Return to Level1Techs.com

Intel FUBAR ... again - Kernel memory leak in nearly every Intel CPU of the last decade (Spectre hits everyone, Meltdown still Intel exclusive)

intel
mega_thread
bug

#1120

@foppe they cant write a working patch but then the claim is “safe sillicon” :man_facepalming:

am more and more wondering how long they knew.

Now this sounds more plausible,sometime 2019 .


#1121

I doubt the safe silicon on intels side. No way to get that into production in less than a year.


#1122

So we gain 25% performance from a chip revision then lose 25% performance from a chip revision.

:smirk_cat:


#1123

Jup but looking back the last 3/4 generations does it make a difference, new arch is lastly just moar clocks no ipc uplift, basicly they just do updates and no ficking arch improfements.

They bet the farm on there “fab” fabs and got bitten in the ass by 10nm, and the 14nm +++ is sorta killing the cillicon by heat and density when clocked up.


#1124

What happened to Nano carbon being used in CPUs ? Or did they ditch that for quantum computing now.

It’s kind of sad funny that those who stuck with a 2500k sandybridge are still pretty much able to say that it was one of the best VFM’s CPU’s around. Looks like there was no serious improvements to be had at this level of nm so the milking continued and features got stripped to eek out small performance improvements.


#1125

It was probably mostly hype.

Like how you hear about this 16 TB SSD COMING OUT IN 2017 or some hyped up bullshit like that? Yeah, and not only do I NOT see it, it still costs $500 for a 2 TB SSD.


#1126

There’s been 16 TB (well, actually 15.6 TB or something) Samsung SSDs for the enterprise since 2016.

/edit 15.36 TB actually. Samsung PM1633a. Of course they’re not exactly cheap but let’s be realistic, it’s going to be some time before we see reasonably priced flash in any kind of product.
Also the 16 TB Intel P4500 series literally just hit so… There’s that.


#1127

Speculative Execution Panel at LCA 2018

Apparently FreeBSD had ~11 days prior notice.


#1128

https://www.phoronix.com/scan.php?page=news_item&px=Spectre-v1-IBPB-Linux-4.16-Pull


#1129

thanks was going to post.


#1130

#1131

#1132

We all knew it was going to be rough for Intel…

The company’s filing also listed three individual Intel shareholders who had filed actions against members of Intel’s board and other managers, alleging that they failed their duties to Intel by failing to take action in relation to alleged insider trading.

It did not specify what allegations of wrongdoing the suit dealt with.

Intel Chief Executive Brian Krzanich sold 889,879 shares in the company on Nov. 29 as per a trading plan adopted on Oct. 30, making roughly $39 million from the sale, well before the details of the flaw were made public.

Bloomberg reported last month that although Intel says Krzanich’s sales were part of a pre-arranged stock plan, securities lawyers believed the larger-than-usual transaction could be examined by the U.S. Securities & Exchange Commission (SEC).

Krzanich said last month that Google researchers told Intel of the flaws “a while ago”.


#1133

https://www.phoronix.com/scan.php?page=news_item&px=FreeBSD-Spectre-Meltdown-Fix

today


#1134

^That’s good news for PfSense.

In other news:

So will this actually fix the problem without being too painful?

Also, where is AMD on Spectre? I have yet to see any BIOS updates on AMD architectures that claim to patch for Spectre. Apparently getting patched is more difficult than I imagined when this was first exposed…


#1135

“… for you to try out”

Yeah, sure. … What could possibly go wrong? :rofl:

And that is exactly what AMD just isn’t doing: rush something out.
They also don’t have to worry that much since it is “only” one version of spectre.


#1136

What I find most funny about this is that they had half a year, planned a release date for the information (which was leaked beforehand), but didn’t have a patch ready for that date either … soo… what?

^For AMD and Intel both that is.


#1137

It’s actually their plan.

Nuke 500,000 CPUs with a bad firmware update and get people to upgrade!


#1138

ya… noticed that. Lets see who starts screaming first ?


#1139

The look on Intel’s face when that backfires and everyone buys Ryzen instead, I hear those APUs are pretty good upgrade right about now.