Linux Kernel 2.2.10ac11 Released

Full details on this page at Linux Today, where Alan Cox says, "A lot of updates here. I've not chased down the lockd bug reports yet, nor the umount nfsd bug that is trapped by the SLAB debugging. This mops up most of the stuff while I've been occupied elsewhere. Chances are given the number of updates there will be a few glitches in it. Have fun testing."

Re:version number

[Anonymous Coward]

Actually, Alan is running the 2.2 tree now. Linus is busy with 2.3.

Re:Alan Cox

[Anonymous Coward]

Not anymore. Linus has said that whatever Alan says for the 2.2 kernels goes (not that Linus wouldn't point out if something is a bad idea, though ;-). So, basically, think of the ac stuff as being like Linus' pre-patches.

Re:Strange..

[stevied]

The big thing that might have been worth pointing out though, is that the file corruptions some people were seeing in 2.2.9, 2.2.10 was allegedly fixed in -ac10 or -ac11. I don't know the details ATM, I lost track of that thread on l-k a long time ago :(

edit your prefs you moron.

[MacJedi]

I shouldn't be biting for such an obvious troll, but if you don't like roblimo you dont have to ever read a post by him again.

It's called log in and edit your prefs.

Re:Strange..

[Roblimo]

It was a newbie mistake on my part. Sorry. I let my personal interests creep in. What caught my eye was that radio modem support is included in the most recent AC patches, and my biggest current hardware desire, er, I mean need, is a Linux notebook equipped with a radio modem. ;-)

Re:DON'T RUN 2.2.8!!

[Anders]

Also, only 2.2.10 has the fix for that latest Denial-of-Service bug in the networking code.

Given how few people has problems with 2.2.10, I would recommend that version - even though it might fail on a few, it has at least as good a chance of success as any previous version.
--

Re:it isn't a mistake.

[Trepidity]

then go read freshmeat

Re:DON'T RUN 2.2.8!!

[Trepidity]

That's why I consider the current Linux release cycle to be a failure. You should not have to have a list of which kernels are stable or not, since 2.2.x is supposedly the stable tree. ALL kernels released in 2.2.x should be 99.9% stable. In other words, more ac prepatches would be nice, and delay the "official" kernel releases until you really have something worthy of the "stable" label.

When you have a stable tree with severe bugs, that doesn't make people confident in your product. I personally wouldn't recommend anybody use anything higher than 2.0.37 for any important applications, since the 2.2.x line is simply not stable enough.

Re:DON'T RUN 2.2.8!!

[Trepidity]

Sure, I understand that there's some limit to the testing that can be done, but more than is currently done is possible. With companies like RedHat, Caldera, and VA having quite a bit of money, they could at least try out 10 or 15 different hardware configurations, not just what they happen to have on their personal boxes at the moment. Also, no new features should be added to 2.2.x. They should go into 2.3.x.

Re:DON'T RUN 2.2.8!!

[Trepidity]

Then perhaps Linux and Alan Cox need to take some lessons from Red Hat. I'd rather run a stable 2.2.5 that has some of the 2.2.6+ features than run a 2.2.10 (with very few additional features) that's a lot less stable.

DON'T RUN 2.2.8!!

[Unknwn]

NOOOOOOOOO!!!!! DO NOT RUN WITH 2.2.8. Sorry to shout there, but 2.2.8 has _severe_ memory issues. So in that case, go back to 2.2.7 :) But, only sporadic people are having fs corruption in 2.2.9/10 so I'd recommend using 2.2.10 (or maybe this patch, haven't looked at it yet).

Re:DON'T RUN 2.2.8!!

[Unknwn]

The only problem is that you don't get the range of testing on a prepatch that you do with an actual kernel release. If the main kernel developers have a problem with a kernel, it generally doesn't get released. But, the sheer variety of PC hardware makes it such that, as Murphy's Law puts it, anything that can go wrong will go wrong. So, unless you have every different possible combination of hardware under realistic load that you're willing to offer for testing, it's not going to get much better.

Anyway, just venting. This gets said a lot by people who just don't understand what the problems boil down to (not that I always do either... but anyways)

Lets get this one tested...

