Announcement

Collapse
No announcement yet.

Bcachefs Looks Like It Won't Make It For Linux 6.6

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • blackiwid I don't see anything unclear about Linus's reply.

    I've read your comments in other threads and it seems like not only you have no experience as software engineer but also assume that CI tests are unimportant and some smartass can by-pass that.

    Please stop spreading these FUDs and stop trying to frame as if CI test is used to stop newcomers from contributing, it's not it's there to ensure quality of code.

    My patience is wearing out by your fanboi of Kent and dismissing code reviews practices are useless, while you are utterly ignorant of how software development work.

    Comment


    • Originally posted by Quackdoc View Post

      lmao, way to represent facts in a completely ass-backwards fashion

      Firstly, who the fuck cares if christan want's it in 6.7? All he will wind up doing is saying wait for 6.8, wait for 6.9.
      First, that's not at all clear that Christain would nak again. bcachefs was for the first time ready during 6.5 (at least with the prereqs) and people wanted a cooling off period. This was after Linus sent an email saying that Kent's arguments with nearly everyone is a showstopper.

      You may not like what James wrote, but it was spot on. It was about how when anyone disagrees with him, he shifts into attacking their motivations. The analogy about little Kents is very clear. Kent is embattling bcachefs developers against core linux developers. This attitude that "screw being polite those guys are lazy morons who want to cause me pain" is clearly his attitude and it will spread to bcachefs developers, just as it's spread on this forum.

      I'm glad at least you agree that some of Kent's comments cross the line, even if we disagree about other things.

      I am a long term patreon sponsor of bcachefs and very much want to see it succeed. I think the best way to go is that Kent has to reset his relationship with Jens and Christoph, and leave the past in the past. He has to eagerly work with iomap/rwsem and the codegen/jit/execmem groups on core improvements that he would be happy with using. If people saw that he can work constructivly on these items they will be a lot happier ACK'ing his submission. Certainly I'd like to see him send an email to bcachefs and fsdevel talking about his plan at least a week before he sends Linus a PR.

      Funny thing is that if he listened to Christoph this time, and immediately submitted to linux-next instead of how he reacted, he might have been able to salvage the pull this cycle.



      Comment


      • Originally posted by blackiwid View Post
        Sorry is this text generated by ChatGPT or are you very young or is that English in easy language or something? This whole paragraph has not 1 punctuation character and who repeats the full name every time (not correctly either but that can happen).
        Its normal for me not to use punctuation at all. Reason the more punctuation I use the more likely I am to type complete sentences dyslexia scrambled. Yes sometimes when it comes to names I end up doing all full names when I am suspect I am not exactly running right. That I had typed linux instead of linus I was right this time to think I am not running exactly right. Spell check only helps so much.

        Originally posted by blackiwid View Post
        I am not sure Linus would have done bad things, he used harsh language so that some people that took that to serious from their side stopped committing, but I am not aware that he did some blocks of medium mistakes if the person was no complete moron and had no chance of ever submitting anything useful to the kernel.
        Linus has done bad things you look at what happens when he took 3 months off that case he ripped massively into a person for something they had not done. The bad language of Linus is in fact symptom of burnout Linus shows. One of the problems with burnout if you don't treat it the problem gets progressively worse.

        Originally posted by blackiwid View Post
        Which despite sugarcoding to a degree happened in this case, too. So the softer language did not help. Again people seem to mix up being (somewhat) friendly and not be direct and clear. If there is a deadline I want to reach, accepting this, I want a mail from Linus that says A? If you do A B C we can add it. Or because of A this time it will not be added because it's to late.
        Not It would be nice to do B and C... (no comment if it has a chance to get added because A is not given). You don't need to be rude to give that information.
        Linus writing polite/sugarcoating is for his own good. Not doing that lets his anger run away with him and has caused him to in past to incorrectly lash out at someone who had done nothing wrong for quite some time the Linux world excepted this as Linus being Linus that was not doing Linus any good in his home life. Emotions are dangerous for person like Linus. The number of people Linus deals with Linus can have the problem of very similar names.

        Sugarcoated does mean if it does email the wrong person the person is more likely not to be highly upset.

        Originally posted by blackiwid View Post
        And even if 80% is Kents fault, the leader of a project or leaders should feel responsible and guilty and try to better themself, self critical, that's how companies operate. Leadership did not work apparently if you want to be a dictator you better be a good leader.
        No the best company leaders in many cases are psychopaths who do not have concept of guilt. Feeling emotions over what you have done wrong increases the speed you burn out in management roles. Reality is everyone is human and everyone will do something wrong the only thing that is important is that you learn from you mistakes and don't repeat it. Never expect upper management in a company to act sorry or apologies other saying I did X wrong like Linus just did.

