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

  • Originally posted by timofonic View Post

    But who uses NILFS2?
    Anyone who wants to, and believe that it meets their use-case. It is free, after all.
    If it doesn't suit you, that's fine. You are not forced to use it.

    Comment


    • Originally posted by blackiwid View Post
      Why doesn't he say "hey Christian told you" so.... why does he say "it's my fault not noticing it" does he try to sugarbread him to keep going because he wants the cake and eat it, too? Be strong but somehow be nice to him (lie) to keep him going because he wants the code?
      I think you're misinterpreting what Torvalds said. It was his error for not noticing bcachefs was not in linux-next from the get-go. If he had noticed, I gather he would have told Overstreet that the first step is getting the code there, and Overstreet's time and effort would not have been wasted trying to fight against a process not designed for code that had yet to go through linux-next.

      Comment


      • Originally posted by blackiwid View Post
        Why doesn't he say "hey Christian told you" so.... why does he say "it's my fault not noticing it" does he try to sugarbread him to keep going because he wants the cake and eat it, too? Be strong but somehow be nice to him (lie) to keep him going because he wants the code?
        So Linus is trying to be nice and you have a problem with it?

        What's fucking wrong with you, you are turning Kent into an Idol and convince yourself everything they do is right

        Comment


        • Originally posted by blackiwid View Post
          That doesn't explain the communication from Linus, why does he say he should add this pgp keys and make it compile when he clearly could say "without linux-next you will not be added this round POINT!" I rather have negative feedback that is on point and honest than some vague mixed messages.
          There is something in particular happening here.


          Linus Torvalds has a incorrect presume. Linus is presume that the stuff his being asked about has gone though normal channels. So why would you as Linus ask if a patch is in Linux-next when everything you get normally is always in Linux-next.

          Common issue for patches in linux-next to move to Linus tree is missing pgp keys. Linus had people look into if it compiler was suspect somehow over these patches because he presumed that Kent Overstreet has obey policy..

          It does happen from time to time for compiler errors to get past the Linux-next kernel bots but it rare.

          This is the problem for Linus Torvalds none of the issues with Kent Ovestreet code was having said straight away that this is something that has disobey policy. Linus Torvald failing to check for code being linux-next when Kent Overstreet went around the VFS maintainer is on linux. Yes this is Linus Torvalds as a mistake.

          Kent Overstreet code with the numbers of compiler errors now shown its now clear that Kent Overstreet was not following the patch submit guidelines and checklist. Yes intel developers have wasted time making sure the 0-day bot on Linux-next was not busted.

          Originally posted by blackiwid View Post
          Why doesn't he say "hey Christian told you" so.... why does he say "it's my fault not noticing it" does he try to sugarbread him to keep going because he wants the cake and eat it, too? Be strong but somehow be nice to him (lie) to keep him going because he wants the code?
          Few pointers Linus Torvalds does not read the communication between VFS maintainer and the people submitting code to VFS maintainer. So Linus did not know that Kent Overstreet had already been given this direction submit to Linux-next by the VFS maintainer when he wrote that post.

          So if your question is when was Kent Overstreet first told to put bcachefs up to Linux-next that june/july from the VFS maintainer. In fact I am not sure that Linus Torvalds has even been told Kent Overstreet was told to put his code in linux-next in june/july but its in the Linux kernel mailing lists.

          Originally posted by blackiwid View Post
          ‚ÄčIs that some new policies to talk like you would talk to asiats to not insult him or something, you can say the unfiltered truth without being rude. I dislike this unclear communication, something is smells wrong and I don't see the full blame on Kent, if that would be the case, be clear about it...
          The full blame is not on Kent Overstreet just most of it. There is blame on Linus Torvald for allowing a direct due to maintainer dispute without checking a basic thing of the code being in Linux-next before he looks at it.

          Yes most likely if Linux Torvald had said no Kent at the start you need it in linux-next before I am looking at it as he should have Kent most likely would have returned with his tail between his legs and go back to the VFS maintainer.

