x86 FPU Optimizations Land In Linux 5.2 That Torvalds Loves But Worries Of Regressions

[phoronix]

Phoronix: x86 FPU Optimizations Land In Linux 5.2 That Torvalds Loves But Worries Of Regressions

As part of the first week of changes for the Linux 5.2 merge window, a patch series providing some x86 FPU optimizations were merged though there is some concern there could be regressions on older hardware...

http://www.phoronix.com/scan.php?page=news_item&px=Linux-5.2-x86-FPU-Optimizations

[milkylainen]

There will always be regressions when changing complex core code.
Personally I'd rather break an egg or two once in a while over polishing a turd once more.

[sdack]

There will always be regressions when changing complex core code.

No, there isn't always a regression. There is no logical reason why a change to code - core code or not, complex or not - would always have to cause a regression. You're probably assuming the theoretical case, but the suggested change may well be harmless.

Even in the event that it does cause a regression, would I try to look at the affected code as well. It may not be an egg you're breaking, but to use your words, it might be a piece of turd you're breaking.

[Termy]

personally i think it's the right thing that Linus pulled it despite the slight chance of regressions on really ancient hardware. It's unlikely that PCs from the Pentium-era need new drivers and most of them probably aren't even connected to the internet i would assume

[rogerx]

To my knowledge, x86 (eg. 32 bit) is alive and well on smaller platforms. Most current and newer software drivers will execute just fine on 32 bit platforms as long as they're re-compiled against the newer kernels. Also, my Dell Inspiron laptop with a Pentium 3 chip runs just fine and still connects to the Internet. (Granted, MS Windows being a security risk.) I welcome the change, as long as the change does not purposely break code. (Ahem.) I also have a 2xP3 750 w/ 1G RAM server, that I ran 24/7 up until a few years ago, now temporarily mothballed as I need room for it, but will likely have it operational again. Nice item to have, a well tested backup computer! (If one builds a computer correctly, no need to toss it in the garbage after only a few years use!)

[ThoreauHD]

I just started getting some 'MODSIGN Can't find UEFI db list' error on 5.1.1. Not sure if it's because they suck or ubuntu mainline sucks, but somebody is sucking.

[schmidtbag]

No, there isn't always a regression. There is no logical reason why a change to code - core code or not, complex or not - would always have to cause a regression. You're probably assuming the theoretical case, but the suggested change may well be harmless.

I think you're taking his statement a bit too literally. He's saying that failures/problems are to be expected when meddling with something complex. Doesn't mean it is inevitable, just very probable and therefore the concept of there being an issue is implied.

Even in the event that it does cause a regression, would I try to look at the affected code as well. It may not be an egg you're breaking, but to use your words, it might be a piece of turd you're breaking.

That I think is a fair interpretation, though at this point, I don't think the FPU code was ever a turd considering how well it has held up over the years. Linux's FPU performance, to my knowledge, has been good.

[creative]

Honestly I would not recommend using a kernel that new with hardware from that era. It's great to still be able to use much older hardware and I like the idea of nothing going to waste especially if it can be used or 'you can't afford to upgrade or don't want to'., but in those use cases you will be wanting to opt for an older lighter kernel, possibly even an older distribution. Or even a new distribution tailored for older hardware.

Even on intel 7th gen, I disable javascript 97% of the time. Why? Why do I need it? So it breaks a bunch of web pages functionalities I don't want? Good! Was just wanting to read text anyway, not see a bunch of fancy graphics, code and images I did not ask for.

[wizard69]

At this point nobody should care if Pentium era hardware fails. You can’t improve Linux if you are tied to old technology, sometimes you just have to move on. I’m pretty bad in this regard too as I just sent an old 486 based machine to recycling. After awhile though you need to say is it really worth it ?

If old hardware issues are found I hope mainline developers have the strength to eat oh well time to upgrade.

[Teggs]

If kernel developers want to know if there are regressions, why don't they contract with Phoronix?