        Lower level management and HR is meant to care about staff emotional state this is an area where the Linux kernel development is still very weak. Not like companies or programming software training courses teach lower level management/HR skills to programmers. You check out the education history of all the Linux kernel maintainers you find only 2 that have had any lower management/HR training. So Linux kernel developers are skilled and trained in reviewing code but they are not trained is personal management.

        Like it or not the higher you are in the management system be it software or company the more dangerous to yourself having emotions is. Remember the problem a project or company run into sometimes the choice that is emotionally wrong is the choice that has to be done that the company/project survives.

        Linus Torvalds is upper management. The high level people like Linus getting the title benevolent dictator is no mistake. Dictator does not need to take stuff emotionally and many cases should not. The title is very good description of what long term upper management in companies look like.

        Yes going over your supervisors head in any big company or project is a very quick way to run into not so nice people. Not being so nice is side effect of job. The problem is always the right choice for the company or project will not always match emotional choice and the person having to make those choices over and over again comes at a mental cost.

        Yes the upper and lower management model is so that the upper management does the choices that are bad for social interaction/staff management.

        There is another problem if you cannot work out lot of the maintainers in the Linux kernel are upper management for the companies they work for with lower management under them at the companies they are at. Linus has said it self Linux kernel development is like hurding cats. Some of this is that you have everyone who is upper management is use to giving orders not receiving them most of the time and they have to get along. Upper management runs on policy and procedures more than anything else.

        To get to what would be classed in most countries for a company to have healthy staff the Linux kernel need between 2 to 3 times as many people doing the maintainership work. Please note that 2 to 3 times more staff is not adding more programmers. Its adding more technical writers to to keep on top of procedure documentation, its adding HR staff to deal with emotional issues and so on . So yes the Linux kernel is critically understaffed so you are not going to have the best interactions all the time.

        Some of the problem people go into something like the Linux kernel thinking is going to function like their work place. Of course they don't notice their work place is correctly staffed and resourced and the Linux kernel is not. Incorrectly staffed and under-resourced means you have to leave your emotions at the door and never try to rush things because if you don't things are going to badly wrong.

        Yes Linux kernel development is the classic over loaded camel that is always one piece of straw away from going horrible wrong for someone and everyone is basically crossing their fingers they are not the person who makes the mistake of placing that piece of straw. One of the straws that can start everything breaking is letting emotions come into play.
        Last edited by oiaohm; 09 September 2023, 09:03 PM.

        Comment


        • Originally posted by oiaohm View Post
          ...
          I am sorry I am very short from put you on by blacklist or whatever it's called. And after around the half I could not make myself read the rest of this rambling.

          You made some very strange comments, I reacted to your claim that he somehow banned or technically prohibited from some systems, not if language can be bad, I never made the claim that angry language can never be bad.

          I also think you either did not read what I wrote or intentionally misunderstood me to straw-man me. You make this strange either or, either you give complete useful complete information or you be friendly and sugarcoat. I never made that, I just asked myself if some strange view on sugarcoating lead to not giving fair information that a person would want or if it's some tactical sugarbread and whip tactic. To keep him doing stuff in the hopes it could be added this round even if it would not happen or if he was to cowards to say it to him which seems to be the case, that if it's not in linux-next it would not happened why not saying that clearly?

          Being clear is not being rude, being rude would be something like "you moron didn't added it to Linux-next, so fuck off that garbage get's not added" that is rude...

          Then you say such people need to sugarcode and be nice but also psychopaths?

          And where did I demand that he feels emotions about doing something wrong? I said he should be critical about himself, that has literally nothing to do with emotions at all, only if you have a big ego or better a small ego you would not separate a mistake you did from yourself and take critique from outside or inside personal.

          Especially self critique should in any healthy human not create any or big emotions, I never said he should make a sob video or something... it's a logical step my actions did give not ideal results, I should change my procedure to be more effective. A human that is literally incapable of having emotions could do that even a AI could do that with critique from outside.

          Comment


          • Originally posted by blackiwid View Post

            I am sorry I am very short from put you on by blacklist or whatever it's called. And after around the half I could not make myself read the rest of this rambling.
            oiaohm Does this almost all of the time, and it doesn't help that its written in barely legible English so most of the time its hard to make any sense of what is being written.

            Comment


            • Originally posted by blackiwid View Post
              Then you say such people need to sugarcode and be nice but also psychopaths?
              There is a catch here. Mitigation method against a person in charge going deeper into psychopath mind set is sugarcoating. This is the problem sugarcoating is not black or white. Person in a high management role writing sugercoatted response means they spent more time thinking about the response they are writing. Horrible as it sounds is a upper management mental health management thing. Lower management is normally trained to deal with it inside major companies then your normal people below them.

