linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Mel Gorman <mel@csn.ul.ie>
To: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
Cc: linux-kernel <linux-kernel@vger.kernel.org>,
	linux-mm <linux-mm@kvack.org>, Adam Litke <agl@us.ibm.com>,
	Nishanth Aravamudan <nacc@us.ibm.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	Eric Whitney <eric.whitney@hp.com>
Subject: Re: [BUG] 2.6.25-rc4 hang/softlockups after freeing hugepages
Date: Fri, 7 Mar 2008 11:48:49 +0000	[thread overview]
Message-ID: <20080307114849.GC26229@csn.ul.ie> (raw)
In-Reply-To: <1204824183.5294.62.camel@localhost>

On (06/03/08 12:23), Lee Schermerhorn didst pronounce:
> Test platform:  HP Proliant DL585 server - 4 socket, dual core AMD with
> 32GB memory.
> 
> I first saw this on 25-rc2-mm1 with Mel's zonelist patches, while
> investigating the interaction of hugepages and cpusets.  Thinking that
> it might be caused by the zonelist patches, I went back to 25-rc2-mm1
> w/o the patches and saw the same thing.  It sometimes takes a while for
> the softlockups to start appearing, and I wanted to find a fairly
> minimal duplicator.  Meanwhile 25-rc3 and rc4 have come out, so I tried
> the latest upstream kernel and see the same thing.
> 
> To duplicate the problem, I need only:
> 
> + log into the platform as root in one window and:
> 
> 	echo N >/proc/sys/vm/nr_hugepages
> 	echo 0 >proc/sys/vm/nr_hugepages
> 

Uncool, I am going to try and find a machine to reproduce this one but
in case I have no luck, can you try setting the following in your
.config which may rattle out something please?

CONFIG_DEBUG_SPINLOCK=y
CONFIG_DEBUG_MUTEXES=y
CONFIG_DEBUG_LOCK_ALLOC=y
CONFIG_PROVE_LOCKING=y
CONFIG_DEBUG_SPINLOCK_SLEEP=y
CONFIG_DEBUG_VM=y

and as you have DEBUG_INFO, can you say what line is ffffffff8027b693 ?

> In my case, N=64.  If I look, before echoing 0, I see 16 hugepages
> allocated on each of the 4 nodes, as expected.
> 
> + then in another window, log in again.  
> 
> Sometimes it will hang during the 2nd login and I'll never see a shell
> prompt. 

My initial guess was that is is something to do with page_table_lock but as
you didn't get to fault in huge pages, it doesn't make much sense.

-- 
Mel Gorman
Part-time Phd Student                          Linux Technology Center
University of Limerick                         IBM Dublin Software Lab

--
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:[~2008-03-07 11:48 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-03-06 17:23 [BUG] 2.6.25-rc4 hang/softlockups after freeing hugepages Lee Schermerhorn
2008-03-06 17:45 ` Ingo Molnar
2008-03-06 18:19   ` Lee Schermerhorn
2008-03-06 17:53 ` Nishanth Aravamudan
2008-03-06 18:17   ` Lee Schermerhorn
2008-03-07 11:48 ` Mel Gorman [this message]
2008-03-07 14:36   ` Lee Schermerhorn

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=20080307114849.GC26229@csn.ul.ie \
    --to=mel@csn.ul.ie \
    --cc=Lee.Schermerhorn@hp.com \
    --cc=agl@us.ibm.com \
    --cc=akpm@linux-foundation.org \
    --cc=eric.whitney@hp.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=nacc@us.ibm.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;
as well as URLs for NNTP newsgroup(s).