From: mel@csn.ul.ie (Mel Gorman)
To: Andrew Morton <akpm@osdl.org>
Cc: Martin Bligh <mbligh@google.com>,
apw@shadowen.org, linux-kernel@vger.kernel.org
Subject: Re: sparsemem panic in 2.6.17-rc5-mm1 and -mm2
Date: Wed, 7 Jun 2006 10:16:37 +0100 [thread overview]
Message-ID: <20060607091636.GA19510@skynet.ie> (raw)
In-Reply-To: <20060606164241.69d55238.akpm@osdl.org>
On (06/06/06 16:42), Andrew Morton didst pronounce:
> On Mon, 05 Jun 2006 17:51:00 -0700
> Martin Bligh <mbligh@google.com> wrote:
>
> > http://test.kernel.org/abat/34264/debug/console.log
>
> What sort of machine is this, anyway?
>
> > WARNING: Not an IBM x440/NUMAQ and CONFIG_NUMA enabled!
>
It's an IBM x440 that is not recognised by that check. Later in the log,
we see
IBM eserver xSeries 440 detected: force use of acpi=ht
> And is it expected that ZONE_NORMAL only has 384MB? That seems awfully low
> for a 64GB x86 machine. Could be that we went oom because we chose to
> allocate really big hash tables, based on the total amount of memory?
>
The log reports 392MB LOWMEM available. As it is 32 bit machine, it started
with about 896MB but consumes much of it with mem_maps. The log shows
calculate_numa_remap_pages() reporting
Reserving 35328 pages of KVA for lmem_map of node 0
Shrinking node 0 from 4456448 pages to 4421120 pages
Reserving 31232 pages of KVA for lmem_map of node 1
Shrinking node 1 from 8650752 pages to 8619520 pages
Reserving 31232 pages of KVA for lmem_map of node 2
Shrinking node 2 from 12845056 pages to 12813824 pages
Reserving 29184 pages of KVA for lmem_map of node 3
Shrinking node 3 from 16777216 pages to 16748032 pages
Reserving total of 126976 pages for numa KVA remap
reserve_pages = 126976 find_max_low_pfn() ~ 229375
That is 400MB gone already which is mapped to lowmem. A little later in the
log, it says
min_low_pfn = 1140, max_low_pfn = 100352, highstart_pfn = 100352
so about 4MB is missing from the beginning (probably the kernel image),
so we're down to 396ish. Not sure where the last 4MB is exactly but you
get the idea.
While it is possible we are going OOM due to the size of lowmem, it's
doubtful to be the only cause.
http://test.kernel.org/functional/index.html shows that elm3b67 (the
machine in question) has passed loads of tests in the past.
--
Mel Gorman
Part-time Phd Student Linux Technology Center
University of Limerick IBM Dublin Software Lab
next prev parent reply other threads:[~2006-06-07 9:16 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-06-06 0:51 sparsemem panic in 2.6.17-rc5-mm1 and -mm2 Martin Bligh
2006-06-06 3:07 ` Andrew Morton
2006-06-06 5:19 ` KAMEZAWA Hiroyuki
2006-06-06 5:36 ` Yasunori Goto
2006-06-06 7:27 ` Andrew Morton
2006-06-07 0:43 ` KAMEZAWA Hiroyuki
2006-06-07 4:58 ` Andrew Morton
2006-06-07 5:36 ` Rusty Russell
2006-06-07 5:50 ` Andrew Morton
2006-06-07 6:49 ` Rusty Russell
2006-06-07 9:26 ` Andy Whitcroft
2006-06-07 16:29 ` Andrew Morton
2006-06-07 16:35 ` Andrew Morton
2006-06-07 17:50 ` Andy Whitcroft
2006-06-15 12:28 ` [PATCH] zone handle unaligned zone boundaries Andy Whitcroft
2006-06-07 17:22 ` sparsemem panic in 2.6.17-rc5-mm1 and -mm2 Andy Whitcroft
2006-06-06 23:42 ` Andrew Morton
2006-06-07 9:16 ` Mel Gorman [this message]
2006-06-07 17:38 ` Andy Whitcroft
2006-06-07 17:41 ` Andy Whitcroft
-- strict thread matches above, loose matches on Subject: below --
2006-06-06 3:50 Chuck Ebbert
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=20060607091636.GA19510@skynet.ie \
--to=mel@csn.ul.ie \
--cc=akpm@osdl.org \
--cc=apw@shadowen.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mbligh@google.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