Bcachefs Merged Into Linux-Next

[Khrundel]

I see quite some enthusiasm here for BcacheFS, but I wonder what people expect to gain.
Why is another experimental filesystem so desirable? What performance jump do you expect to see?
What unique feature is offered hat no other provides?

With respect to filesystems I am in the conservative camp.

First of all, this filesystem inherits a great deal of BTRFS pros. It currently lacks readonly snapshots, an ability to mount subvolume, and backups via "btrfs send". I hope this will be added soon, at least latter.

Second, a tiered storage. You can select devices for read cache, devices for writing files, devices for long time storage. And you have control over this on by-directory or by-file level. For example you can set SSD for read and write cache for all filesystem and then make an exception for Downloads directory, making it keep out from SSD.

Third is write amplification. Unlike bcache or flashcache, bcachefs treats SSD as first class citizen, it not different from HDD except from write target policy. So space allocation is flash-friendly, like the f2fs, bcachefs allocates space by big blocks, combining near simultaneous random writes into one big and contiguous. With later move to "background target" during re-balance, making possible defragmentation. This also makes filesystem SMR friendly, meaning it would be a good default choice in current situation, when most storage drives are either SSDs or SMR HDDs.

[oiaohm]

I don't care about your or their opinion how "management" should behave I don't have to agree with them or you, if they don't defend themself or attack him openly with evidence, why should I pick any side?.

https://lkml.org/lkml/2023/9/6/888

So these kinds of "I'll just ignore _all_ basic rules" kinds of issues
do annoy me.

The reality is Linus and the VFS maintainer post openly with the issues they were having with the bcachefs.

https://lkml.org/lkml/2023/9/5/15 Bad one is like here were Kent calls IOMAP a pet project of Christoph Hellwig.

https://lkml.org/lkml/2023/9/8/219

Care to send a pointer?

Yes then Kent has not talked to Christoph Hellwig with what the issues with IOMAP are with bcachefs. So how does Kent know if the issues with IOMAP are not defects in IOMAP that should have been fixed the answer is Kent does not because he never brought the issues up with the Maintainer of IOMAP.

In fact the last time Kent talked with Christoph Hellwig before this round of submits Christoph was left believing IOMAP would be used as Kent was directed to talk to people like Christoph Hellwig.

blackiwid simple point the evidence to Kent massive miss behaviors is all over the different Linux kernel mailing lists. Yes Kent has been making up excuses why it does not have to talk to people or follow correct process over and over again.

https://lkml.org/lkml/2023/9/7/883

Kent code not being quality is not my own personal option. As Linus put it best Kent was submitting patches that simple result in a kernel tree that failed to build.

Because he defended himself he could be held responsible and he could change.

Again this is not understanding. When the Contributor Covenant Code of Conduct came in it because everything had been studied. To be correct you don't need to defend yourself to be held responsible. Linus Torvalds defending himself when he was was in the wrong came written justifications for others to carry out he same classes of bulling and abuse.

blacklwid bulling like it or not can have the snowballs effect where you have one person being a bully that bully justifies it then more people takes that justification and bullies then justify it ... this keeps on going until majority of your community can have turned in bully in one form or another.

I also find the argument btw not very strong that he did not go to Code of Conduct committee

You need to read again what I wrote.

https://docs.kernel.org/process/code-of-conduct.html Yes under the Linux kernel procedures you believe you are being bullied by Linux kernel developers in the mailing lists you are to bring it up with the Code of Conduct Committee. If you believe someone is being bullied you are to again report it to the Code of Conduct Committee who will use the tools I know about to pull out all correspondence on the mailing lists/git trees and so on and check if miss behavior has happened.

Its everyone responsibility to report bulling to the Code of Conduct Committee.

blackiwid as upper management you never should defend yourself against a charge of bulling. If accused of bulling you send it to be investigated. Not for the legal reason its that if you are in the wrong and you attempt to defend yourself you will just encourage others to copy your bad behavior that you will have to fix latter as upper management. There is no advantage as upper management to defend self against charge of bulling. What you want is a report saying if you were or were not bulling and also you most like will not publish that report that says you are not guilty for the same reason that someone could mess read it and use it as justification to-do the wrong things. Linux kernel has Linus Torvalds and other examples of why you should never ever defend yourself against a charge of bulling.

Yes the reality here if you truly believe Kent has been bullied from what you can find on the Linux kernel mailing lists you should email the Code of Conduct Committee not trying to defend him here. Yes that Committee is not restricted to people who are submitting code to the Linux kernel for raising issues.

blackiwid the problem here you are presuming something should happen when in fact it should not happen. This is one of those things remember the saying "common sense is not that common". Lot of what people class as common sense says you will defend yourself from what you are accused of problem is that can be the totally incorrect answer. Accused of bulling never defend yourself that when you get legal or others to investigate what you have done because you cannot really afford to be wrong if you have bullied or not because either way you will just make matter worse attempting to defend self.

[blackiwid]

Yes the reality here if you truly believe Kent has been bullied from what you can find on the Linux kernel mailing lists you should email the Code of Conduct Committee not trying to defend him here. Yes that Committee is not restricted to people who are submitting code to the Linux kernel for raising issues.

If you can't read what I write and instead nonstop strawman me then I can't help you, I said I don't know it either way, I am NEUTRAL, you make the claim or attack me for that therefor you must know that he was not bullied once in 15 years.

Because if you were not sure about that you just had exactly my opinion and attack me for nothing.

Sure there is this other discussion about processes you have a strong believe in but that are just subjective Opinions no objective facts, the only way that could be objective is if there is a god and he told you so, otherwise it's not objective.

And yes there is cases where not fighting back publicly is the better way, but there are other cases where that is fatal. Even psychologically that can be essential to stay sane and even save a person from suicide. Sure about such a claim probably likely nobody commits suicide, but if we talk about sexual claims tons of people did that, they did not openly fought back and they got wrongfully demonized and they could not take it. So if for nothing else fighting back gives you agency and so on. So generalizing "people can't afford" is bullshit.

The time a legal process can take is so long that nobody can just shut their mouth for years and not take extreme psychological damage that is often unreversable. And because it's extremely unlikely that any legal action will result from that there is even less incentive to not speak up because not speaking up at least often if not always is on the cost of somebodies mental state.

Now I talk here not to specific about this case because you make broad claim that somebody that is accused should never fight back verbally but always only talk to a lawyer if necessary. Of course it's not your job to talk back to accusations that are so vague at least not in this case and sexual stuff get because of gynocentrism and puritanism treated always differently so in this cases you have to defend yourself even against very vague accusations, but that is different.

And the answer is different for different people as a let's say superstar with a lot of following and money to pay good lawyers yes not fight back might be the better option but that is not true for people without all this luxus, and if the accusations are not maybe even criminal, but still defamatory.

Even the "positive" examples like Johnny Depp did not really win, he made his loss slightly smaller with that legal case, but the damage from this accusations is done and him getting kicked from many movies based on lies cost him probably something around 50 millions.

His released sound bites, that got leaked or released before the process of Amber abusing him verbally in their talk sessions they recorded did probably more effective and much cheaper for his image than the whole process.

After listening to 1-2 Minutes of this recordings you know who was the victim and who the perpetrator. But the cases are of course very different, but you always make absolute general claims so I can bring up other cases.

And he could effort to leak that without hurting his legal case apparently. I also don't get why you can't just let it rest I asked if somebody knows more about it, nobody seem to answer, and then I have to suspend maybe temporarily maybe permanently depending on the actors future behavior this "case" and assume innocence, but I assume innocence for everybody. The claim is not did he do something wrong in the kernel submitting process because that is no strong moral wrong doing, and he got punished by his work getting rejected...

