LinuxQuestions.org
View the Most Wanted LQ Wiki articles.
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 04-07-2014, 01:44 PM   #91
Bazzaah
Member
 
Registered: Mar 2007
Distribution: Slackware64-current, Slackware64 14
Posts: 327

Rep: Reputation: 49

Quote:
Originally Posted by brianL View Post
SystemDux? Might be better all-round if P & S created that from scratch and left Linux alone.
Indeed, would that it were so.
 
Old 04-07-2014, 02:08 PM   #92
enorbet
Member
 
Registered: Jun 2003
Location: Virginia
Distribution: Slackware 14 is Main OpSys on Main PC, 2ndary are OpenSuSe 13 and SolydK
Posts: 735

Rep: Reputation: 345Reputation: 345Reputation: 345Reputation: 345
Quote:
Originally Posted by k3lt01 View Post
Why is it these discussions all go like this? Got nothing reasonable to say so someone posts analogies that are ludicrous.

The problem between systemd and the kernel indicates the kernel has a weakness, admittedly a weakness that would probably never have been found unless something like systemd come along but the thing is it has and it has shown this weakness. Now that the weakness has been shown to exist wouldn't it be prudent to fix it? Isn't fixing problems the Linux way? or are we just going to ignore this weakness in the kernel because it is systemd, and by extension systemd's creators, who through no fault of their own have shown it exists?
Linus and others on the kernel dev team stated that kernel debugging was left open to flags that they hadn't guessed at, assuming Linux people would be grateful for the flexibility and responsible for "taking out their own trash". I have yet to replicate the issue but it is my understanding that the output generated by systemd, at least in this case, is absolutely useless data, literally useless, nobody can use it. So no gain but big problem..... much like my "ludicrous" analogy of a rock allowed to stay in the channel that serves no useful function but can sink hapless boats.

What strikes me as truly ludicrous is the apologetic attitude you display toward these two who have a history of utter arrogance and denial, not to mention creating bugs they refuse to fix, blaming everyone else. Furthermore Poettering has referred to Linux as a whole and the kernel dev team as "toys" and "cliques playing with toys" as well as referring to OSX as a "real OS, not a toy like Linux" and lastly compares systemd to dependency-resolving package managers for the init system. Choose what side as you will and for whatever reasons you fancy, even call me or anyone who disagrees with you ludicrous and "nothing reasonable" (hmmmmm... why does that sound familiar?) but we shall see very soon whether you chose well or not, and we all will be judged by that in the community. Hope you considered carefully. I did.

Last edited by enorbet; 04-07-2014 at 02:43 PM. Reason: added a sentance
 
2 members found this post helpful.
Old 04-07-2014, 02:16 PM   #93
Ser Olmy
Senior Member
 
Registered: Jan 2012
Distribution: Slackware
Posts: 2,129

Rep: Reputation: Disabled
Quote:
Originally Posted by TobiSGD View Post
It was a bug and that bug is fixed.
Yes, it was a systemd bug. No-one outside the systemd development team has ever suggested or acknowledged that this could be considered a kernel bug. Claiming otherwise is disingenuous.

Rate-limiting /dev/kmsg offers protection against malicious and defective processes; it's neither a bug fix nor the addition of an obviously-useful-but-previously-overlooked feature.

Any process with root privileges can crash the system, and PID 1 can cause a crash simply by exiting. The next systemd bug that causes a kernel panic will also be a systemd bug, not a kernel bug.
 
7 members found this post helpful.
Old 04-07-2014, 03:19 PM   #94
Drakeo
Senior Member
 
Registered: Jan 2008
Location: Urbana IL
Distribution: Slackware, Slacko,
Posts: 2,658
Blog Entries: 3

Rep: Reputation: 224Reputation: 224Reputation: 224
Quote:
Originally Posted by TobiSGD View Post
Originally Posted by Drakeo View Post
Is Systemd going to become a module in the kernel if I am wrong correct me.
Code:
TobiSGD  No, it is not.
Thanks TobiSGD I always need to get my eyes together. the module loading service http://www.freedesktop.org/software/...d.service.html

Then the wish list. http://www.freedesktop.org/software/...n/systemd.html

Only one company I know personally put a broken operating system on the main communication servers and devices IBM and new that the next update would not fix it. As for my retired friends from Corning keep smiling it does not matter what system IBM uses they like to break it so only they can use it.

