From: chandra seetharaman <sekharan@us.ibm.com>
To: KUROSAWA Takahiro <kurosawa@valinux.co.jp>
Cc: ckrm-tech@lists.sourceforge.net, linux-mm@kvack.org
Subject: Re: [ckrm-tech] [PATCH 0/8] Pzone based CKRM memory resource controller
Date: Thu, 02 Feb 2006 17:01:13 -0800 [thread overview]
Message-ID: <1138928473.3914.9.camel@localhost.localdomain> (raw)
In-Reply-To: <20060203005122.CD94174039@sv1.valinux.co.jp>
On Fri, 2006-02-03 at 09:51 +0900, KUROSAWA Takahiro wrote:
> On Thu, 02 Feb 2006 16:37:37 -0800
> chandra seetharaman <sekharan@us.ibm.com> wrote:
>
> > > > > > - what is the plan to support "limit" ?
> > > > >
> > > > > To be honest, I don't have any specific idea to support "limit" currently.
> > > > > Probably the userspace daemon that enlarge "guarantee" to the specified
> > > > > "limit" might support the "limit", because "guarantee" in the pzone based
> > > > > memory resource controller also works as "limit".
> > > >
> > > > I am not able to visualize how this will work.
> > > >
> > > > In simple terms, sum of guarantees should _not_ exceed the amount of
> > > > available memory but, sum of limits _can_ exceed the amount of available
> > > > memory. As far as i understand your implementation, guarantee is
> > > > translated to present_pages of the pseudo zone (and is subtracted from
> > > > paren't present_pages). How can one set limit to be same as guarantee ?
> > >
> > > The number of pages in the pseudo zones can also be considered as limit
> > > because tasks in a class can't allocate beyond the number of the pages
> > > that are allocated to the pseudo zones.
> >
> > Yes. but, it is true only when limit and guarantee are the same.
> >
> > Consider the following scenario:
> > A system with 1024MB of memory.
> >
> > I want to create 6 classes:
> > - 4 of which has guarantee of 128MB and limit of 512MB
> > - 2 of which has guarantee of 256MB and limit of 768MB
> >
> > We cannot do this with this memrc. Can you explain how a userspace
> > program can help me do this.
>
> Our memrc with a userspace program doesn't help this case.
> If you'd like to setup classes like this, you can select the memory
> resource controller in current CKRM.
That is how we intended guarantee and limit to work. What was your
understanding, and what one can do through the userspace support ?
>
--
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>
next prev parent reply other threads:[~2006-02-03 0:59 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-01-19 8:04 [PATCH 0/2] " KUROSAWA Takahiro
2006-01-19 8:04 ` [PATCH 1/2] Add the pzone KUROSAWA Takahiro
2006-01-19 18:04 ` Andy Whitcroft
2006-01-19 23:42 ` KUROSAWA Takahiro
2006-01-20 9:17 ` Andy Whitcroft
2006-01-20 7:08 ` KAMEZAWA Hiroyuki
2006-01-20 8:22 ` KUROSAWA Takahiro
2006-01-20 8:30 ` KAMEZAWA Hiroyuki
2006-01-19 8:04 ` [PATCH 2/2] Add CKRM memory resource controller using pzones KUROSAWA Takahiro
2006-01-31 2:30 ` [PATCH 0/8] Pzone based CKRM memory resource controller KUROSAWA Takahiro
2006-01-31 2:30 ` [PATCH 1/8] Add the __GFP_NOLRU flag KUROSAWA Takahiro
2006-01-31 18:18 ` [ckrm-tech] " Dave Hansen
2006-02-01 5:06 ` KUROSAWA Takahiro
2006-01-31 2:30 ` [PATCH 2/8] Keep the number of zones while zone iterator loop KUROSAWA Takahiro
2006-01-31 2:30 ` [PATCH 3/8] Add for_each_zone_in_node macro KUROSAWA Takahiro
2006-01-31 2:30 ` [PATCH 4/8] Extract zone specific routines as functions KUROSAWA Takahiro
2006-01-31 2:30 ` [PATCH 5/8] Add the pzone_create() function KUROSAWA Takahiro
2006-01-31 2:30 ` [PATCH 6/8] Add the pzone_destroy() function KUROSAWA Takahiro
2006-01-31 2:30 ` [PATCH 7/8] Make the number of pages in pzones resizable KUROSAWA Takahiro
2006-01-31 2:30 ` [PATCH 8/8] Add a CKRM memory resource controller using pzones KUROSAWA Takahiro
2006-02-01 2:58 ` [ckrm-tech] [PATCH 0/8] Pzone based CKRM memory resource controller chandra seetharaman
2006-02-01 5:39 ` KUROSAWA Takahiro
2006-02-01 6:16 ` Hirokazu Takahashi
2006-02-02 1:26 ` chandra seetharaman
2006-02-02 3:54 ` KUROSAWA Takahiro
2006-02-03 0:37 ` chandra seetharaman
2006-02-03 0:51 ` KUROSAWA Takahiro
2006-02-03 1:01 ` chandra seetharaman [this message]
2006-02-01 3:07 ` chandra seetharaman
2006-02-01 5:54 ` KUROSAWA Takahiro
2006-02-03 1:33 ` KUROSAWA Takahiro
2006-02-03 9:37 ` KUROSAWA Takahiro
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=1138928473.3914.9.camel@localhost.localdomain \
--to=sekharan@us.ibm.com \
--cc=ckrm-tech@lists.sourceforge.net \
--cc=kurosawa@valinux.co.jp \
--cc=linux-mm@kvack.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox