Linus Torvalds Gets Coding To Improve Linux's User-Mode Stack Expansion

[phoronix]

Phoronix: Linus Torvalds Gets Coding To Improve Linux's User-Mode Stack Expansion

Linux creator Linus Torvalds doesn't write as much actual kernel code these days as he used to. These days he's often busy overseeing the upstream kernel development community with reviewing code, managing releases, and chiming in on mailing list discussions. Once in a while though he gets down and dirty with some low-level kernel hacking just as he's done now for Linux 6.5 with improving the user-mode stack expansion code...

https://www.phoronix.com/news/Linux-65-User-Mode-Stack-Expand

[zcansi]

... is this missing a link to the source?

[Michael]

... is this missing a link to the source?

HTML was messed up, should be made more clear now.

[Radtraveller]

So….
who vettes and approves Linus’ changes?
:-p

[timofonic]

I think news like this should ve a lot comprehensive and didactic, so mere users are able to understand the importance of it.

I feel stupid trying to understand what this is useful for and how it works (what's that stack stuff, why this is and improvment, how important it is, principles behind this, etc), like me having severe mental retardation.

Could someone do a comprehensive, didactic, etc description? Please.

I fail to understand fully the stack concept, I just have some glimpse of it.

I think this may:
- Reliability: Reduce stack overflow even for legacy applications
- Resource management: Better performance and resourceoptimization, more efficient memory usage, better dynamic memory management,, resource optimization,
- Flexibility: Enhaced handling of large data structures, deep call stacks, large thread stacks, better multithreading.

Am I correct? Please prove me wrong abd be descriptive!

Did he Implemented strack contraction when needed too? Is this dynamic?

Can be an attack vector? How can be mitigated if it is.

Any other operating systems implementing this?

[timofonic]

Sorry for the grumpy and harsh message, deleted...

[X_m7]

I'm guessing the "newsworthy" part here is really just Linus Torvalds actually writing code instead of the code itself, as far as I can tell it's just a fix for a hack that is no longer working right due to other changes. In fact the listed commits in the merge are also in Linux 6.4.1 (https://lwn.net/Articles/937080/, see Linus' commits).

[schmidtbag]

timofonic Dude... relax. So long as your post isn't spam or too political, it will get approved within a few hours. Case in point: I saw it within 2 hours of you posting it. I'm a premium member and sometimes even the most innocent of posts will still seemingly randomly get unapproved, but it's fine because it won't be long for me to wait. I've had several posts sitting in unapproval limbo that frankly weren't conducive to a healthy community discussion and eventually got approved. Also, a little ironic of you to harp about "moronix", if you @-mention your username...

Anyway... while it's clear that this change has something to do with VMs, I agree that it isn't really clear what the improvement is.

[erniv2]

It´s not about VMs as in Virtual Maschine it´s about the VMM the Virtual Memory Manager and it fixes a locking bug that surfaced with the usage of Maple Trees VMM code so this is a needed bugfix there should be no performance improvements etc. it´s there to prevent userspace from modifying a page in the tree while another process uses it i guess so the lock is now more strict then before.

[timofonic]

timofonic Dude... relax. So long as your post isn't spam or too political, it will get approved within a few hours. Case in point: I saw it within 2 hours of you posting it. I'm a premium member and sometimes even the most innocent of posts will still seemingly randomly get unapproved, but it's fine because it won't be long for me to wait. I've had several posts sitting in unapproval limbo that frankly weren't conducive to a healthy community discussion and eventually got approved. Also, a little ironic of you to harp about "moronix", if you @-mention your username...

Anyway... while it's clear that this change has something to do with VMs, I agree that it isn't really clear what the improvement is.

I deleted the rant, you are right. But the post isn't still approved. Im sorry.

Anyway, I'm pasting the relevant part here:

I think this may improve:
- Reliability: Reduce stack overflow even for legacy applications
- Resource management: Better performance and resourceoptimization, more efficient memory usage, better dynamic memory management,, resource optimization,
- Flexibility: Enhaced handling of large data structures, deep call stacks, large thread stacks, better multithreading.

Am I correct? Please prove me wrong and be descriptive!

Did he Implement stack contraction when needed too? Is this dynamic?