From: Nathan Fontenot <nfont@austin.ibm.com>
To: Andi Kleen <andi@firstfloor.org>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH] memory hotplug disable boot option
Date: Fri, 25 Jun 2010 11:00:10 -0500 [thread overview]
Message-ID: <4C24D28A.30006@austin.ibm.com> (raw)
In-Reply-To: <20100625152819.GC13421@basil.fritz.box>
On 06/25/2010 10:28 AM, Andi Kleen wrote:
>> Yes, this would work to reduce the number of memory sections created. The reason
>
> Note it could be done on the sysfs interface level too, e.g. have one
> entry that gets a bitmap as input/output.
This was sort of the idea I was thinking of for a flat representation in sysfs.
Instead of creating all of the directories for each memory section, have a single
file for the memory section attributes. For example, an 'online' file. Users
could read the file to get online memory section numbers and write to it to
online a specific memory section.
>
>> I have not gone this route is that increasing the memory section size would
>> break DLPAR memory add and remove for powerpc pseries.
>
> How would it break it?
For the powerpc/pseries hardware memory is assigned in LMB size chunks. If the
memory section size is larger than the LMB size then DLPAR remove would have
to hotplug remove more memory than required to DLPAR remove the memory. For
DLPAR memory add, we would fail because we could not hotplug add the larger
memory section since that much memory had not actually been given to the
machine.
>
> But aren't you breaking this anyways by disabling the sysfs entries?!?
This is meant more for allowing people to bot systems without the long boot
times (as mentioned before 8+ hours for a 2TB system) even if it means they
will not be able to memory hotplug or memory DLPAR.
-Nathan
next prev parent reply other threads:[~2010-06-25 16:00 UTC|newest]
Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-06-25 1:06 [PATCH] memory hotplug disable boot option Nathan Fontenot
2010-06-25 2:04 ` KOSAKI Motohiro
2010-06-25 9:19 ` Andi Kleen
2010-06-25 14:51 ` Nathan Fontenot
2010-06-25 14:56 ` Andi Kleen
2010-06-25 15:21 ` Nathan Fontenot
2010-06-25 15:28 ` Andi Kleen
2010-06-25 16:00 ` Nathan Fontenot [this message]
2010-06-28 2:20 ` KOSAKI Motohiro
2010-06-28 4:15 ` Eric W. Biederman
2010-06-28 14:16 ` Andi Kleen
2010-06-28 19:43 ` Eric W. Biederman
2010-06-28 15:02 ` Greg KH
2010-06-28 15:37 ` Nathan Fontenot
2010-06-28 15:44 ` Greg KH
2010-06-29 0:04 ` Dave Hansen
2010-06-29 2:56 ` KOSAKI Motohiro
2010-06-29 15:38 ` Nathan Fontenot
2010-06-30 0:00 ` KOSAKI Motohiro
2010-06-29 16:03 ` Dave Hansen
2010-06-29 18:04 ` Greg KH
2010-06-30 0:32 ` KAMEZAWA Hiroyuki
2010-06-30 15:47 ` Greg KH
2010-07-01 0:31 ` KAMEZAWA Hiroyuki
2010-07-01 3:17 ` Nathan Fontenot
2010-07-01 3:30 ` KAMEZAWA Hiroyuki
2010-07-01 23:28 ` Greg KH
2010-07-01 5:15 ` KAMEZAWA Hiroyuki
2010-07-01 13:23 ` Dave Hansen
2010-07-06 15:20 ` Nathan Fontenot
2010-07-06 15:33 ` Dave Hansen
2010-07-06 15:47 ` Nathan Fontenot
2010-07-01 23:26 ` Greg KH
2010-07-02 5:50 ` KAMEZAWA Hiroyuki
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=4C24D28A.30006@austin.ibm.com \
--to=nfont@austin.ibm.com \
--cc=andi@firstfloor.org \
--cc=kosaki.motohiro@jp.fujitsu.com \
--cc=linux-kernel@vger.kernel.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