Slashdot Mirror
Linux v2.6 Begins Testing
xose quotes Linus from the kernel list: "the naming should be familiar - it's the same deal as with 2.4.0.
One difference is that while 2.4.0 took about 7 months from the pre1 to
the final release, I hope (and believe) that we have fewer issues facing
us in the current 2.6.0. But very obviously there are going to be a
few test-releases before the real thing.
The point of the test versions is to make more people realize that they
need testing and get some straggling developers realizing that it's too
late to worry about the next big feature. I'm hoping that Linux vendors
will start offering the test kernels as installation alternatives, and
do things like make upgrade internal machines, so that when the real
2.6.0 does happen, we're all set." You all know what to do ;) Update: 07/14 17:49 GMT by S : OverNeith writes "Joe Pranevich has done it again! He's written another summary document on what to expect in the new and upcoming 2.6 Kernel!"
What is it I have to do? Send loves and kisses?
Anytime they start work on a new point release it means great things for the Linux OS.
Linus isn't one to just slap another number on there, notice they are usuallly things like 2.4.25-10 and what not.
2.6 should bringbig changes in lots of the core system components. We could get a new way to handle SMP or a new filesystem. I personally can't wait to skim the change logs.
There isn't an ebuild yet, and I'm too lazy to do it the old way...
I know if I start downloading and compiling this kernel I am going to have a 'case of the Mondays' real quick.
Unique signatures are rare.
For us newbies here, what are the relevant differences in the new kernel? Better performance? New hardware support?
All I see is badness coming from this. If someone is good enough with unix to want to use the 2.6 kernel to develop software, odds are they already know how to download and install the kernel themselves. If, on the other hand, we have someone new to Linux see 2.6 and think "that must be better than that old 2.4 kernel POS", and proceed to choose that one, odds are is that the 2.6 kernel is going to result in a less-than-stable system, and is going to look badly upon linux in the future.
OK, yesterday I was testing 2.5.75. Helpfully, the computer locked up and gave me an opportunity to send a bug report. So far so good. Only that I was in X, I wasn't doing anything particularly interesting or demanding (was playing kbounce), the panic report (if there was one) probably went to tty1 and I have no idea why the computer locked up. How do you report a bug when you can't see what went wrong with the kernel?
Gentlemen, you can't fight in here, this is the War Room!
2.5.x was a testing series. The release following 2.5.0 was 2.5.1, not 2.5.0.pre2 or some such. (In other words, not long)
TODO: Something witty here...
And ive just compiled it. I was quite surprised I managed to get it to boot without it panicing. I'm even typing from the new kernel now. But there is a word of warning though. The layout of the /dev folder has been rearranged. As a result some of my programs have broke.
/dev/hda, /dev/hdb/, /dev/hdc now become /dev/discs/disc0, /dev/discs/disc1, /dev/discs/disc2. So you will need to edit /etc/fstab to reflect the changes.
For example.
Steve French:
o NTLMv2 password support and NTLMSSP signing part 1
o ntlmssp signing
o More NTLMv2
I don't understand - why is this in the kernel? No entiendo.
Get your own free personal location tracker
In a lot of development projects, including the linux kernel, the odd ended builds are developmental releases. So the question is not really "how long has it been since 2.5.0 and how long did that take?" , it's how long since 2.4.0 ...? I believe that question was answered in the summary, and AFAIK it was probably answered in the article (I did not RTFA).
Seems like we are actually gonna get tehre ON TIME.
Any way three cheers for 2.6 ...
2.4 is deprecated long live 2.6...
Pls. no falmes on deprecated thing i just wanted it to rhym with king is dead long live the king.
BAIN http://www.devslashzero.com
No more SCSI-Emulation for burning CDs with this.
Should mention that the sound capture seems to cause the problem -- without sound, the capture is smooth under Linux, but adding either ALSA or OSS to the mix guarantees problems.
Try not. Do or do not, there is no try.
-- Dr. Spock, stardate 2822-3.
Downloaded, compiled and installed. Working since 4 hours on a Slackware-9.0-current, asus L8460K notebook (p3/1000, 256mb ram, i440bx, S3 savage/MX, ess allegro) and quite standard compilation options (acpi, alsa, pcmcia, usb, netfilter, no ipv6, preemptible kernel). Applied patch as seen on LKML (see here) for vfsmount.
