:Link error

[killme]

reply to my own question:

I found where did deactivate_page used to be defined: it's in mm/swap.c

it's been removed from the 2.4.14 release. :(

in 2.4.13:
/* Deactivate_page will move an active page to the right
* inactive list, while activate_page will move a page back
* from one of the inactive lists to the active list. If
* called on a page which is not on any of the lists, the
* page is left alone.
*/
static inline void deactivate_page_nolock(struct page * page)
{
if (PageActive(page)) {
del_page_from_active_list(page);
add_page_to_inactive_list(page);
}
}

void deactivate_page(struct page * page)
{
spin_lock(&pagemap_lru_lock);
deactivate_page_nolock(page);
spin_unlock(&pagemap_lru_lock);
}

Re:Link error

[Ranger Rick]

Not a typo, that was the suggested way of doing it (at least used to be). "clean" doesn't clear out dependency files; 'mrproper' is the "super clean" that erases those, along with your .config and other things (makes a pristine tree as if you just untarred it).

Re:Link error

[indigo78]

I'm really thinking that many features aren't checked anymore. Isn't there somebody in the kernel team that checks for a complete build every release?

Re:Link error

[42forty-two42]

How to fix: in any file that is linked in: add: void deactivate_page(struct page *){}

kernel bugs

[Foxxz]

where does one go to report kernel bugs?

-foxxz

This is all good, but...

[da3dAlus]

I've had a problem with my Promise Ultra66 controller since kernel 2.4.6 and above. Apparently, my system hangs at boot whenever it gets to my Maxtor 60GB drives. So far, not a single kernel update has addressed the problem, and I'm a bit pissed. I've looked on Google, Google Groups, irc.linux.com #linuxhelp, and nobody has heard of the problem or has a solution. I really don't know who to report this error to, or what information is needed. Can anyone help? I've been a Linux user for about 5 years now, and this is the first problem that I've been clueless to solve in a very long time.

Re:This is all good, but...

[Ctrl-Z]

Did you try the linux kernel mailing list? It would seem to me that that would be the place to go.

Re:This is all good, but...

[BJH]

Try this:

http://marc.theaimsgroup.com/?l=linux-kernel&w=2&r =1&s=promise+ultra66&q=b

Other people seem to be running it OK, and there's a patch or two that might be related...

Re:This is all good, but...

[Lumpy]

if you downgrade it works perfectly then?

are you using the exact same .config file?

Did you check that you aren't turning on anything different in the new kernels?

have you tried it without all the add on hardware? remove everything but the video card and ram, does it detect and work then or still blow up?

Re:This is all good, but...

[da3dAlus]

Ok, here's more specifics. Thanks for all the replies BTW.

Promise Ultra66 PDC20262
2x Maxtor 96147U8 (hde, hdf)

I can only run 2.4.5 or earlier, all options are the same ones I've used for about a year. I rebuild kernels from the same config file, tweaking new features when neccessary. For the most part, I have the bare minimum options/modules for my system. During the boot, I get about here and then it freezes:
....
PDC20262: IDE controller on PCI bus 00 dev 70
PCI: Found IRQ 11 for device 00:0e.0
PDC20262: chipset revision 1
PDC20262: not 100% native mode: will probe irqs later
PDC20262: ROM enabled at 0xeffe0000
PDC20262: (U)DMA Burst Bit ENABLED Primary PCI Mode Secondary PCI Mode.
ide2: BM-DMA at 0xc000-0xc007, BIOS settings: hde:pio, hdf:DMA
ide3: BM-DMA at 0xc008-0xc00f, BIOS settings: hdg:pio, hdh:pio
hda: WDC WD205AA, ATA DISK drive
hdc: KENWOOD CD-ROM UCR004010 V126E, ATAPI CD/DVD-ROM drive
hdd: SAMSUNG SV2044D, ATA DISK drive

But I should be seeing these, too:
hde: Maxtor 96147U8, ATA DISK drive
hdf: Maxtor 96147H8, ATA DISK drive

Re:This is all good, but...

[mbyte]

I had exactly the same problem, about 1 year ago, appearently it isn't fixed yet. I did exchange a few emails with the IDE maintainer, but he could not help me :(

the sollution was not to put maxtor drives as slaves !?

Re:why is this news?

[solszew]

I can't respond for every slashdot reader, but I have been waiting impatiently for *every* kernel release in the 2.4 series, hoping each time that the VM issues are getting fixed. This release in particular looks like a good one, and I am thankful that slashdot (my primary source of Linux news) continues to announce kernel version updates. Those of you who are not interested can feel free to ignore these postings.

Hmm.. compile error on 2.4.14

[Anonymous Coward]

make[2]: Nothing to be done for `all_targets'.
make[2]: Leaving directory `/usr/src/linux/arch/i386/lib'
make[1]: Leaving directory `/usr/src/linux/arch/i386/lib'
ld -m elf_i386 -T /usr/src/linux/arch/i386/vmlinux.lds -e stext arch/i386/kernel/head.o arch/i386/kernel/init_task.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/char/drm/drm.o drivers/ide/idedriver.o drivers/scsi/scsidrv.o drivers/cdrom/driver.o drivers/sound/sounddrivers.o drivers/pci/driver.o drivers/video/video.o \
net/network.o \
/usr/src/linux/arch/i386/lib/lib.a /usr/src/linux/lib/lib.a /usr/src/linux/arch/i386/lib/lib.a \
--end-group \
-o vmlinux
drivers/block/block.o: In function `lo_send':
drivers/block/block.o(.text+0x8c9b): undefined reference to `deactivate_page'
drivers/block/block.o(.text+0x8cd7): undefined reference to `deactivate_page'
make: *** [vmlinux] Error 1

any quick fix for an non kernel-hacker?

Re:Hmm.. compile error on 2.4.14

[jaju]

Yes, there's an error. Remove the two instances of
deactivate_page from loop.c (linux/drivers/block/loop.c)

It's been posted on the lkml.

Preemptible kernel patch

[tarka69]

Once again it's worth mentioning the preemptible kernel patch. I've been running this for the last couple of releases, and for a developer's desktop system it give noticable results.

Case in point from yesterday: Running gnome desktop, with a kernel build going in the background, while ripping a CD, running mozilla and netscape 4.x, and running and testing a mod_perl/mysql system, all on the same machine, xmms didn't miss a beat.

No, I'm not exaggerating. This was all on a 700MHz Athlon, 256M, IDE.

Re:Preemptible kernel patch

[LSD-OBS]

Does xmms normally skip a beat without the pre-emptive patch? I swear, I've only ever had xmms skip when spinning up a powered-down hard drive, or when my loadavg goes above 15, and that on a much more modest box.

