GrapheneOS / linux-hardened

Minimal supplement to upstream Kernel Self Protection Project changes. Features already provided by SELinux + Yama and archs other than multiarch arm64 / x86_64 aren't in scope. Only tags have stable history. Shared IRC channel with KSPP: irc.freenode.net ##linux-hardened. Currently maintained at https://github.com/anthraxx/linux-hardened.
https://grapheneos.org/
Other
390 stars 102 forks source link

CVE-2017-5753 - Spectre Variant 1 #71

Closed ralaud closed 6 years ago

ralaud commented 6 years ago

Is CVE-2017-5753 already patched ?

Because, screenshot-vuln

With regards.

thestinger commented 6 years ago

There's no complete patch available, so what are you asking for exactly?

ralaud commented 6 years ago

There's no complete patch available

I did not know that there is no complete patch. Thanks for your fast information and for your work in linux-hardened kernel.

thestinger commented 6 years ago

I don't think there ever will be a complete patch available, unfortunately. They can only fix things on a case-by-case basis now.

thestinger commented 6 years ago

The latest release is 4.15.2.a though, so you can update to that, which has more of these fixes.

ralaud commented 6 years ago

So its necessary to upgrade CPU later, if I want a complete fix?

thestinger commented 6 years ago

If Intel or AMD ever feels like making a CPU not vulnerable to it. The kernel can work around it on a case by case basis, which they're doing, but it can never realistically be fully mitigated.

thestinger commented 6 years ago

They won't necessarily be interested in making a CPU without this vulnerability. The performance hit may be too high...

ralaud commented 6 years ago

But big companies like Google, Amazon or Microsoft could force Intel and AMD. They need to focus on security, for their Cloud services.

ralaud commented 6 years ago

Is it possible to use machine learning for automating the case by case basis?

ralaud commented 6 years ago

Case by case method is working :)

screenshot-cve

thestinger commented 6 years ago

KPTI mitigates Variant 3, retpolines nearly fully mitigate Variant 2 (with other more expensive options available) but there isn't going to be a magic bullet for Variant 1.

thestinger commented 6 years ago

Having a few things sanitized doesn't really change that it's an issue everywhere else. It's also an issue for lots of userspace programs too, not just the userspace <-> kernel boundary or remote <-> kernel boundary.