          Human nature here you cheat the system you are getting away with it when you get caught you get very upset and attempt to bully you way though instead of excepting you should not have done it as first human response.

          blackiwid like it or not by what Kent Overstreet did not just waste his time. He has wasted Linus Torvalds time reviewing patches not up to standard. He has wasted Intel 0-day bot developers time checking out that the bot because they assumed bot not was working right because Linus was rejecting way too many patches so something had to be broken that is another incorrect assume.

          Yes it was not only Linus Torvald who missed the problem was coming from Kent Overstreet bypassing linux-next and kernel patch submit policy.

          Please note blacklwid you missed that Linus Torvalds made the mistake with handling this process because he started with one totally incorrect assume and Linus Torvalds in his post directly admitted to that mistake.

          The time before Kent Overstreet when some developer same error of ignoring the patch submit checklist and skipping the Linux-next in front of Linux Torvalds was 2005 when really Linux Torvald was very burnt out and had not worked it out yet and he banned the person from having their code accepted by any maintainer of the Linux kernel for 10 years yes that end up being until Linus got his abusive writing treated along with the cause being the burn out..

          We should be thankful that Linus Torvalds is in a lot better mind state these days Linus of old would have just overreacted canceling Kent Overstreet maintainer ship status on bcache and forbid code submit.

          Kent Overstreet most likely need some of the same counseling so if he wants to take time off to address this its most likely a good idea.

          Comment


          • Originally posted by nadir View Post

            Nobody is "burdening" him. No sudden requirements. The steps he missed are so basic that people didn't even check exactly because they assumed somebody as experienced as him wouldn't miss them.

            Merging code, especially a whole filesystem, is also not about just dumping it in the kernel. It has to interact with other parts and other people. If both the code and the person behind it don't mesh with the rest of the system, it's kinda useless.

            Him having an existential crisis and contemplating not merging already is actually a good indicator he wouldn't be a good citizen.

            People in some fields, especially certain parts of programming, have this idea that work gets done by unique geniuses that should be accommodated by us peons.

            Instead; the collaborative effort of a group of average programmers that work consistently, together, is much more productive, especially long term.

            And because you have to work with people and because they deserve your respect and kindness, you have to treat them with respect and kindness. This respect also includes respecting the systems those people are working under.
            I disagree with you regarding the LK developers. If they truly ASSUMED that Kent should have known, then the LK developers are equally guilty.

            It never hurts to be nice and say, "Here's the checklist on how we do things. Your work is currently at this point on that list. Do you understand?"

            And if someone did point out all that to Kent and Kent ignored it (or kept ignoring it), then Kent should stand up and take responsibility for his own behavior. I'll leave the LKML archaeology to someone else.

            Honestly, I think all LK developers suffer from some degree of "programmer social disability disorder" where they are consistently unaware that they are interacting with other human beings. The systemic cause of this disorder seems to be WAY TOO MUCH TIME focused on the keyboard, the monitor, and the code they are writing / reviewing; they lose their sense of humanity.

            Last edited by NotMine999; 09 September 2023, 09:57 AM.

            Comment


            • Originally posted by NotMine999 View Post

              I disagree with you regarding the LK developers. If they truly ASSUMED that Kent should have known, then the LK developers are equally guilty.

              It never hurts to be nice and say, "Here's the checklist on how we do things. Your work is currently at this point on that list. Do you understand?"

              And if someone did point out all that to Kent and Kent ignored it (or kept ignoring it), then Kent should stand up and take responsibility for his own behavior. I'll leave the LKML archaeology to someone else.

              Honestly, I think all LK developers suffer from some degree of "programmer social disability disorder" where they are consistently unaware that they are interacting with other human beings. The systemic cause of this disorder seems to be WAY TOO MUCH TIME focused on the keyboard, the monitor, and the code they are writing / reviewing; they lose their sense of humanity.
              Yes, I agree in many points.