Happy testing!
I'm fat, you're ugly. I can get slimmer, and you?
Doesn't devfs take care of all this?
UNIX/Linux Consulting
Last time I looked at ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/ must-fix/ there were still some showstoppers. It seems like they were updated about a month ago, so I guess progress must have been made on them...
The biggest problem I have with the newer kernels is probably some ACPI/IRQ routing bug in my board. It's a common problem with the NForce2 chipset (APIC doesn't work, so you have to boot with pci=noacpi or acpi=off). It's not the biggest inconvenience, but it causes half of my unused USB slots not to work...
I must say the snappiness of 2.6 is great! I'm looking forward to beta-testing. AFTER I backed up my drive, of course!
I hope they added a few lines like "This is not SCO code" :-P
OK, So what if I'm a troll?!
Congrats and thanks, Linus. I was not expecting testing to begin so quickly after he expressed his opinion.
To-do List: Receive telemarketing call during a tornado warning. Check.
You know, the firewire disk driver. Man that thing has never worked 100%.
Just try corrupting a large (mine was 90GB) partition on a firewire HD and then fschk it. Eventually it'll start getting timeout errors and all sorts of crap, and will eventually trash the filesystem even worse. Then you can't mount the drive at all.
I usually end up having to go to Windows because it's the only place that I can force a massively corrupted partition to mount (and it has better SBP2 support). From there I can copy everything that is still good off and reformat the drive.
This hasn't just happened once. More like 3 or 4 times (both EXT3 and Reiser partitions) over the last year or so.
The ratio of people to cake is too big
If I remember correctly there's a new Block IO (BIO) layer included too, which should enable IDE CD burning without the need for SCSI emulation. Should speed things up somewhat.
I'm not exactly sure if this is correct - I believe I heard it a the Linux Forum in Denmark back in march. The speaker was Jens Axboe, the current cdrom subsystem maintainer.
zWhat would an EWOULDBLOCK block, if an EWOULDBLOCK could block would? -- me
Are there any distributions out there that are actually using devfs?
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
Two possibly dumb questions (but this is slashdot, after all). (1) Can you change the scheduler default timeslice (10 msec seems a bit long for a multi-GHz CPU). (2) does it do the right thing for hyperthreading? (for hyperthreading, the scheduler needs to understand that one of the CPUs is sorta crippled, so jobs should flop back & forth between both CPUs).
--- Often in error; never in doubt!
Technical achivements aside, the most amusing thing about the 2.6 series of kernels is seeing all the large corporate entities with vested interests deal with the release schedule.
That is to say, there isn't one. I especially liked the quote from Torvalds I recently saw in a CNet news.com that basically said, "it'll be done when it's done - deal with it".
The dangers of knowledge trigger emotional distress in human beings.
Let's hope this supports USB 2.0 "Full Speed" or "High Speed", whichever is faster..
LINUX 2.6 KEY CHANGES
Faster, more predictable performance and new APIs are on tap
[Yay]
Desktop improvements
[Whatever that means]
Universal Serial Bus 2.0 and production Bluetooth support
[Yay]
Pre-emptible kernel with low-latency kernel patches for more user responsiveness and better multimedia performance, even under heavy loads
[Like Windows XP!]
Server improvements
[Whatever that means]
Updated I/O and memory subsystem for faster throughput and scalability
[Yeah, like blast processing]
Faster, more scalable process scheduler
[Pfft]
User-mode Linux to allow multiple system images running on the same box to aid server consolidation and application separation
[Sounds like the minutes of a business meeting]
Asynchronous I/O and completion events--a big improvement for Web servers and databases
[I'll take your word for it]
Support for disks larger than 2 terabytes and for SGI's XFS enterprise file system
[OK]
Faster, POSIX-compliant threading library
[Redundant]
and it still shows nothing on screen if i pass vga=normal during boot, and it took me several atempts before I relized that regular ps/2 keyboard can be left out or compiled as a module. well, this kind of changes were expected.
after i managed to get it working (booting, with keyboard, framebuffer console, et. all) surprise... no DRM on X.
happens that for some reason X doesn't detect working agp when a Radeon 8500LE in inserted in my kt266 based mobo. even with agpgart and radeon modules loaded.
so here's a few sugestions:
leave ps/2 kboard selected by default for x86 architectures, same for a way to display the console on text mode vga and check this radeon issue.
except those minor stuff, the new kernel is great. really fast for regular use.
