All of lore.kernel.org
 help / color / mirror / Atom feed
From: Juergen Gross <jgross@suse.com>
To: George Dunlap <dunlapg@umich.edu>
Cc: xen-devel <xen-devel@lists.xenproject.org>,
	Boris Ostrovsky <boris.ostrovsky@oracle.com>
Subject: Re: Patches for stable
Date: Fri, 6 Apr 2018 12:44:51 +0200	[thread overview]
Message-ID: <2e25337d-103a-7812-acfd-37606807f496@suse.com> (raw)
In-Reply-To: <CAFLBxZbT8=WZokxKBD1KUaQ_br3OLnak0f_PgZSZM+a0V3ZjSg@mail.gmail.com>

On 06/04/18 12:13, George Dunlap wrote:
> On Thu, Apr 5, 2018 at 6:11 PM, Juergen Gross <jgross@suse.com> wrote:
>>> Option 1: Put the RSDP in lowmem unless we know the guest will use the
>>> address in start_info
>>> Pro: Existing Linux instances boot
>>> Con: Existing BSD instances whose memory is an exact multiple of 1 GiB
>>> will have slightly slower TLB miss times.
>>
>> ... whose memory is >=1GiB ...
> 
> No, the performance benefit only applies to sections of ram that are
> complete 1GiB-aligned sections.  If you have 1.75GiB, then the first
> gigabyte will be a 1GiB page in the p2m (and get the performance
> benefit) but the other 0.75GiB will be 2MiB aligned (and not).

Right. Moving the RSDP out of the first GiB will make it possible to
map that area with a single 1 GiB page.

> We can be pedantic and nail down exactly who will get what benefit
> when, but all it will do is demonstrate how limited this performance
> benefit is.

Moving the shared_info page of HVM guests allowing large pages to be
used increased HVM performance by several percent (e.g. munmap of
user pages was up to 2 times faster this way).


Juergen

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel

  reply	other threads:[~2018-04-06 10:44 UTC|newest]

Thread overview: 55+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-04-04 10:38 Patches for stable Juergen Gross
2018-04-04 14:27 ` Greg KH
2018-04-04 14:27 ` Greg KH
2018-04-04 14:30   ` Juergen Gross
2018-04-04 14:46     ` Greg KH
2018-04-04 15:12       ` Juergen Gross
2018-04-04 15:12       ` Juergen Gross
2018-04-04 15:42         ` Greg KH
2018-04-04 16:32           ` Juergen Gross
2018-04-04 16:32           ` Juergen Gross
2018-04-05  6:33             ` Greg KH
2018-04-05  7:02               ` Juergen Gross
2018-04-05  7:14                 ` Greg KH
2018-04-05  7:14                 ` Greg KH
2018-04-05  8:00                   ` Juergen Gross
2018-04-05  8:00                   ` [Xen-devel] " Juergen Gross
2018-04-05 10:06                     ` George Dunlap
2018-04-05 10:06                     ` [Xen-devel] " George Dunlap
2018-04-05 10:15                       ` George Dunlap
2018-04-05 10:15                       ` [Xen-devel] " George Dunlap
2018-04-05 12:19                       ` Juergen Gross
2018-04-05 12:19                       ` [Xen-devel] " Juergen Gross
2018-04-05 13:00                         ` Boris Ostrovsky
2018-04-05 13:00                         ` [Xen-devel] " Boris Ostrovsky
2018-04-05 13:06                           ` Juergen Gross
2018-04-05 13:06                           ` [Xen-devel] " Juergen Gross
2018-04-05 13:42                             ` George Dunlap
2018-04-05 13:42                             ` [Xen-devel] " George Dunlap
2018-04-05 14:09                               ` Juergen Gross
2018-04-05 14:56                                 ` George Dunlap
2018-04-05 14:56                                 ` [Xen-devel] " George Dunlap
2018-04-05 17:11                                   ` Juergen Gross
2018-04-05 18:33                                     ` Boris Ostrovsky
2018-04-06  8:00                                       ` Juergen Gross
2018-04-06  9:38                                         ` George Dunlap
2018-04-06  9:49                                       ` George Dunlap
2018-04-06 10:02                                         ` Juergen Gross
2018-04-06 10:07                                           ` George Dunlap
2018-04-06 10:57                                             ` Juergen Gross
2018-04-06 11:13                                               ` George Dunlap
2018-04-06 13:12                                                 ` Juergen Gross
2018-04-06 13:33                                                   ` George Dunlap
2018-04-06 14:10                                                     ` Boris Ostrovsky
2018-04-06 15:15                                                       ` Juergen Gross
2018-04-06 10:13                                     ` George Dunlap
2018-04-06 10:44                                       ` Juergen Gross [this message]
2018-04-05 14:09                               ` Juergen Gross
2018-04-05 20:20                 ` Thomas Backlund
2018-04-05 20:20                 ` Thomas Backlund
2018-04-05  7:02               ` Juergen Gross
2018-04-05  6:33             ` Greg KH
2018-04-04 15:42         ` Greg KH
2018-04-04 14:46     ` Greg KH
2018-04-04 14:30   ` Juergen Gross
2018-04-04 10:38 Juergen Gross

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=2e25337d-103a-7812-acfd-37606807f496@suse.com \
    --to=jgross@suse.com \
    --cc=boris.ostrovsky@oracle.com \
    --cc=dunlapg@umich.edu \
    --cc=xen-devel@lists.xenproject.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.