I am sure kernel org love big blue. And I am sure Linux Foundation goes where the money is. And I am sure Linus likes being paid every one should. So let the Dev's have fun http://www.linuxfoundation.org/news-...-power-systems But don't forget. The last Time a bicycle was actually made for the customer was 1955. After that it was all production.

Big bucks nothing is free monetarily.

Code:
In 1935 we had a 14.9 pound bike made of the finest hand crafted steel tubing in the world. It cost 350 dollars then. 
Today that bicycle cost 3,000 dollars to make. 
The Average English french bike of the 80's was 350 to 800 dollars.
 built the same way with different  high quality tubing.
Today we have a carbon fiber composite bike that cost much less to make, No Steel to roll, no need to hand braze like we do with formula racers still to this day. no putting them in a oven and taking all the molecules up to a certain temperature around 750 degrees so the molecules all settle the same. no need for a trained craftsman to follow every step and see it through. Not a big thing just your life.

Quote:
No the average $2,500.00 is a composite mesh in a mold. made so fast that the time from mold to box is about 15 min human labor. But I am sure automation has failed the consumer. There are very few Hand wrapped carbon fiber bikes it is many days of hard work. My dear Friend Jay place third ROA Race across America with his team mate. In Indiana going off the Road it broke the Top of the Line Trek bike. He used his Magnesium frame bike we put together years ago. But the fresh out of the box top of the top trek bike broke
What does this have to do with Systemd developers and a Lead kernel man that cusses like a Sailor. It is about quality and building it for the User.

Quote:
Oh Trek refused to live up to the life time warranty on the $6,000.00 dollar bike, said he was racing a racing bike.

Last edited by Drakeo; 04-07-2014 at 04:30 PM. Reason: dyslexia is a bitch some times.
 
Old 04-07-2014, 04:37 PM   #95
Smokey_justme
Member
 
Registered: Oct 2009
Distribution: Slackware
Posts: 400

Rep: Reputation: 113Reputation: 113
Quote:
Originally Posted by TobiSGD View Post
I can only repeat myself here: The bug was noticed by systemd developers and fixed before that bug report was even issued. Now I could turn around and say "So much for reading up on the issue before commenting", but let us just assume that I won't do that, because to err is human.Sorry, but no, it does not.
Funny story actually :P
https://bugs.freedesktop.org/show_bug.cgi?id=76935#c22

It seems there was a bad assertion in their code that should (again, quote) have been fixed before the bug got to the kernel devs.. And that was causing all the spamming I thought, like you did, that it was fixed before hand (read it somewhere).. It seems it wasn't.. The non-bug, this is expected behavior thing was an improper assertion that probably, if not for all this Linus debate, would have remained unfixed.. :s That's kind of scary.

LE:
Ahh, and the fun part (quote from the above link):
Quote:
Originally Posted by David Strauss
No one claimed that assertion failure wasn't a bug, and no one hesitated to fix it (see comment #17).
And the bug history of the bug report: https://bugs.freedesktop.org/show_activity.cgi?id=76935

Last edited by Smokey_justme; 04-07-2014 at 04:47 PM.
 
Old 04-07-2014, 05:22 PM   #96
ReaperX7
Senior Member
 
Registered: Jul 2011
Location: California
Distribution: LFS-7.6, Slackware 14.1, FreeBSD 10.1
Posts: 3,849
Blog Entries: 15

Rep: Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189
I'm not actually convinced there was a bug with the kernel, and before anyone says otherwise let me plead my case:

The debug flag was always set to allow an uninterrupted debugging output feedback from the kernel kmsg system. This was to allow continual debugging either to a verbose console or to a log file to track problems and bad code.

For 20 years this was never changed to where kernel debug output was ever questioned regardless of the init system. Every init system that has been around all have worked flawlessly with kmsg and debug as it was intended.