What ? Me, worry ?
It will be nice to see some articles in the mainstream press showing that Linux is still marching on regardless of SCO's drum beating.
Boffoonery - downloadable Comedy Benefit for Bletchley Park
Also, with RedHat's aggressive "enterprise" campaign, as well as IBM's openness towards Linux, many application vendors are now buliding their software for Linux. Having easy access to test kernels will make it much easier for them to qualify their products against 2.6, allowing them to get their products "blessed" for 2.6 as soon as it's generally available.
I know Slashdot isn't a support forum, but could someone point out a good tutorial for compiling and installing a new kernel? I'd like to give 2.6 a try, but I don't know where to begin.
DLed it last night, and built it. Looked fine - I like that the make xconfig is no longer really REALLY ugly, but xinerama seemed to confuse it (;
Anyway, I couldn't get the nvidia viddeo drivers to build for it, and it WAS 4am, so I'm back to 2.4.20, and maybe I'll play with it later. Hoping someone already did it and feels like posting. (:
In the land of the blind, the one-eyed man is kinky.
In the GNU tradition:
Linucs Is Not Using Code from SCO
Accountability on the heads of the powerful.
Power in the hands of the accountable.
LinuxThreads sucked. NPTL puts Linux on par with Solaris in terms of multithreaded performance. This will benefit all multithreaded programs - especially Java.
I see No great things. But it is the kernel. May be we are overestimating the importance of the kernel. Al least, my whishlist are not related with the kernel appart from the hardware support with is huge problem.
--
One by one the penguins steal my sanity...
Re: Linux v2.6.0-test1
The whole thread is here Linux v2.6.0-test1
I thought 2.6 was suppose to have crashdump support? If not, that's to bad, because often that is what is required to fix problems in the real world. Often the technical person isn't the same person who is using the machine. There needs to be a way for the technical person to figure out what went wrong after the fact. OOP's are about as useful as the BSOD data. Plus, unless its a repeatable problem usually by the time the machine crashes its a little to late to run out and hook up a serial console.
% lilo
Fatal: Kernel vmlinuz is too big
Where's the version without windows and gadgets?
i don't like style guides
Security should be taken seriously. This is truly informative.
To test this issue out, run Sawfish, and bind a key like Ctrl-Alt-B to a black-background xterm. Launch X, and run Sawfish. Hit Ctrl-Alt-B once and see what happens. It's consistant here across about 6 machines, all different hardware.. a 3-4 second delay, then anywhere from none to 4 xterms will open up. On 2.4.anything, it opens the xterm instantly, and only opens one of them, not 3, not none.
The other issue is that there's some underlying change in the TCP stack/net drivers that cause rsync and anything running over ssh/ipsec to fail with weird dropped-socket errors from the applications using them. Again, on 2.4, it works flawlessly.
It's very annoying, and both of these are blockers for me and most of the machines I'd be running this on. It happens with anything that involves keyboard shortcuts; menu accels, launched applications, keybindings, everything.
Changing to the different schedulers does not help; deadline, as, or cfq. 2.5.68 worked perfectly, and didn't have these anomalies, but every single kernel since that time, has had it. I've diffed, and I can't tell which of the dozens of changes actually broke this.
If anyone has a solution, I'm all ears.
FWIW, 2.6 has ksymoops built in now. Not sure about a full-on debugger - I lost track of where that idea went. Last I checked, anyway (yesterday). The thing that will get most people (I bet) is needing to have the right config options enabled for the console and for kernel debugging.
C|N>K
The 100 was simply a good compremise between desktop and batch. In the past, the schedular was fast, but there was still overhead (faster than other OSs for process switch, but slow on threading switch). Now, that the schedular is on O(1) and the interuptable patch is applied, it is easier to move it downwards and still get major work done.
I prefer the "u" in honour as it seems to be missing these days.
I remember reading that a new kernel configuration utility/script (ie "make menuconfig") were to be in place for 2.6 .
