linux-kernel - Re: [RFC PATCH] x86/msr: Filter MSR writes

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 12 Jun 2020 19:48:01 +0200
From:   Borislav Petkov <bp@...en8.de>
To:     Linus Torvalds <torvalds@...ux-foundation.org>
Cc:     x86-ml <x86@...nel.org>, lkml <linux-kernel@...r.kernel.org>
Subject: Re: [RFC PATCH] x86/msr: Filter MSR writes

On Fri, Jun 12, 2020 at 10:20:03AM -0700, Linus Torvalds wrote:
> Since you already added the filtering, this looks fairly sane.
> 
> IOW, what MSR's do we expect people to maybe write to normally? You
> added MSR_IA32_ENERGY_PERF_BIAS as an allowed MST, maybe there are
> others?

Right, this MSR is being written by cpupower in tools/. My search was
confined within the kernel source only so there very likely are others.

If we want to add others, though, I think the criteria would be, if
writing to a MSR is safe - and this is where it becomes fuzzy but to use
the above example: if all it does is give hints to the hw but the hw can
ignore those hints and there's no other effect on the hw, then those
writes are fairly safe.

I'm pretty sure we'll massage that definition of "safe" over time.

> So I'm not against this, but I suspect the whitelist should be thought
> through more,

Yap.

> and then maybe the "allow_writes" parameter should be
> yes/no/default/<msr-list>, where "default" is that list of
> known-normal MSR's.

So I know of another out-of-tree tool which has a whole whitelist
management of MSRs: https://github.com/LLNL/msr-safe

Question is, what use cases we want to support. My thinking is to start
small and then keep extending it based on use cases we want to support.

> And I suspect it's a lot longer list than your single one. IIRC,
> people were working around CPU bugs or features by doing MSR writes at
> startup.
> 
> So the first phase might be to introduce this, but have the default
> for non-recognized MSR's be "log", not "deny".

Ok. How are we going to "learn" about those non-recognized MSRs? Ask
people to send us a note to lkml so that we can fix the list once they
see the logged message?

Thx.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette