Bcachefs Fixes Pull Once Again Frustrates Linus Torvalds - Two Choices Offered

[toves]

>
it is linus's project and tree. he has the right to do whatever he wants, especially when people are refusing to play by his rules. kent isn't just entitled to be in linus's tree.

Linus' ballpark. Anyone is free to take their bat and ball and go home.

I rather suspect Linus could implement a file system.
One half of the people reading Phoronix could implement the System V boot file system (bfs), the residual half would probably have difficulty with shoe laces - it is a really basic file system that makes FAT-12 appear clever.

[jacob]

Linus' ballpark. Anyone is free to take their bat and ball and go home.

I rather suspect Linus could implement a file system.
One half of the people reading Phoronix could implement the System V boot file system (bfs), the residual half would probably have difficulty with shoe laces - it is a really basic file system that makes FAT-12 appear clever.

To be fair, FAT-12 was clever. As a filesystem designed for floppies, with an initial capacity of 180K IIRC, and where 99.99% of all files were only a few kilobytes in size, it was space-efficient and reasonably performant. Remember, this was an era where every CPU cycle counted and dumb algorithms handling trivial cases were more efficient than advanced algorithms and data structures capable of dealing with stuff that was waaaaay beyond the hardware's capabilities.

[Developer12]

Bcachefs has what's likely the best CI setup among all Linux filesystems.

Then why the fuck does Kent keep repeatedly posting patches last-minute to the mailing list that break other people's shit?

[microchip8]

Oh man, this Kent guy. What and asshole!

[Summanis]

It seems like a lot more than the "Experimental" flag is gonna get removed in the next year.

>"If you're so convinced you know best, I invite you to start writing your own filesystem. Go for it."

If Kent knows best, why doesn't he write his own kernel? No one is stopping him.

[DanL]

Kind of a jerk move to accept the pull request and then whine about it.

No, the jerk move would be to not accept it and make users suffer bugs in 6.12. The only person he's being a "jerk" to is the person who can't understand that the kernel release schedule doesn't revolve around his project as the center of its universe.

Based on Kent's reaction, I hope Linus kicks it out of the kernel for 6.13. I think it would be better for everyone involved.

[horizonbrave]

Time to fork Linux

[gmokki42]

When AMD GPU driver was having trouble keeping things stable, Linus demanded that their feature PRs have to wait one kernel cycle in staging. That gives each new feature an extra few months to stabilize.

For a 10 year old filesystem having new features delayed by few months while being tested in linux-next branch would not do harm, jus improve the filesystem.

[Ferrum Master]

The greatest problem is, that such personas as Kent are in so many and different seats, it just produces shit code, thinks he's an untouchable princess and really nothing happens, the shit code just piles up.

[Quackdoc]

I do think a lot of these issues would be, just not an issue if kent would support DKMS for testing stuff. Kent has stated multiple times to not use DKMS for bcachefs, however it seems like a way better idea to ship things like this as a DKMS and then wait for the next RC or even release schedule.