linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Christoph Lameter <clameter@sgi.com>
To: Andi Kleen <ak@suse.de>
Cc: Lee Schermerhorn <lee.schermerhorn@hp.com>,
	linux-mm@kvack.org, akpm@linux-foundation.org, nacc@us.ibm.com
Subject: Re: [PATCH/RFC 0/11] Shared Policy Overview
Date: Tue, 26 Jun 2007 20:25:11 -0700 (PDT)	[thread overview]
Message-ID: <Pine.LNX.4.64.0706262019470.24504@schroedinger.engr.sgi.com> (raw)
In-Reply-To: <200706270042.27365.ak@suse.de>

On Wed, 27 Jun 2007, Andi Kleen wrote:

> > You are sure that this works? Just by looking at the description: It
> > cannot work. Any allocator use of a memory policy must use rcu locks
> > otherwise the memory policy can vanish from under us while allocating a
> > page. This means you need to add this to alloc_pages_current
> > and alloc_pages_node.  Possible all of __alloc_pages must be handled
> > under RCU. This is a significant increase of RCU use.
> 
> I've been actually looking at using RCUs for the shared policies 
> too to plug the recent reference count issue.  I don't think it's a problem 
> because the RCU use can be limited to when policies are actually
> used. Besides rcu_read_lock() is a nop on non preemptible kernels
> anyways and users of preemptible kernels will probably not notice
> it among all the other overhead they have anyways.

If a system policy is set then it will be used all of the time.
Could be a signficant increase in RCU use.
 
> > If we can make this work then RCU should be used for all policies so that
> > we can get rid of the requirement that policies can only be modified from
> > the task context that created it.
> 
> Huh? RCU doesn't give you locking against multiple writers. Just existence
> guarantees. And you can have those already by just holding the reference 
> count.

If you want to replace one policy by another then RCU ensures 
that the old policy can still be used for the remainder of the rcu period.

If RCU is not used then the updating of a policy is not possible since
there is currently no locking and there may be concurrent uses of the 
policy or the zonelist generated by a policy. One thread may acquire
the pointer to a policy while another changes the policy.

If the old policy is immediately freed then the first thread may access 
invalid data.



--
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>

  reply	other threads:[~2007-06-27  3:25 UTC|newest]

Thread overview: 48+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-06-25 19:52 Lee Schermerhorn
2007-06-25 19:52 ` [PATCH/RFC 1/11] Shared Policy: move shared policy to inode/mapping Lee Schermerhorn
2007-06-25 19:52 ` [PATCH/RFC 2/11] Shared Policy: allocate shared policies as needed Lee Schermerhorn
2007-06-25 19:52 ` [PATCH/RFC 3/11] Shared Policy: let vma policy ops handle sub-vma policies Lee Schermerhorn
2007-06-25 19:52 ` [PATCH/RFC 4/11] Shared Policy: fix show_numa_maps() Lee Schermerhorn
2007-06-25 19:52 ` [PATCH/RFC 5/11] Shared Policy: Add hugepage shmem policy vm_ops Lee Schermerhorn
2007-06-25 19:53 ` [PATCH/RFC 6/11] Shared Policy: Factor alloc_page_pol routine Lee Schermerhorn
2007-06-25 19:53 ` [PATCH/RFC 7/11] Shared Policy: use shared policy for page cache allocations Lee Schermerhorn
2007-06-25 19:53 ` [PATCH/RFC 8/11] Shared Policy: fix migration of private mappings Lee Schermerhorn
2007-06-25 19:53 ` [PATCH/RFC 9/11] Shared Policy: mapped file policy persistence model Lee Schermerhorn
2007-06-25 19:53 ` [PATCH/RFC 10/11] Shared Policy: per cpuset shared file policy control Lee Schermerhorn
2007-06-25 21:10   ` Paul Jackson
2007-06-27 17:33     ` Lee Schermerhorn
2007-06-27 19:52       ` Paul Jackson
2007-06-27 20:22         ` Lee Schermerhorn
2007-06-27 20:36           ` Paul Jackson
2007-06-25 19:53 ` [PATCH/RFC 11/11] Shared Policy: add generic file set/get policy vm ops Lee Schermerhorn
2007-06-26 22:17 ` [PATCH/RFC 0/11] Shared Policy Overview Christoph Lameter
2007-06-27 13:43   ` Lee Schermerhorn
2007-06-26 22:21 ` Christoph Lameter
2007-06-26 22:42   ` Andi Kleen
2007-06-27  3:25     ` Christoph Lameter [this message]
2007-06-27 20:14       ` Lee Schermerhorn
2007-06-27 18:14   ` Lee Schermerhorn
2007-06-27 21:37     ` Christoph Lameter
2007-06-27 22:01       ` Andi Kleen
2007-06-27 22:08         ` Christoph Lameter
2007-06-27 23:46         ` Paul E. McKenney
2007-06-28  0:14           ` Andi Kleen
2007-06-29 21:47           ` Lee Schermerhorn
2007-06-28 13:42         ` Lee Schermerhorn
2007-06-28 22:02           ` Andi Kleen
2007-06-29 17:14             ` Lee Schermerhorn
2007-06-29 17:42               ` Andi Kleen
2007-06-30 18:34                 ` [PATCH/RFC] Fix Mempolicy Ref Counts - was " Lee Schermerhorn
2007-07-03 18:09                   ` Christoph Lameter
2007-06-29  1:39           ` Christoph Lameter
2007-06-29  9:01             ` Andi Kleen
2007-06-29 14:05               ` Christoph Lameter
2007-06-29 17:41                 ` Lee Schermerhorn
2007-06-29 20:15                   ` Christoph Lameter
2007-06-29 13:22             ` Lee Schermerhorn
2007-06-29 14:18               ` Christoph Lameter
2007-06-27 23:36       ` Lee Schermerhorn
2007-06-29  1:41         ` Christoph Lameter
2007-06-29 13:30           ` Lee Schermerhorn
2007-06-29 14:20             ` Andi Kleen
2007-06-29 21:40               ` 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=Pine.LNX.4.64.0706262019470.24504@schroedinger.engr.sgi.com \
    --to=clameter@sgi.com \
    --cc=ak@suse.de \
    --cc=akpm@linux-foundation.org \
    --cc=lee.schermerhorn@hp.com \
    --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