Although Linus Torvalds didn’t meet a goal to release the upcoming 2.6 version of the Linux kernel in June, the Linux leader said a test version could come as soon as next week. A day after Linux founder Linus Torvalds joined the Open-Source Development Lab, the group has hired another top programmer, Andrew Morton, the developer who will assume responsibility for the upcoming 2.6 version of Linux.
Linux 2.6 Beta Coming; Linux Lab Hires Second Guru
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
I read in the release notes that it has frame buffer support… does this mean X windows is gone now!? YAY!!!
“read in the release notes that it has frame buffer support… does this mean X windows is gone now!? YAY!!!”
Um. The Linux kernel has had framebuffer support for a very long time. I don’t know what you’re talking about.
Also I hope they add plug and play. Last time I tried Redhat none of my windows device settings were regocnized
This thread is off to a great start…
Framebuffer has been there for quite some time, as Shawn pointed out. Possibly a new video card has framebuffer support, I haven’t checked.
Also, RedHat (and most other distributions) has been an easier, more settings-free install for a few versions now.
Plug and play is working very well, however I would like to see improvements in adding devices the kernel programmers did not include support for.
On a good note, X should seem more responsive when using 2.6 with the new scheduling refinements.
windows device settings were regocnized
??? This make me confius
All I care about is the numerous standardized improvements in a native Linux audio interface that is supposed to exist in 2.6? Does anybody know if this kernel would offer lower latencies for multimedia production?
As far as I know ALSA is now in the kernel so you no longer have to compile to get sound support
someone needs to blackhole cable.rogers.com
I have been using Con Kolivas’ patches and yes, it dramatically improves performance ( O(1), Preemptible kernel, etc). I just hope Mr. Marc Boucher or a fellow Linux community member writes a patch for the connexant drivers for the HCF/HSF modems so that they support the preemptible kernel soon.
Great work Linus and other gurus! I’m a die-hard Linux fan and no matter how many friends call me a geek, Linux is my choice and I will never let it go- yes, I was wild enough to run Linux as my primary OS for a year at my prev. Univ and even now that I have a laptop, I’m running Linux on it 😀
YES! I can’t wait! tho it’ll probally be a half year or so before I even try 2.6
LET THE BETA TESTING BEGIN!
i bet if you use windows, nobody will call you a geek
so it is more like linux gives you the special status – being called a geek
Is ACPI support any better? IMHO this is quite bad news if you can’t easily type acpi suspend, or set up auto power management.
ACPI is still pretty borked. The problem is that the Windows ACPI implementation is really broken, and all the hardware manufacturers have implemented non-complient devices so it works on the broken implementation. The Linux ACPI guys (which use code from the Intel open source ACPI implementation) refuse to be bug-for-bug compatible with the Windows ACPI implementation. Hopefully, once MS gets its act together, less broken hardware will be released.
PS> This has bitten me personally. My Inspiron’s ACPI tables are broken. ACPI works fine on Windows, and not so fine in Linux.
ALSA is in the kernel, and its a high performance API. With the preemptible patches, and the new scheduler improvements, Linux maximum latencies are down to less than a millisecond, which is beyond BeOS and into OS X CoreAudio territory. The audio latency depends on the driver you are using. Some pro-level audio hardware has vendor-supported drivers, so depending on what you have, you can get very good results.
The problem is that the Windows ACPI implementation is really broken, and all the hardware manufacturers have implemented non-complient devices so it works on the broken implementation.
Rather usefully this shows how broken software can have consequences for hardware, and unlike web pages can’t be fixed with a simple tweak and an upload.
Yes, ACPI is definately not as good as windows. But, it is definately getting much better with this beta kernel series (as compared to 2.4) and the acpi patches kick butt. Alas, it will never work as consisently across all hardware as it does on windows, because of the above reasons.
But, one of the sleep states, hibernation (which is when you dump all contents of RAM onto disk, and _completely_ turn off power, but resume quickly back to the same state when power is turned on) will be available via the “swsusp” feature of the kernel, which should work (in theory, once it’s more stable) on all systems, and does not rely on traditional power management (just software routines). So that is nice… but “standby” mode will not work as well I presume, because it is based on ACPI and is apparently quite hard to implement, especially when you have to fight with non-compliant hardware.
“The open society, the unrestricted access to knowledge, the unplanned and uninhibited association of men for its furtherance – these are what may make a vast, complex, ever growing, ever changing, ever more specialized and expert technological world, nevertheless a world of human community.”
— J. Robert Oppenheimer
Fight against DRM, TCPA, PALLADIUM, and other implements of evil whenever you find them.
Computers and software do not have to be militarized.
Believe in a better future, act for a better future, and it shall come to be.
At least since the 2.5.50 series all the way to 2.4.73 (just downloaded 2.5.74…), I also use Andrew Morton’s mm patch series, mostly for the anticipatory i/o scheduler – which ROCKS. My thoughts so far; suprisingly stable – the last three releases especially so. All my devices work 100% now, no data corruption so far….fingers crossed. What do I like? Performance! Its noticable everywhere, for me its most noticable in i/o. Morton’s AS scheduler rocks, launching multiple apps while listening to mp3/oggs and copying large files to and fro on the file system, at once, is very very smooth. Try it with 2.4 – stutter and drag. AS gives preference to reads, so even if your doing heavy copying in the background, apps and whatever else needs to read from disk will launch much faster then in 2.4. Also the process scheduler is nice, although it “hiccups” a bit compared to 2.4. Their aware of that, its being worked on I hear. PMCIA drivers still seem borked, else I’d use it as my default kernel on my laptop – I have an aging 266 mhz PII laptop that is much snappier with 2.5, but no PMCIA (NIC, in my case) then I can’t stay in it if I want on the net. Ah well. Oh, and downloads seem faster – nice. I really am impressed so far, Linux gets faster and faster…. 😀 Seems like the last of our major trouble areas are being worked on, whatever performance advantages FreeBSD had seem suddenly gone – now all the kernel is slick, the VM, scheduler, threading (NPTL), I/O, Async I/O, NUMA, and more. Yum. 🙂 Maybe I speak too soon, but aren’t we now approaching genuine Solaris level performance? Certainly not 100% there yet, but we’re much closer then ever before.
> Maybe I speak too soon, but aren’t we now approaching genuine
> Solaris level performance? Certainly not 100% there yet, but we’re
> much closer then ever before.
I agree with you, all of the kernels past 2.5.70 anyway feel very nice on the desktop. 2.5.72 can lag, staying responsive but requiring a lot of clock cycles for seemingly insignificant operations (more so than 2.4.0) under certain circumstances but AFAIK this has been resolved in 2.5.73 and 2.5.74. Probably just debug code as these are classic side effects.
> more so than 2.4.0
Oops, I meant the 2.4.x series.
(Not having a good day!) 🙁
If they get it right, the 2.6 kernel will really fuel Linux adoption even more. Especially at the enterprise level, where the new scalability features will be greatly appreciated. Given how well Linux has fared with the 2.4 kernels, it can only get a lot better with 2.6. All those things that were massively patched in to 2.4 by Red Hat et al will not be needed any more.
And it even helps desktop users with the interactivity and new scheduling and latency stuff. It’s all part of Linux maturing into a fully grown-up operating system. SCO is never going to stop it now 😉
I’m very happy to hear that Andrew will talk over maintainership of 2.6. I was a bit afraid that we would have to live with Marcello for yet another stable kernel series.
I often play around while dowloading stuff, playing some of the minor Linux games.
Very jittery performance. At times I feel like throwing the computer out the window.
If 2.6.x can give flawless games performance, that’ll be an even bigger reason for games developers to migrate. I can’t wait!
YES! I can’t wait! tho it’ll probally be a half year or so before I even try 2.6
LET THE BETA TESTING BEGIN!