I have heard only good things about the pre-emptive (or pre-emtible, whatever the correct name is), but let's not forget that xmms uses a sound buffer several seconds long, of which it would play 64k chunks at a time, which at worst means the kernel needs to service and swap DMA buffers 3 times per second (65536/(44100*2*2) = 372ms) with basically no decoding/mixing overhead for up to several seconds at a time.

If you're getting a latency longer than 350ms, you've got bigger problems. Just my £0.02

Re:Preemptible kernel patch

[LukeyBoy]

I'm running the 2.4.13-ac7 with NO preemptive patch, doing pretty much the same stuff as you - except I'm running KDE 2 with the resource draining Mosfet Liquid theme, and Oracle running crap in the background. XMMS doesn't skip a beat here either.

Oh yeah, and it's on a P2-300 with 256 megs of RAM. I've never tried the pre. patch - it may be that the VM in the Alan Cox series is a hell of a lot better. Just my two cents.

Re:Preemptible kernel patch

[tarka69]

> Does xmms normally skip a beat without the > pre-emptive patch?

It does if you try and rip a CD at the same time :)

Admittedly, I hadn't tried under the 2.4 series without preempt patch, but under 2.2 a CD rip caused an occasional audio dropout.

Take a look at the results of the audio latency benchmark at the project page. That states that you'll start seeing dropouts at 5.8ms latency.

Re:Preemptible kernel patch

[tarka69]

This sounds CPU intensive. You need to hammer the IO subsystem to really see the effects of latency.
It's not a question of the CPU being overloaded as a case of the the scheduler denying access to the CPU at critical times.

Of course, I'm not a kernel hacker, so YMMV.

Re:Preemptible kernel patch

[LSD-OBS]

Well, for the record I have regularly ripped CD's using 2.2.xx, 2.3.xx and 2.4.xx kernels, with xmms playing, and not had a skip.

Firstly, are those tests done with DMA enabled on the IDE devices?

Secondly, where on earth did they get the 5.8ms figure from? Some math:
0.0058s * 44100 * 2 * 2 = 1024 byte audio buffer.

Sir, that is rubbish. The global audio buffer is 65536 bytes (unless somebody does funny things with ALSA or something). I have had enough experience programming sound cards on the hardware driver level to understand the ins and outs, and unless I'm missing something seriously fundamental (I'm probably lacking in linux-kernel specific knowledge), those figures are quite skewed.

Re:Preemptible kernel patch

[LSD-OBS]

With the 2.2 series kernels, you have to manually enable DMA transfers, using 'hdparm'. In 2.4, you can compile the IDE driver with the option to enable DMA automatically.

Try doing: hdparm -d1 -m16 /dev/hda (and do the same to your other drives).

Then try again. If you're not using DMA mode, the IDE IO subsystem really takes a knock when reading/writing, and will slow your whole box to a halt. I can happy encode a DVD (see drip.sourceforge.net) while compiling a kernel and whatever else, while listening to silky smooth audio. (And I'm not special or anything :)

Re:Preemptible kernel patch

[evil_one]

Here's a quote from one of Andera's (author of the new VM) posts on LKML that may interest you: indeed, the only thing that PE can really change is the mean latency but everbody only cares about worst case latency and nominal performance, possibly except realtime signal processing (not multimedia playback like listening mp3).

Re:Preemptible kernel patch

[destiney]

You guys must be doing something wrong... I play my mp3s on a P1 133 /w 32 mbs of ram... got my whole 13 gig collection on there. Never skips whatsoever. I can even play them through samba across the network, no skipping or anything.

The secret? _the_ patch: http://www.openwall.com/

Re:Preemptible kernel patch

[achurch]

No, I'm not exaggerating. This was all on a 700MHz Athlon, 256M, IDE.

Hm, so do you suppose it'll help my 75MHz Pentium, 32M laptop?

Re:Preemptible kernel patch

[quelrods]

eh i run the "normal" kernel and fail to see your problem...running enlightenment, 6 virtual screens...running several mozilla windows, consoles, xmms, vmware running openbsd and 1 vmware running win2k to watch the star wars ep2 trailer...while that happens i can compile a kernel and downsample mp3s for my player...this is a 1.2ghz athlon 512mb ram, ide as well...so um what was your point ?

Re:Preemptible kernel patch

[10Ghz]

OK, help out a newbie-loser: how does one install the patch?

Re:Preemptible kernel patch

[kervel]

i have written a small script that does something like this:
sox -blah /dev/dsp -blah | lame -V 3 > out.mp3
for realtime recording to mp3's.
without the preemptible kernel patch those mp3's are mostly fine, but with the patch, they contain oddities like skips and so.. i wonder why.

Re:Preemptible kernel patch

[wedg]

Hey, I'm doing the same thing, except on a pII 400, 128M, IDE, and without the patch!

Re:Changes/Improvements

[brer_rabbit]

The ChangeLog definitely needs some verbosity. The word to the wise is "only update if you need to", yet viewing the changelog leaves one wondering if they need the newer software.

What exactly do entries like "shrink dcache/icache more aggressively" or "random bugfix" mean? The former I'd guess has to do with data and instruction caches, but what aggressively shrinking does to them I have no clue without a more context. And the latter, "random bugfix", I hit my coworkers over the head when I see that in our CVS logs.

So is my current kernel effected? Am I missing out by not having my dcache/icache shrunk more aggressively? Or maybe those random bufixes effect me? A little more verbosity in the ChangeLog would go a long way. Having to follow the hacker's mailing list is not an option, this should be included in the release notes.

Check this one out....

[crt]

From pre6 in the changelog -
Tigran Aivazian: make the x86 microcode update driver happy about hyperthreaded P4's

I thought hyperthreaded P4's were just a rumor at this point... does this mean some of the linux developers already have dev versions that they are testing with? Whoot! A hyperthreaded version of the P4 should finally make that processor worthwhile.

Re:Check this one out....

[orz]

Neither, I think.

It's not a rumor anymore, since Intel put out a press release about it a while ago. I think the press release said that it was going to show up in the server (Xeon) chips in 2002, and in the desktops in 2003.

My presumption is that Intel provided some kind of documentation to Linux kernel developers about how to be compatible with HyperThreading, but I don't really know.

No..

[mandolin]

After all, isn't the point of an OS to run *apps*?

Nope, sorry, the chief reason for the existense of the kernel is to provide fodder for /. articles like this one. It also happens to a good compiler benchmark/regression suite, and the uptime dicksize wars it generates are fun, but really, the main reason is useless /. articles.

You're the same person that likes to rag on Jon Katz every time he posts, aren't you? I don't think the editors were listening to you then either..

Re:call it what it is

[Arandir]

Linus really, really, REALLY needs to start the 2.5.x branch.

