Torvalds' Comments On Linux Scheduler Woes: "Pure Garbage"

[Volta]

@michael
That's wrong! It's not about spinlocks from Linux. The developer who blamed the kernel has written his own spinlock in userspace without necessary knowledge. He did his own locking wrong and was measuring the wrong numbers. Spinlocks from Linux kernel are completely out of scope here.

While I'm not native English speaker I sometimes doubt Michael knows this language better.. Google developer sucks btw.

[dpanter]

"be aware that the likelihood that you know what you are doing is basically nil"

Oh Linyos Torovoltos, you glorious Soviet computer hacker bastard.
This is gold Michael, gold! We need this on a T-shirt.

[DanL]

The "new" Linus reminds me of Colonel Potter in M*A*S*H - "Mule muffins! Cow cookies! Horse hockey!"

[mb_q]

Some more interesting insights about user-space spinlocks:

https://matklad.github.io/2020/01/02/spinlocks-considered-harmful.html

https://matklad.github.io/2020/01/04/mutexes-are-faster-than-spinlocks.html

[ssokolow]

How is he rude in this case? I actually never thought he was rude, it can be frustrating working with people when you're simply not good at it, so that's why his comments used to be the way they were. And he always confronted the problem, not just the person.

I feel like I know exactly where he's coming from, and Scar from The Lion King emoted it best: I'm surrounded by idiots.

[bug77]

Well, this confirms what we were suspecting in the previous thread: that the Google dev was talking mostly out of his rear. I think the very issue of spinlocks in userspace was brought up.

And again, seriously: complaining about "the Linux scheduler" when Linux' scheduling is not only pluggable, but already has several implementations? Try (and tune) them all and if none of them is up to your standards, write your own damn implementation.

[Creak]

Whether you like it or not, Linus is quite rude in his answer. No matter how right or wrong the Google dev is, Linus sets an example of how you will be received if you tell something on the mailing list that you believe is right. This doesn't set a friendly environment and doesn't encourage new comers to share their issues, even if theirs might be real ones.

[Neraxa]

Well, this confirms what we were suspecting in the previous thread: that the Google dev was talking mostly out of his rear. I think the very issue of spinlocks in userspace was brought up.

And again, seriously: complaining about "the Linux scheduler" when Linux' scheduling is not only pluggable, but already has several implementations? Try (and tune) them all and if none of them is up to your standards, write your own damn implementation.

The only way to switch out CPU scheduler AFAIK is to use a patch set. Thats not what I consider pluggable. They could have made it so a CPU scheduler could be switched out at compile time with no overhead by using static linking or a macro that could make it a compile time option, but this does not seem to be the case. Only i/o schedulers can be swapped in runtime.

[menasaw683]

Whether you like it or not, Linus is quite rude in his answer. No matter how right or wrong the Google dev is, Linus sets an example of how you will be received if you tell something on the mailing list that you believe is right. This doesn't set a friendly environment and doesn't encourage new comers to share their issues, even if theirs might be real ones.

I disagree; wholeheartedly even. And here is why -- Assume for a moment you're at at your job and you're performing your assigned task to a sub-par standard. At some point or another, your manager will have to make you aware of this fact. And there really comes a moment at which point negative words, such as 'poor', 'no' and other such words do come into play. This is perfectly normal and always has been. Suggesting otherwise is just SJW-foolishness.

The nature of language requires this balance between both positive language and negative language. Whether the word is 'poor' or 'un-good', the point remains the same and always has remained the same. There literally is no way to create a natural language that does not, at some point, use words that describe things in a negative manner.

And the point I am making is this -- Read WHAT Linus is saying, not HOW he is saying it. The MESSAGE is important. Not the language used to convey said message. And the message itself is quite clear and hardly rude on its own. It simply is criticism of the methodology used by the Google developer to arrive at, obviously, incorrect conclusions. Again, at some point you will fail. Everyone does. And one has to be made aware of this failure. There is no way around this and no way to improve understanding other than being made aware of things we're doing wrong.

So, no... let's stop this excessive SJW-bullshit already. The Google developer works for Google... I think they're quite aware of the concept of criticism and being made aware of different points of view. They do not white knights protecting them from Linus' supposedly harsh language (which I never agreed with, I actually always found him quite tame to be truthful, kind of boring actually), they need to take what is being said and use that to improve their own methodology.

The goal is important. Not the words used to get to that goal. Substance over form, please.

[Volta]

Whether you like it or not, Linus is quite rude in his answer. No matter how right or wrong the Google dev is, Linus sets an example of how you will be received if you tell something on the mailing list that you believe is right. This doesn't set a friendly environment and doesn't encourage new comers to share their issues, even if theirs might be real ones.

Maybe it will encourage them to make some research before posting such bullshit claims. The Google developer showed himself as a fool.