Evidently you don’t know much about the Linux kernel. I’ve been using it since 2.5.6x (the 2.5 being the 2.6 alphas) about 2 months ago) and it’s been solid as a desktop and much faster than the 2.4 kernels.
I’m currently on one of the 2.5.69-mm and will upgrade to a 2.5.74-mm.
>> I was a bit afraid that we would have to
>> live with Marcello for yet another stable
>> kernel series.
And that would be because??? What have you got against Marcello’s work???
Re: Importance of Standard
Imagine if acpi was an absolutely critical function. A big bully can easily use a thing like that to hurt the competition.
Morton’s patches, from what I could gather, are mostly performance and optimization stuff. This is very good news.
I’m not the original poster, and I have nothing against Marcello, but I also have to agree that he could’ve done a better job with the 2.4.x kernel, changing the VMM mid-series was an incredibly stupid thing to do, not to mention several data corruption issues that crept up in the middle of the series. The 2.4.x kernel is supposed to be the “stable” kernel, yet it was very late in the 2.4.x series before things really got stable. Don’t get me wrong…I’m running 2.4.21 on my primary machine and it’s great, but some of the mid-series kernels were kinda scary.
>> changing the VMM mid-series was an incredibly stupid thing to do, not to mention several data corruption issues that crept up in the middle of the series.
>>
Linus was the one responsible for the VMM change. Marcello wasn’t even maintaining the stable branch as of the time the decision was taken, I believe Alan Cox was in-charge back them. And like Cox, Marcello disagreed with Linux decision.
Overall, I think the young man has done a wonderful job. He has been pretty careful not to get entangled in kernel politics, and that alone merits a thumbs-up.
Framebuffer has been there for quite some time, as Shawn pointed out. Possibly a new video card has framebuffer support, I haven’t checked.
I think what this is is framebuffer support for the NVIDIA GeForce3/4 card w/128MB + Athlon + 1GB RAM combination. Due to a strange bug, machines with such a combo could not use the framebuffer and displayed a black screen instead (which made using graphical installers quite difficult).
There’s been a patch available for a few weeks now – good to see it will be included in the next kernel, even though it is a somewhat exotic problem.
Linus was the one responsible for the VMM change. Marcello wasn’t even maintaining the stable branch as of the time the decision was taken, I believe Alan Cox was in-charge back them. And like Cox, Marcello disagreed with Linux decision.
I appologize to Marcello then for my comment. I was unaware that Linus had made that decision.
Well, then he did really well considering what Linus did to him.
A fine tuned Linux has a lower latency than Windows, BeOS or MacOS… 😉
You could get a latency below 1ms, should be good enough.
http://www.linuxdj.com/audio/lad/faq.php3#latency
i have nvidia gforce4 + amd tbird + 1gbram and i haven’t had a problem i’m not sure what you’re refering to…
any chance you could link to something about that problem?
thanks
anyone know which kernel version is going to natively support 802.11g? or which cards already work with 2.5.x?
thx
My mistake, it doesn’t seem to have to do with the processor type, only with the video card (GeForce4 Ti and supposedly Radeons as well, both with 128MB+ of on-board memory) when used on a system with 1GB of memory, in vesafb framebuffer mode.
Here is a bug describing the problem on Mandrake’s Bugzilla:
http://qa.mandrakesoft.com/show_bug.cgi?id=3198
There are also a few threads about this on the linux-fbdev-devel list:
http://sourceforge.net/mailarchive/forum.php?thread_id=1895772&foru…
http://sourceforge.net/mailarchive/forum.php?thread_id=2109303&foru…
http://sourceforge.net/mailarchive/forum.php?thread_id=2482221&foru…
From the Linux-kernel mailing list, an early thread about the problem, and another thread with the suggested patch:
http://www.ussg.iu.edu/hypermail/linux/kernel/0303.3/1098.html
http://www.cs.helsinki.fi/linux/linux-kernel/2003-17/0535.html
The patch was eventually included in kernel v2.4.21-rc2:
http://lwn.net/Articles/31701/
I don’t know if that’s the one they’re talking about here, though (for 2.6, I mean).