Has it been merged in yet and/or if not will it be in an AC kernel? If not for 2.6, will it be in a future 2.x kernel or will it be in 3.0?
For a not so newbie, but not technically savvy, will any functionality be lost? The latter tends to have more of an impact on my day-to-day.
"Consensus" in science is _always_ a political construct.
As for the sound card angle, I've got two sound devices in the system -- one motherboard provided, one Sound Blaster Live. I'm using the SBLive! for dumping with Alsa 0.9.2 (which also drives the other device), however I have tried the other device with alsa9 and the SBLive! with the emu10k1 OSS driver provided with the kernel with similar results. There was also a method that escapes me at the moment for dumping audio directly from a chip on the capture board, which was unsatisfactory.
Dumping video without sound yields exactly the level of quality I'm seeking: no frame drops, and only an extremely slight regular periodic jog on the news ticker that is probably the artifact from de-interlacing NTSC frames. However, no method I've tried of dumping the sound with the video escapes frame dropping (severe at the start of the application tapering off to one or two blocks of frame dropping per second as the capture continues.) As with the video codec, I've tried various settings for audio recording -- lame, PCM, '-oac copy' without dice.
I live by hdparm, and also directly patched the kernel a few versions back to recognize my UDMA133 chip. I do use cheapass hard drives (and frequently backup) but have varied the relative CPU/HDD strain on encoding by testing the spectrum of video codecs with similar results. It does drop frames quite mightily after about twenty seconds of MPEG2 encoding under ReiserFS (I'm using ext2 for my real video-dump drive, though). For what I'm doing right now, I'm cool with single-pass XVID.
Sadly, NuppelVideo gives me similar issues. MythTV was one of the first things I tried out on this card.
PhracturedBlue's scenario sounds pretty close to mine, but the mentioned tweaks didn't seem to work. Will try them again, though; I'm using a newer kernel and there's always the potential something's been changed.
Thank you all for the advice. Maybe I'll delve into it again one of these nights with the new kernel and try profiling driver calls (or drivers...)
Try not. Do or do not, there is no try.
-- Dr. Spock, stardate 2822-3.
Well, that's good - 2.6 is testing. That means that it will only be, what, like 4 or 5 years until Debian puts 2.6 in their stable tree? :P And yet, I still won't switch to something else. It's just too good (plus, I have the advantage of not necessarily needing to be 100% stable :).
Support for USB 2.0! Great!
Wait a minute... is that HIGH SPEED, or FULL SPEED?
-j
Though honestly, much of MS software is also sold shrinkwrapped. This gives a latency between the final build, documentation print run, CD pressing, packaging and distribution that doesn't exist with something like the linux kernel. During this time development continues, which is why you can have patches for a game or application avaliable before said product is even in wide distribution.
And again honestly, I don't think you can argue that the linux stable series are released as "full quality" and don't need patches right away. History does not support such claims.
Indeed, Linus knows this; he sees the problem that the unstable series doesn't get tested well enough and that only slapping a 2.even.0 number on it increase the testing crosssection several orders of magnitude.
In conclusion, I think you should read "it'll be released when it's done" not mainly as an assurance of release quality, but as a short form for "it'll be released when I think we'll have a good enough chance of getting people to test this without getting too badly burned".
Belief is the currency of delusion.
holy sweet jesus, huge noticeable performance difference on my athlon 650, going to 2.6.0-test1 with the new scheduling algorithms and the preemptive kernel mod... much, much better performance under heavy loads than it was with 2.4.20
"I'm hoping that Linux vendors will start offering the test kernels as installation alternatives, and do things like make upgrade internal machines, so that when the real 2.6.0 does happen, we're all set."
;-)
That would be useful in getting these new kernels debuged quickly. The install should default to the latest stable release however, and should be very clear that the optional test kernels are infact not a final product release.
Otherwise we'll have made the same mistake as Microsoft - shipping incomplete products under the guise of a polished solution - and having our paying customers debug and test them for us.
I'd rather be a conservative nutjob than a liberal with no nuts and no job.
Ok, Now this is the kind of sh*t we like to hear.
Proper testing by a proper Linux user hitting the machine hard and heavy.
Nice one, Urchlay!
