Well, you weren't following the rules when they didn't exist, so how can I follow those rules? That's just unfair oppression of me. Your project was young and unstable and niche decades ago, so I must be allowed to have my young and unstable code in your mature and stable and widely used codebase now. Of course, only I should be allowed that because I want to my code to be used by an old mature widely used project, and if everyone does what I'm doing then people will stop using it.
Linus invited a bunch of woke narcissists into kernel development at the expense of most of his tried and true devs. He made his bed, now he can lie in it.
linus has 100x more patience than me. every single reply would be more than enough to say "you're done, work out of mainline" but you have to hand it to linus that he really gave kent many many chances
Linus invited a bunch of woke narcissists into kernel development at the expense of most of his tried and true devs. He made his bed, now he can lie in it.
There must be an entire University department of specialists working 24/7 on supporting Linus, we would have seen so many murders on LKML otherwise by now
"I invite you to write your own filesystem," he says to the man that literally wrote the first filesystem for Linux. And also the (content-addressable) filesystem that they're using to merge these bcachefs patches. Yikes.
Linus invited a bunch of woke narcissists into kernel development at the expense of most of his tried and true devs. He made his bed, now he can lie in it.
@@SpiderUnderUrBed_Alt He implemented the MINIX fs, I don't think he ever designed his own. The latter "content-addressable filesystem" the original comment is talking about is Git lol
Kent Overstreet does not seem to understand the very concept of the Linux kernel's merge window. As Linus said back before 6.1: "The whole 'do an all-nighter to get the paper in the day before the deadline' is something that should have gone out the window after high school. Not for kernel development."
thinking about school still gives me PTSD. my language teacher saying yes i know all of you have a two week school break but here is your homework write an entire book for your homework thanks.
@@SolitaryCZ Linus invited a bunch of woke narcissists into kernel development at the expense of most of his tried and true devs. He made his bed, now he can lie in it.
I was thinking of that. If Kent had the audacity to tell Linus to build a filesystem (which he already did, but that's besides the point) then he should try to build his own 30 million code kernel and see how easy is to review thousands of lines of code dumped all at once expecting to be done in one day. And not realize that if everybody did the same, you'll get to to 100k lines of code to be reviewed in 10-20 hours. Really smart of you Kent, really smart!
7:20 If I counted right, there are like 60 filesystems on the kernel. So, I think Linus know one thing or two about what it takes to ship a filesystem.
@@kuhluhOG Mainline kernel is about having a stable release. If you want to develop your filesystem driver fast and stabilize it later, it doesn't belong to mainline kernel. And Overstreet keeps arguing that his method is so good and the code is close to perfect, and at the same time he argues that he cannot wait for comments to his patches because critical bugs need to be fixed ASAP. You cannot have it both ways. You cannot have both "close to perfect code" and lots of critical bugs at the same time!
Mr. Overstreet seems to have forgotten that the very act of observing yourself writing high-quality bug-free code damages your ability to write high-quality bug-free code. Even if you have a 10-year streak of zero bugs and amazing quality, the instant you look at that spotless track record and think "you know what, I think I _am_ good at this" you start introducing bugs. Stay ignorant to how good you are. It's quantum physics and kernel astrology, baby.
looks a lot like rust in the kernel, not playing well with others. the solution for both is simple. do the work by themselves outside the mainline, and as bits stabilise or need to update other parts of the kernel you get someone who can play nice specifically to work on getting that code from where it works and is stable to where it is in a shape that upstream is happy with. thats how the real-time branch got upstreamed, so it can work just as well for rust and bcachefs.
@@KoopstaKlicca they are two sides of the same coin. in the case of kent and his codevelopers it is exactly as you describe. in the case of rust in the kernel, there is a set of work to be done, and major disputes as to the cost and who is going to pay it, with the rust developers thinking it is minimal and wanting the maintainers to pay, while the maintainers think it is significant, and want the rust guys to do it. in both cases it is some project which on the kernel scale is relatively minor thinking that most of the playing well with others should be done by the kernel maintainers adapting to the needs of their project, with the maintainers resisting. note: i am not saying all the rust in the kernel developers are like this, just enough of them to be harming their own case.
Honestly, if it was this, he'd better be pulling back immediately and doing everything possible to please Linus. I get in arguments with management all the time and you almost never win them. Its easier to just play ball and deal with the headaches. In arguments like this, you waste *so much time* doing nothing instead of moving forward
Plausible. But if that's true, arguing with Torvalds is the dumbest thing he could have done. Torvalds is the only man who can single-handedly get your project in or out of mainline.
At the very least, he's sold investors on the project partly because it's mainlined into the kernel now. I have no idea if bcachefs is the greatest fs ever to exist or not, but it seems like this developer is definitely not ready for primetime yet.
@@jshowaoa person like this is almost certainly not self-aware enough to realise this. His thinking is probably “not only am I a genius, I now have money, what has Linus got?”
I was honestly waiting for it to happen with every passing reply. if someone argued with me like that especially after Linus's third* reply where he was beyond clear I'd lose hope for any communication.
Linus should do this. At this point the guy isnt learning. He can come back when he can play nice. Im sure an apology and a demonstration of change would really go over well with Linus.
Giving it's Linus complaining about it, and him being an asshole top to bottom, it wouldn't matter if he didn't... he's still being an ass instead of a leader. Shouldn't have had this conversation in the open to begin with.
After "let that sink in" and "supporting my user base", he doesn't seem to realize that Linus is supporting the ENTIRE(!!!!) Linux user base! Now waiting for the following from Kent: "I'm creating this amazing file system, what did you ever do for the kernel?"
At this point, I wonder if Kent thinks, the Kernel he submits to is "just another Linux distro". But it is the very heart of (almost) every linux distro.
And then insists that you have to let him rearrange your furniture even though he didn't talk to anyone in your house about it first, because he has his own methods and a vision.
@@parad0xheart Look, if you want to have some input on where the furniture goes, I understand, but it has to be a discussion, not just you yelling about it.
It's more like, he builds his house and them comes over unannounced and starts to smash YOUR walls, because "they are not compatible", while in his house, the 2nd floor just collapsed.
"A file system that Linux can be proud of" and "OMG, we have to bypass all the rules and get this fix out ASAP or people will lose their data!" don't seem compatible. What would I do? Decide that this file system is not ready for the kernel yet.
Props to Linus for working on his aggressive communication, I really mean that, but if there was ever a time for an off-the-rails rant it was half way through this conversation.
I never dug too much into past complaints about Linus's "off the rails" rants, but for those that I did, they always seemed to be similar to Kent's situation. Warnings were given, ignored, and reprimanded without concern for political correctness.
"No issues to date have been found in the stuff you kicked back, which tells me my process is just fine" is the same argument as "I drive drunk all the time and haven't crashed yet, I guess I'm just a really good driver"
Kent has such a massive ego. He should be removed feom the kernel. I read the whole conversation. He didn't just trash Linus Torvalds. He also heavily trashed BTRFS so much that the BTRFS lead developer had to step in and say that even Meta uses BTRFS. It's like Kent thinks he has made a perfect filesystem, which "only" has to have critical data loss bugfixes every week lol.
I've been running bcachefs for 3 years and only hit one data loss bug and that was only in the tree for a few days, and that was before it had a CI, bcachefs runs basic fsck on every commit, and will fire off a kernel panic if it detects any bugs. Kent is still a garbage human though.
@@woobilicious. i been using btrfs since 2018. I see many people complaint about btrfs. But I never had a problem with it. Is there any specific use case that cause the loss of data to happen?
@@matjer2800 there are notorious write holes in raid5 and raid6 setups, and those have been there for years at this point, it should recieve major fixes in near future since they are currently working on major redesign of some of the internals. Big part of this is also their userland tooling being riddled with footguns and in general unintuitive UX, which sometimes leads people to loosing data when they do recovery procedures incorrectly.
Posting it here as well because of the fact that not everyone will see my comments on comments: He stepped back a while ago to actually get help with anger management and it seems like it's been going great for him honestly. He's been going to therapy and seems like he's achieving his goal which I'd say is quite nice
@@AstoundingAmeliadude managed some truly absurd levels of personal growth in a ridiculously short time. I kept expecting him to blow up, fuck knows I would have on reply 3.
Man, Kent is really a self-centered mf, probably even at a narcissitic level lol You can't just work in the kernel, not follow the process EVERY DEVELOPER has to follow, and expect your changes to even be accepted because your filesystem is "good" and your process is "way better". Props to Linus, he is really patient with this guy, I would have blacklisted him a long time ago
Linus invited a bunch of woke narcissists into kernel development at the expense of most of his tried and true devs. He made his bed, now he can lie in it.
I think he knows, he just doesn't seem to care at all, which tends to happen with people who deem themselves a god aka better than anyone else. At least that's what I get from it. There is no way he won't get kicked off the kernel in the next month or two, I am not convinced people like this can change (or at least, it doesn't happen often).
I think the proof is where he is flabbergasted that Linus is using such "strong" language. Expecting/demanding linus to treat him with kid gloves makes no sense.
@@ethancampbell2422y'know how everyone was proud of Linus for deciding to step back and work on himself, become a better person, grow and develop patience, etc? Could we get, just, like a hint of the old Linus here for a second... Not forever, just to deal with this shit show. Would have made this much less *AAAAARRRRRRRRGGGGGHHHH!*
Thought BcacheFS was cool, but I'm definitely sticking to Btrfs. Kent really needs to get over himself. The kernel is a collaborative project, not a hill to climb disregarding any fence or warning.
Seems he got some more bigger funding and can hire some people to work with him, I really hope this happens and they get some people to handle the 'upstreaming' part (sending it to Linus) of the project.
@@thewhitefalcon8539 In all my years of using btrfs, this has never happened. So Idk what you're doing with your drives to make the data corrupt but it's not a widespread issue from what I can see
As a developer, not having your code compile is a big oops for me. The fact that he then doubles down with arguments like "well my testing is better than your testing" means he can't be trusted. FOSS is built on trust. This aint it.
To be fair, it was on an obscure architecture. But yes he needs to get that automated compile testing set up asap, and stop downplaying the embarrassment.
what we see from a Linus perspective is a corporate ordeal being struck by someone who only worked in startups. When you are already on the market, your develompent team consists of literally tens of thousands of people following clear rules (and especially keeping the docs) is more important than fast-paced development. Even if I am working with startups I emphasize importance of proper workflow. Otherwise, the havoc starts in unexpected moments.
This is proof that people can change and become better. From all accounts, Linus used to be impatient and rude when people were being idiots. Now, he's showing a ton of patience. Good job Linus
What I don't understand, is why there isn't a write up somewhere that Linus could have linked Ken. It's clear Ken fundamentally misunderstands that he's expected to follow kernel development conventions, and that document should exist for all contributors to see. It should have been an open and shut case of RTFM on governance.
Bro really went with the Donald Trump 'we do so much testing, the greatest tests. Nobody does testing better then me' immediately after sending completely untested code to upstream. Dudes an actual comedy act
"And let me tell you, folks. Linus (and I call him minus. Because that's how little he tests his code). Linus, he wasn't testing that well in o'4. Didn't know what testing was. Just adding into RC, he was. That's true, that's what he did. And they don't want you to know that"
That's a very good point and a scary one at that...However Kent here is the founder of bcachefs and has always been insufferable at communicating. So, there is no clear break in behaviour that would suggest that his account has been hijacked by a sinister party, like it was the case with the the xz-library.
Of course my comment was made with tongue firmly in cheek and I don't think there's any chance that Kent is the next Jia Tan, but that episode has shown that a sufficiently motivated actor is willing to play the long game, implied trust counts for less than we thought, and independent code scrutiny is vitally important. If a contributor can't play by the rules to allow that, then they shouldn't be playing.
"just because you don't see it doesn't mean it doesn't exist" That's the entire point though. Like Kent points out, Linus isn't involved with the project, and doesn't know what's going on in the project. So in order for him to be able to pull the code confidently he needs the receipts. That is how it is judged whether or not your code is safe to pull.
I doubt many people would complain, I mean people that *do* use it aren't exactly shit out of luck. If I was him I would've already told Ken to shove his filesystem up his ass.
Haven't you seen videos of him cussing the hell out of his other colleagues in similar dev discussions and showing the middle finger to Nvidia? Right now he's currently very chill than what he used to be lol
Credit to Linus for being a hell of a lot more patient than I would have been with him. I don't know what Kent doesn't understand, but it is clear he doesn't and his project shouldn't be in the mainline anymore.
honestly, if i would have someone like that guy contribute to any more my< works i would refer to the projects liscence and tell him to go fuck off and do his own thing. if he cant make it work after a few tries then im done working with them. I may not get angry like some other people but i certainly also dont have that big of a patience like some people show towards the worst of people.
I am geniunely brain dead when it comes to development. However, I know the difference here. This dev is treating the Kernal Upstream like its a Highschool group project, where you discuss and come to agreements on how to do the project... Its been a while since I've seen a dev so far up their own ass they actually made a loop. I don't even need to know how any of this works to know: You either work with Linus and the Kernal devs how they expect you to or you go away. Because what is the point of being in that project if you can't operate under its rules, its not like you *HAVE* to be in it... And given their reasons for being in it are "Pride and Accomplishment" I hope they either fix their shit or get off the project.
If you had as much experience managing opensource communities you would know how to build it instead of breaking it down. Its really hard work and takes more tact then we give Torvalds credit for.
@@2greenifyLinus has argued with this guy incessantly now and for what? So his pet FS will be integrated? Nobody cares and Linus should put his foot down right now and stop this chaos from infecting mainline
@@2greenify The kernel developers are employees motivated by a paycheck and replaceable. What you wrote about "community" open source project is true, but kernel is not that. Hasn't been for a very long time.
@@honza970i would strongly disagree with the idea that kernel developers are replaceable, as most of the are doing critical work in important but obscure areas of development. as for them getting paid, so what if some company wants to pay them, with the idea of basically putting bug bounties on problems affecting their company. you still have a number of kernel contributors who are not tied to big companies. as to it being a community or not, i don't see any way you can define that word which would not cover 15,000 people trying to not get in each others way for years. if you mean the kernel project not being community driven like centos used to be and rocky is now, very few large projects or distributions fit that model. it certainly does not fit how systemd is developed, nor the kernel, and not gnome either.
Honestly seeing all this has drained any excitement I might have had in the file system right out of me. This is not the behavior of a developer I want working on something that I trust my data to
He really is daft,isn't he? Linus doesn't care if there are bugs or not in bcachefs. He needs to know for sure the code doesn't break the kernel or some other bits you may have touched or interact with. It's not about what your project offers, it's about how it fits in with all the other bits.
Kent is a genuinely pathetic person. The level of narcissism he displays while trying to talk down to Linus Torvalds of all people is absurd. He needs to be removed from mainline ASAP, because he is absolutely the kind of person to break stuff out of spite.
Would've lost it immediately after 19:15. This isn't even just a bad philosophy in coding, this is just awful in general. It's entirely a counterproductive mindset to have in any team environment.
The part that got me was he literally admitted to understanding what linus wanted and refused to do it because...its not criticism from within his project? Huh? What?
@@woobilicious. in both cases there are cloud instances available that can be used for automated testing. No need to actually purchase and maintain dedicated hw for CI pipelines.
I'm baffled that Linus clearly cannot see the stability and robustness of this critically important filesystem and that he abuses Kent so much.... Seriously though maybe Kent thinks his toy is so important because Linus is so patient to his shenanigans?
Imagine if linus had to have a conversation with every single contributer about processes every time something gets merged, this is what happens when you are too polite with people.
check the response from Daniel Hill at the (current) end of the thread and make an appendix video ;) whole conversation looks like a bunch of people are trying to cooperate with a skilled developer, while all the advises he is receiving like daggers, so he is trying to fight back. i hope that there somewhere accessible logs from bcachefs irc channel as looks like it would be a much deeper hole to sink in...
If only ZFS could be in-tree so that people could have a good, stable filesystem. It being out of tree means you have to be "safe" and keep /boot and the like on something else.
@@LunarLambda isn't there a risk on kernel update it not having a working zfs module? Last I tried /boot on ZFS was when Ubuntu was pushing it and I've just avoided it since.
@@blazewardog I run several hosts with root on ZFS and have encountered two issues. 1) With Debian Testing the kernel can get ahead of available ZFS versions resulting in failure to build using DKMS. Solution is to hold the newer kernel until the modules are available. (RpiOS has also messed this up and solved by using backports for ZFS.) 2) At one point Grub had a problem booting ZFS. The fix was to update Grub. Not entirely w/out problems but IMO worth the extra effort.
I put zfs on my home drive (4TB mirror x2 hdd) when I rebuilt my system with 16-core ryzen and 128GB ram. kept getting wicked lag spikes with user input and IDEs that restart processes to analyze code. found hundreds of timeout messages in dmesg logs related to zfs driver. started migrating files from zfs to btrfs (on an nvme) and lag spikes are much less frequent. don't know what problem I have with zfs, but it's not good/stable enough for me to continue using it.
@@charetjc ZFS is brilliant, but not very straightforward to use. You need stuff like ZIL's (commit cache) and ARC's (read cache) (and not to mention at least 16GB of RAM) to get the most out of ZFS. I've not done anything with ZFS on Linux lately (I've only used it on FreeBSD and OpenSolaris), but weren't there some license incompatibilities why ZFS could not run in Kernel Space (due not being part of the linux kernel tree due to the license issues) but had to run in User Space and thus not getting a privileged amount of CPU slices?
The mainline Linux kernel gets tens of thousands of contributors. If everyone followed Overstreet's "system" of pushing whatever commits, whenever, without coordinating with anyone outside their tree, then it would be complete and utter chaos. You can't just trust that everything _should_ work. Not at the kernel level. If bcachefs is so unstable that it requires these rapid, unchecked commits to be reliable, then it shouldn't be mainlined. Period. Fix that up on your own, then come back later. If bcachefs truly is worthy of being mainlined, then its maintainers should have no problem following the same rules as everyone else. Otherwise, see above. Why should bcachefs get special treatment? There's just too much going on in mainline for this archaic kind of coordination.
I am sure this response will help Kent with the fundraising /s. Literally replying nothing would have been better. He needs to take a breather and think two weeks before his next action.
Awhile back he posted a kind of whiny complaint on patreon, so I unsubscribed. I think the ideas are neat, but for filesystems the execution is the entire ballgame. In this case working well with others isn’t just important, it is essential.
Anyone using a filesystem that depends on chaotic patch processes to ‘not lose their data’ needs to evaluate just how important their data is. This developer needs an ego check. He simultaneously thinks his filesystem is god’s gift to the world and thinks it’s so broken that the world is going to break if he can’t just Willy nilly submit patches out of process. He also clearly thinks he has way more users than he does. No one except hobbyists use an experimental filesystem in production, and anyone who is seeing this exchange is likely scrambling to get off of it anyway
I just noticed the subtitles puts in "big Indian" instead of "big endian" 😂. You're right, Kent, people typically don't run Linux on big Indians. Or any Indians.
BRO, Linux isn't YOUR PROJECT, you DON'T get to make the rules! Linus should issue an official warning, and if you don't go away and at least THINK about why what you did was wrong, thats it, its easily 3 strikes by now, i count at least 2 here.
"Linus is a saint" I would say that this is the understatement of the year within this context. He is showing a level of patience with this person that others would have a hard time to muster. I had a bit of a laugh when I read the part about him wanting Linux to have a filesystem to be proud of, one that is robust and stable. So why not contribute to existing filesystem codebases instead of rolling your own? XFS has been around for ages and is both stable and robust and performs rather well to boot. The same goes for ext4, although that is not my personal favourite, and while I have design philosophy issues with ZFS and btrfs, even I have to admit that they are both robust and stable. And as Linus points out: mistakes happen. Bugs happen. Build failures happen. You can't have a project as vast and complex as a kernel without having them pop up on a regular basis. That i s why all these procedures and processes exist: to minimize the number of times they pop up.
@@itskdog yeah the good old RAID5/6 write hole, RAID 1 and 10 have been stable for me for close to 6 years, even started as a HDD pool and got turned into a SSD pool over time, hecc, it started as a RAID1, then I added 2 more drives and converted it to RAID1, I just love that flexibility. But that conversion was tense, not because it didn't work, but I was expecting it to fail, but nah, all good
"That has to work both ways" No, it doesn't. One of you is the main guy and maintainer of the Kernel. The other is not. Linus is being more than accommodating but ultimately he doesn't have to care about how you do things, you do have to care about how he and every other developer working on the kernel does things.
Back when Linux only ran on floppies and one particular type of hard drives, and was ENTIRELY written by Linus Torvalds, there was a filesystem. This means that back in the 1990s, Linus ACTUALLY wrote a filesystem - not necessairly a powerful, efficient, or innovative one, but still a filesystem. In fact, the very first thing you need for a kernel is a filesystem.
We've been using btrfs for a while now on the main work laptop. We did not blame it for a folder disappearing last year. We keep backups, so yeah. It seems reliable enough. We are enjoying the dynamic allocation of space (so no more fixed partition size). But haven't bothered with snapshots, subvolumes, or RAIDs. Not sure what good it will do to add the chcksums and compression of bcachefs to the list of things we already don't use.
I'm impressed how Linus improved his communication skills and his patience those last years. I think a few years ago there would have been a flamethrower involved, at the very least.
Linus: Why aren't you following the rules?
Kent: How dare you imply my code is bad!
This is the kdbus fiasco all over again.
Well, you weren't following the rules when they didn't exist, so how can I follow those rules? That's just unfair oppression of me.
Your project was young and unstable and niche decades ago, so I must be allowed to have my young and unstable code in your mature and stable and widely used codebase now.
Of course, only I should be allowed that because I want to my code to be used by an old mature widely used project, and if everyone does what I'm doing then people will stop using it.
kent, kent, where have I heard that name associated with obnoxious person. Oh yeah Real Genius! movie
@@timroach5898 I assume you are referencing a certain Kent "annoying Linux users since 2018"?
Linus invited a bunch of woke narcissists into kernel development at the expense of most of his tried and true devs.
He made his bed, now he can lie in it.
linus has 100x more patience than me. every single reply would be more than enough to say "you're done, work out of mainline" but you have to hand it to linus that he really gave kent many many chances
Dude the amount of Patience Linus has, like WHY IS HE ENTERTAINING THAT MORON!? XD
while i'd love to see this guy get his ass kicked, linus being this chill makes him even more cringe by comparison
He's changed a lot compared to the Torvalds of old. I'll give him credit he definitely changed his toxic behavior
Linus invited a bunch of woke narcissists into kernel development at the expense of most of his tried and true devs.
He made his bed, now he can lie in it.
@@mattkeith530
Effectiveness is "toxic"?
Linus gets milder in his mid 50s 😄
i was literally shocked about his response. is he okay?
At list he doesn't suggest people to remove themselves anymore.
@@penguin86bitals I guess that's too much, but his rants about how he doesn't know how someone survived till his *insertnumber*'s were pretty funny.
@@faxen123he is going to the therapist is all
he got sent to retraining camp, i wonder what they really did to him. maybe lobotomy? it explains the allowance of rust though.
linus' therapist is kicking in joy on how much of a good job he's done with the way linus is handling this situation
There must be an entire University department of specialists working 24/7 on supporting Linus, we would have seen so many murders on LKML otherwise by now
I don’t know much about him, but how would he have reacted back then?😂
This is a low bar to set for celebration.
@@tamino3777 up to 15 years
You mean Linux, not Linus
"I invite you to write your own filesystem," he says to the man that literally wrote the first filesystem for Linux. And also the (content-addressable) filesystem that they're using to merge these bcachefs patches. Yikes.
Linus invited a bunch of woke narcissists into kernel development at the expense of most of his tried and true devs.
He made his bed, now he can lie in it.
@@JPs-q1oThis is at least the fourth time you copy-pasted this garbage. Get better.
What was the name of the original kernel filesystem he wrote? Curious.
@@SpiderUnderUrBed_Alt He implemented the MINIX fs, I don't think he ever designed his own.
The latter "content-addressable filesystem" the original comment is talking about is Git lol
@@JPs-q1owhat's the woke part? Please be specific
Kent Overstreet does not seem to understand the very concept of the Linux kernel's merge window. As Linus said back before 6.1: "The whole 'do an all-nighter to get the paper in the day before the deadline' is something that should have gone out the window after high school. Not for kernel development."
thinking about school still gives me PTSD. my language teacher saying yes i know all of you have a two week school break but here is your homework write an entire book for your homework thanks.
I think he understands it very well, but thinks it doesn't apply to him,
@@SolitaryCZ
Linus invited a bunch of woke narcissists into kernel development at the expense of most of his tried and true devs.
He made his bed, now he can lie in it.
@@JPs-q1oSomeone needs to disable the script on your account that makes you spam the same messages throughout multiple comment threads.
@@JPs-q1o Funnily enough you're just as unhinged as he is
before watching, I guess Bcachefs
bingo
guessing wasn't even needed 🙂
same
I have dearrow, no need to guess 😄
Linus should just tell him to write his own kernel for his filesystem at that point
Breaking news: Kent Overstreet becomes HURD maintainer ;)
@@SugarBeetMC I was going to say that he should take over TempleOS.
I was thinking of that. If Kent had the audacity to tell Linus to build a filesystem (which he already did, but that's besides the point) then he should try to build his own 30 million code kernel and see how easy is to review thousands of lines of code dumped all at once expecting to be done in one day. And not realize that if everybody did the same, you'll get to to 100k lines of code to be reviewed in 10-20 hours. Really smart of you Kent, really smart!
Maybe he can become Clara, put on the sissiest "programmer socks" and work with the Hurd people...
"Take your toy and go home"
😂😂😂😂
He has such a way with words.
I could feel my blood pressure rising with everything Kent said. Linus is MUCH more patient than I am.
LOL the insults I blurted out as a reaction are... not exactly age-appropriate...
7:20 If I counted right, there are like 60 filesystems on the kernel. So, I think Linus know one thing or two about what it takes to ship a filesystem.
How to ship it, yes.
How to develop a modern one? Well, Linus himself said a few years ago that he is too out of the loop for that about the topic.
@@kuhluhOG Mainline kernel is about having a stable release. If you want to develop your filesystem driver fast and stabilize it later, it doesn't belong to mainline kernel.
And Overstreet keeps arguing that his method is so good and the code is close to perfect, and at the same time he argues that he cannot wait for comments to his patches because critical bugs need to be fixed ASAP.
You cannot have it both ways. You cannot have both "close to perfect code" and lots of critical bugs at the same time!
Mr. Overstreet seems to have forgotten that the very act of observing yourself writing high-quality bug-free code damages your ability to write high-quality bug-free code.
Even if you have a 10-year streak of zero bugs and amazing quality, the instant you look at that spotless track record and think "you know what, I think I _am_ good at this" you start introducing bugs.
Stay ignorant to how good you are. It's quantum physics and kernel astrology, baby.
It's just bizarre anyone making a filesystem would do or say any of this. The most important place to have no bugs.
@@kuhluhOGhe doesn’t need to know how to develop one.