              Originally posted by blackiwid View Post
              And where did I demand that he feels emotions about doing something wrong?
              Where you wrote:
              the leader of a project or leaders should feel responsible and guilty
              This is emotions you are asking the leader to feel something.

              Something to remember why felling guilty is such a bad thing for those in high management roles is sod law. A stack faults in procedures can all turn up one after the and if the manager takes is personally like with guilt and responsibility this is how you end up with a high skilled manager being dead or mentally broken down completely.

              Originally posted by blackiwid View Post
              I said he should be critical about himself, that has literally nothing to do with emotions at all, only if you have a big ego or better a small ego you would not separate a mistake you did from yourself and take critique from outside or inside personal.
              Linus already this this with the my bad for not picking up that the code was not in linux-next.

              Yes being logically critical of yourself to make sure you don't repeat the same mistakes does not require having to feel guilty or responsible. First time a problem happen you are fixing a procedure error in fact you should be feeling the exact other way as a manager proud that you are fixing this error so it never happen again.

              Originally posted by blackiwid View Post
              Especially self critique should in any healthy human not create any or big emotions, I never said he should make a sob video or something... it's a logical step my actions did give not ideal results, I should change my procedure to be more effective. A human that is literally incapable of having emotions could do that even a AI could do that with critique from outside.
              Yes this is a catch any human wanting to do management and not burnout is going be using logical self critique and leave guilt and responsibility and emotions in general out of it.

              Of course companies have PR people who job is to take the emotional stripped management stuff insert emotional stuff before its exposed to people expecting emotional stuff. Linux kernel development you are seen deep inside inside a very complex company model to sections that normally would be hidden from people not trained to work in those areas.

              Of course you get people working in the Linux kernel who don't understand that since they are dealing in upper management don't be expecting the upper management to be emotional that not what they are required to be..

              Remember how I have mentioned lack of resources. If Linux core developers had unlimited resources you would not be able to disobey Linux kernel patch submission checklist because automated bot would have checked everything and would be rejecting anything that does not pass the . So kent was only able todo what he did due to lack of resources.

              blackiwid the culture of upper management in large companies is it own beast for very critical reasons yes with it own specialist parts in HR, PR and so on. When a company has spent billions dollars at times training upper management personal having them take their lives because the let emotions into their reasoning is not profitable. It does kind of worry me that in future we could have AI upper management in lots of companies because AI could be more compatible at doing it.

              Yes something wrong with Linux kernel development is due to is complexity you need upper and lower management but upper management of the Linux kernel is missing many specialist parts the should exist to smooth over issues.

              Comment


              • Originally posted by blackiwid View Post
                Being clear is not being rude, being rude would be something like "you moron didn't added it to Linux-next, so fuck off that garbage get's not added" that is rude...
                Well, Linus literally said this.
                When you scold someone's fuckup in public saying like "this is basics", you literally call him a moron. When you tell this code isn't compiling you literally call it garbage. And even "fuck of" part is present in a form "better not show up for another decade".
                The funny part is one of the most cited reason not to merge bcachefs was Kent's being ill-tempered and unable to work in a team. In a team where boss can publicly give such beating. And afterward Linus admitted his "big thing" was actually an excuse: he was asked beforehand, he haven't answered directly but have continued reviewing process what in all common sense practices means "I'm OK with it". Maybe you can call this misunderstanding and pitiful accident, but you can't pretend it was gross contempt to the rules from one side. So what was all this "You are not special, mr, Potter" speech about?
                Originally posted by blackiwid View Post
                Especially self critique should in any healthy human not create any or big emotions, I never said he should make a sob video or something... it's a logical step my actions did give not ideal results, I should change my procedure to be more effective. A human that is literally incapable of having emotions could do that even a AI could do that with critique from outside.
                That depends. You may wait a little, calm down and start again in case it was really some conflict over nothing. But what if nobody wants your filesystem? What if they just use bureaucratic nitpicking to make themselves look more welcoming than they are? You can satisfy their whims for years and will never achieve anything, moreover every failed attempt will be used against you, why to merge this garbage when it's developer is so dumb, he couldn't fix his code for 5 years despite help of kernel developers.

