Linux Kernel 2.3.41

sdriver writes "For those of us who enjoy *panic*, *oops*, and suddenly seeing their video BIOS... the newest version is out! Be the first on your block to submit a new patch! ;) " If you don't know where to get it, you probably should stick to your warm and cuddly 2.2.x kernel *grin*. Now outta my way, I wanna crash my laptop!

Comments

A new Linux development kernel.

[cmc]

Well why don't we also post each new FreeBSD 4.0-CURRENT snapshot and each new NetBSD-Current snapshot and each new OpenBSD-Current snapshot to Slashdot, too? It's pretty much the same thing.

On the subject of kernels

[kcarnold]

I have recently been required to describe what the heck a "kerenl" is to some people in close relation to myself. Of course there is the obvious for us geeks, but there are two other possible interpretations that I can think of:

1. Kernel as a rank in the military -- Why would I want one of them on my computer? And I guess the military has very high expectations of these kernels, because they come out with new kernels every couple of weeks. For some reason, some higher-ranking officers don't replace their kernels that often; maybe they like them. But the new kernels have lots of new personality features, and they are even more tough on those below them. There even development kernels, the kernels who have not yet completed their training but are available for use by those of higher rank who might want to try out some of the new features before everyone else, but there is the possability of an occasional slip-up, possibly causing an "oops!" or even a "panic". Worse yet, these development kernels can cause corruption in the system below them. Some kernels, for some officers, have run for a very long time in some of the military systems, and I am very suprised that they don't get tired and fall to the ground. I guess that's the problem withh kernels in the Bill Branch of the military, but the Open Source troops seem to have little problem with this. Well the Open Source commanding officers get to have an in-depth look at how their kernels are trained, so it is possible that this allows the officers at higher ranks to make better decisions about them.
2. Corn kernels -- Yeah, that fits on my hard drive. I spend endless hours in front of my computer monitor^D^D^D^D^D^Dagnifying glass, examining the various ridges and structures on and in the kernel. I occasionally tap at the kernel to see how the structure responds to stresses. It seems to occupy a lot of people for many hours at a time examining and prodding at these kernels. Sometimes, at lower magnification, the surface of the kernel begins to form the shapes of letters, when viewed by the trained eye. Untrained people are cluless. At higher magnifications pictures sometimes become visible, appearing similar to pictures found in everyday life. Through careful prodding through the prodboard one is sometimes able to coax these pictures into changing. The "mouse" is a fairly new addition to the interface used to study these kernels; it is unknown to us how exactly it influences the kernel without eating it. Micros kernels have always showed a lot of detail initially, but viewing the kernel at a higher magnification showed minute impurities that would cause the kernel to split apart relatively soon after the investigation began. The kernel has remained much the same despite many advances in the apparent detail and complexity as viewed at low magnification. Macintosh kernels started out rathar bland-looking, but they contained many useful features from the beginning that allowed studying to be significantly more efficient. However, until recently, the smaller details of the Macintosh kernels were almost as bad as the Micros kernels. In recent years Macintosh kernels have developed a much more detailed structure at any magnification, and as an added bonus when replacing kernels, some have even begun to taste good (Yum.) Linukernels are regularly updated, but older kernels are still so interesting to look at that people have kept them around for a long time. The development kernels sometimes fall apart due to structural problems, but in general they have offered much more detail. People staring at Linukernels are generally happy with how stable their kernels are and how nice it is to be able to know how the kernels were made. Some people tailor their Linukernels to their specific tastes, something which is usually not possible for investigators of other kernels. The report must be cut short here becuase we are still investigating the role of the "printer" in the kernel investigations. Preliminary findings indicate that often the printer is used to capture a view of the kernel to share with others, but often the output of the printer is much different than the view on the magnifying glass. This may be attributed to a special scanning structure that is able to examine different parts of the kernel and print the results. That is all the information we have at this time.

And we make Beowulf clusters out of these things?

``Unstable''?

[David A. Madore]

What exactly does ``unstable'' mean? Or, more accurately, what does ``stable'' mean? The 2.2.13 and 2.2.14 kernels (supposedly ``stable'') have rather nasty inode leaks. And 2.2.11 and 2.2.12 had a memory leak (which is why SGI based their SGI Linux 1.1 kernel patch on 2.2.10). Hmm... Looks like stability ``leaves somewhat do be desired''.