              Everyone was wrong and acted wrong here.

              This must be taken in consideration and change many important things so next time goes better.

              I consider it must be done ASAP, I consider current situation hinders Linux Kernel development and promote toxicity.

              There must me a review of personal attitudes from all involved parties. I would like specialiced mental health professionals would check what happened in a mandatory process.

              There should be better automation of certain tasks, including the miss of linux-next and other issues.

              Management must be inproved by fixing human behaviors and automate them too, to make them time consuming. Maybe Git should adopt certain features to make it easier.

              There's also a urgent need to properly document the involved processes and be sure is totally clear to EVERYONE. Currently, many of them are "traditions", and that may cause severe confusion to many people.

              Also, keep in mind that lots of very talented geeks are not neurotypical at all. Many have Asperger Syndrome and other peculiarities that may improve their technical skills but impair others, including social ones.

              I don't think ostracism to them isn't the way, despite it may be involuntary because conflicts caused by the genius weirdo that's good at code but awful at people. That's a waste of talent and massively suffering to them.
              Last edited by timofonic; 09 September 2023, 10:27 AM.

              Comment


              • Originally posted by oiaohm View Post
                VFS maintainer had told him to put the code in next branch 4 times before ghosting him.
                can you send the links to them? I went back through the public bcachefs with good old grep, I've seen one reccomendation do to it, but I have yet to see where anyone has actually told him to

                Comment


                • Originally posted by NotMine999 View Post


                  Honestly, I think all LK developers suffer from some degree of "programmer social disability disorder" where they are consistently unaware that they are interacting with other human beings.
                  Or they're sick of other people's shit and stopped caring to get some actual work done.

                  if you want to be part of a community it's your duty to learn the rules not sit there and complain that noone has noticed you whining on the sidelines.
                  This seems to be some generational thing too: everyone nowadays expects instant gratification, participation trophies, being spoon-fed everything and talked to in the most sugarcoat way possible lest they unleash some nonsensical rant about toxicity/unwelcoming/genderbias/ in their social media bubble du jour.

                  Comment


                  • Originally posted by NotMine999 View Post
                    I disagree with you regarding the LK developers. If they truly ASSUMED that Kent should have known, then the LK developers are equally guilty.

                    It never hurts to be nice and say, "Here's the checklist on how we do things. Your work is currently at this point on that list. Do you understand?"
                    There is a reason why you have to deal with Maintainers because if everyone who could respond did you don't want to see you mailbox.

                    There is almost 2000 active developers every release. So this is case if everyone ask that question it could well and truly hurt. 1000+ emails because everyone is suspect not going to work.

                    The entry maintainers like the subsystem maintainers they are meant to ask if person submitting code has done to the patch checklist. Yes those entry level maintainers when they submit to Linux-next will get told off to replace if over again triggering bots and it appears the checklist is not being done.

                    Once you get to Linus Torvalds level in the Maintainer-ship he should have asked if it was Linux-next not if the checklist has been done because there is another issue here. When the code is submitted to next if the checklist is not done the bots will reject the patch because they will automatically do most of the tests of the patch checklist. Yes you have very low odds of not getting something wrong and passing the next branch bots if you have not done the checklist. Mostly because Linus historically as not been able to take people at their word they have done the checklist.

                    Basically Linus Torvalds at his level is not meant to be taking your word that you have done the checklist instead he should be requiring you to submit your code to the Linux next branch and when it fails as maintainer give you serous please explain and be question if all code you have submitted should be re-reviewed and possible your maintainer-ship status revoked.

                    Going to Linus Torvald the way Kent did is like going to a company CEO to complain about something only to have the CEO find that its your fault because you have been attempting to ignore company policies.

