Linux Kernel Developers Fed Up With Ridiculous Bugs In Systemd

[DeepDayze]

This whole tug of war between systemd and kernel devs has become quite hilarious especially when Father Linus has to jump in. I am also in agreement with a better way to keep each subsystem's debug switches separate and not stomping all every thing else.

Seems these devs have contracted spring fever thus this latest hoo ha

[Akka]

This whole tug of war between systemd and kernel devs has become quite hilarious especially when Father Linus has to jump in. I am also in agreement with a better way to keep each subsystem's debug switches separate and not stomping all every thing else.

They can already do that. The discussion is what the generic debug should do. As I understands it the systemd people think it should be for end user that debug boot. The kernel devs think it should be for kernel developers becouse that has traditionally been the usercase.
Logically I think the systemd dev is right. Its logical debug is a generic debug option and systemd.debug and kernel.debug is used for the specifik case of debugging.

[DeepDayze]

They can already do that. The discussion is what the generic debug should do. As I understands it the systemd people think it should be for end user that debug boot. The kernel devs think it should be for kernel developers becouse that has traditionally been the usercase.
Logically I think the systemd dev is right. Its logical debug is a generic debug option and systemd.debug and kernel.debug is used for the specifik case of debugging.

Yes, using something like systemd.debug would only set debugging mode on for systemd and not the kernel and this sort of namespacing should help reduce confusion on what subsystem you want to debug while leaving the others alone.

The kernel is not the only thing to debug, there's other areas that can be debugged too such as drivers, systemd, etc.

[Nobu]

Because your estimates are significantly off.

On the musl+tinyX11 system I'm running, libc is 716k, libssl is 372k, libjpeg is 416k, libX11 is 248k, libpng is 188k, libpcre is 148k, libgpm is 144k, libexpat is 136k, libz is 104k, and most of the X libs are less than 100k. (I'm using size based on du |sort -rn, which is very approximate.) libXm, libstdc++, and libcurl are the only libs over 2 megs, and libiconv and libcrypto are the only other libs over 1 meg.
I have a static busybox that has almost everything enabled, weighing in at 808208 bytes. That includes busybox httpd and ftpd, along with bits of runit.
And a static Xfbdev is smaller than that.
dropbear, which is an SSH server and client, is ~140k.

So I could have runit for init mangement, an X server, SSH, and a nearly complete Linux commandline environment in ~2 megs.

Meh, if you want small, there's DSL, Puppy, and Tiny Core Linux. If you want smaller, you are probably custom making your own distribution. Either way, you don't have to deal with systemd, and if the smaller distros eventually switch to systemd (I doubt it'll be soon, as they have something that works and is small now), I doubt they'll use every feature of systemd anyway.

To each his own. (I had a longer post, but firefox decided to lock up. :/ )

[ihatemichael]

Yes, using something like systemd.debug would only set debugging mode on for systemd and not the kernel and this sort of namespacing should help reduce confusion on what subsystem you want to debug while leaving the others alone.

The kernel is not the only thing to debug, there's other areas that can be debugged too such as drivers, systemd, etc.

What with the programs parsing 'debug' from /proc/cmdline and turning debug that way?

Doesn't systemd do that atm?

[profoundWHALE]

Meh, if you want small, there's DSL, Puppy, and Tiny Core Linux. If you want smaller, you are probably custom making your own distribution. Either way, you don't have to deal with systemd, and if the smaller distros eventually switch to systemd (I doubt it'll be soon, as they have something that works and is small now), I doubt they'll use every feature of systemd anyway.

To each his own. (I had a longer post, but firefox decided to lock up. :/ )

I'd recommend Crunchbang for a fast and light distro It's my favourite at least.

[Guest]

WARNING! This post isn't meant to be serious, don't react to it with actually useful or insightful replies #YOUVEBEENWARNED

Different Points of View:

1. PoV of user:
-- hey, this should be fixed, lets report on systemd / kernel bugzillas/mailing list

2. PoV systemd developer:
-- its true that my code spams machine to the point of halting, but it's linux fault

3. PoV of kernel developer:
-- not my problem, your code is crap, dont bother me

4. PoV of (the) kernel maintainer:
-- oh f*ck, this f*cking Kay Sievers guy AGAIN? I better make statement for tabloids so he leave us alone for a while.

BEST PROGRAMMERS ITW (or at least most well known due to their scandalous natures).

Lennart Poettering && Kay Sievers corp a.s.

• creator of systemd and pulseaudio (possibly more)
• #dedication #ifitdoesntworkyouarenotcodinghardenough
• #prophet
• "upstream first, but if we can't control upstream, we better make sure WE ARE THE UPSTREAM"
• #worlddomination

brought upon this world pulseaudio, nearly 10 year old project that is only kinda usable for last year or so, since it was buggy beyond usable and one might argue also harming (making people go insane and their hairs falling out)
Once his evil deeds were done, he focused on world domination via 'pid 1', init system, that is just slightly better than all the alternatives which gives lennart and sievers license to kill (or extend systemd by adding some core daemon that can't be removed and replaces something worthless, like network manager - systemd-networkd).




J?rg Schilling

• "star is better than tar"
• "star does that since 1919"
• "this is already implemented in star"

the star creator "An enhanced (world's fastest) tar, as well as enhanced mt/rmt", he comments on every issue in gnu/tar mailing list with 'nya nya nya, star already does it since 2001' attitude.

Linus Torvalds

• "f*ck you! (nvidia)"
• "what you are doing is mental masturbation"
• "nya nya nya, how about pulling from me"

-has ego of a small planet ond goal of world domination. We all know and <3 Linus.


RMS

• #theonewhoshallnotbenamed
• enforcing freedom (it's not oxymoron in his case)
• evil corporations are out to getcha, hells bells, yeeee

Evil corportations are out to get your freedom, so we will ENFORCE freedom in all the software you write, you can sell it, but it's free already.
Freedom is worth more than feeding your children.




Forgot any infamous programmer??? (sure I did :-)

[profoundWHALE]

The simplest way that I can understand systemd is: systemd targets compatibility with the Linux kernel and stuff, and everything else targets systemd. In the end we get excellent performance and compatibility across distributions.

[ryao]

The simplest way that I can understand systemd is: systemd targets compatibility with the Linux kernel and stuff, and everything else targets systemd. In the end we get excellent performance and compatibility across distributions.

Given that the software systemd starts is supposed to work on several platforms, that is not happening. Anyway, not much will ever require systemd. gdm is a rare exception.

[NotMine999]

Talk About &quot;head in the clouds and no feet on the ground&quot;

The simplest way that I can understand systemd is: systemd targets compatibility with the Linux kernel and stuff, and everything else targets systemd. In the end we get excellent performance and compatibility across distributions.

You see a rather idealized world don't you?

Have you ever considered this fact: If/when every Linux distribution is using "systemd" by default, the uniqueness of Linux distributions will disappear?

What do you have then? Well, you have that now and just don't see it do you? It's called "Microsoft Windows". What choice will we have then?

Redhat and who else? SuSE? Ubuntu? I think that Debian will hang on due to it's developer community. Gentoo will survive just so it can quietly annoy the others just by it's very existence.

In your idealized world we can expect many Linux distributions to fall into the depths of obscurity because there will be nothing to differentiate them...other than a fancy GUI...and that is due to the fact they will all be very very similar "underneath" since "similarity == compatibility".