Quotes of the week

[Posted January 5, 2022 by corbet]

And no, we don't add "benchmark tuning Kconfig questions" to the kernel. We leave those kinds of games to companies that need to fake their benchmark numbers.

— Linus Torvalds

The Linux kernel has X.509 processing inside it.
We're all going to die.

— Peter Gutmann

It could be worse. Your CPU could have a hidden x.509 processing code *and* a web server, that isn't getting regular security patches, and was probably written by firmware engineers.
Oh, wait....

— Ted Ts'o

Quotes of the week

Posted Jan 6, 2022 22:07 UTC (Thu) by flussence (guest, #85566) [Link] (1 responses)

The thread containing Linus' comment reminds me that the i915 driver has an even worse list of totally opaque tweakable numbers that shows up under EXPERT=y, and given the code's scope creep over the years the defaults are probably poor for a lot of devices.

Things like this are silly. Can you imagine if everyone had to answer questions about which RAID hash function to use instead of the kernel autoselecting the right one at boot?

Quotes of the week

Posted Jan 6, 2022 22:46 UTC (Thu) by pebolle (guest, #35204) [Link]

> Things like this are silly. Can you imagine if everyone had to answer questions about which RAID hash function to use instead of the kernel autoselecting the right one at boot?

So true.

Is it hyperbole to say that most Kconfig options shouldn't exist? Those options should be runtime tweaks only and even better - I want a pony - automagic runtime adjustments if possible.

Quotes of the week

Posted Jan 9, 2022 9:29 UTC (Sun) by CChittleborough (subscriber, #60775) [Link]

I like John Ioannidis’s reply to the last comment above:

We are already dead, it's just that the cache coherence circuitry hasn't caught up with us.