[Ami Ganguli]

Yeah, Slashdot probably shouldn't make a habit out of announcing 'ac' releases, but it would be good to pound on this one a little. Given all the glitches with the 2.2.x tree, it's important that we get a rock solid 2.2.11. The best way to make that happen is to try 2.2.10ac11 and report any bugs.

2.2.10-ac* and SMP?

[demon]

Has anyone else noticed that the 2.2.10-ac series kernels don't seem to like being built as SMP? I've tried 2.2.10-ac3, ac10 and ac11 on 2 different SMP systems (built as SMP) and it dies quickly (doesn't even finish the kernel-boot phase) with a NULL pointer dereference. Anyone else seen this?

Re:2.2.10-ac* and SMP?

[demon]

I'll try to run the stack dump through ksymoops and see what it has to say. (I've been playing with getting Linux up on a new Proliant 3000 at work, so I'll mess with that tomorrow.)

I'm guessing you're Alan, and you just forgot to log in...

Re:2.2.10-ac* and SMP?

[demon]

*hiding his face in shame* Ooops... I guess enabling the pre-alpha LVM support was a bad idea... 2.2.10-ac11 works fine now... :)

more freelance moderation

[unitron]

In a previous post I said
"in a post moderated down for other reasons bugg asked "Anyone know about the Intel Etherexpress 10mbit cards? I guess this is off topic, but i just got some and i wanna know if any of the latest kernels include it.
-bugg" "
As I said, the original post was moderated down for reasons unconnected to the question about NIC cards, but I felt that in a discussion of a new kernal a question about what hardware it supports was perfectly valid, so I reposted it under my name (minus the first post garbage that got it knocked down in the first place) in order to "promote" it so that someone who knew the answer might see it. As someone else observed, this thread seems to have attracted a record number of "off-topic" (to put it politely) posts, most of which had no content worth sifting out. Why someone felt it worthwhile to "burn-up" 20% of their moderation points to knock it back down when there were so many other more worthy targets I'm sure I don't know.

Re:version number

[kronos]

If you see a linux kernel numbered with an ac__ at the end, like "2.2.10ac11", it means that it's based on 2.2.10, patched with the eleventh patch written for that kernel by Alan Cox (ac). Since it's not an official "Linus-blessed" version of linux, he increments his own counter after the "ac". Usually, most or all of alan's patches get included in the following version of linux (in this case, that'd be 2.2.11).

For freedom,
kronos.

Re:Interesting ...

[Alex Zepeda]

Here's a few that might interest you:

From: Bill Paul
To: [e-mail], [e-mail]
...
Log:
This commit adds driver support for the SysKonnect SK-984x series
gigabit ethernet adapters. This includes two single port cards
(single mode and multimode fiber) and two dual port cards (also single
mode and multimode fiber). SysKonnect is currently the only
vendor with a dual port gigabit ethernet NIC.
From: Bill Paul
To: [e-mail], [e-mail]
...
Log:
Make the Winbond ethernet driver work on FreeBSD/alpha. Also added
bridging support while I was in the area.
From: Kirk McKusick
To: [e-mail], [e-mail]
...
Buffer flushing has been reorganized. Previously buffers were flushed
in the context of whatever process hit the conditions forcing buffer
flushing to occur. This resulted in processes blocking on conditions
unrelated to what they were doing. This also resulted in inappropriate
VFS stacking chains due to multiple processes getting stuck trying to
flush dirty buffers or due to a single process getting into a situation
where it might attempt to flush buffers recursively - a situation that
was only partially fixed in prior commits.
...
A small race condition was fixed in getpbuf() in vm/vm_pager.c.

Submitted by: Matthew Dillon
From: Kirk McKusick
To: [e-mail], [e-mail]
...
Log:
These changes appear to give us benefits with both small (32MB) and
large (1G) memory machine configurations. I was able to run 'dbench 32'
on a 32MB system without bring the machine to a grinding halt.
... Submitted by: Matthew Dillon <[e-mail]>

Re:version number

[smileyy]

Anonymous Coward, obviously.