                    Originally posted by NotMine999 View Post
                    And if someone did point out all that to Kent and Kent ignored it (or kept ignoring it), then Kent should stand up and take responsibility for his own behavior. I'll leave the LKML archaeology to someone else.
                    You look into the VFS maintainers you see he was told by the vfs maintainers who should they should told him yes both the patch checklist and the linux-next tree. Yes Kent was asked about the checklist and it had answer it was done as well. Then VFS developer said for him to submit to Linux-next then he tried coming up for reasons that the code should not enter next then VFS end up washing hands of him and he then took his problem to Linus. That the flow of this problem is not good.

                    Originally posted by NotMine999 View Post
                    Honestly, I think all LK developers suffer from some degree of "programmer social disability disorder" where they are consistently unaware that they are interacting with other human beings. The systemic cause of this disorder seems to be WAY TOO MUCH TIME focused on the keyboard, the monitor, and the code they are writing / reviewing; they lose their sense of humanity.
                    Big one companies pay for Linux kernel developers to write code for them. Companies don't pay for the HR mental health staff or the Documentation writers to keep everything in nice social order.

                    I do not dispute that the Linux kernel developers and maintainers are not in the healthiest mind set most of the time. But this illness is not restricted to Linux kernel developers you find most open source projects have developers suffering from it you also find in many countries without good employee protection laws those working on closed source programs also suffer from it.

                    There is a cultures problems and lack of mental health care across most of software development.


                    There is some of software development that worse than others.

                    Linux kernel development like it or not not good but it one of the better software development environments because of the code of conduct and other things like it. Please note I am not saying Linux development space could not be better. But you have to think how much disrupt has been required to get the Linux kernel this far that a bungie class problem is not going to happen any more.

                    NotMine999 lot of the problem is lack of the required support staff. Support staff don't write code so companies don't want to spend money there. Some ways to improve Linux kernel culture requires laws to require companies to care more about their staff mental health. Like it or not the issue with Linux development is only a tip of a very big iceberg of problems as global culture we need to address.

                    Number of people paid to look after all the Linux kernel documentation processes and procedures two person that improvement on the first 15 years of no body paid. People wonder why the documentation is not the best written there is not the staff to review and improve the documentation.

                    HR mental health sub department does not exist at all with the Linux kernel. No one is checking how hours maintainers are really working and so on.

                    NotMine999 said part this problems are not restricted to the Linux kernel.

                    One of the realities that eats into maintainers and developers in general is the repeat case of thinking they have something fixed only to find out they have fixed x problem but now created y and this can go on for months to years at time. Software development with the goal of producing quality code by its nature is very human nature damaging this is some of the reason why Linux kernel developers are considering rust in the hope that improved programming language will reduce some of the mental damage. Yes that damage shows up as many different toxic workplace problems.

                    A software company that produces quality software code that does not have some toxic workplace problems just does not exist. Some places do invest correctly in mental health support be be dealing with the toxic workplace problems before they get out of hand.(please note I said out of hand not fixed/cured the software development process is very anti-human)
                    Last edited by oiaohm; 09 September 2023, 11:11 AM.

                    Comment


                    • Originally posted by mlau View Post

                      Or they're sick of other people's shit and stopped caring to get some actual work done.

                      if you want to be part of a community it's your duty to learn the rules not sit there and complain that noone has noticed you whining on the sidelines.
                      This seems to be some generational thing too: everyone nowadays expects instant gratification, participation trophies, being spoon-fed everything and talked to in the most sugarcoat way possible lest they unleash some nonsensical rant about toxicity/unwelcoming/genderbias/ in their social media bubble du jour.
                      This is one of the best comment here. However my impression is that in the past the things were more simpler. I remember very well my when first patch was accepted very fast (ok, it was one line patch :-) ), but it was 2005...

                      Frankly speaking I don't see any drama here. I saw a general ACK with minor notes, like passing via linux-next (which is a reasonable request) and adding a GPG signature to the commit. Arguing if this was told [not] in a good way or if Kent should be [not] aware of that being an historical developers is ridiculous.

                      Comment

                      Working...
                      X