It was ONLY systemd that did so otherwise by continuously spamming kmsg. OpenRC, BSDinit, sysvinit, Runit, s6, daemontools, etc. none of them had any issues with kmsg and debug.

This leads me to believe this was not accidental, but intentional. For 20 years, one kernel developer feature remained as it was unchanged until systemd came along. This only proves that either systemd was broken by design, yet again, or the kmsg spam bug was not a bug but an intentional deliberate act to force the kernel developers to change something on their end that has never been changed for any reason.

I think, in my opinion, Linus needs to halt kernel developments completely and perform a full audit of the code to not only test this against other init systems, but equally against systemd and if systemd is found to be the only probable cause then all those involved with this action need to be immediately suspended from kernel developments and contributions until this can be investigated to prove otherwise, and if they are found guilty of attempting to tamper with kernel code, they and all involved need to be permanently banned from kernel code contributions, and then inquiries made into currently implemented code to exorcise it from the kernel and reverse out, possibly for the first time, code from the Linux kernel.
 
3 members found this post helpful.
Old 04-07-2014, 05:23 PM   #97
TobiSGD
Moderator
 
Registered: Dec 2009
Location: Hanover, Germany
Distribution: Main: Gentoo Others: What fits the task
Posts: 15,653
Blog Entries: 2

Rep: Reputation: 4095Reputation: 4095Reputation: 4095Reputation: 4095Reputation: 4095Reputation: 4095Reputation: 4095Reputation: 4095Reputation: 4095Reputation: 4095Reputation: 4095
Quote:
Originally Posted by Ser Olmy View Post
Yes, it was a systemd bug. No-one outside the systemd development team has ever suggested or acknowledged that this could be considered a kernel bug. Claiming otherwise is disingenuous.
http://lkml.iu.edu//hypermail/linux/...4.0/02645.html

At least Steven Rostedt and Theodore T'so consider that this can be seen as kernel bug.

It is pretty simple: /dev/kmesg is the earliest way (before writable storage is mounted) to get logging working, which is a good thing for both, kernel and userspace developers. Bugs can happen all the time on both sides and it is good to have an early logging feature to actually find the bugs. If this logging feature crashes the whole system because something (even by accident) is logging to it then that defeats totally the purpose of this feature and it can be considered broken (aka: a bug is present).

This by the way is not the first time that logging problems appear due to spammy sub-systems, it happened before with some buggy in-kernel drivers. Guess what happened: the logging feature got a rate limit and nobody cared.
 
1 members found this post helpful.
Old 04-07-2014, 05:27 PM   #98
ReaperX7
Senior Member
 
Registered: Jul 2011
Location: California
Distribution: LFS-7.6, Slackware 14.1, FreeBSD 10.1
Posts: 3,849
Blog Entries: 15

Rep: Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189
Quote:
Originally Posted by TobiSGD View Post
http://lkml.iu.edu//hypermail/linux/...4.0/02645.html

At least Steven Rostedt and Theodore T'so consider that this can be seen as kernel bug.

It is pretty simple: /dev/kmesg is the earliest way (before writable storage is mounted) to get logging working, which is a good thing for both, kernel and userspace developers. Bugs can happen all the time on both sides and it is good to have an early logging feature to actually find the bugs. If this logging feature crashes the whole system because something (even by accident) is logging to it then that defeats totally the purpose of this feature and it can be considered broken (aka: a bug is present).

This by the way is not the first time that logging problems appear due to spammy sub-systems, it happened before with some buggy in-kernel drivers. Guess what happened: the logging feature got a rate limit and nobody cared.
Most buggy drivers were always labeled as such Tobi, or they were assigned to the Staging drivers, given Unstable and/or Experimental status.
 
