Hi,
can someone please the differences between the original kernel Aptosid (slh kernel) and the Liquorixkernel explain? Where are the significant differences?

I would be interested in a comparison test of the two kernel, if there is one.

bluelupo
Back to top

I wouldn't call the slh kernel 'the original kernel', it's just another kernel out there that a few people use, made by some guy out there in the ethersphere. Very few use it now, judging by the distrowatch stats, by the way.

The liquorix kernel is certainly striving to be faster, smoother, and more focused on desktop optimizations than standard kernels, as well as striving to support as much normal user hardware as possible. Sometimes it has issues when damentz is trying new patch sets that trigger issues on other users' systems, so it's a bit more experimntal/cutting edge than the slh stuff.

Other differences I've noted is that damentz will actually respond to bug reports even if they are triggered by some non free thing, whereas slh will ignore bugs if the trigger is a non free item, even if the bug itself is easily fixed or diagnosed. Just look at these recent forum threads for examples of that.

The more important differences would be between for example the current debian experimental kernel branch, that would be information that is slightly interesting to see, if one is interested in such arcana.

Just as an example, just now, as firefox was loading over a hundred tabs, I was able to enter one tab and scroll it. Never seen that before. As I was typing this to you, another application was running at 100% cpu, yet I barely noticed any hang at all doing something else.
Back to top

Going back in history, Liquorix was an early fork of slh's kernel. Now it only shares parts of the kernel package... everything else is radically different.

The one major difference is the priorities of the kernel. Liquorix' first objective is to provide a desktop kernel that's both stable and fast using whatever methods reasonable and within my understanding to achieve this goal. Both patches and configurations come into play here and that's where Liquorix shines most.

Simply tuning the configuration is not enough to fix bottlenecks introduced by new kernels. Usually there's a faulty method or algorithm that causes performance problems that you can see, hear, and feel; your 3D applications are jerky, your audio pops and clicks when system load increases, the whole system feels slow when certain things happen on your system.

When things like this happen and gain enough attention, developers on the LKML (Linux Kernel Mailing List) are usually already aware and thinking of new ways to solve these problems generically. I'll try to retrieve early versions of these patches and patch them into the Liquorix source (now essentially a mirror of the Zen-Kernel sources since I'm the only active developer there at the moment). So, you can think of Liquorix as a the mmotm of the stable kernel series.

The second priority, but equally important, are the users. My initial frustration that brought me to maintaining a new kernel was that none of my arguments were taken seriously in the days of sidux (now aptosid). At some point, they simply lied in order to ignore my suggestions and recommendations. This really ticked me off and so I saved a copy of a forum thread before deleting it: liquorix.net/slh-lock-retardation/powertop-recommendation.html. This thread is kept available as a reminder to me of things distasteful and harmful to projects and communities.

Needless to say, almost every single suggestion for the kernel on sidux resulted in the same lies and deception that resulted in zero progress. Because of this, I never want to see the same thing happen to anyone who asks a question on my forum, ever. I may be stubborn sometimes and prolong an argument that sounds completely silly to you, but I will at least try to admit my mistakes at the end if I find out that I was wrong the whole time.
Back to top

Spot on, I for one can't see slh or anyone on aptosid taking *valid*
suggestions seriously and I remember that thread too. No wonder why serious users of Debian are switching to a kernel that's better and takes full advantage of performance.
Back to top

Hi Damentz...
Really appreciate the work you are doing.. Keeping abreast with the rate of change of development and tracking them, can be quite a daunting task.

I also appreciate the fact that you take time to reply on most forums personally!

I just wanted to know a couple of things:

-1-
How do you keep yourself up to date with the patches that are released on different modules, before integrating them into the liquorix kernel sources?
-2-
I am using the 2.6.38-3.dmz.1-liquorix-686 kernel on ubuntu. I know you use bleeding edge technology... so i found out that on my synaptic, the source is labelled - 'unstable/main (liquorix.net)' Do you have something like a stable version? not that what I'm using now is not stable.
-3-
Can I obtain just the sources in addition to the binary of the liquorix kernel? I was not able to find it, after searching quite a bit..
-4-
How did you get your knowledge on linux? you'd have had to start somewhere? You have something like a bio, where I can learn about your linux heritage?