LKML is slowly becoming my favorite drama show
the wayland sitcom is still my favorite, lkml is the second one lol
Kent is the type of guy whose skill and talent is wasted on an obstinate and entitled personality. These are the most frustrating people to deal with
a lot of talent and skill on show here for sure lol
A frustratingly common occurrence in the kernel now that I think about it lol
looks a lot like rust in the kernel, not playing well with others.
the solution for both is simple. do the work by themselves outside the mainline, and as bits stabilise or need to update other parts of the kernel you get someone who can play nice specifically to work on getting that code from where it works and is stable to where it is in a shape that upstream is happy with.
thats how the real-time branch got upstreamed, so it can work just as well for rust and bcachefs.
@@grokitall NO ONE brought up rust. This is schizophrenic. We're talking about one dude named Kent
@@KoopstaKlicca they are two sides of the same coin. in the case of kent and his codevelopers it is exactly as you describe.
in the case of rust in the kernel, there is a set of work to be done, and major disputes as to the cost and who is going to pay it, with the rust developers thinking it is minimal and wanting the maintainers to pay, while the maintainers think it is significant, and want the rust guys to do it.
in both cases it is some project which on the kernel scale is relatively minor thinking that most of the playing well with others should be done by the kernel maintainers adapting to the needs of their project, with the maintainers resisting.
note: i am not saying all the rust in the kernel developers are like this, just enough of them to be harming their own case.
I've got a theory: he got VC money that was dependent on bcachefs being in te mainline kernel. Anyways, he seems fun to work with.
Honestly, if it was this, he'd better be pulling back immediately and doing everything possible to please Linus.
I get in arguments with management all the time and you almost never win them. Its easier to just play ball and deal with the headaches. In arguments like this, you waste *so much time* doing nothing instead of moving forward
Plausible. But if that's true, arguing with Torvalds is the dumbest thing he could have done. Torvalds is the only man who can single-handedly get your project in or out of mainline.
Yes, the “I have staff lined up” is irrelevant to the conversation but it’s also concerning
At the very least, he's sold investors on the project partly because it's mainlined into the kernel now.
I have no idea if bcachefs is the greatest fs ever to exist or not, but it seems like this developer is definitely not ready for primetime yet.
@@jshowaoa person like this is almost certainly not self-aware enough to realise this. His thinking is probably “not only am I a genius, I now have money, what has Linus got?”
Every time he argues back i get worried the next reply from Linus is gonna be telling him his filesystem is removed from the kernel
If I was Linus I would have done so after the 3rd reply where he kept dodging the threat.
I was honestly waiting for it to happen with every passing reply. if someone argued with me like that especially after Linus's third* reply where he was beyond clear I'd lose hope for any communication.
Linus should do this. At this point the guy isnt learning. He can come back when he can play nice. Im sure an apology and a demonstration of change would really go over well with Linus.
just about time.
Giving it's Linus complaining about it, and him being an asshole top to bottom, it wouldn't matter if he didn't... he's still being an ass instead of a leader. Shouldn't have had this conversation in the open to begin with.
After "let that sink in" and "supporting my user base", he doesn't seem to realize that Linus is supporting the ENTIRE(!!!!) Linux user base!
Now waiting for the following from Kent: "I'm creating this amazing file system, what did you ever do for the kernel?"
At this point, I wonder if Kent thinks, the Kernel he submits to is "just another Linux distro". But it is the very heart of (almost) every linux distro.
@@Sypakawhat Linux distro doesn’t use Linux?
@@kraio-sfuYou can or could install Debian with the BSD kernel if you want(ed)
@@kraio-sfufreebsd
@@Myname-l3h freebsd isn’t linux disto tho? It’s actual unix, and a whole unix OS at that.
Ken goes to other peoples houses and starts rearranging their furniture
And then insists that you have to let him rearrange your furniture even though he didn't talk to anyone in your house about it first, because he has his own methods and a vision.
@@parad0xheart Look, if you want to have some input on where the furniture goes, I understand, but it has to be a discussion, not just you yelling about it.
It's more like, he builds his house and them comes over unannounced and starts to smash YOUR walls, because "they are not compatible", while in his house, the 2nd floor just collapsed.
More like: starts breaking it - perhaps without even knowing, perhaps while just not caring... Oh well.
He plonks an untested sofa in your living room, completely breaking the feng shui, and wonders why you're upset about it.
"A file system that Linux can be proud of" and "OMG, we have to bypass all the rules and get this fix out ASAP or people will lose their data!" don't seem compatible.
What would I do? Decide that this file system is not ready for the kernel yet.
Props to Linus for working on his aggressive communication, I really mean that, but if there was ever a time for an off-the-rails rant it was half way through this conversation.
I never dug too much into past complaints about Linus's "off the rails" rants, but for those that I did, they always seemed to be similar to Kent's situation. Warnings were given, ignored, and reprimanded without concern for political correctness.
"No issues to date have been found in the stuff you kicked back, which tells me my process is just fine" is the same argument as "I drive drunk all the time and haven't crashed yet, I guess I'm just a really good driver"
Except that one time last week when I crashed... but that was just once and had nothing to do with me being drunk.
not applicable. drunk driving is a skill. people are just not encouraged to get good at it
Kent has such a massive ego. He should be removed feom the kernel. I read the whole conversation. He didn't just trash Linus Torvalds. He also heavily trashed BTRFS so much that the BTRFS lead developer had to step in and say that even Meta uses BTRFS. It's like Kent thinks he has made a perfect filesystem, which "only" has to have critical data loss bugfixes every week lol.
I've been running bcachefs for 3 years and only hit one data loss bug and that was only in the tree for a few days, and that was before it had a CI, bcachefs runs basic fsck on every commit, and will fire off a kernel panic if it detects any bugs.
Kent is still a garbage human though.
@@woobilicious. You can be totally garbage and a genius. Google Schrödinger... He was much worse than Kent but changed Physics.
@@vidal9747exactly I can't believe Schrödinger was an opponent of NSDAP
@@woobilicious. i been using btrfs since 2018. I see many people complaint about btrfs. But I never had a problem with it. Is there any specific use case that cause the loss of data to happen?
@@matjer2800 there are notorious write holes in raid5 and raid6 setups, and those have been there for years at this point, it should recieve major fixes in near future since they are currently working on major redesign of some of the internals.
Big part of this is also their userland tooling being riddled with footguns and in general unintuitive UX, which sometimes leads people to loosing data when they do recovery procedures incorrectly.
The amount of patience Linus is showing here is crazy, I don't know how he managed to stay so calm.
I was thinking the same thing
Posting it here as well because of the fact that not everyone will see my comments on comments:
He stepped back a while ago to actually get help with anger management and it seems like it's been going great for him honestly. He's been going to therapy and seems like he's achieving his goal which I'd say is quite nice
@@AstoundingAmeliadude managed some truly absurd levels of personal growth in a ridiculously short time. I kept expecting him to blow up, fuck knows I would have on reply 3.
@@Briggsby yeah he has and honestly, He deserves to be proud of himself for that massively
Man, Kent is really a self-centered mf, probably even at a narcissitic level lol
You can't just work in the kernel, not follow the process EVERY DEVELOPER has to follow, and expect your changes to even be accepted because your filesystem is "good" and your process is "way better".
Props to Linus, he is really patient with this guy, I would have blacklisted him a long time ago
en.wikipedia.org/wiki/Narcissistic_personality_disorder
Linus invited a bunch of woke narcissists into kernel development at the expense of most of his tried and true devs.
He made his bed, now he can lie in it.
@@JPs-q1oYou need therapy.
@@JPs-q1oStop copy-pasting messages, and keep your politics away from this topic.
@@RealMephresI swear anti-woke is as annoying or worse than woke.
I'm starting to think Kent has no clue who Linus Torvalds is.
I think he knows, he just doesn't seem to care at all, which tends to happen with people who deem themselves a god aka better than anyone else. At least that's what I get from it. There is no way he won't get kicked off the kernel in the next month or two, I am not convinced people like this can change (or at least, it doesn't happen often).
I think the proof is where he is flabbergasted that Linus is using such "strong" language. Expecting/demanding linus to treat him with kid gloves makes no sense.
@@heinzerbrew Especially considering that Linus is already treating him with kid gloves. He's been uncharacteristically patient and diplomatic.
And he uses Mercury as VCS
@@ethancampbell2422y'know how everyone was proud of Linus for deciding to step back and work on himself, become a better person, grow and develop patience, etc?
Could we get, just, like a hint of the old Linus here for a second... Not forever, just to deal with this shit show. Would have made this much less *AAAAARRRRRRRRGGGGGHHHH!*
Thought BcacheFS was cool, but I'm definitely sticking to Btrfs. Kent really needs to get over himself. The kernel is a collaborative project, not a hill to climb disregarding any fence or warning.
and then to get uppity when the king of the hill says "hey cut that shit out"
Seems he got some more bigger funding and can hire some people to work with him, I really hope this happens and they get some people to handle the 'upstreaming' part (sending it to Linus) of the project.
He’s a great programmer, but unlike Linus, not a great leader nor manager
Btrfs still corrupts data?
@@thewhitefalcon8539 In all my years of using btrfs, this has never happened. So Idk what you're doing with your drives to make the data corrupt but it's not a widespread issue from what I can see
As a developer, not having your code compile is a big oops for me. The fact that he then doubles down with arguments like "well my testing is better than your testing" means he can't be trusted. FOSS is built on trust. This aint it.
It's the first rule of kernel development. Never break the build.
To be fair, it was on an obscure architecture.
But yes he needs to get that automated compile testing set up asap, and stop downplaying the embarrassment.
This is the convict arguing with the judge at the sentencing hearing. Never a smart thing to do.
Could you please zoom out when you do these 80-column tty reads? The text is still too readable on my phone.
Thank you.
Savage.
That's a new one
what we see from a Linus perspective is a corporate ordeal being struck by someone who only worked in startups. When you are already on the market, your develompent team consists of literally tens of thousands of people following clear rules (and especially keeping the docs) is more important than fast-paced development. Even if I am working with startups I emphasize importance of proper workflow. Otherwise, the havoc starts in unexpected moments.
Imagine that ego top
i am writing the FS of the future :>
I've always thought that Linus had short fuses. Well, his fuses for sure are longer than mine.
He seems to have calmed down with age.
He saves the best for last! Like a cat playing with a dead mouse
@@vidal9747and therapy too
Ego of ken is masssiiivveee good lord. Rules must be followed so kernel doesn’t break ON A RC, WOW who would have guessed!?
"It wasn't a break that mattered, because no one actually uses big-endian"
This is proof that people can change and become better. From all accounts, Linus used to be impatient and rude when people were being idiots. Now, he's showing a ton of patience. Good job Linus
Ye, but I feel like he's tempted to show his old self to thus m'fucker lmao
@@garrettrinquest1605 his words have even more weight now that he focuses his wrath more on the objective and impersonal. It's basically a superpower.
What I don't understand, is why there isn't a write up somewhere that Linus could have linked Ken. It's clear Ken fundamentally misunderstands that he's expected to follow kernel development conventions, and that document should exist for all contributors to see.
It should have been an open and shut case of RTFM on governance.
Bro really went with the Donald Trump 'we do so much testing, the greatest tests. Nobody does testing better then me' immediately after sending completely untested code to upstream.
Dudes an actual comedy act
"And let me tell you, folks. Linus (and I call him minus. Because that's how little he tests his code). Linus, he wasn't testing that well in o'4. Didn't know what testing was. Just adding into RC, he was. That's true, that's what he did. And they don't want you to know that"
Exactly what I thought xd
Brilliant
Top comment right there.
hahaha, that's a quote i exactly would imagine trump doing, very funny.
man, i been following this drama for a while and the ego still hurts my eyes
Remember the last time an impatient developer pressured an overworked project maintainer into accepting patches without adequate review?
No. *jordan_peele_sweating.gif*
That's a very good point and a scary one at that...However Kent here is the founder of bcachefs and has always been insufferable at communicating. So, there is no clear break in behaviour that would suggest that his account has been hijacked by a sinister party, like it was the case with the the xz-library.
Of course my comment was made with tongue firmly in cheek and I don't think there's any chance that Kent is the next Jia Tan, but that episode has shown that a sufficiently motivated actor is willing to play the long game, implied trust counts for less than we thought, and independent code scrutiny is vitally important. If a contributor can't play by the rules to allow that, then they shouldn't be playing.
@@LarixusSnydes relevance?
@@LarixusSnydes The threat vector wasn't a hijacked account. It was a slow boil, a long con.
- Follow the guidelines pls
- NO U
Guy desperately needs to hire someone else to do communication with Linus and do the patch formatting for bcachefs
Kent: This is a stable and useful filesystem!
Also Kent: I WILL MAKE ABSOLUTELY CERTAIN THAT NO ENTERPRISE LINUX DISTROS WILL EVER USE MY FILE SYSTEM!
Imagine telling Linus that he just doesn't see it, rather than giving a decent explainantion.
If Linus doesn't *see it* then it's not getting pulled.
Interacting with all these naruto main characters must be so tiring
Lmao
Yapping no justu
"just because you don't see it doesn't mean it doesn't exist" That's the entire point though. Like Kent points out, Linus isn't involved with the project, and doesn't know what's going on in the project. So in order for him to be able to pull the code confidently he needs the receipts. That is how it is judged whether or not your code is safe to pull.
Also, now that he entered mainline, "The Project" no longer means bcachefs. It means Linux now and I don't think Kent quite understands that.
@@Poldovico That entire conversation just sounds like Kent thinks that Linus is now working for his project and not the other way around like it is...
I too wonder how Linus hasn't just punted bcachefs from the kernel yet. Like this man has the patience of a saint.
I doubt many people would complain, I mean people that *do* use it aren't exactly shit out of luck. If I was him I would've already told Ken to shove his filesystem up his ass.
A 100% reasonable and deserved Linus rant.