                Comment


                • Originally posted by jacob View Post
                  Generally speaking the idea, pervasive among many Linux folks, that desktop is somehow less important and non critical is a giant fallacy. There is a huge amount of extremely valuable data, from source code to graphics to multimedia to stuff like marketing material etc. that's only ever processed and modified on desktops. Desktops are also usually much more prone to power failures, system crashes (especially for kernel developers) etc. and that's where the integrity guarantees of a CoW filesystem really shine.
                  But speaking of "power failures" and "system crashes", there are stories floating around, about how BTRFS is vulnerable to power failures and not safe to use without UPS battery or laptop battery. People used BTRFS, power interrupted, the system reboot and found the partition refuse to mount. They file system may still be repairable but it is causing a huge loss in availability. Then fanboys defended that BTRFS is just playing "better safe than sorry" when there are any tiny risk of data loss. But then you hear victims lost a lot of files after partition repair, because BTRFS think it can't guarentee the content of those files are 100% correct, thus should discard them even for any risk of just 1-bit of corruption. Then fanboys defended you should have done regular backups for important files already. We are playing "better safe than sorry" for any tiny risk of assuming corrupted file as authentic.

                  Hearing those stories make me fear BTRFS. It makes totally no sense for me for any filesystem to lose files it is not actually writing on right at the time of power interruption. And for self-proclaimed CoW filesystem, I rightfully expect it should give me back a version of intact files no matter what unfortunate timing the power interruption occurs.

                  Comment


                  • Originally posted by billyswong View Post
                    But speaking of "power failures" and "system crashes", there are stories floating around, about how BTRFS is vulnerable to power failures and not safe to use without UPS battery or laptop battery. People used BTRFS, power interrupted, the system reboot and found the partition refuse to mount.
                    If you're going to base your choice of filesystem on anecdata, I can tell you I've been using Btrfs for ten years across three computers in a house which has relatively unreliable power (minor power failures happen about once a month), and have never lost data or time due to Btrfs failing to recover gracefully.

                    If it ever happens I'll consider it a very small price to pay for the convenience of snapshots.

                    Comment


                    • After reading everything, I think everyone here is somewhat right about their opinions.

                      Linux kernel management has to substantially improve, it's extremely incomplete. Documentation is incomplete, not every procedure is completely documented. There's massive holes in documentation that causes lots of confusion and conflicts. There should be full-time paid managers to cope with certain tasks, such as training and conflict solving.

                      Kent did it wrong. He took it too personally. It's somewhat understandable he failed at this, because he did an immense effort at Bcachefs. But he wasn't aware of the environment and it's defects, even less about his own ones. He also victimized himself and said things in a very wrong way too. Despite he may be right in some points, that makes the situation worse.

                      Linus is still passive-aggressive, like most programmers unfortunately. He doesn't insult directly, but he fails at communicating and being more constructive. On the other side, he's an extremely talented develope. He fails at managing stuffand that makes him to be overloaded too.

                      Where's Andrew Morton, Greg Kroah-Hartman and others here? Can anyone please provide a proper organization diagram with the duties of them? I'm curious about it. Any URLs about it other than the so 80s MAINTAINERS file?

                      Maintainers failed too. They weren't constructive to the debate and took it very personally too. They were too bureaucratic. They must have be an a lot more constructive and pedagogical attitude, instead of being so pedantic.

                      The structural organization of Free Software projects differs a lot from proprietary software development. Everything is mostly OPEN, trash is a lot more visible. I'm certain proprietary software organizations have a lot of crap that is even worse, specially at corporations. I agree that this needs to improve, it causes Linux to not evolve and is the most important cause even compared to issues such as IBM/RedHat and Canonical dictatorship attitudes towards FOSS ecosystem.

                      When a project has such massive size such as Linux kernel, there are essential resources that must be provided or failures like this happen and the situation will get worse over time.

                      Let's be realistic. As much I love FOSS and despise proprietary software, FOSS success is because a lot lower development costs. But certain structural issues must be solved to make it a real alternative. There's a very urgent need to improve documentation, analyzing issues from both technical and psychological and create good literature about it so others learn these lessons and avoid failing at the same all the time.

                      FOSS isn't just coding! It's about communities and human interactions too! It must more seriously take into account the psychological part to have a really healthy environment. This is the only way to continues improveñement and innovation. If not, it will continually stagnate in a regressive way. That is happening these days, unfortunately.

                      I'm totally sure this will be solved over time, at least to the point of accept merging. But it will happen again with Bcachefs or any other effort if the real causes aren't really solved.

                      I have hopes Kent Overstreet will really think about this and find ways to improve the situation. But I'm still not sure all involved parties will learn what's really needed to make situations like this not happen AGAIN. There have been a zillion of similar situations like this.

                      Code Covenant is good, but needs to improve and there's an urgent need to write more documentation. There's a unmet urgent need of constantly mainaimed documentation and not just coding, this is causing lots of conflicts too. I see it in too many FOSS projects.

                      And Reiser4 merging issues are just one example. It wasn't just about that he murdered his wife, there were lots of crap before and some people said in this forum thread.
                      Last edited by timofonic; 10 September 2023, 06:58 AM.

                      Comment

                      Working...
                      X