This guy should do reviews methinks!
"I used to have that really cool,funny sig
I see several great things. What exactly did you want from this release?
Here's the big list for me:
* O(1) scheduler
* Customizable elevators for load tweaking
* CPU Affinity
* Preemptible kernel
* I forget the name, but it's an IO improvement based on doing nothing
* Use of IDE CD Writer without ide-scsi, which allows for DMA usage
* Other things I forgot
Anyway, it looks to be an excellent release. What exactly were you wanting?
Engineering and the Ultimate
Yes, every product has quality issues. But Microsoft products have way more than their share.
If I read the changelog properly, this kernel is what you were looking for. It has all of the necessary plumbing to support that - the keyboard is loaded as a module, etc. The majority of the work for that would need to go into XFree86 and the distributions.
Engineering and the Ultimate
Anyone who thinks that the link proposed is a summary has waaay too much time on their hands!
But he didn't ask you how long ago was 2.5.x, did he? He asked you how long has it been since a specific version, namely 2.5.0.
Instead of giving him some bullshit, which only shows that you know how to count, why not answer his question and say that 2.5.0 came out on the 22nd of November, 2001?
None are more hopelessly enslaved than those who falsely believe they are free. Johann Wolfgang von Goethe.
It would be enough if some kind soul provided some .config files pre-written so that us non-hackers could get a kernel build going quickly, that's suited to our distro.
I'm glad 2.5 is working good with you. But I'm experiencing laggy mouse response.
I have RH9 + XD2 under 2.5.75 - when I load a heavy web page(i.e. slashdot news with lotso comments) under my Galeon(or Mozilla) web browser my mouse goes laggy for a bit as the web page loads.
I solved this problem mostly by using Con Kolivas's desktop patch suite under 2.4. I don't know why the laggy mouse, but using the Variable HZ patch helped.
I thought 2.6 was suppose to have crashdump support?
Say it isn't so! A feature that commercial Unix has had for years isn't yet implemented in Linux?
Folks, when those Ask Slashdot questions come up that sound like "why doesn't everyone switch from commercial Unix to Linux right now? Unix suxxx.... Linux r0x0rz!!!"
This is why.
"When the president does it, that means it's not illegal." - Richard M. Nixon
From the article:
:)
And finally, Linux 2.6 will include improved 64-bit support on block devices that support it, even on 32-bit platforms such as i386. This allows for filesystems up to 2TB.
2 TB? Is this really the filesystem limit in the still unreleased 2.6? Wouldn't a RAID of 7 300GB drives be hitting this limit already, using hardware availiable today? 2 Petabytes or Exabytes would be a nicer limit
There a whole bunch of features that exist in commercial unix distro's (and NT too, crashdump has been there for a long time) that are fully mature and stable that don't exist in linux. Anyone who isn't completly biased, quickly finds feature after feature that is half baked or non existant.
Of course, this requires one to stop reading slashdot and accually try to get something done.
I like this feature this most. Won't this give large SMP activity a boost on Linux, letting it scale up even better to a larger number of processors with less of a penalty?
Let us know when 2.6.18 comes out. That's one version past when 2.6.17, which is the first time someone with get the wise idea to test it on a production machine and it will go *whooooof* <- sound of a sysadmin running past your cube.
Wasn't it about 2.4.7 before you could reboot without the drive getting corrupted?
2.6.x will rock, for high values of x.
- Adam L. Beberg - The Cosm Project - http://www.mithral.com/
If you want the new scheduler, preemption and other goodies on the stable 2.4.21 kernel, try the CK patch.
Escher was the first MC and Giger invented the HR department.
In the "summary document" I read a lot of references to support of hardware and competing systems that most of us have trashed long ago. Fine with me, but how about de floppy disk driver? That hasn't worked reliably for me since 1.2 or so, but was perfect in early Linux versions.
It is not so much of a problem these days as floppies are rarely used anymore, but I've always wondered why it has been in such a sad state for so long...
This is not a hardware-related problem, I have seen floppy trouble on many different systems. Read errors or unable to format on floppies that read or formatted perfectly in DOS or Windows on the same system. When I need to write a couple of bootfloppies I usually format them on Windows, saves a lot of grief...
Simple IDE timing tells it all: 2.4.21 /dev/hda:
Timing buffer-cache reads: 128 MB in 0.45 seconds =284.44 MB/sec
Timing buffered disk reads: 64 MB in 2.57 seconds = 24.90 MB/sec
2.6-test1 /dev/hda:
Timing buffer-cache reads: 128 MB in 0.51 seconds =252.01 MB/sec
Timing buffered disk reads: 64 MB in 31.07 seconds = 2.06 MB/sec
Go to the FAQ, or look further down on the left for specific FAQ questions
specifically: How do I compile a kernal?
If that don't do it, note the link at the bottom of the box, the "more indepth tutorial".
I spotted that about an hour ago, and now I'm an expert on finding it. :-)
My other car is a 1984 Nark Avenger.
It's not true, as the article claims, that making one process look like two doesn't buy you much. The reason is that cache misses are getting more and more expensive: without hyperthreading, a cache miss might cause the processor to wait a hundred cycles. With hyperthreading, we simply switch to the other process, and pay a far smaller cost.
There's a project looking at moving _everything_ disk/disc related under the SCSI layer. Supposedly it streamlines quite a lot of things. Read the latest Kernel Traffic if you're interested. Not ready for Prime Time yet, but looking for testers it seems.
Lump lingered last in line for brains, and the ones she got were sorta rotten and insane.
Hyperthreading is the ability for a single processor to actually masquerade as two (or more) processors from the operating system perspective. What is absolutely the most amazing thing about this feature is that Linux was the first OS to bring the features to market, despite compatible processors being released by Intel almost a year ago.
I find this odd, since my FreeBSD kernel has had an option for enabling HyperThreading support in the kernel since 4.8 (option HTT). FreeBSD 4.8 was released on the fourth of april this year. Linux 2.6 is not out yet. I hardly think this is a first for Linux.
It does seem to be a common belief amongst Linux users that Linux and Windows are the only two operating systems in the world. Guys, there are other options out there. I hear even a little company called SCO has some kind of Linux-like OS...
I am TheRaven on Soylent News
Is there working Linux/PPC support? I got through menuconfig, and make choked on time.o because two variables were defnied as both time_t and int. Since time_t is an int (via a typecast), I changed them to time_t. Now make is dead (after a lot of progress) on drivers/ide/ppc/pmac.o. I'll see if I can fix it, but it looks a little more involved. Oh well. Any PPC people know what's going on here!?
My other car is first.
Does this kernel fail to compile on riscom8, isicom, or specialix for anyone else, I'm scared to report it to the list becuse i'll probably be laughed off for it being something thats my fault, i tried googling it but couldn't find any answers.
Does anyone know the status of Ingo Molnar's Exec shield patch?
One of the new things in the kernel is User Mode Linux. And I run a hosting company that runs UML servers. So this release is a very cool milestone.
Sadly, the 'official' tree is not keeping up with Jeff Dikes patches. Which after being initially accepted are now gathering dust in some bitkeeper dump.
This from Jeff on the UML mailing list:
Hopefully, Jeff will get his patches compiling against the latest 2.6 kernel and Linus or Andrew will accept them. This is a cool piece of technology. And what's currently in the 'official' kernel has become old and crufty. (Workaround being to apply Jeff's patches).
Anyway, I got around to downloading and compiling the latest 2.5 kernel with the UML patches and running up a few root file systems. It seems to work well. Er, I mean, boot up without panicing.
If anyone is interested in tinkering with a Linux server running the 2.5 kernel, feel free to contact me. It'd beat having a malicious alpha kernel corrupting all the data on your PC.
- Peter
RimuHosting - VPS Servers. Now with 2.6 Kernels
For anyone interested in having a tinkering/testing Linux server
Well, the first fun thing I've noticed is simply in the configuration stage of building a kernel. I always stick with the classic ncurses `menuconfig' but noticed `gconfig' in the README. Jeeze, there's a GTK2 interface on the sucker, and it soaked up my /boot/config-2.4.21 (never noticed if 2.4 does that). The menu hierarchy is cleaner too. Filesystems are classified by nature. nForce2 support where it counts. Preempt, etc... looks great.
(For more fun, grep for the word "fuck" in the kernel tree.)