Stable branches need to be stable. Do all of the new feature experimentation in unstable. It's gotten to the point now that "stable" has become a meaningless word in linuxland.

Next time around, let's fork off 2.7.0 at the same time 2.6.2 is released. Or maybe Linux needs to split into three branches: "flimsy" for experimentation and VM wars, "unstable" for up-to-date hardware support but no new features, and "stable" which only gets bug and security fixes.

What excitement!

[Anonymous Coward]

I'd be more excited about Kernel Sanders doing the kernel releases!

Re:What excitement!

[iamsure]

EEEEYE! With his WEEE BEADY EYES!

Thinks he's so special..

Re:What excitement!

[Phexro]

...aooOOhH, you're gonna mod my post up, oOOhH.

From the changelog

[SilentChris]

"- me: clena up page dirty handling"

English speak first language, yes?

Honestly, anyone have an idea what he's saying?

Re:From the changelog

[AnimeFreak]

No offense dude, but attacking one's english skills is sort of weak. I think he meant "I need to clean up the dirty page handling".

Re:From the changelog

[AnimeFreak]

Mon francais c'est terrible. Je parle langue preféré anglais. Parle-tu anglais?

Re:From the changelog

[WNight]

It's not a problem with his English skills, it's a problem with his typing.

Even if it wasn't, not everyone speaks or writes English fluently. Even those who grew up with it. And you know, nobody really cares, other than you, and nobody (I mean that, not even your mother) cares what you think. If they did, you'd have better things to do than flame Linus over a typo.

Re:From the changelog

[SilentChris]

Haha.

I didn't know Linus's wife posted on Slashdot... :) "Good defense, honey!"

Re:From the changelog

[mandolin]

"- me: clena up page dirty handling"

Honestly, anyone have an idea what he's saying?

Translation: "I tried to clean up dirty page handling, but also introduced a subtle bug that ocrrupts my keyboard buffer". Hope it doesn't bite me!

Re:From the changelog

[evil_one]

He's a fin. English is not his first language.

Re:From the changelog

[mr3038]

"- me: clena up page dirty handling"
English speak first language, yes?

AFAIK, "me" refers to Linus and his first language isn't English. Though I think that Linus knows English well enough to write correctly if needed. This has probably more to do with the fact that he doesn't care. In short: If you cannot decode this you don't need to know...

English isn't my first language either.

Re:From the changelog

[FrostedWheat]

arh! It's Jar-Jar Binks !!

I guess they didn't include him in the next movie .. had to do something with his spare time.

Re:why is this news?

[Anonymous Coward]

I agree How many people post that were rejected so this can be posted. I have posted article I think are more important then a kernel release. It is a MINOR Upgrade. Ohh should their be a message for every windows update. How about a message for every time the kernal writter type a line of code.

MAX wrote to tell us that in line 223 of nfs.c they added the line if (x=fork()) { What stuff will the come out next.

2.4.14 not ready yet either

[(startx)]

looks like this one isn't ready yet either. loop.c needs some sergery, or else the whole thing won't even compile, dying at

drivers/block/block.o: In function `lo_send':
drivers/block/block.o(.text+0x894f): undefined reference to `deactivate_page'
drivers/block/block.o(.text+0x8999): undefined reference to `deactivate_page'
make: *** [vmlinux] Error 1
it's allready been posted to the lkml, look for a 2.4.15-pre1 or at least a loop.c patch to come around soon.

Re:2.4.14 not ready yet either

[(startx)]

it's in loop.c, so you have to have compiled in loopback device support in order for this to effect you.

Re:2.4.14 not ready yet either

[Hoonis]

I got that on 2 boxes here as well, had to turn
it off to get a clean build

Re:2.4.14 not ready yet either

[Kynde]

Or to be exact. This loop.c linker error will only manifest if you compile it in, that is the 'make bzImage' will fail _OR_ you compile it as a modules, when the compiler will issue a mere warning of 'implicit declaration' during 'make modules', which will result in an ok kernel, but modprobe for loopback device will fail making loopback device unusable, but the kernel will work otherwise as well as without loopback device.

My advice to you and many others is if you value your time more than you love kernel hacking stick to the older kernels. If it aint broken dont fix it.

If you love kernel hacking you should read the kernel mailing lists anyway, these things get quite enough coverage there...

Re:2.4.14 not ready yet either

[loopkin]

actually, there are already two bugs. the one for loop.c is by far the worst.

as for loop.c, Linus gives this (not very detailed)answer:
http://www.uwsg.indiana.edu/hypermail/linux/kernel /0111.0/1084.html
(just delete the two lines were deactivate_page appear).
But what makes me worried about that is Andrea's mail:
http://www.uwsg.indiana.edu/hypermail/linux/kernel /0111.0/1088.html

second bug is using I2C Philips PARPORT:
http://www.uwsg.indiana.edu/hypermail/linux/kernel /0111.0/1089.html
it should impact that many people, but it's a great one.

so actually, that means Linus didn't try to compile the kernel enabling all modules before releasing it. waow :-((

Re:Life on the edge is too stressful

[EvlG]

Yes they should release.....open source methodology is release early and often.

Re:Life on the edge is too stressful

[TeknoHog]

THE PREVIOUS VERSION SHOULDN'T HAVE BEEN RELEASED UNTIL BUGS GOT IRONED OUT.

Looks like someone missed the point of open source development. How do you think the bugs got caught so quickly? Besides, with OSS you get what you pay for... :-)

Those who worry about production servers are not forced to upgrade. YOU are not forced to upgrade. Many others, however, are ready to take the chances and do their share of development by giving feedback to the actual hackers.

Remember, virtually every software has bugs, and if you don't like the idea of free bugfixes, you can just happily ignore them for whatever noble reason you have.

here's a patch for it

[(startx)]

yeah, I know replying to ones' self if lame, whatever. Here is a patch from the lkml.

diff -X /home/chris/dontdiff -Naur linux-2.4.14/drivers/block/loop.c
+linux-2.4.14-loop/drivers/block/loop.c
--- linux-2.4.14/drivers/block/loop.c Thu Oct 25 13:58:34 2001
+++ linux-2.4.14-loop/drivers/block/loop.c Mon Nov 5 17:06:08 2001
@@ -207,7 +207,6 @@
index++;
pos += size;
UnlockPage(page);
- deactivate_page(page);
page_cache_release(page);
}
return 0;
@@ -218,7 +217,6 @@
kunmap(page);
unlock:
UnlockPage(page);
- deactivate_page(page);
page_cache_release(page);
fail:
return -1;

use at your own risk, or some of that legal junk.

Useful Information

[TheMeld]

Linux has more or less zero control over data/instruction caches in the cpu. That's all done in hardware, maybe cpu microcode.

I don't know for sure, but I suspect that dcache/icache refers to dentry and inode caches, which are internal kernel structures representing files & directories, iirc.

Random bugfix really does deserve a cranial whallop, though.

Re:Useful Information

[mickwd]

Unless, of course, "random bugfix" refers to correcting the kernel's generation of random numbers.

Re:Changes/Improvements

[BJH]

It's not easy to bring the summaries down to the level of your average user; certainly, there's plenty of stuff that I don't undertand fully. But anyway, here's some additional explanations for the last two updates:

final:
- David Miller: sparc/scsi scatterlist fixes
SCSI on Sparc machines is now more stable.
- Martin Mares: PCI ids, email address update
The kernel now identifies some PCI hardware more precisely.
- David Miller: revert TCP hash optimizations that need more checking
Remove some changes to the TCP stack, as they're not quite ready for primetime
- Ivan Kokshaysky/Richard Henderson: alpha update (atomic_dec_and_lock etc)
Changes to Alpha-specific, er, stuff
- Peter Anvin: cramfs/zisofs missing pieces
Properly merge the cramfs/zisofs changes

pre8:
- Andrea: fix races in do_wp_page, free_swap_and_cache
Virtual memory handling is now more stable
- me: clena up page dirty handling
This also improves VM
- Tim Waugh: parport IRQ probing and documentation fixes
Automatic IRQ assignment for parallel ports works better
- Greg KH: USB updates
USB subsystem improved
- Michael Warfield: computone driver update
Erm... dunno. What's a computone?
- Randy Dunlap: add knowledge about some new io-apics
Improve the kernel's handling of particular chipsets' IRQ assignment
- Richard Henderson: alpha updates
Who knows?
- Trond Myklebust: make readdir xdr verify the reply packet
Since it's Trond, it's probably a RAID or VFS update...
- Paul Mackerras: PPC update
Bring Linus's kernel more up-to-date with respect to the PPC tree
- Jens Axboe: make cpqarray and cciss play nice with the request layer
Probably RAID stuff...
- Massimo Dal Zotto: SMM driver for Dell Inspiron 8000
Make the kernel work better on a particular type of laptop
- Richard Gooch: devfs symlink deadlock fix
Fix a bug in devfs's handling of symlinks
- Anton Altaparmakov: make NTFS compile on sparc
Allow people to read NTFS filesystems on Sun Sparc hardware
I'm subscribed to l-k, so if I actually bothered to read the 200-odd messages coming in each day, I could probably give a better summary... anybody know how to stretch a day to 30 hours?

Schweeeeeeet! But is there an .rpm?

[green pizza]

I can't find an RPM of 2.4.14 in any of the usual places. Anyone know where one is available? I had problems compiling 2.4.13 last time and had to wait for an RPM.
I'd like to get this installed on the workstations and maybe some of the servers tomorrow morning.
Thanks in advance!

Re:Whats an SMM driver?

[philipsblows]

System Management Mode?

Re:Link error (yes it is broken)

[Crow-]

I unpacked a fresh tree and tried to build it. Got the same error. Don't they test these fucking things? Jesus christ.

Great patch

[roystgnr]

Undefined symbol? Just delete it! If nothing breaks, then it couldn't have been that important...

Okay, I'm mostly joking, but have you tried rebooting with the new (patched) kernel and using the loopback device yet?

Re:Great patch

[(startx)]

yes, and it seems to work fine, I can read loopback filesystems just fine. Reading some posts further down in the lkml, "just delete it" is Linus's advise, because those two lines never should have been there in the first place.

Re:Great patch

[Harik]

It was left out... the .14 patch deleted deactivate_page from other places, but missed loop.c. Since it was a straight deletion everywhere else (not replaced with anything) that's the correct
answer.

--Dan

Re:Schweeeeeeet! But is there an .rpm?

[m0nkyman]

You're putting .rpm kernels on a server! are you on crack? It's (barely) acceptable on a workstation, but on a server it's just short of absolute stupidity. and asking for a link to one on slashdot is past stupidity and into luser-dom.
Hand-compiling a kernel not only enhances performance, it reduces the number of potential problems in a kernel.
Asking for a link to an .rpm is beyond dumb....
Sure here's my link : http://www.rootsrus.net/kernel-2.4.14.rpm.

New ./ feature idea

[Atilla]

Ok, how about this -

maybe major kernel release posts should go in their own category, so whiners could set their user prefs to not display them.

Re:New ./ feature idea

[scumdamn]

Maybe whiners should go in their own category so we could set out user prefs to not display their whining.

Re:New ./ feature idea

[Anonymous Coward]

Maybe major kernel release posts should be posted to the front page, and MINOR RELEASES like 2.4.14 shouldn't be posted at all.

Good gravy!

[CAIMLAS]

I can't hardly get a kernel compiled before the next version is released! You'd think they'd wait a little longer to go to the next version. Seems to me like the 2.4 strain got up to .14 much, much faster than 2.2 did... or is it just me?

Re:Good gravy!

[Mudhiker]

You compiling on a 386 it takes you that long? ;-)

*

Re:Good gravy!

[zmooc]

Actually it's not that must faster. I don't have the exact dates (looked at the dates at ftp.nl.kernel.org), but 2.2 took about 14 months to get to .14 and 2.4 took about 11 months. And then there was this kernel that was withdrawn so actually one can say we're at .13 now:) And besides: you don't have to compile every new kernel. If it works fine, don't touch it.

Re:Good gravy!

[budgenator]

If they don't post a lot of minor stuff the activity percentile drops like a rock on sourceforge. I don't agree with this, I feel that a CVS ci should have a reasonable chance at running on a typical machine. I'm not saying that every command combination needs to be checked, but it should reasonably work.

Everday people get realy turned off if they co the latest and greatest from CVS and the program doesn't even compile, or segfaults on first try. It makes OSs developers look real unprofessional.

It used to be that a even minor number on the kernal was considered stable, and at least was in beta status. I guess that that is out-the-window now. I can't figure out why both Linus and AC are working on 2.4, I thought Linus was supossed to jump from 2.3 to 2.5, when 2.3 was stable enough to become 2.4, then 2.4 versions increments were little details that turned up during more widespread usage, and features from 2.5 that became stable enough.

Re:Good gravy!

[CAIMLAS]

I've been waiting for a kernel that has had all-around good reports after a couple weeks of being in the wild. 2.4.14 did this time, but go figure, I can't get it to compile - it has problems with the block device section during compilation. :-/

2600, those Microsoft hackers

[cant_get_a_good_nick]

5.1.2600

So Microsoft is now in league with those phreaking 2600 guys? I always knew....

2600 is for 2600 kHz, the carrier wave for the sounds that payphones used to signal the home office that someone dropped in a shekel or the call. All the kool kids had electronics that could make that tone and bingo, free calls. Was called phreaking at one time. Now the kids on to bigger and better things, like finding holes in telnetd.