Just out of curiosity what distros are people using to run 2.6 on? Is redhat 9 capable of running it out of the box, or does it take modifications to glibc or modutils?
Would anyone know of any good instructions for compiling 2.5.x, becuase I heard there’s a slightly different way of compiling it vs the 2.4.x series. Thanks.
Minimal software requirements for 2.5.X from the README of linux-2.5.74 http://www.kernel.org :
o Gnu C 2.95.3 # gcc –version
o Gnu make 3.78 # make –version
o binutils 2.12 # ld -v
o util-linux 2.10o # fdformat –version
o module-init-tools 0.9.9 # depmod -V
o e2fsprogs 1.29 # tune2fs
o jfsutils 1.0.14 # fsck.jfs -V
o reiserfsprogs3.6.3 # reiserfsck -V 2>&1|grep reiserfsprogs
o xfsprogs 2.1.0 # xfs_db -V
o pcmcia-cs 3.1.21 # cardmgr -V
o PPP 2.4.0 # pppd –version
o isdn4k-utils 3.1pre1 # isdnctrl 2>&1|grep version
o procps 2.0.9 # ps –version
o oprofile 0.5.3 # oprofiled –version
o nfs-utils 1.0.3 # showmount –version
I used to run 2.5.x kernel with RedHat 8.0. I presume it’ll work with RH9. The only other package I needed to install were the new modutils needed by the new kernel. Essentially, old programs like insmod would be mv’d to insmod.old and the new insmod would be able to detect whether a 2.[56] kernel was running or a 2.4 and run the appropriate routines.
I was even running it with an NVidia GeForce4 Ti5600 card with patched drivers from NVidia and it ran pretty well (gaming-wise).
After all that talk of hardware manufacturers making products which conformed to MS’s broken ACPI implementation but not to standard, is there a list of hardware which DO implement ACPI correctly? I’d like to make sure that my next hardware purchase (this week) will be ACPI- and Linux- (which means standards-)compliant.
Direct from Kernel Traffic #220 for 27 Jun
http://kt.zork.net/kt20030627_220.html
Someone explained why it was so hard to get ACPI right, saying, “its a moderately bad spec that your vendor implemented to deal with a horrid redmond interpreter. (And, to make things worse, the linux-acpi team specifically insists on implementing the spec, not the reality. “We refuse to be bug-for-bug compatible with the other major implementation.” So linux-acpi is “right” but redmond-acpi is tested and actually works.)”
What good is being right if it doesn’t actually work?
The Linux kernel has no qualms about supporting other shoddy implementations of well defined specs, such as buggy IDE controller cards and the like.
What’s the big idea here? And what do they test their implemenation against so that they are so sure they are “right”?
“Seems like the last of our major trouble areas are being worked on, whatever performance advantages FreeBSD had seem suddenly gone – now all the kernel is slick, the VM, scheduler, threading (NPTL), I/O, Async I/O, NUMA, and more. Yum. :-)”
Yes, this is partially true. In terms of kernel capabilities, Linux probably will be at least equal most of the time and ahead more often than behind. However, there are still other reasons to use FreeBSD. One is that it is a fully developed Operating Environment. It has both source based installation and binary packages installation. No Linux distribution can match it. RedHat is too commercial and only supports a handful of packages. Debian is too outdated, the 2.8 kernel will come out before they release a 2.4 kernel based distro. Gentoo is too source based to be considered in the enterprise. Another reason to use FreeBSD is that it is BSD based rather than System V based. It has two run levels, full and single user. Whenever you need to modify the system, drop into single user mode, modify it, and then kick it back into full mode. Furthermore, the majority of the system configuration are held in one file so you don’t have to go looking through the /etc directory looking for one or two line files.
