All of lore.kernel.org
 help / color / mirror / Atom feed
From: Joel Schopp <jschopp@austin.ibm.com>
To: Dave Hansen <haveblue@us.ibm.com>
Cc: Mel Gorman <mel@csn.ul.ie>,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	lhms-devel@lists.sourceforge.net
Subject: Re: [PATCH 4/7] ppc64 - Specify amount of kernel memory at boot	time
Date: Fri, 17 Feb 2006 15:31:24 -0600	[thread overview]
Message-ID: <43F640AC.6060600@austin.ibm.com> (raw)
In-Reply-To: <1140196618.21383.112.camel@localhost.localdomain>

>> This patch adds the kernelcore= parameter for ppc64.
>>
>> The amount of memory will requested will not be reserved in all nodes. The
>> first node that is found that can accomodate the requested amount of memory
>> and have remaining more for ZONE_EASYRCLM is used. If a node has memory holes,
>> it also will not be used.
> 
> One thing I think we really need to see before these go into mainline is
> the ability to shrink the ZONE_EASYRCLM at runtime, and give the memory
> back to NORMAL/DMA.
> 
> Otherwise, any system starting off sufficiently small will end up having
> lowmem starvation issues.  Allowing resizing at least gives the admin a
> chance to avoid those issues.
> 

I'm not too keen on calling it resizing, because that term is 
misleading.  The resizing is one way.  You can't later resize back. It's 
like a window that you can only close but never reopen.  We should call 
it "runtime incremental disabling", or RID.

I don't think we need RID in order to merge these patches.  RID can be 
merged later if people decide they want a special easy reclaim zone that 
could disappear at any moment.  I personally fall in the camp of wanting 
my zones I explicitly enabled to stay put and am opposed to RID.

If only somebody had presented a solution that was flexible enough to 
dynamically resize reclaimable and non-reclaimable both ways.
http://sourceforge.net/mailarchive/message.php?msg_id=13864331





WARNING: multiple messages have this Message-ID (diff)
From: Joel Schopp <jschopp@austin.ibm.com>
To: Dave Hansen <haveblue@us.ibm.com>
Cc: Mel Gorman <mel@csn.ul.ie>,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	lhms-devel@lists.sourceforge.net
Subject: Re: [PATCH 4/7] ppc64 - Specify amount of kernel memory at boot	time
Date: Fri, 17 Feb 2006 15:31:24 -0600	[thread overview]
Message-ID: <43F640AC.6060600@austin.ibm.com> (raw)
In-Reply-To: <1140196618.21383.112.camel@localhost.localdomain>

>> This patch adds the kernelcore= parameter for ppc64.
>>
>> The amount of memory will requested will not be reserved in all nodes. The
>> first node that is found that can accomodate the requested amount of memory
>> and have remaining more for ZONE_EASYRCLM is used. If a node has memory holes,
>> it also will not be used.
> 
> One thing I think we really need to see before these go into mainline is
> the ability to shrink the ZONE_EASYRCLM at runtime, and give the memory
> back to NORMAL/DMA.
> 
> Otherwise, any system starting off sufficiently small will end up having
> lowmem starvation issues.  Allowing resizing at least gives the admin a
> chance to avoid those issues.
> 

I'm not too keen on calling it resizing, because that term is 
misleading.  The resizing is one way.  You can't later resize back. It's 
like a window that you can only close but never reopen.  We should call 
it "runtime incremental disabling", or RID.

I don't think we need RID in order to merge these patches.  RID can be 
merged later if people decide they want a special easy reclaim zone that 
could disappear at any moment.  I personally fall in the camp of wanting 
my zones I explicitly enabled to stay put and am opposed to RID.

If only somebody had presented a solution that was flexible enough to 
dynamically resize reclaimable and non-reclaimable both ways.
http://sourceforge.net/mailarchive/message.php?msg_id=13864331




--
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>

  parent reply	other threads:[~2006-02-17 21:32 UTC|newest]

Thread overview: 48+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-02-17 14:15 [PATCH 0/7] Reducing fragmentation using zones v5 Mel Gorman
2006-02-17 14:15 ` Mel Gorman
2006-02-17 14:16 ` [PATCH 1/7] Add __GFP_EASYRCLM flag and update callers Mel Gorman
2006-02-17 14:16   ` Mel Gorman
2006-02-17 14:16 ` [PATCH 2/7] Create the ZONE_EASYRCLM zone Mel Gorman
2006-02-17 14:16   ` Mel Gorman
2006-02-17 14:16 ` [PATCH 3/7] x86 - Specify amount of kernel memory at boot time Mel Gorman
2006-02-17 14:16   ` Mel Gorman
2006-02-17 14:17 ` [PATCH 4/7] ppc64 " Mel Gorman
2006-02-17 14:17   ` Mel Gorman
2006-02-17 17:16   ` Dave Hansen
2006-02-17 17:16     ` Dave Hansen
2006-02-17 19:03     ` Mel Gorman
2006-02-17 19:03       ` Mel Gorman
2006-02-17 19:17       ` Dave Hansen
2006-02-17 19:17         ` Dave Hansen
2006-02-17 19:36         ` Mel Gorman
2006-02-17 19:36           ` Mel Gorman
2006-02-17 21:31     ` Joel Schopp [this message]
2006-02-17 21:31       ` Joel Schopp
2006-02-21 14:51     ` Mel Gorman
2006-02-21 14:51       ` Mel Gorman
2006-02-21 17:35       ` Dave Hansen
2006-02-21 17:35         ` Dave Hansen
2006-02-22 16:43         ` Mel Gorman
2006-02-22 16:43           ` Mel Gorman
2006-02-23 16:42           ` Dave Hansen
2006-02-23 16:42             ` Dave Hansen
2006-02-23 17:19             ` Mel Gorman
2006-02-23 17:19               ` Mel Gorman
2006-02-23 17:38               ` Dave Hansen
2006-02-23 17:38                 ` Dave Hansen
2006-02-23 18:01                 ` Mel Gorman
2006-02-23 18:01                   ` Mel Gorman
2006-02-23 18:15                   ` Dave Hansen
2006-02-23 18:15                     ` Dave Hansen
2006-02-24  0:15                     ` KAMEZAWA Hiroyuki
2006-02-24  0:15                       ` KAMEZAWA Hiroyuki
2006-02-24  9:04                     ` Mel Gorman
2006-02-24  9:04                       ` Mel Gorman
2006-02-23 17:40               ` Mike Kravetz
2006-02-23 17:40                 ` Mike Kravetz
2006-02-17 14:17 ` [PATCH 5/7] At boot, determine what zone memory will hot-add to Mel Gorman
2006-02-17 14:17   ` Mel Gorman
2006-02-17 14:17 ` [PATCH 6/7] Allow HugeTLB allocations to use ZONE_EASYRCLM Mel Gorman
2006-02-17 14:17   ` Mel Gorman
2006-02-17 14:18 ` [PATCH 7/7] Add documentation for extra boot parameters Mel Gorman
2006-02-17 14:18   ` Mel Gorman

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=43F640AC.6060600@austin.ibm.com \
    --to=jschopp@austin.ibm.com \
    --cc=haveblue@us.ibm.com \
    --cc=lhms-devel@lists.sourceforge.net \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mel@csn.ul.ie \
    /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.