I just want to say

[IntlHarvester]

Thank god for the Compaq Smart2 array driver support. I've got a few "outdated" Proliants just looking for something to run.
--

Re:DON'T RUN 2.2.8!!

[mrsam]

Well, Red Hat does sort of like that. They take a kernel that they've tested on a bunch of hardware, and believe to be stable. Usually, by the time they're done, they're two or three releases behind. So they just backport all the minor patches from the later kernel revisions. Their current 2.2.5 production kernel really contains patches from 2.2.6 and 2.2.7, as well as the SNMP DOS patch from 2.2.10.
--

2.2.10, out since June 13

[crow]

2.2.10 was released on June 13, based on the timestamps I see at my kernel.org mirror site. However, there are some obscure file corruption bugs in 2.2.9 and 2.2.10 that they've been working on finding, so I would consider 2.2.8 to be the latest stable version. That bug is the reason you haven't seen 2.2.11 in over a month.

Personally, I live on the edge with 2.3.11-pre5. So what if FAT file system support is broken? That's what development kernels are all about.

Alan Cox

[crow]

Alan may be mostly running the 2.2 tree. He certainly ran the 2.0 tree long after Linus had moved totally to 2.1. However, when it comes time to release the next official version, Alan still gets Linus's approval first.

In many cases, the AC patches include things that Linus has specifically said won't make it into his kernel, or at least not in the 2.2 kernel. I'm not sure if that is still the case, though, now that 2.3 is underway--most of those patches that Linus wasn't ready to accept into 2.2 are now in 2.3.

Anyway, we can expect to see Alan working on both 2.2 and 2.3 until 2.4 comes out.

Re:DON'T RUN 2.2.8!!

[crow]

Thanks. I missed that.

This brings up an important point: We need a good resource to check to see which kernel versions are stable.

Imagine the following fictional site (infomation is fictional, too):

www.kernelbugs.org/2.2/

2.2.0
Has severe problem with the wigglenet driver.

2.2.6
Improved memory management for better SMP performance.

2.2.8
Severe memory system problems

2.2.9
Mysterious and rare file corruption bug, possibly actually due to faulty hardware.

2.2.10
File corruption bug still not fixed.

Most stable version: 2.2.7
High performance version: 2.2.10

The idea is to have a review of different kernel releases at a higher level than just what features have been tweaked. Something where an experienced user can figure out easily what kernel makes the most sense.

Re:version number

[Black Parrot]

> And (someone correct me if I'm wrong), eventually his modifications get merged in with Linus' work.

IIRC, his doc for 2.2.10-ac10 said that most of the changes had already been incorporated into the 2.3 tree.

Re:version number

[nd]

The AC stands for Alan Cox, everyone's second favorite Linux kernel developer. He has his own patches to the kernel that he distributes. And (someone correct me if I'm wrong), eventually his modifications get merged in with Linus' work. Personally, I don't know why this kernel was posted as a story. AFAIK no other ac patches have been posted, regardless of how major the update was.

Like this?

[PhoboS]

So what you want is something like Kernel Newsflash [csiro.au]?

version number

[Omar Djabji]

What does the ac11 in the version number stand for?

2 Gig file limit?

[slieberg]

Does this mean that this patch fixes the 2 gig file limitation now? Where else can I find info about this?

Strange..

[Junta]

I have not seen an Alan Cox enhancement patch posted here, well, never to my recollection... Is there something especially big about this one?

Easily fixable

[ElJefe]

Then turn off those stories in your preferences. Or just ignore them (like I do to most Linux stories).

-ElJefe

Hmmmm ..

[Pingo]

It looks like FreeBSD is doing some heavy stuff. Gigabit ethernet adapters sounds cool.

Very interesting and amusing.

//Pingo

Interesting ...

[Pingo]

I would like to see more of what the *BSD guys are doing. I know that they are also doing great stuff. Some is better than Linux and some isn't.

It would need an update to the slashdot site. Split slashdot into departments. One Techie department where Linux/*BSD can discuss things close to their hart. Stuff that is to heavy for the causal visitors to slashdot.

I guess the guys running slashdot pretty well knows what the other departments should be.

//Pingo