There are reasons to use FreeBSD, but those ain’t it.
RedHat is too commercial and only supports a handful of packages.
>>>>>
RedHat is too commercial to be used in the Enterprise?
Debian is too outdated, the 2.8 kernel will come out before they release a 2.4 kernel based distro.
>>>>>>>
First of all, kernel 2.4 is already an option in Debian 3.0. Second of all, most enterprise users prefer something that is highly tested, if a bit stale, to something that is bleeding edge. If you really need something recent, there is always the testing branch, which in reality is about as stable as most distros.
Another reason to use FreeBSD is that it is BSD based rather than System V based. It has two run levels, full and single user.
>>>>>>>
SysV has 5 runlevels. What’s your point?
Whenever you need to modify the system, drop into single user mode, modify it, and then kick it back into full mode.
>>>>>>>>
Can do this in SysV too.
Furthermore, the majority of the system configuration are held in one file so you don’t have to go looking through the /etc directory looking for one or two line files.
>>>>>>>>>>>
It’s a matter of taste. If you know the directory layout already, than its much quicker to go directly to the directory you want than digging through a file hundreds of lines long. If you don’t know the layout already, its easier to just search through one file. Most experienced sysadmins should know the layout, so I don’t think it would be an issue.
I wish all major distros would settle on a specific kernel version for atleast a few months to a year, instead of each having different ones. Wouldn’t this solve a lot of problems by avoiding compiling drivers as modules can be made pre-available?
>> However, there are still other reasons to use FreeBSD. One is that it is a fully developed Operating Environment.
>>
I often wonder about it. Does the distribution model help or hurt linux? If linux was fully Operating environment, perhaps standards would be a lot easier to stick to, and developers would have a much easier time developing for the platform. But would it make life for the distributions any easier? Would they still be able to have a business? Because, when you think about it, all the dirty user-level, non-kernel and app enhancements are done by the distributions. No being a full environement could either be harming us, or fueling the growth of linux. Me wonders.
Ive been running 2.5.xx on and off on Gentoo. Most recently Ive been running off of 2.5.73 mm-sources (Andrew Morton). Performance is fantastic. Stability is really getting good (we are still pre-beta mind you). Many things are just ridiculously faster than the 2.4 series, and I especially like the built in ALSA. So far I havent had any big problems at all. A couple of very minor glitches due to being early pre beta kernel, but certainly no show stoppers. I had some trouble with vmware 4 building its modules, but a guy over in gentoo forums hacked em to work with any 2.5.x kernel and that took care of that.
I basically see betas coming really soon and final release shortly thereafter. I think it will be another few months after the final when people have had time to refine things like drivers etc that things are really gonna be rock solid. Bring it on!!! This is a HUGE step forward for Linux.
>Yes, this is partially true. In terms of kernel capabilities, >Linux probably will be at least equal most of the time and >ahead more often than behind.
Yes, the Linux kernel is doing outstandingly well for it’s age. The sky is the limit.
>However, there are still other reasons to use FreeBSD. >One is that it is a fully developed Operating Environment.
You actually mentioned the reason I don’t use FreeBSD. I see it become very bloated and unmanageable in say, 10 years. I like Linux’ modular approach to functionality.
->Seperate kernel from tools
->Allow user to streamline kernel to suit users need.
->Allow user to select tools to suit users need.
->Allow user efficient scalability.
FreeBSD’s approach to it’s Operating Envronment reminds me so much of the annals of Microsoft Windows. Yes, it’s being a long journey from DOS to Windows 3.1 to XP. A long voyage of bloat. However, I prefer the overall management and unition of FreeBSD over GNU/Linux. And integration does have it’s benefit.
>It has both source based installation and binary packages >installation.
Linux users have access to both source based and binary packages, irrespective of distros.
>No Linux distribution can match it.
Gentoo is the FreeBSD equivalent of Linux. FreeBSD has been matched.
>RedHat is too commercial and only supports a handful of >packages.
