kvm.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Avi Kivity <avi@redhat.com>
To: "Nadav Har'El" <nyh@il.ibm.com>
Cc: kvm@vger.kernel.org, "Roedel, Joerg" <Joerg.Roedel@amd.com>,
	owasserm@redhat.com, abelg@il.ibm.com
Subject: Re: [PATCH 0/10] nEPT: Nested EPT support for Nested VMX
Date: Thu, 10 Nov 2011 14:26:30 +0200	[thread overview]
Message-ID: <4EBBC2F6.8050903@redhat.com> (raw)
In-Reply-To: <1320919040-nyh@il.ibm.com>

On 11/10/2011 11:57 AM, Nadav Har'El wrote:
> The following patches add nested EPT support to Nested VMX.
>
> Nested EPT means emulating EPT for an L1 guest, allowing it use EPT when
> running a nested guest L2. When L1 uses EPT, it allows the L2 guest to set
> its own cr3 and take its own page faults without either of L0 or L1 getting
> involved. In many workloads this significanlty improves L2's performance over
> the previous two alternatives (shadow page tables over ept, and shadow page
> tables over shadow page tables). Our paper [1] described these three options,
> and the advantages of nested EPT ("multidimensional paging").
>
> Nested EPT is enabled by default (if the hardware supports EPT), so users do
> not have to do anything special to enjoy the performance improvement that
> this patch gives to L2 guests.
>
> Just as a non-scientific, non-representative indication of the kind of
> dramatic performance improvement you may see in workloads that have a lot of
> context switches and page faults, here is a measurement of the time
> an example single-threaded "make" took in L2 (kvm over kvm):
>
>  shadow over shadow: 105 seconds
>  ("ept=0" forces this)
>
>  shadow over EPT: 87 seconds
>  (the previous default; Can be forced now with "nested_ept=0")
>
>  EPT over EPT: 29 seconds
>  (the default after this patch)
>
> Note that the same test on L1 (with EPT) took 25 seconds, so for this example
> workload, performance of nested virtualization is now very close to that of
> single-level virtualization.
>
>

This patchset is missing a fairly hairy patch that makes reading L2
virtual addresses work.  The standard example is L1 passing a bit of
hardware (emulated in L0) to a L2; when L2 accesses it, the instruction
will fault and need to be handled in L0, transparently to L1.  The
emulation can cause a fault to be injected to L2, or and EPT violation
or misconfiguration injected to L1.

-- 
error compiling committee.c: too many arguments to function


  parent reply	other threads:[~2011-11-10 12:26 UTC|newest]

Thread overview: 49+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-11-10  9:57 [PATCH 0/10] nEPT: Nested EPT support for Nested VMX Nadav Har'El
2011-11-10  9:58 ` [PATCH 01/10] nEPT: Module option Nadav Har'El
2011-11-10 12:23   ` Avi Kivity
2011-11-10 14:21     ` Nadav Har'El
2011-11-10 14:38       ` Avi Kivity
2011-11-10 15:14         ` Nadav Har'El
2011-11-10 15:21           ` Avi Kivity
2011-11-10  9:58 ` [PATCH 02/10] nEPT: MMU context for nested EPT Nadav Har'El
2011-11-10 10:31   ` Avi Kivity
2011-11-10 12:49   ` Avi Kivity
2011-11-10 14:40     ` Nadav Har'El
2011-11-10 15:19       ` Avi Kivity
2011-11-10 20:05         ` Nadav Har'El
2011-11-12 10:39           ` Avi Kivity
2011-11-12 21:37             ` Nadav Har'El
2011-11-13  9:10               ` Avi Kivity
2011-11-13 11:30               ` Orit Wasserman
2011-11-13 14:32                 ` Avi Kivity
2011-11-13 18:26                   ` Orit Wasserman
2011-11-14  8:25                     ` Avi Kivity
2011-12-08 15:21                       ` Nadav Har'El
2011-12-06 12:40                   ` Nadav Har'El
2011-12-06 13:07                     ` Avi Kivity
2011-11-23 15:06                 ` Nadav Har'El
2011-11-23 15:44                   ` Nadav Har'El
2011-11-24 13:36                     ` Avi Kivity
2011-12-07  9:06                 ` Nadav Har'El
2011-12-07 10:10                   ` Avi Kivity
2011-11-10  9:59 ` [PATCH 03/10] nEPT: Fix cr3 handling in nested exit and entry Nadav Har'El
2011-11-10  9:59 ` [PATCH 04/10] nEPT: Fix page table format in nested EPT Nadav Har'El
2011-11-10 10:37   ` Avi Kivity
2011-11-10 11:03     ` Nadav Har'El
2011-11-10 12:21       ` Avi Kivity
2011-11-10 12:50         ` Avi Kivity
2011-11-10 13:07   ` Orit Wasserman
2011-11-10 10:00 ` [PATCH 05/10] nEPT: Fix wrong test in kvm_set_cr3 Nadav Har'El
2011-11-10 10:00 ` [PATCH 06/10] nEPT: Some additional comments Nadav Har'El
2011-11-10 10:01 ` [PATCH 07/10] nEPT: Advertise EPT to L1 Nadav Har'El
2011-11-10 10:01 ` [PATCH 08/10] nEPT: Nested INVEPT Nadav Har'El
2011-11-10 12:17   ` Avi Kivity
2011-12-11 14:24     ` Nadav Har'El
2011-12-11 14:37       ` Avi Kivity
2011-11-10 10:02 ` [PATCH 09/10] nEPT: Documentation Nadav Har'El
2011-11-10 10:02 ` [PATCH 10/10] nEPT: Miscelleneous cleanups Nadav Har'El
2011-11-10 12:26 ` Avi Kivity [this message]
2011-11-13  8:52   ` [PATCH 0/10] nEPT: Nested EPT support for Nested VMX Nadav Har'El
2011-11-13  9:21     ` Avi Kivity
2011-12-12 11:37       ` Nadav Har'El
2011-12-12 13:04         ` Avi Kivity

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4EBBC2F6.8050903@redhat.com \
    --to=avi@redhat.com \
    --cc=Joerg.Roedel@amd.com \
    --cc=abelg@il.ibm.com \
    --cc=kvm@vger.kernel.org \
    --cc=nyh@il.ibm.com \
    --cc=owasserm@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).