linux-kernel - Re: [PATCH 2/2] KVM: VMX: Add Force Emulation Prefix for "emulate the next instruction"

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]

Message-ID: <20180327051840.GF16974@char.us.oracle.com>
Date:   Tue, 27 Mar 2018 01:18:40 -0400
From:   Konrad Rzeszutek Wilk <konrad.wilk@...cle.com>
To:     Wanpeng Li <kernellwp@...il.com>
Cc:     LKML <linux-kernel@...r.kernel.org>, kvm <kvm@...r.kernel.org>,
        Paolo Bonzini <pbonzini@...hat.com>,
        Radim Krčmář <rkrcmar@...hat.com>,
        Andrew Cooper <andrew.cooper3@...rix.com>
Subject: Re: [PATCH 2/2] KVM: VMX: Add Force Emulation Prefix for "emulate
 the next instruction"

On Tue, Mar 27, 2018 at 01:03:15PM +0800, Wanpeng Li wrote:
> 2018-03-27 12:55 GMT+08:00 Konrad Rzeszutek Wilk <konrad.wilk@...cle.com>:
> > On Tue, Mar 27, 2018 at 12:40:20AM -0400, Konrad Rzeszutek Wilk wrote:
> >> On Mon, Mar 26, 2018 at 07:12:15PM -0700, Wanpeng Li wrote:
> >> > From: Wanpeng Li <wanpengli@...cent.com>
> >> >
> >> > This patch introduces a Force Emulation Prefix (ud2a; .ascii "kvm") for
> >> > "emulate the next instruction", the codes will be executed by emulator
> >> > instead of processor, for testing purposes.
> >>
> >> Can you expand a bit ? Why do you want this in KVM in the first place?
> 
> Please refer to the original discussion(Force Emulation Prefix part).
> https://lkml.org/lkml/2018/3/22/220

That really should be part massaged in this patch as part of the description.
> 
> >> Should this be controlled by a boolean parameter?
> >
> > .. per guest. That is instead of a global one, have a per guest one?
> 
> As Paolo pointed out offline:
> > Testing without the hacks being done by emulator.flat (TLB mismatch between instructions and data).

And also this above.

> I think a global module is enough for testing.

If so, perhaps have it wrapped with #ifdef DEBUG?

No need to put code gadgets that won't be utilized 99% of time.

> 
> Regards,
> Wanpeng Li