Re:2600, those Microsoft hackers

[wik]

2600Hz, maybe. I doubt the crappy verizon payphones could pick up 2.6MHz "tones". Heck, they barely pick up anything except my pocket change.

Re:2600, those Microsoft hackers

[Dahan]

So Microsoft is now in league with those phreaking 2600 guys? I always knew....

Yup, XP was codenamed "Windows Whistler". Captain Crunch, anyone? :)

Re:2600, those Microsoft hackers

[42forty-two42]

2.6KHz, you mean...

Looking for these drivers?

[PM4RK5]

Actually, I've just compiled my kernel, and I recall seeing an option (CONFIG_BLK_DEV_ATARAID_PDC for the RAID device, and CONFIG_BLK_DEV_PDC202XX for the Promise 202xx chipset).

I think these are new as of 2.4.14, but maybe not. Try compiling your kernel with those enabled and see what happens.

chrisd?

[KidSock]

Who's chrisd? Is that short or Christine? Finally, a female editor. But I still think this is discriminatory because a Kernel release is trivial. Let's see what Christie can really do! Let her post something real goddamnit!

Re:chrisd?

[Electrum]

Or it could be short for Christopher...

Re:chrisd?

[quannump]

or maybe it's the chris daemon.

Re:chrisd?

[The Madpostal Worker]

chrisd == Chris Dibona

Re:chrisd?

[Denor]

It's the daemon that listens for connections on port 'chris'. Kinda like httpd.

A kernel hacker's first language...

[horza]

... is surely C, with English coming second?
Anyway, if you can't decode a single transposition
in an English sentence you will be pretty useless
at debugging.

Phillip.

Re:A kernel hacker's first language...

[ninewands]

Coder's type/spell well because a compiler is totally anal about spelling. If you want to see how well a coder REALLY types/spells, look at the comments ans CVS log, not the code.

Re:call it what it is

[Enigma2175]

Or maybe Linux needs to split into three branches: "flimsy" for experimentation and VM wars, "unstable" for up-to-date hardware support but no new features, and "stable" which only gets bug and security fixes.

We already have the "stable" branch, it's version number is 2.2.x.

Re:why is this news?

[Anonymous Coward]

It's news because the editors and owners of Slashdot have decided it's newsworthy. You're certainly entitled to disagree with its newsworthiness, but how hard is it to understand that the editors wanted to post it because they thought it was important?

Re:Schweeeeeeet! But is there an .rpm?

[fodi]

Why such a mentality? It's the same mentality that labels IT professionals as geeks, without any personality !!

Do you really believe that your way is the only way to do something? Sure, he might not be following the most secure/best method for a kernel update, but if that's how he wants to go about it ...it's his choice!

How about offering some advise, instead of Trolling him. That's probably one reason Linux hasn't caught onto the mainstram market; Everytime someone asks a simple question, they get bagged. You're a fool. Fuck off!

Re:Wow -- 3rd Kernel release in a month or somethi

[bleechack]

"Release early and release often"

See "The Cathedral and the Bazaar", esr

http://www.tuxedo.org/~esr/writings/cathedral-baza ar/cathedral-bazaar/x147.html

Re:Schweeeeeeet! But is there an .rpm?

[gotan]

There are rpms for the Kernel-source out there (not to be confused with source rpms) which will put a linux-2.4.X-distZ directory or somesuch in a convenient place and let you compile (although i ran out of luck with some mandrake ones since it apparently mangles symbol versioning). That completely invalidates your argument about the necessity of hand-compiling a kernel, since you get to hand-compile the kernel with those. Also most distributors like to include their flavor of kernel-patches, for example supermount with mandrake, or ext3 with redhat (you can get that from the ac-patches too). Especially stuff relying on kernel-patches that didn't (yet) make it into the linus kernel (supermount, lm_sensors) will only work if you get the kernel-rpm, or apply the patches yourself (or compile the modules from a convenient package) and then you probably have to tell rpm to ignore some dependencies from kernelversions. Sometimes the rpms contain quite vital stuff, For example Debian (where it's .deb, and not .rpm) patching a hole in 2.2.xx (19 i think it was).

So there's some good reason for someone wanting to run a kernel compiled from a kernel-source-rpm, and if, for example, you use gcc3.0 as opposed to gcc2.96/gcc2.95, and also don't have those handy, you might even be better off using a precompiled version (most distris put out a server- and a desktop version) than compiling your own. The main reason why the original statement is funny, is asking for an rpm when the kernel is just out. The bashing was not appropriate, i think.

Re:Schweeeeeeet! But is there an .rpm?

[Jburkholder]

make-kpkg --revision=customSlashdot.1.0 kernel-image

Re:Schweeeeeeet! But is there an .rpm?

[wholesomegrits]

You Were Trolled. See, the post you replied to was a Troll of a Troll. A Meta Troll if you will.

You were doubly trolled, as getting trolled by a meta troll is worth the number of trolls following the parent. Your Troll Factor is 2.

Re:Schweeeeeeet! But is there an .rpm?

[G Money]

I was having the same problem with mandrake kernel source rpms until someone told me to disable "Set version on all modules symbols." After that it worked like a charm, although I have noticed significantly quicker bootup times when using a stock kernel (guess all those extra patches they include slow things down some...)

Re:why is this news?

[scumdamn]

Well put. We do need to keep in mind the things we have in common (such as Linux).
I usually find out about the latest Mozilla releases here that I wouldn't normally go looking for, but since it's so easy to just link to it from my default page I usually go grab it and jack with it for a while.
My point being that the really important releases should be reported on because these projects bring us together.

Re:Changes/Improvements

[Saint Aardvark]

Excellent...thanks so much, this is exactly what I was looking for. And yeah, how do those kernel guys do it?

Re:Schweeeeeeet! But is there an .rpm?

[Paul Jakma]

ehmmm... try this:

untar the kernel source
make config
make rpm

rpm -i /usr/src/redhat/RPMS/i386/kernel\t\t\n

aye... stupid.

Re:Changes/Improvements

[psamuels]

- Michael Warfield: computone driver update
Erm... dunno. What's a computone?

Multi-serial-port thing, I think. Useful for modem banks if you're a pre-broadband ISP.

- Trond Myklebust: make readdir xdr verify the reply packet
Since it's Trond, it's probably a RAID or VFS update...

No, Trond does NFS. xdr (External Data Representation) is the low-level NFS wire protocol.

- Jens Axboe: make cpqarray and cciss play nice with the request layer
Probably RAID stuff...

Yeah, two drivers for hardware RAID boards. We all know the saying: if you don't know whether or not you have one, you don't....

- Massimo Dal Zotto: SMM driver for Dell Inspiron 8000
Make the kernel work better on a particular type of laptop

Specifically, System Management Mode lets you control stuff like video output switching, volume control, etc (you know, those hot-key functions on most laptops). I think.

Re:call it what it is

[psamuels]

Linus really, really, REALLY needs to start the 2.5.x branch.

Everyone is saying this, but there are issues involved. They were saying the same thing before 2.2 was even released - "give us a 2.3 so we can play". Linus doesn't work that way, because he believes (unlike, say, the Debian folks) that the resulting "brain drain" due to people hacking on the next unstable release unacceptibly hinders the bug-fixing effort in the current stable.

Sure, it's at least 80% psychological, but it does work, for all that. If you read his posts, you'll see that Linus has never been one to shy away from using psychology!

The unfortunate fact is, until recently, 2.4 did not have a usable VM. We can argue all day about whether Linus was too hasty in integrating Andrea's VM (I think he was, which is why I've switched to -ac kernels for now) but it is hard to argue that 2.5 should have forked before the 2.4 VM shaped up ... at least given Linus's brain drain philosophy.

So, while the 2.4-only series has almost certainly gone on too long, there are valid reasons for it. Perhaps if you had contributed all your VM fixes six months ago, we could have had a 2.5 by now. (:

Re:why is this news?

[dattaway]

Its a celebration of our operating system's core. Not much, really. Just a big party.

Compile and grab a cold beer. Then join us.

Re:call it what it is

[Jagasian]

It's gotten to the point now that "stable" has become a meaningless word in linuxland.

If you truely want stability in "linuxland", then you should use the current stable version of Debian Linux, the community based Linux Distro. It is very upsetting that the majority of Linux users haven't tried Debian Linux, considering how its one of, if not the best Linux Distro, and it is developed by the Linux community, not some corporation.

Wondering why your shiny new P4 system won't boot?

[alexburke]

pre4:
- Mikael Pettersson: fix P4 boot with APIC enabled

Hey guys. Cut him a break.

[RealityThreek]

I'm pretty sure he was just kidding. And anyway, it was pretty amusing. I'd bet even Linus would chuckle. ;)

does it bother anyone else

[kraada]

that new kernel releases are being dropped like flies? not that i don't like having the kernel as up to date as possible, but it seems that the number of releases this fall has been astounding . . . i'd rather see the numbers stay low but each new kernel be a significant step . . .
just my .02 . . .

Re:[OT]Hardware requirements [was Re:Version nicen

[RealityThreek]

I'd bet the real problem isn't that you are installing RH7.2, but that you are installing a new window manager. Try installing one of the older WMs. A Linux command line runs great, even on a 486. ;)

Re:[OT]Hardware requirements [was Re:Version nicen

[Jon Abbott]

Ack, I wouldn't try running the latest Linux distros on older hardware, as the latest distros are obviously "optimized" for newer hardware. Red Hat 6 should run fine, although there are a number of different paths you could take. I've been using Debian, and it installs on everything from a 386 to an Itanium, and runs smoothly. It's a bit harder to install, but then installing and upgrading software on it is a piece of cake later on, using apt-get.

Another route you could take would be to try FreeBSD, as even the newest versions still run very nicely on old hardware. I recently installed FreeBSD 4.4 on a P75 with 16 MB of RAM, and it has been chugging along, happily playing MP3s via NFS in our living room. I can even be compiling software while playing MP3s, and it plays flawlessly.

The issue you mentioned about Tuxracer going so slow is because the S3 Trio64+ doesn't have OpenGL hardware acceleration (to my knowledge, at least), which means that Tuxracer is going to do all of the OpenGL via software. This equates to incredibly low frame rates, until a better video card is used. Your best bet, if you're really looking into running OpenGL applications, would be 1) buy a better video card, 2) install XFree86 4.1.0 (see xfree86.org), and 3) enjoy.

I'm not sure about the SoundIII drivers you are asking about -- search google for "SoundIII and Linux" and see what comes up.

Re:Changes/Improvements

[jmd!]

Yep. Linus' ChangeLog's are incredibly awful. Useless to non-kernel developers, and more often then not, useless to kernel developers as well.

It's not uncommon for him to mention things that didn't actually get in, and what he does mention are typically only half the changes.

It's too bad he's hell-bent on this current development model he has going. Might have worked in '91, but it's quite apperant he's in over his head at this point, managing it all through email, and just deleting his mail spool when it gets too full (seriously! You're appearanly supposed to just keep sending in patches if it doesn't make it into his release).

Re:call it what it is

[jsse]

Why stop there. Why not fork off odd and even kernel no. . develop them seperately in parallel, and release only one of the branch at a time?

Wait, put down your troll tag, take it easy... That's what Intel did, in the past, in the x86 processor development.

Unslashdottable

[cafelatte]

I'm having serious deja vu here. I can vaguely recall an article posted previously on slashdot at sometime in the past before now, that mentioned something about a new linux kernel release. It seems that the kernel.org server(s) are unslashdottable. I'm downloading at 27kb/s. Why would they need mirrors? The server specs must be pretty good. I wonder what operating system they'd be using.

Re:Unslashdottable

[sydneyfong]

isn't mirrors supposed to be used to reduce the load of the network?

i mean, when the kernel.org server is located on the other side of the planet of the person downloading it, then perhaps there's a good reason to download it off a mirror physically located near you. It would help reduce the cost of the lines, i think
Besides, i want 800kb/s! ;-)
(my line is fast enough to do that, but for most US sites i don't usually go above 100kb/s)

Re:Unslashdottable

[Paul Komarek]

I've seen kernel.org fill up. I think it did sometime late last week, in fact. But they do have a 100Mbit/sec connection, which is pretty darn fast. 27kb/sec is kind of slow for them; they usually max out my DSL link at around 650kbit/s or 700kbit/s (== about 84kbyte/sec).

-Paul Komarek

Re:Unslashdottable

[Andreas(R)]

"I'm downloading at 27kb/s. "

I'm downloading at 70 Kb/s... Sorry, I just had to brag of my connection too : )

Re:Unslashdottable

[PastaAnta]

Kernel.org is running a VArServer 700 with a total of 400 GB of storage on a 100 Mbit/s line.

The OS is Linux with an Apache Webserver.

This is all according to http://www.kernel.org/powered.html. So the server is NOT running Windows and IIS as some (stupid) AC posted here!

No information about kernel version though....

Re:why is this news?

[Dahan]

I want to see important Linux topics. Not a stream of minor kernel revisions.

Re:call it what it is

[dorward]

How about unstable, testing, and stable? And give them code names taken from characters from a popular film.

(Upgrading became so much easier when I moved over to Debian Unstable)

Put all new releases in their own category

[Jeppe Salvesen]

If a kernel release is uninteresting, why should a Gnome or KDE or Mozilla release be any more interesting?

