LinuxQuestions.org
Review your favorite Linux distribution.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Slackware
User Name
Password
Slackware This Forum is for the discussion of Slackware Linux.

Notices


Reply
  Search this Thread
Old 01-16-2008, 03:00 AM   #1
boler
Member
 
Registered: Jan 2008
Location: London
Distribution: slackware64-current
Posts: 46

Rep: Reputation: 16
Slackware vanilla kernel philosophy?


Is the kernel released in Slackware distributions a pure vanilla kernel with no modifications to it?

I know that the majority of the well known distros apply kernel hacks but I don't understand the reasoning. Surely it would be a better philosophy for every distro to use exactly the same kernel for reasons of simplicity and standardisation (or am I being short sighted)?

Thanks, Tim.
 
Old 01-16-2008, 03:51 AM   #2
arubin
Senior Member
 
Registered: Mar 2004
Location: Middx UK
Distribution: Slackware64 15.0 (multilib)
Posts: 1,349

Rep: Reputation: 75
Yes

and

IMHO yes
 
Old 01-16-2008, 04:23 AM   #3
H_TeXMeX_H
LQ Guru
 
Registered: Oct 2005
Location: $RANDOM
Distribution: slackware64
Posts: 12,928
Blog Entries: 2

Rep: Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301
Yes it is a vanilla kernel.

Yes, in fact many of the problems other distros have come from the fact that they patch the crap out of the kernel. Why ? I don't know. Probably to add a new feature or a hack or just cuz they think they know better than Linus T.
 
Old 01-16-2008, 06:51 AM   #4
duryodhan
Senior Member
 
Registered: Oct 2006
Distribution: Slackware 12 Kernel 2.6.24 - probably upgraded by now
Posts: 1,054

Rep: Reputation: 46
I don't see anything wrong with them patching a kernel.

Linus T is not the last word or GOD . People have gone into a lot of hardwork to make their distros and if they think they can make it better then so be it. Don't many slackbuilds apply patches to the source of programs? (not the same thing I know but still ...)

applying patches to kernel is not child's play .. ppl do it only if they think its really necessary.
 
Old 01-16-2008, 06:58 AM   #5
H_TeXMeX_H
LQ Guru
 
Registered: Oct 2005
Location: $RANDOM
Distribution: slackware64
Posts: 12,928
Blog Entries: 2

Rep: Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301
Quote:
Originally Posted by duryodhan View Post
applying patches to kernel is not child's play .. ppl do it only if they think its really necessary.
If that were true I also would not have a problem with patching kernels (the kind that distros do). However, the patches they apply do NOT typically add more stability or security (which and the typical functions of a patch), they usually just try to implement new features and don't do a very good job of it. I don't even think they should be named patches, they should be named 'mods' or 'hacks'.
 
Old 01-16-2008, 08:01 AM   #6
mcnalu
Member
 
Registered: Dec 2006
Location: Glasgow, UK
Distribution: Slackware current
Posts: 423

Rep: Reputation: 73
Am I correct in thinking that installing propriety ATI or nVidia drivers
involves a hack/patch to the kernel?

I felt like a naughty schoolboy when the "...taints kernel" whizzed by during bootup.

Andrew
 
Old 01-16-2008, 08:32 AM   #7
duryodhan
Senior Member
 
Registered: Oct 2006
Distribution: Slackware 12 Kernel 2.6.24 - probably upgraded by now
Posts: 1,054

Rep: Reputation: 46
mcnalu : Yes.

and ofcourse they are mods/hacks. So what ? You had that same thing when you needed ntfs-3g and older kernels with no fuse/old fuse (well not exactly, similar). Hibernate / suspend support was only just added to the kernel. People used other software for that. And you know the whole deal with ATI/nVidia drivers. Everywhere people are patching the kernel.

I am saying that many of the decisions by linus are worth contending, but he made the damn thing so he has every right to do whatever he wants. Similarly, the other distros have every right to do whatever they want
 
Old 01-16-2008, 09:23 AM   #8
H_TeXMeX_H
LQ Guru
 
Registered: Oct 2005
Location: $RANDOM
Distribution: slackware64
Posts: 12,928
Blog Entries: 2

Rep: Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301Reputation: 1301
The ATI and nvidia proprietary drivers have caused some controversy because they do 'taint' the kernel. But, they are NOT in violation of the GPL v2 because they don't modify kernel code. They are however a part of the kernel once installed, that's why they 'taint' it. I don't consider that a patch or hack either.

Quote:
Originally Posted by duryodhan View Post
I am saying that many of the decisions by linus are worth contending, but he made the damn thing so he has every right to do whatever he wants. Similarly, the other distros have every right to do whatever they want
yes, but it makes them less stable. I can vouch for this from personal experience, for example when I was using Fedora Core (it's quite unstable as you may know), it was recommended to me by someone that compiling a vanilla kernel will solve many instabilities, while at the same time removing some features. I did as they said and it suddenly became much more stable. Was it just that the official kernel was not compiled properly or that it had tons of useless patches that destabilize the kernel. It's true that many of the kernel modules that come with the kernel are unstable themselves, so really the smaller and more exclusive a kernel that you compile, theoretically the more stable it is (also excluding modules known to be unstable).
 
Old 01-16-2008, 12:03 PM   #9
duryodhan
Senior Member
 
Registered: Oct 2006
Distribution: Slackware 12 Kernel 2.6.24 - probably upgraded by now
Posts: 1,054

Rep: Reputation: 46
Quote:
so really the smaller and more exclusive a kernel that you compile, theoretically the more stable it is (also excluding modules known to be unstable).
you said it!

Thats true for slackware too btw. Nothing gets my system as charged up as a slimmed down kernel.
 
Old 01-16-2008, 02:40 PM   #10
jong357
Senior Member
 
Registered: May 2003
Location: Columbus, OH
Distribution: DIYSlackware
Posts: 1,914

Rep: Reputation: 52
That's a good point and I wish it was brought up more. With the 2.4 kernels, I used to be able to get it down to 600kb.. The 2.6 kernels around 1mb... The more crap you can remove and specific tweaking you do for your machine, the better it will run. Period.

Only build in what you need and only build the modules you use. I'm still trying to decide which is better, monolithic or modular..

I've spent 3 hours before in 'make menuconfig' refining kernels... It really is well worth the effort.

There is something about a pure vanilla system, in general, that I've never been able to put my finger on. Basically it just radiates stabiltity, if you know what I mean... Well, I'm sure you guys do, that's probably why you use slackware.

The number of patches you'll find in Slackware FTP are miniscule compared to what you'll see in Rawhide. THAT is the number one reason why Fedora gives you an uncomfortable feeling. That 'bastardized' kind of feeling that you associate with instability...

That's my 2 cents even tho it sounds a little crazy. You Slackers know what I'm talking about tho... Try delving into DIY Linux... Your end result puts out that same wierd 'vibe' as Slackware only more so. Very strange, I've always thought, but makes sense... It's certainly not 100% psychosomatic from spending the time doing it yourself, altho I will agree that it may play a part..

Last edited by jong357; 01-16-2008 at 03:00 PM.
 
Old 01-16-2008, 03:47 PM   #11
gargamel
Senior Member
 
Registered: May 2003
Distribution: Slackware, OpenSuSE
Posts: 1,839

Rep: Reputation: 242Reputation: 242Reputation: 242
Quote:
Originally Posted by H_TeXMeX_H View Post
If that were true I also would not have a problem with patching kernels (the kind that distros do). However, the patches they apply do NOT typically add more stability or security (which and the typical functions of a patch), they usually just try to implement new features and don't do a very good job of it. I don't even think they should be named patches, they should be named 'mods' or 'hacks'.

H_TeXMeX_H, that's a bit unfair. Not all other distributions are unstable. Eg, SuSE is used by some of the largest ISPs in the world, with the SuSE default kernel including SuSE's patches. As I subscribed to one of these, myself, I can tell that the availability of their service is excellent. No signs of instability.

Of course, it is true, that every modification of something that is known to work, bears the risk to make it less stable. But not all "vanilla" kernels showed the same degree of stability. Vendor patches can help to remove vulnerabilities, increase scalability and stability, improve the performance of the overall system, or support new hardware that is not yet supported by the standard kernel. And: Many of the patches you find in SuSE are backports from the development version of the standard kernel. As SuSE usually tests thoroughly, this helps more than it hurts, IMHO.

Eg, SuSE has contributed a lot to the kernel to make SAP's ERP software run on Linux, which was an important milestone for Linux, as from that moment on it got "management attention".

You may argue, that you are not interested in business use cases for Linux, but as a matter of fact, it helps all Linux users, that companies large and small contribute to the Linux development, by paying for consulting and releasing the results under the GPL2, and offer jobs to people with Linux skills.

Coming back to "vendor patches vs. stability": In many cases the patches added by the distributors are integrated into the standard kernel. To the benefit of all Linux users, including us Slackers. Therefore I think it is quite ok, that those who drive the Linux development make their achievements available to the public as and when it seems appropriate to them, even it is in advance of the patch being integrated into the standard Linux kernel.

I have my reasons for using Slackware, and one of them is that it avoids distribution specific patches. While I haven't found any instabilities in SuSE kernels, I found some incompatibilities in some command line programs, that made it impossible to compile a ROCK Linux target, eg, on SuSE (years ago). On Slackware this just worked. But it had nothing to do with the kernel.

However, even Slackware isn't all purity, here: As Slackware doesn't support LSB completely, and doesn't follow the file system structure used by most "modern" distributions, some packages, in particular for programs offering services, are patched for Slackware, in order to make them installable, at all. Has anyone noticed any instability due to this?

Let's face it, and be fair: Slackware is kind of a parasite in the Linux universe. It doesn't drive the technological advancement of Linux, and it doesn't contribute lots of patches. But it uses what others contribute.

Instead of bashing "the others", IMHO it would be more appropriate to honor their efforts, as without them Slackware wouldn't be what it is: Simply the best!

gargamel
 
Old 01-16-2008, 09:53 PM   #12
folkenfanel
Member
 
Registered: Sep 2004
Location: formerly Fanelia and Zaibach
Distribution: Slackware-current !
Posts: 342

Rep: Reputation: 59
Talking

Quote:
I've spent 3 hours before in 'make menuconfig' refining kernels... It really is well worth the effort.
Only 3 hours? I must have spent about 5 days playing with that.

Yes, I kinda enjoyed it.
 
Old 01-17-2008, 12:47 AM   #13
Alien_Hominid
Senior Member
 
Registered: Oct 2005
Location: Lithuania
Distribution: Hybrid
Posts: 2,247

Rep: Reputation: 53
Interview with Patrick Volkerding

Slackware has done a lot of work - it offers new x-windows, gnome 1.4 and kde 2.2.2 etc.. what will be next steps, eg. new kernels 2.4.14/2.2.20, do you think you (slack) will be possible to offer patched kernels (as choice)?

I'll offer the newest tested kernels that I can, but except in rare cases the Slackware policy is to not patch the kernel source Linus delivers.
Personally, I think it's not good to roll out kernels that contain lots of untested patches to get additional hardware support or features. People are going to be using these kernels on production machines, so it's better to wait until they're actually approved in the standard kernels.


EDIT: Link fix.

Last edited by Alien_Hominid; 01-17-2008 at 10:58 AM.
 
Old 01-17-2008, 01:07 AM   #14
rvdboom
Member
 
Registered: Jul 2007
Distribution: Slackware
Posts: 235

Rep: Reputation: 30
Honestly, I think Pat's decision has a lot to do with keeping maintainance of kernels reasonable enough for himself. Going out after patches and drivers, making them compile and run stable with the stock kernel is not fun and can consume a lot of time. For instance, I'm eagerly waiting for the 2.6.24 kernel to get the rt2X00 wifi drivers in it, as I often have issues in finding a proper version of this driver at each new kernel release. I'd like madwifi to be in it too. Currently, I need to compile those by myself, and it's not much fun.
And this is for drivers. Patches that go into the internals of the kernel should be avoided in any case.
 
Old 01-17-2008, 01:26 AM   #15
acummings
Member
 
Registered: Jul 2004
Distribution: Slackware
Posts: 615

Rep: Reputation: 50
Doesn't CentOS and Debian Etch do some sort of additions to the kernel?

I've had no stability probs with either of those two distros.

I've yet to run CentOS 5.0

But I do have a CentOS 4.X box (it's now update/upgraded to 4.6).

When CentOS 4.0 first came out, I installed it on a friend's box. They use it regularly as a desktop machine on the internet.

It's now going on way way upwards of two years and it's only had security updates (I use yum) and the updates that come from CentOS mirror server. Five or six version changes -- from 4.0 to now it is 4.6 (who knows how many kernels through all of that).

My friends find it hard to believe that the OS has not yet ever frozen (they're used to Win XP). *Very* rarely an desktop app gets stuck, needs killed then start again.

--
Alan.
 
  


Reply

Tags
kernel, vanilla


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off



Similar Threads
Thread Thread Starter Forum Replies Last Post
What is Vanilla kernel ? suriyamohan Linux - General 8 08-07-2008 09:34 AM
Red Hat kernel customizations compared to vanilla kernel Z98 Red Hat 1 09-16-2007 02:27 PM
Vanilla Gnome 2.18.3 for Slackware 12 jong357 Slackware 166 08-29-2007 03:01 PM
Vanilla kernel RichardSimmons Linux - General 2 03-21-2004 12:25 PM
slackware 8.0 vanilla 2.4.18 ethernet problems bluskyambition Linux - Networking 8 08-24-2003 12:12 PM

LinuxQuestions.org > Forums > Linux Forums > Linux - Distributions > Slackware

All times are GMT -5. The time now is 11:58 PM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Open Source Consulting | Domain Registration