That’s the point. It is intentionally designed to be commercial. In the Linux world, each distro serves its purpose. Welcome to the world of choice.
>Debian is too outdated, the 2.8 kernel will come out before >they release a 2.4 kernel based distro.
Agreed, but Debian is perhaps the most reliable Linux distribution. It is respected for that.
>Gentoo is too source based to be considered in the >enterprise.
Better put, Gentoo is ahead of its time. When 5 GHz become the norm, say 5yrs from now, I wonder if you’d prefer to compile your packages from source, or you’d prefer and alien do that for you. Besides, Gentoo also supports binary based packages.
>Another reason to use FreeBSD is that it is BSD based >rather than System V based. It has two run levels, full and >single user. Whenever you need to modify the system, >drop into single user mode, modify it, and then kick it back >into full mode.
Yes, all unices I know of use multiple run levels. Some, like Linux, are even more advanced than two run levels. Linux has a run level for shutdown, reboot, graphical interface, console, console with no network and single-user mode. Much more flexible and obviously more sophisticated than a two coined run level. Other unices have more.
>Furthermore, the majority of the system configuration are >held in one file so you don’t have to go looking through the >/etc directory looking for one or two line files.
All of the system configurations are held in one directory /etc. This, I believe, is universal among most Linux distributions. Frankly, I prefer that to long winding file.
When it comes to the Unices, it is a matter of personal preference, but I can confidently assert that GNU/Linux is advancing the fastest in the OSS world and is silently considered its Leader.
Regards
Mystilleef
Insofar as FreeBSD goes, I’m certainly not one of those anti-bsd folks, but I am glad to see Linux is now closing the gap if not surpasing in performance areas. Naturally there are a multitude of reasons to use BSD of Linux, mostly over taste. I rather like BSD myself, I use NetBSD on my old sparc – hardware that Linux sucks on, from what I hear. But my prerence remains Linux, mostly because it was my first Unix experience more then anything else. Oh, and I vastly prefer the Linux method of kernel compilation the BSD’s – text files?!? Ick, it takes me much longer to go through two config files (GENERIC and LINT) to see what I need, what I don’t and hope I get the syntax right then just to run through menuconfig or xconfig and very quickly choose what I need and get on with it. Preference, is all.
Lots of folks will continue to prefer BSD, thats dandy – any open source o.s. is fine by me.
URL for the mm patch? (cannot find it on kernel.org)
Just some points, my original post on reasons to use FreeBSD were purely preferential rather than technical reasons. I think anyone would be hard pressed to find any technical reasons now not to run Linux with the 2.6 kernel. That being said, I would like to make some comments on some of the follow-ups.
“RedHat is too commercial to be used in the Enterprise?”
No, RedHat is perfect for “Enterprise” businesses. However for small business and home use, it is becoming too expensive. While, it is still free for download now, I do not wish to constantly play the upgrade game. And I don’t trust RedHat’s intention. Also, it looks like RedHat is trying to take away what makes Unix fun by adding a whole bunch of Windows junk/logic/feel.
“SysV has 5 runlevels. What’s your point? ”
You made my point. Most people almost never change to different runlevels. In my opinion, having 5 runlevels actually creates an extra security risk because now people have to audit what services start on each runlevel.
“Gentoo also supports binary based packages.”
I did not know this. Maybe I need to look into Gentoo more.
. Do they have precompiled packages on their ftp server. It my opinion, it makes sense to compile some things, doesn’t make sense to bother with other things. But I definately like the fact that FreeBSD gives you the option.
“Lots of folks will continue to prefer BSD, thats dandy – any open source o.s. is fine by me.”
Agreed. I personally don’t care if it is open or closed source as long as there is competition in the market.
*RedHat rant*
So then don’t use it, RH!=Linux!
“SysV has 5 runlevels. What’s your point? ”
You made my point. Most people almost never change to different runlevels. In my opinion, having 5 runlevels actually creates an extra security risk because now people have to audit what services start on each runlevel.