Re:Put all new releases in their own category

[sydneyfong]

it's not really uninteresting, but there are so many kernel releases in a short time that some people get bored seeing "Kernel 2.4.xx is out!" stuff all the time.

Re:Odd/Even Releases

[hughk]

Maybe a development fork would be too much at the moment. How about embedding functionality into even releases and bug-fixes into odd.

This has worked well for manufacturers and could certainly help the kernel's meander towards maturity here. In particular it would mean that we always use the odd-numbered release where we want stability.

Re:why is this news?

[aliebrah]

I like many who read Slashdot am a linux enthusiast. I appriciate having a place where my OS of choice is reported on with the zeal and timeliness that Microsoft gets in the main stream press.

I like using Windows as my primary OS, and yet I read Slashdot on a daily basis, and consider myself very knowledgable about computers and operating systems. I like you would appreciate it if Slashdot had timely reporting about Windows. Most Slashdot editors don't know the first thing about Windows and proudly claim never to touch it. I have yet to understand how people can ever claim to be in touch with the computing world without being in touch with Windows (which is different from liking it). When Windows makes news it hardly ever makes Slashdot (unless of course it's bad news). God forbid that a Windows Service Pack would ever make news on Slashdot. But a Linux kernel released ... oh that's so different.

Re:why is this news?

[Erik Hensema]

It's a matter of personal interest. The editors don't care about Windows, hence they know little about it. Offcourse they're using it from time to time -- almost everybody does, but that doesn't say they know anything about it.

That is, if the editors are anything like me.

Token Ring Cards & DHCP kaputt since 2.4.0?

[buglord]

I held a Linux class last weekend and was in for some surprises. It turned out to be really 'hands on', because the Token Ring cards in the machines refused to work with DHCP.

Of course the kernel ignored my cajoling and the increasingly nervous class... But wait, that wasn't a problem with 2.2er kernels!!
Anybody know since when this was an issue? I'm not exactly anxious about embarassing myself again in front of Linux Newbies.

Re:Token Ring Cards & DHCP kaputt since 2.4.0?

[Dwaine Garden]

I have never had a problem with Token Ring cards.
I use it at work.

I use the TR PCI card.

Re:Token Ring Cards & DHCP kaputt since 2.4.0?

[buglord]

which kernel version? *that* would be interesting!

Re:call it what it is

[Anonymous Coward]

Or maybe Linux needs to split into three branches: "flimsy" for experimentation and VM wars, "unstable" for up-to-date hardware support but no new features, and "stable" which only gets bug and security fixes.

BSD has been doing this for years. Are the Linux developers too full of themselves to consider such a development process a viable option to balancing development/stability?

Re:why is this news?

[flex941]

Good comment. My thoughts are with you :)

Re:why is this news?

[GdoL]

Yes, it's a very fundamental thing as a community to have on Slashdot a discussion of some of the most important developments of the LINUX/GNU world. Like the Kernel but also about Mozilla, X, Xmms, Emacs, GCC.

It's the first place I look to see the news and I really want continue to see that here.

Re:call it what it is

[Ben-Oh]

> Or maybe Linux needs to split into three branches:
> "flimsy" for experimentation and VM wars,
That will be 2.5 (Though I agree it could have been started a little earlier)

> "unstable" for up-to-date hardware support but no new features,
That is 2.4

> and "stable" which only gets bug and security fixes.
And that is 2.2

In short: we have what you describe (most of the time anyway)

Re:why is this news?

[jonathan_ingram]

Get MPlayer... it plays DivX movies very well (on x86 machines).

obligatory old-timer post

[Phexro]

"I still don't know why 2.2.x had so few releases. only 17 before 2.4.0..."

why, back in my day, we got nine stable kernel releases, and that was it! oh, sure, some times we'd get as much as thirteen - though those cursed releases from 1.2.10 to 1.2.13 were for the ignorant masses! anyone who knows their stuff never uses a kernel with a patchlevel of 10 or above.

happily posting with 2.4.9, and you should too!

Erm... I'm having a little trouble

[TheMMaster]

**DISCLAIMER** I've searched google for this but couldn't find it.
Can someone tell me how to use a patch cut'n'pasted from a website? when I use the normal command patch -p1 it gives me some errors.
thanx in advance

Re:Erm... I'm having a little trouble

[GC]

Just go into the file and comment out the deactivate_page() calls.

That should fix it for you.

That patch was a very long winded way of saying "remove the lines".

I helps if you can read patches.

Re:Erm... I'm having a little trouble

[crsm]

Beware of the evils of HTML formatting !

Only belive in the one and true Linux Kernel Mailing List and cut'n paste from here into a file.

Then this should do it:

cd linux patch -p1 < /path/to/patch-file

Re:Changes/Improvements

[Kynde]

What exactly do entries like ... "random bugfix" mean?
And the latter, "random bugfix", I hit my coworkers over the head when I see that in our CVS logs.

I wouldn't mention this otherwise, if it weren't for the growing worldwide interest in cryptography and related issues. Many developers for linux platforms aren't fully aware how fantastic devices the /dev/random and /dev/urandom actually are.

As an example, I work in a company that makes slot machines and evidently the random number generator is a great concern to us. Having read the kernel sources for the random number generator device I must say, other than few measurement issues (i.e. clumsy word/byte/bit handlings), it is rather fantastic service. After months of looking into countless generations of earlier random number generators I'm now only checking wether I need to patch even a bit or not.

Those of you in need of good random number services I suggest you lay a glance on /dev/random and /dev/urandom.

Re:Changes/Improvements

[Pieter-Bas]

I'm subscribed to l-k, so if I actually bothered to read the 200-odd messages coming in each day, I could probably give a better summary... anybody know how to stretch a day to 30 hours?

A good night's sleep seems to pass in a minute. So for just one minute of good sleep, you currently invest six or more hours. Therefore the problem can be solved by sleeping just this one minute. The hours you win by doing this can be added to your day, effectively yielding a 30-hour day.

I especially like...

[caryw]

pre6:
- me: remember to bump the version number ;)

Re:I especially like...

[42forty-two42]

*Slaps his forehead* Of course! How could I have forgotten...

The 2.4 releases are the normal kernel now

[hta]

says the Linux Counter systemstats.
Granted, this has about a month's lead time, and is hardly "representative" (more enthusiasts than real users), but it shows 2.4 at 58.6% of the 1142 machines registered. 2.4.14 is 0.4%, 2.4.12 is 10.8%, beating everything but 2.2.19 (at 14.6%).
btw, the 5 registered 2.4.14 kernels are all prereleases - pre5aa1 being the most popular one.
This is a fun view - but that's no reason not to get counted!

Harald, counter, compulsive.