Old 04-07-2014, 05:36 PM   #99
TobiSGD
Moderator
 
Registered: Dec 2009
Location: Hanover, Germany
Distribution: Main: Gentoo Others: What fits the task
Posts: 15,653
Blog Entries: 2

Rep: Reputation: 4095Reputation: 4095Reputation: 4095Reputation: 4095Reputation: 4095Reputation: 4095Reputation: 4095Reputation: 4095Reputation: 4095Reputation: 4095Reputation: 4095
Quote:
Originally Posted by ReaperX7 View Post
Most buggy drivers were always labeled as such Tobi, or they were assigned to the Staging drivers, given Unstable and/or Experimental status.
Just because there might be a bug the drivers don't immediately get moved to a different branch, rate-limiting is necessary anyways. Or as Linus put it:
Quote:
Steven, Borislav, one thing that strikes me might be a good idea is to
limit the amount of non-kernel noise in dmesg. We already have the
concept of rate-limiting various spammy internal kernel messages for
when device drivers misbehave etc.
http://lkml.iu.edu//hypermail/linux/...4.0/01488.html
 
Old 04-07-2014, 06:29 PM   #100
saulgoode
Member
 
Registered: May 2007
Distribution: Slackware
Posts: 257

Rep: Reputation: 121Reputation: 121
Quote:
Originally Posted by TobiSGD View Post
Just to make that clear, Linus has no problem with userspace software using the generic debug option, systemd is not the only userspace software that does that. At the point that discussion came up he simply didn't know that that bug was already fixed. ...
According to the LKML archive, Linus posted on April 2nd.
Quote:
Originally Posted by TobiSGD View Post
At the time that bug was reported the bug in systemd that caused the spamming was already fixed in their git. All Kay Sievers had to do was to tell the bugreporter: "We fixed that already, please use a newer version" and marking that bug as fixed. Why he didn't is something I (and many other people) can't explain at all.
According to Bugzilla, the report was made on April 2nd.
Quote:
Originally Posted by TobiSGD View Post
I can only repeat myself here: The bug was noticed by systemd developers and fixed before that bug report was even issued.
Perhaps the systemd developers noticed the bug before April 2nd, nonetheless, according to the FDO GIT log, a fix was not committed until April 6th.
Quote:
Originally Posted by TobiSGD View Post
Now I could turn around and say "So much for reading up on the issue before commenting", but let us just assume that I won't do that, because to err is human.
Let us assume what we will.

Last edited by saulgoode; 04-07-2014 at 06:32 PM.
 
Old 04-07-2014, 08:00 PM   #101
Richard Cranium
Senior Member
 
Registered: Apr 2009
Location: Carrollton, Texas
Distribution: Slackware64 14.1
Posts: 1,578

Rep: Reputation: 468Reputation: 468Reputation: 468Reputation: 468Reputation: 468
Like many other things in life, a cooperative tasking model can get more done per cycle than a preemptive one.

Users of the kernel logging module for 20 years applied the cooperative tasking model: Don't assume that you are the only user of the resource.

The systemd users assumed otherwise.

Make of that what you will.

From my experiences in a similar environment, that would be due to untrained monkeys writing code using a framework that they did not attempt to understand. Make of that what you will.
 
5 members found this post helpful.
Old 04-07-2014, 09:38 PM   #102
ReaperX7
Senior Member
 
Registered: Jul 2011
Location: California
Distribution: LFS-7.6, Slackware 14.1, FreeBSD 10.1
Posts: 3,849
Blog Entries: 15

Rep: Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189Reputation: 1189
When you throw away all the gathered and consolidated knowledge that has been collected, and pieced together painstakingly since the beginning and trash all the standards just to say you can do it and try to do things without thinking on the consequences... this is the end result.

A thousand monkeys at a thousand keyboards...

See attachment...
Attached Images
File Type: jpg monkeys.jpg (40.3 KB, 57 views)

Last edited by ReaperX7; 04-07-2014 at 09:41 PM.
 
Old 04-07-2014, 11:29 PM   #103
Smokey_justme
Member
 
Registered: Oct 2009
Distribution: Slackware
Posts: 400

Rep: Reputation: 113Reputation: 113
Quote:
Originally Posted by saulgoode View Post
Perhaps the systemd developers noticed the bug before April 2nd, nonetheless, according to the FDO GIT log, a fix was not committed until April 6th.

That's not the bug fix.. That's only the change for the default target for logging after the early boot phase.

Last edited by Smokey_justme; 04-07-2014 at 11:32 PM.
 