Hope you'll answer these...
Regards.
Roy.
Back to top

Roy, thank you for the warm compliments! Let me answer some of your questions.

1. If you mean out-of-tree module - I typically look for new updates whenever I'm rebasing the next kernel version. This is good and bad as some authors will release updates halfway through the development cycle of the kernel in RC. This usually encompasses aufs and ck, but that means all I do is remember to periodically look for updates from either of those. Reiser4 is another, which this post just reminded me to check and merge their latest patch.

So, there are holes in my method, lots of holes, but that's because I didn't put any thought into it. Most likely, it'll be easier to create a shared document of missing modules and take a peek at it whenever I tinker around with the Zen Kernel sources.

2. Sort of, I do have a kernel in the past series, which I've been unfortunately neglecting (it's running the stable patch, 2.6.37.5), but a lot of these I'll update if someone asks me. You're busy with life, I'm busy with life too, so it's easy to put things on the back burner if no one mentions it. However, there is an exception - if I recognize that the new stable kernel is actually pretty unstable, then you'll see me updating the old stable kernel more frequently. In my opinion, this is all subjective because I've tested this on only one piece of hardware, but 2.6.38 feels very stable to me and I'd rather put all my energy into what's current rather than what's being actively deprecated on the LKML.

This is a method I read about in an OpenBSD video on youtube. It talks about keeping everyone on the newest code, all the time, reducing the necessity for a QA group, other than yourself. Now granted, you should test your patches if you get the hunch they have the possibility of introducing new instability, but otherwise, if everyone's on the newest code, there will be no sudden surprises and regressions that bubble up organically will be caught much earlier from patches that you would think are harmless.

3. There's two or three places: you can get the patch blob and configuration from liquorix.net/sources, you can retrieve the package source like any debian package using apt-get source (you can read tutorials online about that), and if you want the absolute newest code that may not be in liquorix yet, you can get what will be the newest code on git.zen-kernel.org/zen-stabe/

4. My history started in 2004/2005 where I began dual booting Kanotix so that I could ween myself off of PC games. I thought that by switching to an operating system that had less game availability, it will push me to do something more productive and creative with my life than play games all day. That sort of worked, but once I got to the point that I could play many of the games I loved to back in the day, I already lost my inertia to play games for several hours straight. Now I'm an arcade type of gamer, will just play for about 30 minutes to an hour at most and then go do something else.

Then I got involved in the Kanotix, then sidux community and helped in the forums, wrote the initial quick start guide on sidux so people know right off the back the unique things they should know about their distribution and how to maintain it. That's just an itch of mine I get whenever I try a new distribution - the unique ways to manage your distribution are usually not centralized into one quick read so get you quickly acquainted with your system. That's usually why people get frustrated when they distro hop, it's not obvious which letters they needs to jump on when they're trying to get from the letter A to Z.

I'm also a little OCD about performance. When there's an opportunity to get free performance, free features, for almost nothing, then I'll do whatever I can to get it. That's when I started building my own kernels and testing different configurations and patches. At some point I felt that mine were better than the sidux ones, but I didn't think it would be very logical for me to fork off of his and build a new one that people had to choose between. That's when I started asking questions and providing suggestions in the forums for changes to the kernels. I think all of them got ignored or debated using false information, it was kind of sad.

This is where I pulled the slh kernel package source and began tinkering with the values and understand how the source really worked. Now I use a type of fork that lets me use a branding, kind of like a stamp of quality, that lets me easily distribute it through a website (liquorix.net!).

None of this would have been possible though if h2, or Harold Hope, wasn't around. He's almost like a father figure to me; he gives deep insights, gives strong advice, and even provided a platform for me to experiment with the idea of a debian repository. Don't think it's any coincidence when he responds to any of your posts here and gives you lots of information to digest. He's a very warm character that really cares about people's needs and what they really want. I don't think smxi would exist if h2 was any less considerate.

I don't have a bio anywhere that I know of. Maybe this is it, this whole forum could be my bio, depending on how you look at it.

But hey, thanks for asking. I hope this answers your questions.
Back to top

Hey thanks for your quick & detailed reply..
yep...and it did answer my questions.. :)
I hope you continue with your work...I hear you're the only developer in this effort..
Back to top