Re:[OT]Hardware requirements [was Re:Version nicen

[-douggy]

SUSE 7.2 on a 233 laptop with 32 MB ram. Runs fine with Windowmaker as the main GUI.

Slakware 8 did install on the p133 but it got changed to smoothwall.

Thanx

[TheMMaster]

Thanx guys,

It compiles!!

I'll throw in my own plug, then!

[jd]

For those wanting to try the pre-emptible kernel patch, the HP scheduler-plugin, compressed memory hardware, STP, XFS, JFS, Linux on an old VMS box, any one of a number of VME crates, serial-based network controllers, or the various latency clean-ups, then you could always try the FOLK kernel seris. FOLK 2.3.0 is stable (gasp!) and provides more today than the first fifty 2.5.x kernels are likely to.

(And by the time those come out, FOLK will be comparable to Linux 2.7.0 in terms of features & performance.)

Re:I'll throw in my own plug, then!

[Some Dumbass...]

Yes, but unfortunately FOLK 2.3.0 appears to be a patch to kernel 2.4.10. So if you want a stable VM, you may have to look elsewhere (for now).

Re:I'll throw in my own plug, then!

[jd]

It's based off Alan Cox' -ac5, so the VM should be not too bad. However, if you want to be a bit more sure, I'm in the process of composing a new FOLK patch, based on 2.4.13-ac8. (Version on release may be different, as incremental patches become available.)

Re:Changes/Improvements

[BJH]

No, Trond does NFS. xdr (External Data Representation) is the low-level NFS wire protocol.


Yeah, I realized my error just after posting it... thanks for the correction.

Alan Cox?

[jhittner]

Does anyone know what this entails?
Alan Cox: more merging

Is there another change log for his additions, or are they so minor that they can be overlooked?

hmmm... I sort of agree

[Ender Ryan]

While I do enjoy all the kernel updates on slashdot, I think slashdot should also write about some of the great userspace apps as well.

All those apps you listed are most definately news for nerds, and they really are stuff that matters, specifically for Linux.

Anyone who hasn't tried out those apps, give em a whirl! Specifically, for me, mplayer is absolutely great. It is hands down the best video player for linux.

So, while freshmeat lists everything, why not list the important things on slashdot? I just think that would be so useful.

Re:why is this news?

[NullSynapse]

Offtopic, but you might try xine.

oh, quit whining :)

[hawk]

We get a new release in about two months, and *we're* past 4. See, these Linux newbies haven't even caught up with that Gates character's 3.1.

While the strange affirmative actions really shouldn't happen, it's really more important that FreeBSD and our little brother Linux form a united front againat the *real* enemy we have in common: NetBSD.

:)

hawk

Re:why is this news?

[pgilman]

there are already _so_ many other sources of windows information. if i need to learn about windows service packs i go to microsoft.com.

if slashdot reported on everything, it wouldn't be good at anything; without focus, it would be useless.&nbsp i like it the way it is, thanks.

p3t3

101 reasons not to patch and unpatch repeatedly

[Rares Marian]

1. You'd get about the same result from abusing Babelfish.
2. ditto
3. ditto
..
101. ditto

Re:101 reasons not to patch and unpatch repeatedly

[KidSock]

On the contrary they are talking about doing precisely this. It would not be possible under the current build process but kbuild for 2.5 is redesigned in such a way that it supposedly does this anyway.

Changelog Sucks!

[dbretton]

Ok, I'll say it.
I know I am not alone here. The changelog has only the slightest modicum of vutility to the average linux user/sys admin/etc.
I don't follow the kernel list intimately. I don't want to. All I want is a description of updates that distinguishes the new kernel from the old one.
"more merging" is complete shit (pardon the language, but it is the most fitting word). There isn't a single person who can read a statement like that and understand what has been changed/added/fixed in the kernel.

Since I cannot sit here and whine about things without offering something better, how about this:

Every developer has a template document that lists the important things their "patch" does. There could be 4 categories:
New Driver
Bug Fix
New Feature
Caveats/Known Bugs

Each submitter would send this document to the "maintainer" of the various portion of the kernel, who would then merge the documents. These merged documents would go to Alan/Linux/etc., who would merge the list again.
So what if it's huge. At least it would be informative. I'd rather see a lengthy list of changes than hardly anything at all.
-Given the choice, would you prefer to hear your local news station go on about the details of the news in the day, or simply say, "And today, more stuff happened"?

Wouldn't it be nice to skim through the changelog and see if firewire support for your HD has been added? (for example)

I wish something like this would exist...it wouldn't even be difficult to do, either!

-D

Re:Changelog Sucks!

[Cro Magnon]

Linus Changelog: 2.4.15-pre1 fixed stuff.
AC Changelog: 2.4.14-ac1 security fix. See www.anywhere_but_us.com for details.

Re:Changelog Sucks!

[sunset]

I agree.

Me Too! Is She Hot?

[JLucien]

Just thought I'd ask...

Re:Schweeeeeeet! But is there an .rpm?

[JCCyC]

Especially stuff relying on kernel-patches that didn't (yet) make it into the linus kernel (supermount, lm_sensors) will only work if you get the kernel-rpm, or apply the patches yourself

ext3 is not in the main strain either. I'm "make menuconfig"-ing 2.2.14 right now, with experimental options enabled, and it doesn't show up. Guess RH7.2 users like me will have to wait for -ac1.

Re:why is this news?

[festers]

Thanks, this needs to be said every once and a while. I enjoy the kernel posts and the subsequent discussion from people who know a lot more about the Linux kernel than I do.

Chris DiBona

[Sits]

Looks like we'll have to wait a little longer for a female /. editor.
Chris DiBona was one of the editors on open sources. So unless someone made a mistake with his gender there...

Re:Link error (yes it is broken)

[den_erpel]

I got the same thing this morning, and was actually checking /. to see if I was the only one with this problem. LOL.
I agree that *at least the actual source code* should compile from scratch when giving a new release (being a programmer myself).
I at least hope the 2.4.11 and 2.4.14 probs are not a preview of things to come (and we have these kinds of problems every 2 versions).
The 2.5 branch should really be started: these kinds of problems belong in the development branch and not in the 'stable' tree.

this weeks kernel is....

[nixxy]

I sit just me or are they releasing a new kernel each week, could this be cause for the past few revisions there has not been enough testing, I highly doubt that a week's worth of testing is sufficient for a new core to an operating system.

Not even m$ release "hotfixes" this fast its starting to get (well already is) very annoying.

Linux

[Man Eating Duck]

Linux is supposed to have been made so simple.. I am a newbie in Linux, but has lots'a experience since the C64 days... Still, while installing the drivers for my G400 Dualhead mgapdesk fucked up the XF86Config-4 file 4 (four) times, had to reinstall every time...