linux-kernel - Re: Improve retpoline for Skylake

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]

Date:   Mon, 15 Jan 2018 10:10:00 -0800
From:   Andy Lutomirski <luto@...capital.net>
To:     David Woodhouse <dwmw2@...radead.org>
Cc:     Andrew Cooper <andrew.cooper3@...rix.com>,
        Jon Masters <jcm@...masters.org>,
        Henrique de Moraes Holschuh <hmh@....eng.br>,
        Andi Kleen <andi@...stfloor.org>, tglx@...utronix.de,
        x86@...nel.org, linux-kernel@...r.kernel.org, pjt@...gle.com,
        torvalds@...ux-foundation.org, gregkh@...ux-foundation.org,
        peterz@...radead.org, thomas.lendacky@....com,
        arjan.van.de.ven@...el.com
Subject: Re: Improve retpoline for Skylake



> On Jan 15, 2018, at 10:07 AM, David Woodhouse <dwmw2@...radead.org> wrote:
> 
>> On Mon, 2018-01-15 at 10:06 -0800, Andy Lutomirski wrote:
>> 
>>> Refill or not, you are aware that a correctly timed SMI in a leaf
>>> function will cause the next ret to speculate into userspace, because
>>> there is guaranteed peturbance in the RSB?  (On the expectation that the
>>> SMM handler isn't entirely devoid of function calls).
>> 
>> Couldn't firmware fill the RSB with a some known safe address, maybe
>> even 0, and then immediately do RSM?
> 
> Why don't we just unconditionally declare that *all* firmware that uses
> SMI for anything at all is broken?
> 

Because then Intel would need a new reference design for EFI authenticated variables.  Of course, this is totally doable and should be done anyway.

Also, Paolo reports that some misdesign or other makes it important to use SMI on CPU hotplug.