I admire Linus' patience with this guy
How the Hell does he put up with it?
Haven't you seen videos of him cussing the hell out of his other colleagues in similar dev discussions and showing the middle finger to Nvidia? Right now he's currently very chill than what he used to be lol
Linus crossdresses now, so that probably explains it
He's Finnish, we are naturally very patient.
@@polinskitom2277 The most useless comment here.
@@polinskitom2277 Does he really?
Credit to Linus for being a hell of a lot more patient than I would have been with him. I don't know what Kent doesn't understand, but it is clear he doesn't and his project shouldn't be in the mainline anymore.
imagine telling linus that he isnt involved enough with linux to be in a position to talk back on people lol
honestly, if i would have someone like that guy contribute to any more my< works i would refer to the projects liscence and tell him to go fuck off and do his own thing.
if he cant make it work after a few tries then im done working with them.
I may not get angry like some other people but i certainly also dont have that big of a patience like some people show towards the worst of people.
I am geniunely brain dead when it comes to development.
However, I know the difference here. This dev is treating the Kernal Upstream like its a Highschool group project, where you discuss and come to agreements on how to do the project... Its been a while since I've seen a dev so far up their own ass they actually made a loop.
I don't even need to know how any of this works to know: You either work with Linus and the Kernal devs how they expect you to or you go away. Because what is the point of being in that project if you can't operate under its rules, its not like you *HAVE* to be in it... And given their reasons for being in it are "Pride and Accomplishment" I hope they either fix their shit or get off the project.
By this logic this dude's basically a Klein bottle at this point.
If I was Linus, the whole argument would have happened _after_ i removed his FS from the tree..
If I were in his shoes on at most the third time I would have have removed it and any arguing from him would be to A wall.
If you had as much experience managing opensource communities you would know how to build it instead of breaking it down.
Its really hard work and takes more tact then we give Torvalds credit for.
@@2greenifyLinus has argued with this guy incessantly now and for what? So his pet FS will be integrated? Nobody cares and Linus should put his foot down right now and stop this chaos from infecting mainline
@@2greenify The kernel developers are employees motivated by a paycheck and replaceable. What you wrote about "community" open source project is true, but kernel is not that. Hasn't been for a very long time.
@@honza970i would strongly disagree with the idea that kernel developers are replaceable, as most of the are doing critical work in important but obscure areas of development.
as for them getting paid, so what if some company wants to pay them, with the idea of basically putting bug bounties on problems affecting their company.
you still have a number of kernel contributors who are not tied to big companies.
as to it being a community or not, i don't see any way you can define that word which would not cover 15,000 people trying to not get in each others way for years.
if you mean the kernel project not being community driven like centos used to be and rocky is now, very few large projects or distributions fit that model. it certainly does not fit how systemd is developed, nor the kernel, and not gnome either.
Ugh, frustrating, show me another developer that says "new development is exempt from the release process."
Kent giving off Stockton Rush vibes
If I relied on Kent for important software, I would be terrified right now
Does this guy genuinely not understand how to behave in a collaborative project, or is he really being as deliberately obstinate as it appears?
It really seems like he doesn't view himself as part of the kernel project at all. Which would be fine if he wasn't, you know, in the kernel.
Linus should suggest to Kent Overstreet to write his own kernel instead.
Bcachefs the fire system that's going to save Linux, that no one uses...
No one uses it because it's not stable
Honestly seeing all this has drained any excitement I might have had in the file system right out of me. This is not the behavior of a developer I want working on something that I trust my data to
Exactly. Beyond the self-centeredness, it's the total lack of responsibility.
Too bad, I wanted to try Bcachefs once it matured.
@@jasondoe2596 Wait for the dev to mature first.
brodie should bring the bcachefs dev to tech over tea...
He really is daft,isn't he? Linus doesn't care if there are bugs or not in bcachefs. He needs to know for sure the code doesn't break the kernel or some other bits you may have touched or interact with. It's not about what your project offers, it's about how it fits in with all the other bits.
Kent is a genuinely pathetic person. The level of narcissism he displays while trying to talk down to Linus Torvalds of all people is absurd. He needs to be removed from mainline ASAP, because he is absolutely the kind of person to break stuff out of spite.
Would've lost it immediately after 19:15. This isn't even just a bad philosophy in coding, this is just awful in general. It's entirely a counterproductive mindset to have in any team environment.
I think Linus is making him look like an idiot in front of everyone in order to teach future kernel developers what NOT to do.
It’s comedic relief :)
Communication on this is done in public because that's what the process is. I don't think it's special treatment
The part that got me was he literally admitted to understanding what linus wanted and refused to do it because...its not criticism from within his project? Huh? What?
I think if it happens again, it'll be gone. I think Linus might keep him around just to show others what not to do.
Not even building your code before pushing is craazy
it was failing on big endian systems
Which proves the point. FS are not webapps. (and I say this as a webapp dev)
Guy could hire someone to set up CI with all that investor money.
@@SugarBeetMC The issue is getting big-endian hardware, targetting arm64 MacOS is also an issue, you can't put Mac Mini's in to a Rack too easy.
@@woobilicious. in both cases there are cloud instances available that can be used for automated testing. No need to actually purchase and maintain dedicated hw for CI pipelines.
kernel devs don't have time for this. i'd say drop it and move on.
wait for the code and its developer to mature.
@@JorgetePanete important comment, very useful and welcome
I'm baffled that Linus clearly cannot see the stability and robustness of this critically important filesystem and that he abuses Kent so much.... Seriously though maybe Kent thinks his toy is so important because Linus is so patient to his shenanigans?
Imagine if linus had to have a conversation with every single contributer about processes every time something gets merged, this is what happens when you are too polite with people.
While he's not old old, I am starting to become concerned about the future of linux post-Torvalds.
check the response from Daniel Hill at the (current) end of the thread and make an appendix video ;)
whole conversation looks like a bunch of people are trying to cooperate with a skilled developer, while all the advises he is receiving like daggers, so he is trying to fight back.
i hope that there somewhere accessible logs from bcachefs irc channel as looks like it would be a much deeper hole to sink in...
Honestly the way Kent is pushing bcachefs this way feels like a breeding ground for XZ utils all over again.
people have speculated that that is his goal
Linus is famous for having very strong opinions, but the difference is he's earned the right to those.
He just doesn't get "teamwork."
If only ZFS could be in-tree so that people could have a good, stable filesystem. It being out of tree means you have to be "safe" and keep /boot and the like on something else.
my /boot is on ZFS. no issues
@@LunarLambda isn't there a risk on kernel update it not having a working zfs module? Last I tried /boot on ZFS was when Ubuntu was pushing it and I've just avoided it since.
@@blazewardog I run several hosts with root on ZFS and have encountered two issues. 1) With Debian Testing the kernel can get ahead of available ZFS versions resulting in failure to build using DKMS. Solution is to hold the newer kernel until the modules are available. (RpiOS has also messed this up and solved by using backports for ZFS.) 2) At one point Grub had a problem booting ZFS. The fix was to update Grub.