Old 04-08-2014, 03:16 AM   #104
k3lt01
Senior Member
 
Registered: Feb 2011
Location: Australia
Distribution: Debian Wheezy, Jessie, Sid/Experimental, playing with LFS.
Posts: 2,899

Rep: Reputation: 623Reputation: 623Reputation: 623Reputation: 623Reputation: 623Reputation: 623
Quote:
Originally Posted by Ser Olmy View Post
To those who keep claiming this was somehow a kernel bug, here's a car analogy that works pretty well
I'm not even going to start where that analogy is flawed, let me just say there are many and you really shouldn't use cars as an anology unless you have more than a basic idea about them.
Quote:
Originally Posted by genss View Post
it is really easy to break the kernel
i broke it yesterday by playing with alsa
is it an alsa-s flaw that i am an idiot ?

to compensate for what i did would make alsa less efficient as it would have to check things a lot,
and what i did is very much bad practice
so should i report it ?
2 completely different scenarios. If you stuffed your system playing with Alsa, knowing you shouldn't, that is already part of the system then it's your fault. If systemd showed a flaw in the kernel before it has been finalised then it appears the kernel needs a bit of patching to ensure the fault does not rear its ugly head again.

Quote:
Originally Posted by enorbet View Post
So no gain but big problem..... much like my "ludicrous" analogy of a rock allowed to stay in the channel that serves no useful function but can sink hapless boats.
Which again begs the question why even make the analogy?

Quote:
Originally Posted by enorbet View Post
What strikes me as truly ludicrous is the apologetic attitude you display toward these two who have a history of utter arrogance and denial, not to mention creating bugs they refuse to fix, blaming everyone else.
Crank up the dictionary again Apologetic. Let's get this perfectly clear, at no stage have I shown an apologetic attitude for "these two" as you call them. What I have done is displayed an attitude that reflects the ideals of FOSS. If something breaks or has a problem then fix it, simple. Apart from that I have not gone and said either side in this arguement is right or wrong. I agreed with Tobi and that was it. If you have a problem with a regular Linux user suggesting a problem gets fixed then guess what, it's your problem not mine.

Quote:
Originally Posted by enorbet View Post
Furthermore Poettering has referred to Linux as a whole and the kernel dev team as "toys" and "cliques playing with toys" as well as referring to OSX as a "real OS, not a toy like Linux" and lastly compares systemd to dependency-resolving package managers for the init system.
I do not deny the man has a problem but I'm not going to go on a "rantage" about it. I'll leave that to you.

Quote:
Originally Posted by enorbet View Post
Choose what side as you will and for whatever reasons you fancy, even call me or anyone who disagrees with you ludicrous and "nothing reasonable" (hmmmmm... why does that sound familiar?) but we shall see very soon whether you chose well or not, and we all will be judged by that in the community. Hope you considered carefully. I did.
You're telling me to make a choice or be judged by the community! People are dying on this planet and countries are being pulled apart, possibly leading to another Eurpoean war and you're carrying on like this is the last choice anyone will make. Sorry but you've just shown yourself to be an extremist. I have no time for that.
 
Old 04-08-2014, 03:23 AM   #105
jtsn
Member
 
Registered: Sep 2011
Location: Europe
Distribution: Slackware
Posts: 908

Rep: Reputation: 447Reputation: 447Reputation: 447Reputation: 447Reputation: 447
Could this thread be moved out of the Slackware sub-forum? I don't see, where this is relevant to Slackware. Slackware doesn't use Systemd and is not affected by this issue. The "debug" kernel parameter works just fine and dandy.
 
5 members found this post helpful.
  


Reply


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
LXer: Linus on Linux: The Linus Torvalds Interview Part 2 LXer Syndicated Linux News 0 04-27-2009 02:00 PM
LXer: Linus on Linux: The Linus Torvalds Interview Part 1 LXer Syndicated Linux News 0 04-23-2009 04:10 PM
Linus Torvalds rakris General 1 11-14-2006 03:07 AM
Linus Torvalds divorce3 General 8 02-10-2005 02:34 AM
Linus Torvalds linuxzouk General 7 05-16-2004 11:17 PM


All times are GMT -5. The time now is 07:39 AM.

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
identi.ca: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration