[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <d5f7d098-25ec-a112-7bd4-b9712348d795@citrix.com>
Date: Thu, 4 Jan 2018 15:37:40 +0000
From: Andrew Cooper <andrew.cooper3@...rix.com>
To: Paolo Bonzini <pbonzini@...hat.com>,
"Woodhouse, David" <dwmw@...zon.co.uk>,
"pavel@....cz" <pavel@....cz>
CC: "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"tim.c.chen@...ux.intel.com" <tim.c.chen@...ux.intel.com>,
"torvalds@...ux-foundation.org" <torvalds@...ux-foundation.org>,
"tglx@...utronix.de" <tglx@...utronix.de>,
"andi@...stfloor.org" <andi@...stfloor.org>,
"aarcange@...hat.com" <aarcange@...hat.com>,
"gnomes@...rguk.ukuu.org.uk" <gnomes@...rguk.ukuu.org.uk>,
"dave.hansen@...el.com" <dave.hansen@...el.com>,
"gregkh@...ux-foundation.org" <gregkh@...ux-foundation.org>
Subject: Re: Avoid speculative indirect calls in kernel
On 04/01/18 15:32, Paolo Bonzini wrote:
> On 04/01/2018 16:29, Woodhouse, David wrote:
>> Adding that for KVM is in the Linux IBRS patch set that I've seen.
>> Didn't we already have a conversation about how the Linux patch set
>> does it as an atomically-switched MSR while you've done it manually in
>> Xen because it's faster?
> I'm also doing it manually in the RHEL versions of the KVM patches, for
> what it's worth.
Actually, I did it manually in Xen because I was expecting IBRS on AMD,
and there are no MSR load/save lists for PV or SVM guests. (Also yes,
I've been reliably informed that manually is much faster than VT-x
load/save lists.)
~Andrew
Powered by blists - more mailing lists