Not entirely w/out problems but IMO worth the extra effort.
I put zfs on my home drive (4TB mirror x2 hdd) when I rebuilt my system with 16-core ryzen and 128GB ram. kept getting wicked lag spikes with user input and IDEs that restart processes to analyze code. found hundreds of timeout messages in dmesg logs related to zfs driver.
started migrating files from zfs to btrfs (on an nvme) and lag spikes are much less frequent. don't know what problem I have with zfs, but it's not good/stable enough for me to continue using it.
@@charetjc ZFS is brilliant, but not very straightforward to use. You need stuff like ZIL's (commit cache) and ARC's (read cache) (and not to mention at least 16GB of RAM) to get the most out of ZFS. I've not done anything with ZFS on Linux lately (I've only used it on FreeBSD and OpenSolaris), but weren't there some license incompatibilities why ZFS could not run in Kernel Space (due not being part of the linux kernel tree due to the license issues) but had to run in User Space and thus not getting a privileged amount of CPU slices?
The mainline Linux kernel gets tens of thousands of contributors.
If everyone followed Overstreet's "system" of pushing whatever commits, whenever, without coordinating with anyone outside their tree, then it would be complete and utter chaos. You can't just trust that everything _should_ work. Not at the kernel level.
If bcachefs is so unstable that it requires these rapid, unchecked commits to be reliable, then it shouldn't be mainlined. Period. Fix that up on your own, then come back later.
If bcachefs truly is worthy of being mainlined, then its maintainers should have no problem following the same rules as everyone else. Otherwise, see above.
Why should bcachefs get special treatment? There's just too much going on in mainline for this archaic kind of coordination.
Wow… bcachefs needs an intermediatery developer that ingests Kent’s fixes, and then injects them in a timely manner to the Kernel project for review
I am sure this response will help Kent with the fundraising /s.
Literally replying nothing would have been better.
He needs to take a breather and think two weeks before his next action.
Awhile back he posted a kind of whiny complaint on patreon, so I unsubscribed. I think the ideas are neat, but for filesystems the execution is the entire ballgame. In this case working well with others isn’t just important, it is essential.
Anyone using a filesystem that depends on chaotic patch processes to ‘not lose their data’ needs to evaluate just how important their data is. This developer needs an ego check. He simultaneously thinks his filesystem is god’s gift to the world and thinks it’s so broken that the world is going to break if he can’t just Willy nilly submit patches out of process. He also clearly thinks he has way more users than he does. No one except hobbyists use an experimental filesystem in production, and anyone who is seeing this exchange is likely scrambling to get off of it anyway
"I need my bug fixes to ship faster"
"Just don't release buggy builds"
"But then how would my bug fixes make it out quickly?"
Kent talks like he is the creator of the linux kernel, to the creator of the linux kernel
I just noticed the subtitles puts in "big Indian" instead of "big endian" 😂. You're right, Kent, people typically don't run Linux on big Indians. Or any Indians.
fixed
Though big indians run linux
BRO, Linux isn't YOUR PROJECT, you DON'T get to make the rules! Linus should issue an official warning, and if you don't go away and at least THINK about why what you did was wrong, thats it, its easily 3 strikes by now, i count at least 2 here.
I get the impression that kent has no enterprise experience.
"Linus is a saint"
I would say that this is the understatement of the year within this context. He is showing a level of patience with this person that others would have a hard time to muster.
I had a bit of a laugh when I read the part about him wanting Linux to have a filesystem to be proud of, one that is robust and stable. So why not contribute to existing filesystem codebases instead of rolling your own? XFS has been around for ages and is both stable and robust and performs rather well to boot. The same goes for ext4, although that is not my personal favourite, and while I have design philosophy issues with ZFS and btrfs, even I have to admit that they are both robust and stable.
And as Linus points out: mistakes happen. Bugs happen. Build failures happen. You can't have a project as vast and complex as a kernel without having them pop up on a regular basis. That i s why all these procedures and processes exist: to minimize the number of times they pop up.
Imagine being so narcissistic that you speak to Linus Torvalds that way...
btrfs doesn't work? I know it has its inconveniences, but it seems reliable enough for me to use it for years.
Another comment mentioned that it struggles with RAID5/6, I think.
@@itskdogThat’s why I don't use raid5/6
I use ZFS, it’s been extensively tested and improved upon. We just need the proper license so it can be in the Linux kernel.
@@itskdog yeah the good old RAID5/6 write hole, RAID 1 and 10 have been stable for me for close to 6 years, even started as a HDD pool and got turned into a SSD pool over time, hecc, it started as a RAID1, then I added 2 more drives and converted it to RAID1, I just love that flexibility. But that conversion was tense, not because it didn't work, but I was expecting it to fail, but nah, all good
@@itskdog I've read that RAID56 is stable in more recent kernel versions, but still labelled as unstable for production.
Ken identifies himself with the code he wrote. Criticizing the code means criticizing him as a human being.
"That has to work both ways"
No, it doesn't. One of you is the main guy and maintainer of the Kernel. The other is not. Linus is being more than accommodating but ultimately he doesn't have to care about how you do things, you do have to care about how he and every other developer working on the kernel does things.
Back when Linux only ran on floppies and one particular type of hard drives, and was ENTIRELY written by Linus Torvalds, there was a filesystem.
This means that back in the 1990s, Linus ACTUALLY wrote a filesystem - not necessairly a powerful, efficient, or innovative one, but still a filesystem.
In fact, the very first thing you need for a kernel is a filesystem.
11:59 My man is complaining about people reporting bugs on a stable release when he doesn't follow the stability requirements
The Tony Hawk analogy was great
I think that Linus needs to get rid of all other 100 file system that are unnecessary.
Embrace the one true filesystem: NTFS
@@CptJistuceit’s already in linux
@@Myname-l3h But it isn't the only file system, just one of many.
Nothing says "good and reliable file system" like releasing patches ASAP even if I break something.
the filesystem is called "bca chefs" from now on
Chefs belonging to the Board of Culinary Arts?
FFS this was painful.
If i was Linus, would have been like, fuck it, and dropped bcache from mainline.
Really interesting stuff honestly. It's reminding em slightly of templeOS, but nothing will reach that high
No other developer has been endowed with divine intellect.
We've been using btrfs for a while now on the main work laptop. We did not blame it for a folder disappearing last year. We keep backups, so yeah. It seems reliable enough. We are enjoying the dynamic allocation of space (so no more fixed partition size). But haven't bothered with snapshots, subvolumes, or RAIDs. Not sure what good it will do to add the chcksums and compression of bcachefs to the list of things we already don't use.
Linus has to maintain a kernel and a full-fledged asylum, which is harder, I wonder...
i am impressed that linus even has time to deal with that guy in such a long chain of messages.
Main character syndrome.. aka "rockstar" ego.
I'm impressed how Linus improved his communication skills and his patience those last years. I think a few years ago there would have been a flamethrower involved, at the very least.