linux-kernel - Re: [PATCH] Ability to read the MKTME status from userspace

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] [day] [month] [year] [list]

Message-ID: <5bc72926-2193-79af-0139-97bd63857129@intel.com>
Date:   Fri, 19 Jun 2020 08:36:02 -0700
From:   Dave Hansen <dave.hansen@...el.com>
To:     Richard Hughes <hughsient@...il.com>
Cc:     Daniel Gutson <daniel@...ypsium.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        x86@...nel.org, "H. Peter Anvin" <hpa@...or.com>,
        Arnd Bergmann <arnd@...db.de>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Peter Zijlstra <peterz@...radead.org>,
        "David S. Miller" <davem@...emloft.net>,
        Rob Herring <robh@...nel.org>, Tony Luck <tony.luck@...el.com>,
        Rahul Tanwar <rahul.tanwar@...ux.intel.com>,
        Xiaoyao Li <xiaoyao.li@...el.com>,
        Sean Christopherson <sean.j.christopherson@...el.com>,
        Dave Hansen <dave.hansen@...ux.intel.com>,
        linux-kernel <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] Ability to read the MKTME status from userspace

On 6/19/20 8:02 AM, Richard Hughes wrote:
>> Someone does 'cat /proc/mktme' (or whatever) and it says "1" or
>> whatever, which means yay, encryption is on.  What do they do?
> I think "is my memory encrypted" for Intel has to be a superset of:
> 
> 1. TME in CPU info
> 2. not disabled by the platform
> 3. not using unencrypted swap
> 4. not using a memory accelerator
> 5. entire DRAM area is marked with EFI_MEMORY_CPU_CRYPTO

Also realize that this can all be true at one point in time, but can
change if memory is added.

> It seems the only way to answer the questions and make it easy for the
> consumer to know the answer is to ask the kernel for each of the 5
> different questions. At the moment we can only get 1, 3, maybe 4, soon
> to be 5, but not 2.

Actually, the accelerators I had in mind would show up in the memory map
and would have EFI_MEMORY_CPU_CRYPTO properly set by the firmware.

In any case, if we do something like this, I think it fundamentally
needs to be more fine-grained than the whole system.  It probably needs
to be on a per-NUMA-node basis.  That's really the only way for us to
provide meaningful promises about encryption to end users.