Oh, that certainly is confusing… hey you just talked about advantages of keeping most configuration in ONE single text file. That’s not supposed to be more confusing?
If you can’t memorize 5 runlevels, and have no will to look them up in man – then it’s up to you only to remember multi-gui and single mode.. the rest won’t get in your way.
…It my opinion, it makes sense to compile some things, doesn’t make sense to bother with other things. But I definately like the fact that FreeBSD gives you the option.
And so does both gentoo and debian.
I find it perfectly acceptable that you like fbsd better, it IS a good OS indeed. But please don’t come here ranting about advantages that fbsd doesn’t even have over linux.
“So then don’t use it, RH!=Linux! ”
Well, my whole original point was that I’ve tried many Linux packages and none of them satisfy me. I personally like Debian because it has support for about 10,000 packages, but it is too old and there is no reason why it should be so old.
“Oh, that certainly is confusing… If you can’t memorize 5 runlevels, and have no will to look them up in man ”
Read before posting away… I said 5 run levels is pretty useless. And having 5 run levels is a security risk because you have to audit all the run levels. I didn’t say it was “confusing”. Where in my whole article did I mention anything about anything being confusing?
“Gentoo also supports binary based packages.”
I did not know this. Maybe I need to look into Gentoo more.
. Do they have precompiled packages on their ftp server. It my opinion, it makes sense to compile some things, doesn’t make sense to bother with other things. But I definately like the fact that FreeBSD gives you the option.
AFAIK, only some major packages are precompiled, like Xfree, OpenOffice, Mozilla… However, they ain’t as up-to-date as the other ones. Gentoo still have some work to do with portage (especially with the uninstallation… the “stable” branch ins’t that stable either), but I think it’s getting better and better.
http://www.kernel.org/pub/linux/kernel/people/akpm/patches/
Personally I prefer it over Linux. I’m not anti-linux, however, I like the fact that everything is developed under the same roof in co-operation with each other to ensure things work smoothly.
As for FreeBSD 5.2, IMHO, they should want till January/February 2004. I know that is a VERY big release gap, however, they really need to get 85% of the code multi-threaded. BusDMA development needs to be put into full forward motion.
I would much rather have a late operating system than an early and feature incomplete release. If that means I have to wait another 7-8 months, I am willing to wait. As the saying goes, good things take time.
This goes for the Linux kernel as well. There is no use trying to rush it out at the expense of security of stability. Windows XP is the prime example of this when release dates take priority over fixing bugs and security holes.
I’m not anti-linux, however, I like the fact that everything is developed under the same roof in co-operation with each other to ensure things work smoothly.
Yeah, I think that’s one of the best strength of FreeBSD, and probably the major weakness of Linux. I think choice is great, but too much is like not enough. IMO, there’s a lack of focus in the community.
and then, the kernel does not compile (2.5.74)
I’m running Slack 9 -> kernel compiles fine without it
Hmmm….
I’m yet to see strong reasons why FreeBSD is a better operating system than GNU/Linux is. All the proposed reasons have been disturbingly cosmetic.
‘I want two run levels not 5’. ‘I want one configuration file not a directory’. ‘I prefer an operating system not a kernel and tools’. ‘Umm… tools developed under one roof work better than those developed under roofs different from the kernel.’
I’m perplexed as to how the above-quoted complaints affect the performance of either operating systems over the other. How any of the above further translates into a GNU/Linux weakness bewilders me more.
If I was to mention a sphere where GNU/Linux clearly surpases FreeBSD, I’d probably debate on hardware support. Yes, it’s a cog in the advancement of FreeBSD as compared to GNU/Linux. And it is also a technical disadvantage against it.
I’m not against a healthy debate on which operating system is arguably better from a technical and logical perspective. I just do not condone individuals shoving their personal preferences and tastes down my narrow throat, like we all don’t have ours.
Regards,
Mystilleef