Featurefulness perhaps? Well, one patch I find extremely useful is the ext3 filesystem. (Now that seems stable enough, despite the frightening version number: 0.0.2c) But Stephen Tweedie hasn't finished porting that to 2.3.x. Another useful patch is the international crypto patch, and that doesn't come for 2.3.x... Or the Openwall security patch: ditto. Hmm... Looks like I'll stick to 2.2.x for some time. (And, no, it isn't exactly ``warm and cozy''.)

I'd like to try the 2.3.x kernels by using User mode Linux, but so far the only thing I've got from that is a core dump.

Sigh...

Re:``Unstable''?

[David A. Madore]

Can you give me your email address so we can discuss this somewhere else than on /.? Else you can write to me.

Re:``Unstable''?

[jdike]

Which version are you using?

Can I have a look at the console output?

As far as I know, the latest (2.3.36) user-mode kernel is stable. If you know otherwise, then tell me what's happening, and I'll fix it.

Jeff

Re:Unstable warning

[Krakken]

I've been running dev kernels for 2 years now and have never had to hack 'real C'. Yes things get broken, but I watch the linux-kernel mailing list and submit a bug report if I can't find the answer there. Some times its not the code thats broken, but new undocumented proceedures for setup. I test and report, and get a jump new setups. With lilo setup to boot multi versions of the kernel, I can fall back if things are badley broken.

If you can test, do! We all benifit in the end.


Krakken

Hotsying visor over USB on 2.3.41

[doomy]

Due to the generious coding of Greg and Co at Visor Linux USB project, anyone who owns A handspring visor would be able to sync with the USB cradle without having to buy a serial cradle.

Enjoy.
--

Re:Unstable warning

[Captain Zion]

As long as the user is aware that the "possible instability (or worse)" could be "filesystem trashing". I had 2.3.33 to 39 triggering BUG() calls in ext2, 2.3.40 had no such problem so far.

I'm also trying to get ALSA 0.5.1 working but it keeps throwing up due to problems to allocate contiguous memory for DMA buffers. It doesn't seem to be kernel related however, since others have it working in the 30's and 40's. The CPiA (webcam II) video4linux driver doesn't like my system as well.

I/O Completion Ports

[cpeterso]

async i/o code is now in the tcp/ip stack with a feature called i/o completion ports, ... Uh, I got the i/o completion ports info from a guy inside Microsoft.

I/O Completion Ports were introduced in Windows NT 3.51, not Windows 2000.

...

[David Ham]

for those who actually *don't* know where to get it (heh)...

i am a link
--
in a world of deceit, open your eyes

USB support?

[vtMan2024]

Anyone want to comment on how stable/usable USB support is right now?

Re:USB support?

[tao]

Well, it all depends on what you plan to do with
USB. Keyboard/Mouse/Joystick/Serial ports should work fine,
together with some less common stuff (some cameras,
for instance.) All in all, the USB-support itself
seems quite stable, but the problem is the lack
of drivers for everything but the most common things
(that is, those things that share a common standard or
one of the USB-developers own...)

If you have some device that isn't supported, why not either write a driver for it (IF you know how to do so, of course),
or contact the makers of the device and ask them for a Linux-driver (despite what many companies seem to believe, the consumer is always right. Well, apart from those that buy Micro$oft products, of course; they can't be right in their heads...)

Re:USB support?

[Syberghost]

actually this is the best place to start.


Why? The primary recommendation of that link is to go out and buy the book I linked.

Would you prefer a Fatbrain link instead?

Re:USB support?

[Syberghost]

If you have some device that isn't supported, why not either write a driver for it (IF you know how to do so, of course),

Hell, write one even if you *DON'T* know how. You'll learn a shitload.

Start here:

Linux Device Drivers

Re:USB support?

[mohaine]

My Epson 740 (hooked up by USB) works great.
No driver yet for my USB CompactFlash reader though :(


BTW I've been useing kernel 2.3.29 since it came out with no problems. No real reason to upgrade.

AIX code in Linux

[jamienk]

I read here that IBM is offering Linux kernel developers code from AIX to integrate into Linux.

What does AIX have to offer Linux? Is there anything worth mining? Anything that could help the move towards 2.4?

v2.3.41 -- test it and help us!

