linux-kernel - Re: [PATCH] x86: kvm: Avoid guest page table walk when gpa

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:   Tue, 2 May 2017 16:44:19 -0500
From:   Brijesh Singh <brijesh.singh@....com>
To:     Paolo Bonzini <pbonzini@...hat.com>,
        Radim Krčmář <rkrcmar@...hat.com>
CC:     <brijesh.singh@....com>, <joro@...tes.org>, <kvm@...r.kernel.org>,
        <linux-kernel@...r.kernel.org>, <tglx@...utronix.de>,
        <mingo@...hat.com>, <hpa@...or.com>, <x86@...nel.org>,
        <Thomas.Lendacky@....com>
Subject: Re: [PATCH] x86: kvm: Avoid guest page table walk when gpa_available
 is set

Hi Paolo,

>>
>> I've tried the below patch and it seems to work fine. This does not
>> consider
>> PIO case and as you rightly pointed PIO should trigger #NPF relatively
>> rarely.
>> At least so far in my runs I have not seen PIO causing #NPF. If this sounds
>> acceptable approach then I can submit v2 with these changes and remove
>> gpa_val
>> additional.
>
> PIO can certainly cause #NPF, albeit rarely, so this is not a viable one.
>

For SEV guest cases, we anyway want to avoid any guest page table walks
(even if it happens rarely). Sounds like we have to depend on gpa_val variable.
I will refresh the patch to address one of Ramdim's feedback
(move vcpu->arch.gpa_val assignment from handle_exit to pf_interception).

Brijesh