linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Dan Magenheimer <dan.magenheimer@oracle.com>
To: Anthony Liguori <anthony@codemonkey.ws>
Cc: npiggin@suse.de, akpm@osdl.org, jeremy@goop.org,
	xen-devel@lists.xensource.com, tmem-devel@oss.oracle.com,
	kurt.hackel@oracle.com, Rusty Russell <rusty@rustcorp.com.au>,
	linux-kernel@vger.kernel.org, dave.mccracken@oracle.com,
	linux-mm@kvack.org, chris.mason@oracle.com,
	sunil.mushran@oracle.com, Avi Kivity <avi@redhat.com>,
	Schwidefsky <schwidefsky@de.ibm.com>,
	Marcelo Tosatti <mtosatti@redhat.com>,
	alan@lxorguk.ukuu.org.uk,
	Balbir Singh <balbir@linux.vnet.ibm.com>
Subject: RE: [Xen-devel] Re: [RFC PATCH 0/4] (Take 2): transcendent memory ("tmem") for Linux
Date: Wed, 8 Jul 2009 16:31:29 -0700 (PDT)	[thread overview]
Message-ID: <ac5dec0d-e593-4a82-8c9d-8aa374e8c6ed@default> (raw)
In-Reply-To: <4A55243B.8090001@codemonkey.ws>

Hi Anthony --

Thanks for the comments.

> I have trouble mapping this to a VMM capable of overcommit 
> without just coming back to CMM2.
> 
> In CMM2 parlance, ephemeral tmem pools is just normal kernel memory 
> marked in the volatile state, no?

They are similar in concept, but a volatile-marked kernel page
is still a kernel page, can be changed by a kernel (or user)
store instruction, and counts as part of the memory used
by the VM.  An ephemeral tmem page cannot be directly written
by a kernel (or user) store, can only be read via a "get" (which
may or may not succeed), and doesn't count against the memory
used by the VM (even though it likely contains -- for awhile --
data useful to the VM).

> It seems to me that an architecture built around hinting 
> would be more 
> robust than having to use separate memory pools for this type 
> of memory 
> (especially since you are requiring a copy to/from the pool).

Depends on what you mean by robust, I suppose.  Once you
understand the basics of tmem, it is very simple and this
is borne out in the low invasiveness of the Linux patch.
Simplicity is another form of robustness.

> For instance, you can mark data DMA'd from disk (perhaps by 
> read-ahead) 
> as volatile without ever bringing it into the CPU cache.  
> With tmem, if 
> you wanted to use a tmem pool for all of the page cache, you'd likely 
> suffer significant overhead due to copying.

The copy may be expensive on an older machine, but on newer
machines copying a page is relatively inexpensive.  On a reasonable
multi-VM-kernbench-like benchmark I'll be presenting at Linux
Symposium next week, the overhead is on the order of 0.01%
for a fairly significant savings in IOs.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  reply	other threads:[~2009-07-08 23:20 UTC|newest]

Thread overview: 35+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-07-07 16:17 Dan Magenheimer
2009-07-07 17:28 ` Rik van Riel
2009-07-07 19:53   ` Dan Magenheimer
2009-07-08 22:56 ` Anthony Liguori
2009-07-08 23:31   ` Dan Magenheimer [this message]
2009-07-08 23:57     ` [Xen-devel] " Anthony Liguori
2009-07-09  0:17       ` Jeremy Fitzhardinge
2009-07-09  0:27         ` Anthony Liguori
2009-07-09  1:20   ` Rik van Riel
2009-07-09 21:09     ` Dan Magenheimer
2009-07-09 21:27       ` Rik van Riel
2009-07-09 21:48         ` Dan Magenheimer
2009-07-09 21:41       ` Anthony Liguori
2009-07-09 22:34         ` Dan Magenheimer
2009-07-09 22:45           ` Rik van Riel
2009-07-09 23:33           ` Anthony Liguori
2009-07-10 15:23             ` Dan Magenheimer
2009-07-12  9:20               ` Avi Kivity
2009-07-12 16:28                 ` Dan Magenheimer
2009-07-12 17:27                   ` Avi Kivity
2009-07-12 20:59                     ` Dan Magenheimer
2009-07-12 13:28               ` Anthony Liguori
2009-07-12 16:20                 ` Dan Magenheimer
2009-07-12 17:16                   ` Avi Kivity
2009-07-12 19:34                     ` Anthony Liguori
2009-07-13 20:17                       ` Chris Mason
2009-07-13 20:38                         ` Anthony Liguori
2009-07-13 21:01                           ` Chris Mason
2009-07-13 21:17                             ` Anthony Liguori
2009-07-26 15:00                               ` Avi Kivity
2009-07-13 20:38                         ` Anthony Liguori
2009-07-12 20:39                     ` [Xen-devel] " Dan Magenheimer
2009-07-12 20:43                       ` Avi Kivity
2009-07-12 21:08                         ` Dan Magenheimer
2009-07-13 11:33                           ` 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=ac5dec0d-e593-4a82-8c9d-8aa374e8c6ed@default \
    --to=dan.magenheimer@oracle.com \
    --cc=akpm@osdl.org \
    --cc=alan@lxorguk.ukuu.org.uk \
    --cc=anthony@codemonkey.ws \
    --cc=avi@redhat.com \
    --cc=balbir@linux.vnet.ibm.com \
    --cc=chris.mason@oracle.com \
    --cc=dave.mccracken@oracle.com \
    --cc=jeremy@goop.org \
    --cc=kurt.hackel@oracle.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mtosatti@redhat.com \
    --cc=npiggin@suse.de \
    --cc=rusty@rustcorp.com.au \
    --cc=schwidefsky@de.ibm.com \
    --cc=sunil.mushran@oracle.com \
    --cc=tmem-devel@oss.oracle.com \
    --cc=xen-devel@lists.xensource.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