Linux 6.7-rc2 Released: "Slightly Larger Than Average"

[phoronix]

Phoronix: Linux 6.7-rc2 Released: "Slightly Larger Than Average"

Linus Torvalds just released Linux 6.7-rc2 as the second test release of what will become Linux 6.7 stable around the end of the calendar year...

https://www.phoronix.com/news/Linux-6.7-rc2-Released

[fitzie]

kent is getting salty.

https://lore.kernel.org/linux-bcachefs/[email protected]/T/#m981e6274611d154fa4b05d6409e12b83e7e09ec5

> 10) Anything you think an article about BCacheFS should absolutely
> mention?

Would personally love to see some non-phoronix benchmarks

the generous interpretation is that he wants an independent confirmation of where bcachefs is, performance wise, but it's easy to interpret this as an attack, which kent is well known for.

[AlanTuring69]

kent is getting salty.

https://lore.kernel.org/linux-bcachefs/[email protected]/T/#m981e6274611d154fa4b05d6409e12b83e7e09ec5

the generous interpretation is that he wants an independent confirmation of where bcachefs is, performance wise, but it's easy to interpret this as an attack, which kent is well known for.

His interactions with other people have been jarring and he doesn't know how to handle criticism or recognition. It's a really interesting project and he's a clearly talented developer but if he wants to remain involved he's going to need to grow up. Most kernel development is done by professionals and there will come a day where people will refuse to work with him.

[timofonic]

His interactions with other people have been jarring and he doesn't know how to handle criticism or recognition. It's a really interesting project and he's a clearly talented developer but if he wants to remain involved he's going to need to grow up. Most kernel development is done by professionals and there will come a day where people will refuse to work with him.

He needs to be a lot more cautious, IMHO.

He seems to have ADHD traits. I have absolutely nothing against it and I know many extremely talented geeks with that disorder, but it requires coping and extreme improvement of impulse control.

I may think certain people are idiots because I'm angry or it's true, but that feedback may not be pragmatic and cause severe conflicts.

[MastaG]

I'm really waiting for the AMD preferred core enablement (by default of course) and the realtime scheduling patches.
Hope they'll both land in the near future.

[Old Grouch]

kent is getting salty.

https://lore.kernel.org/linux-bcachefs/[email protected]/T/#m981e6274611d154fa4b05d6409e12b83e7e09ec5

the generous interpretation is that he wants an independent confirmation of where bcachefs is, performance wise, but it's easy to interpret this as an attack, which kent is well known for.

I prefer the generous interpretation.

It is rational to prefer to have multiple independent tests. In this way, you can have more confidence in the results if/when they agree, and find areas for investigation and clarification if/when they do not.

I would be more suspicious of, "Phoronix has spoken, no more tests are necessary".

[marios]

kent is getting salty.

https://lore.kernel.org/linux-bcachefs/[email protected]/T/#m981e6274611d154fa4b05d6409e12b83e7e09ec5

the generous interpretation is that he wants an independent confirmation of where bcachefs is, performance wise, but it's easy to interpret this as an attack, which kent is well known for.

What are you talking about? The entire response was on point and far from an attack. He was not afraid to point out limitations and future improvements. How can you take the phoronix statement in bad faith? Anyone who cares about performance wants as many benchmarks as possible and currently only phoronix has ran bcachefs benchmarks. Is it bad to want more? They might narrow down the performance issues, or trigger more bugs. It is natural to want all the feedback you can get, in order to fix (potential) issues.

[varikonniemi]

kent is getting salty.

https://lore.kernel.org/linux-bcachefs/[email protected]/T/#m981e6274611d154fa4b05d6409e12b83e7e09ec5

the generous interpretation is that he wants an independent confirmation of where bcachefs is, performance wise, but it's easy to interpret this as an attack, which kent is well known for.

Follow that message thread and you get to an ironic bug report by the reporter where he had his BTRFS volume get corrupted beyond repair while playing around with the pre-rc1 kernel to test bcachefs

[timofonic]

Follow that message thread and you get to an ironic bug report by the reporter where he had his BTRFS volume get corrupted beyond repair while playing around with the pre-rc1 kernel to test bcachefs

Ironic? It's quite normal, Btrfs is an extremely overvalued filesystem.

An XFS developer already collaborates in Bcachefs, I wonder why.

I think Michael didn't check thoroughly the benchmarking process. I would hope Phoronix gets more financing and get someone to help.

Kent is a bit weird and he needs to learn to be a lot more diplomatic. Sometimes a friendly lie can be better than the pure truth. Another developer may be an idiot, but saying it may be worse than not saying it at all.

[varikonniemi]

Ironic, as in happening contrary to expected and turning out amusing. When testing bcachefs you would not expect btrfs, the mature COW fs, to get corrupted.

I don't think phoronix did anything wrong with the bcachefs benchmark, he simply ran it with default settings. Now bcachefs has changed the defaults, which probably should have happened before submitting to linux