AMD provides update on Spectre and Meltdown patches

[HEXUS]

And Nvidia has patched its drivers to mitigate for host CPU (not its GPU) and OS issues.

Read more.

[spacein_vader]

So Nvidia GPUs aren't vulnerable but they're releasing an update anyway? Eh?

[scaryjim]

So Nvidia GPUs aren't vulnerable but they're releasing an update anyway? Eh?

It would appear that certain calls in the nvidia GPU driver could be an avenue of attack against vulnerable CPUs. Presumably some software contains more instructions and pathways that could be used as a channel for the attacks, and since GPU drivers will cross the user and kernel memory spaces to some degree it's plausible to me that you could tweak a GPU driver to make it less vulnerable to use as an attack vector.

CAT-THE-FIFTH has pointed out elsewhere that nvidia moved a lot of the scheduling for their GPUs to software a few generations ago, which may be a partial explanation for why nvidia have been particularly pro-active in optimising their drivers.

[DanceswithUnix]

CAT-THE-FIFTH has pointed out elsewhere that nvidia moved a lot of the scheduling for their GPUs to software a few generations ago, which may be a partial explanation for why nvidia have been particularly pro-active in optimising their drivers.

I would hope that scheduling would be in a user mode driver, so no impact.

[Corky34]

I would hope that scheduling would be in a user mode driver, so no impact.

True but doesn't it, at some point, have to do a context switch so the Kernel can dispatch the data to the GPU?

[DanceswithUnix]

True but doesn't it, at some point, have to do a context switch so the Kernel can dispatch the data to the GPU?

No, the GPU is memory mapped and so during set up can be mapped such that a user mode process can write to it.

Kernel needs to do support stuff like route interrupts and initial hardware setup, but the main fast path you would hope will be user only.

Other hardware I have worked on functions that way, I presume Nvidia would do the same.

[DanceswithUnix]

Interesting article: https://arstechnica.com/gadgets/2018...t-performance/

an unpatched Linux system can switch about 5.2 million times a second. Dual page tables slashes that to 2.2 million a second; dual page tables with PCID gets it back up to 3 million.

Which I think puts some context on this. If your workload is trying to push the CPU to do over a million kernel context switches per second, like some of the benchmarks do, then you could see some slowdown.

[kalniel]

Ah, that explains why my new westmere hasn't particularly improved (meltdown specifically affected) things with PCID support (alone), need Haswell's INVPCID as well :/

[DanceswithUnix]

The big question I still want an answer to: Why is the Spectre in the logo holding a stick?
Edit: Oh, it's a branch isn't it. *foreheadslap*

[scaryjim]

... Edit: Oh, it's a branch isn't it. *foreheadslap*

Although at a stretch you could argue it's a reference to (Out of) Order of the Stick (execution)...

[Iota]

Ah, that explains why my new westmere hasn't particularly improved (meltdown specifically affected) things with PCID support (alone), need Haswell's INVPCID as well :/

Yes, annoyingly earlier Sandy Bridge has PCID, but not INVPCID, which means;

Windows will use PCID if the hardware supports INVPCID—that means Haswell or newer. If the hardware doesn't support INVPCID, then Windows won't fall back to using plain PCID; it just won't use the feature at all.

So you end up with more overhead due to TLB flushing on older CPUs. It also probably explains why Intel has cherry picked which CPUs it shows performance impacts on, because they know the fix for older CPUs introduces more overhead in certain scenarios. It really does depend on what you are using your PC for though.

[ohmaheid]

On the other hand....Complete silence from Intel on their mitigation scenarios with specifically, the Meltdown problem (apart form " We take customers security seriously") Reading 90% of the tech press - you would be forgiven for thinking that there is comparable culpability from InteL, AMD, and ARM.
There is not. Intel still has the Meltdown architecture problem that affects ALL Intel products. Maybe that's why their CEO sold off all his stock in October, when the problem was going to be made public. But kept only the bare minimum that is allowed legally.

[deadcatwalking]

Maybe instead of worrying so much about viruses, protecting yourself from them and clogging up your computer with rubbish to combat it, perhaps just don't click on stupid things if you don't know what they do, don't open emails you don't know where they came from and don't visit unknown websites?

Reminds me of people that buy into weight loss gimmicks when all you EVER need to do is make sure you're using up more caloric energy than you take in.

It's not rocket science.

[kompukare]

Maybe instead of worrying so much about viruses, protecting yourself from them and clogging up your computer with rubbish to combat it, perhaps just don't click on stupid things if you don't know what they do, don't open emails you don't know where they came from and don't visit unknown websites?

Well, plenty of well-known website have had problems with malicious things being served with their adverts etc., so it's naïve to imagine that people only get viruses when visiting dodgy sites.

[Tabbykatze]

Maybe instead of worrying so much about viruses, protecting yourself from them and clogging up your computer with rubbish to combat it, perhaps just don't click on stupid things if you don't know what they do, don't open emails you don't know where they came from and don't visit unknown websites?

Reminds me of people that buy into weight loss gimmicks when all you EVER need to do is make sure you're using up more caloric energy than you take in.

It's not rocket science.

As a Security Professional i cannot begin to list how many different kinds of wrong you are.

A large percentage of hacks are performed as a drive by while you're visiting legitimate benign sites. This is through malicious cookies or a hackvertisement backdooring through your browser via scripts.

The internet is not safe, wherever you go

[Mr_Jon]

Maybe instead of worrying so much about viruses, protecting yourself from them and clogging up your computer with rubbish to combat it, perhaps just don't click on stupid things if you don't know what they do, don't open emails you don't know where they came from and don't visit unknown websites?

Reminds me of people that buy into weight loss gimmicks when all you EVER need to do is make sure you're using up more caloric energy than you take in.

It's not rocket science.

Yes...with viruses - but the whole spectre/meltdown hoo-hah goes behind that, in that it's out of the hands of the end user.

Couldn't agree more though when it comes to being careful what one clicks on. I've had one virus in 25 years of computing. The work computer, on the other hand, has a virus/malware/pup every other week due to its principle users clicking on crap and insisting on using an unprotected browser (which in itself will have installed even more bumph without them realising).