[tao]

When it comes to these latest v2.3.xx kernels
(38 and on), I've not experienced ANY disk-corruption
whatsoever. There are, however, lost of other bugs,
both known and unknown. Why? Because us (relatively)
few developers can't possibly try every hardware
combination. We have a couple of new subsystems
in v2.3 which needs a lot of testing (USB (even if
it exists in v2.2, this one has been rewritten
quite extensively), FireWire, PCMCIA, I2C, I2O), as well as a lot of new drivers for soundcards, videocards, TV/Radio-cards, disk-controllers etc. The list can be made much longer.

Oh, and if you have an SMP-machine, you should definitely try v2.3.xx; a lot of SMP-related changes has been made, to improve the performance.

So please, unless you have production-machines, give the v2.3.41 or upcoming developmental kernels a try. You will certainly help both yourself and the Linux-community out in the long run.

If (when) you find bugs, submit them to linux-kernel@vger.rutgers.edu.

[PATCH] New Kernel Module: Uptime.o

[MrHat]

I would like to take this time to announce a new kernel module, recently ported to Linux 2.3.41 - uptime.o.

Uptime.o creates a /dev/uptime entry with major number 199, minor number 0. Want to upgrade to the latest development kernel but don't want to lose your precious 325 day uptime? Do you (gasp) not have a 325 day uptime? Uptime.o is the kernel module for you!

To use, simply run insmod uptime, and then echo the number of seconds of your uptime to /dev/uptime. The kernel module sets your system uptime accordingly. For example, echo "28080000" > /dev/uptime restores your 325-day uptime. Version 0.0.2 will be enhanced with scientific notation support and support for 512-byte uptime values. An example of what is possible with uptime.o:

10:38am up 50,000,000 days, 1 user, load average: 0.04, 0.05, 0.07
49 processes: 46 sleeping, 3 running, 0 zombie, 0 stopped
CPU states: 3.5% user, 7.1% system, 0.0% nice, 89.2% idle

Never become the target of low uptime jokes again! Download uptime.o today - amaze your friends, terrify your enemies, and become king of all that is Linux!

- Hat (with too much time on his hands)

Re:Unstable warning

[logicTrAp]

No, actually Linus himself has said that he *wants* as many people to use development kernels, *even if they're not developers*. So long as you don't mind putting up with some possible instability (or worse), there's no reason not to play with development kernels.

Tulip driver support

[Ophelan]

Recently, the tulip ethernet card driver support was added (I believe in 2.2.14)...support for the newer version, that is...previously using certain tulip cards required compiling your own kernel module. Is anyone aware of whether this support is in the 2.3.x series?

Daniel

---

yay

[matticus]

that's one step closer to 2.4. I'm using 2.3.39 right now, and i'm very impressed by some of the new stuff they are doing. In light of Windows 2000 already needing service packs, it's good to know that the kernel is advancing so rapidly. When i hear about 2.4, i like to evaluate someone's conversion to linux from *other* os's as a gradual process--finding that there are alternatives or equals for every task you can do in windows or whatever. When i turned to linux, there were a few things i didn't think i could do without windows, namely, quake3, terminal server client, use my palmpilot, use icq, write cds, use my parallel port ls-120 drive, and so on. i didn't even know that you could access dos hard drives at that time. but through gradual steps including a voodoo3, reading man pages, getting great new software (jpilot comes to mind), i have eliminated every one of these. this was a 2 year project or so. what is my point? well, i know many people who refuse to use linux because of its poor USB support. i had a friend who bought a computer and wanted to install linux on it, and to his dismay found out his keyboard and his mouse were USB, and his motherboard didn't have ps/2 ports. linux was important enough to him, however, that he bought a new motherboard. we need to be able to help even these people without money-spending on hardware. although linux is free, that is often some area that it is not-almost everyone i know has spent a little extra money to get a new one of whatever peripheral was windows-only, or just didn't work as well in linux. the more the kernel advances, the closer we get to ubiquity;-).

"look at that. breach hull, all die. even had it underlined."-crow

Re:Excuse the heck out of me...

[Phaid]

Did you not follow the development of previous kernel versions? There were 95 versions of 1.1->1.2, 99 versions of 1.3->2.0 followed by 14 2.0pre, and 132 versions of 2.1->2.2. So 41 would be an incredibly short development cycle given the trend. If you want to help hurry it along, run the development kernels on a spare box (or your main box if you like) and help squash bugs.

Volume Manager, baby

[Tim Behrendsen]

I would love to have AIX's volume manager, which has everything you want in a partition manager, and more.

Unfortunately, I think they license it from someone else (not sure who), so I don't think they can release the source code. :(

--