For me that is for now and maybe permanently (I have a feeling not really but who knows) done. I don't know why you need from me a declaration that he is Satan and everybody else is definitely completely innocent?

And I would btw not even primarily say the person accused should fight back, but others if they see no reason could tell him to either name concrete evidence or take that claim back. That is btw also typical if a women makes vague accusations you ask them to become very specific to be able to proof it any way, you don't want only vague accusations that are not falsifiable or provable or make a false impression.

"he assaulted me" -> concrete "well he made a sexual joke"... you can manipulate with using broad terms to not say specific strong things but the listener starts to think the most extreme things and then in the end you can say "but I didn't say it" (I just made everybody think this, by the use of manipulative language".

[hoohoo]

See? That wasn't so hard was it?

[oiaohm]

I said I don't know it either way, I am NEUTRAL, you make the claim or attack me for that therefor you must know that he was not bullied once in 15 years.

No this is means to put words in person mouth. What you are doing here is strawman. Say you don't know but raise a point you have no evidence for.

Now I talk here not to specific about this case because you make broad claim that somebody that is accused should never fight back verbally but always only talk to a lawyer if necessary.

Its not everyone its upper management.

Even psychologically that can be essential to stay sane and even save a person from suicide.

Upper management correctly setup has psychological resources in HR to support them this majorly changes things. Linus and all the core maintainers at the Linux foundation has this resource these days and that changes how they should respond. So they are unlikely to suicide by not fighting back because they have psychological councilors to talk so handle the stress to ever since Linus was corrected. Now Linux Maintainers fight back against someone like Kent who does not have this resource.... Fighting back and being wrong can cause a person to commit suicide.

The reality here in the Linux kernel case and most properly setup upper management systems its totally in valid for upper management to attempt to defend themselves against claims they will pass this problem off to the HR area setup for this with the Linux kernel that the "Code of Conduct Committee" and talk to there councilors.

As you said you could not find where the Linux Maintainers would defend themselves but that the correct thing for them todo. The reality here the Linux core maintainers are not going to because with the resources they have it not valid for them to defend themselves. Deformation charge if it proven by investigation would be the Linux foundation in may cases who would push that because its their staff/their brand that being defamed by attacking maintainers falsely. This is the upper management way they don't defend themselves the company defends them if they are in the right.

https://www.psychologytoday.com/au/blog/resilience-bullying/202209/how-stop-false-accusations-in-their-tracks

Yes claiming to be Neutral you just did where the other party has to prove that the accusation is false is a common way a party does a false accusation. You are the one that raised the 15 year claim its you that need to put up the evidence that it true not me put evidence that is false. Stop attempting to use the NEUTRAL claim as means attempt to place accusation without having to provide the evidence.

[mdedetrich]

blackiwid I would just stop responding, when oiaohm gets into this mental state its just never ending arguing and strawmans with no end.

[grahamperrin]

ZFS, block cloning

… upcoming …: https://github.com/openzfs/zfs/pull/13392

Re: ZFS block cloning in FreeBSD 14.0, see https://www.mail-archive.com/stable@.../msg01359.html | https://lists.freebsd.org/archives/f...er/004700.html

https://www.freebsd.org/releases/14.0R/ for draft release notes, and schedule. https://github.com/freebsd/freebsd-s...ts/releng/14.0 includes zfs-2-2-release.

I assume that BETA5 and BETA6 will be added to the schedule (thread at matrix-dev.freebsd.org).

[LinAdmin]

a) I see quite some enthusiasm here for BcacheFS, but I wonder what people expect to gain.
b) What performance jump do you expect to see?
c) What unique feature is offered hat no other provides?
d) With respect to filesystems I am in the conservative camp.

Ad a): An alternative to BtrFS with CRC of data and metadata and the possibility to repair Raid-1 using scrub.
Ad b): None.
Ad c): None, but irrelevant question.
Ad d): Me too.