All of lore.kernel.org
 help / color / mirror / Atom feed
From: Christoph Lameter <cl@linux.com>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>,
	Johannes Weiner <hannes@cmpxchg.org>,
	Michal Hocko <mhocko@suse.cz>, azurIt <azurit@pobox.sk>,
	linux-mm@kvack.org, cgroups@vger.kernel.org,
	linux-kernel@vger.kernel.org,
	Christian Casteyde <casteyde.christian@free.fr>,
	Pekka Enberg <penberg@kernel.org>
Subject: Re: [patch 2/2] fs: buffer: move allocation failure loop into the allocator
Date: Wed, 4 Dec 2013 15:17:08 +0000	[thread overview]
Message-ID: <00000142be2f1de0-764bb035-adbc-4367-b2b4-bf05498510a6-000000@email.amazonses.com> (raw)
In-Reply-To: <20131203180717.94c013d1.akpm@linux-foundation.org>

On Tue, 3 Dec 2013, Andrew Morton wrote:

> >  	page = alloc_slab_page(alloc_gfp, node, oo);
> >  	if (unlikely(!page)) {
> >  		oo = s->min;
>
> What is the value of s->min?  Please tell me it's zero.

It usually is.

> > @@ -1349,7 +1350,7 @@ static struct page *allocate_slab(struct kmem_cache *s, gfp_t flags, int node)
> >  		&& !(s->flags & (SLAB_NOTRACK | DEBUG_DEFAULT_FLAGS))) {
> >  		int pages = 1 << oo_order(oo);
> >
> > -		kmemcheck_alloc_shadow(page, oo_order(oo), flags, node);
> > +		kmemcheck_alloc_shadow(page, oo_order(oo), alloc_gfp, node);
>
> That seems reasonable, assuming kmemcheck can handle the allocation
> failure.
>
>
> Still I dislike this practice of using unnecessarily large allocations.
> What does it gain us?  Slightly improved object packing density.
> Anything else?

The fastpath for slub works only within the bounds of a single slab page.
Therefore a larger frame increases the number of allocation possible from
the fastpath without having to use the slowpath and also reduces the
management overhead in the partial lists.

There is a kernel parameter that can be used to control the maximum order

	slub_max_order

The default is PAGE_ALLOC_COSTLY_ORDER. See also
Documentation/vm/slub.txt.

Booting with slub_max_order=1 will force order 0/1 pages.


WARNING: multiple messages have this Message-ID (diff)
From: Christoph Lameter <cl@linux.com>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>,
	Johannes Weiner <hannes@cmpxchg.org>,
	Michal Hocko <mhocko@suse.cz>, azurIt <azurit@pobox.sk>,
	linux-mm@kvack.org, cgroups@vger.kernel.org,
	linux-kernel@vger.kernel.org,
	Christian Casteyde <casteyde.christian@free.fr>,
	Pekka Enberg <penberg@kernel.org>
Subject: Re: [patch 2/2] fs: buffer: move allocation failure loop into the allocator
Date: Wed, 4 Dec 2013 15:17:08 +0000	[thread overview]
Message-ID: <00000142be2f1de0-764bb035-adbc-4367-b2b4-bf05498510a6-000000@email.amazonses.com> (raw)
In-Reply-To: <20131203180717.94c013d1.akpm@linux-foundation.org>

On Tue, 3 Dec 2013, Andrew Morton wrote:

> >  	page = alloc_slab_page(alloc_gfp, node, oo);
> >  	if (unlikely(!page)) {
> >  		oo = s->min;
>
> What is the value of s->min?  Please tell me it's zero.

It usually is.

> > @@ -1349,7 +1350,7 @@ static struct page *allocate_slab(struct kmem_cache *s, gfp_t flags, int node)
> >  		&& !(s->flags & (SLAB_NOTRACK | DEBUG_DEFAULT_FLAGS))) {
> >  		int pages = 1 << oo_order(oo);
> >
> > -		kmemcheck_alloc_shadow(page, oo_order(oo), flags, node);
> > +		kmemcheck_alloc_shadow(page, oo_order(oo), alloc_gfp, node);
>
> That seems reasonable, assuming kmemcheck can handle the allocation
> failure.
>
>
> Still I dislike this practice of using unnecessarily large allocations.
> What does it gain us?  Slightly improved object packing density.
> Anything else?

The fastpath for slub works only within the bounds of a single slab page.
Therefore a larger frame increases the number of allocation possible from
the fastpath without having to use the slowpath and also reduces the
management overhead in the partial lists.

There is a kernel parameter that can be used to control the maximum order

	slub_max_order

The default is PAGE_ALLOC_COSTLY_ORDER. See also
Documentation/vm/slub.txt.

Booting with slub_max_order=1 will force order 0/1 pages.

--
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>

WARNING: multiple messages have this Message-ID (diff)
From: Christoph Lameter <cl-vYTEC60ixJUAvxtiuMwx3w@public.gmane.org>
To: Andrew Morton <akpm-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org>
Cc: Joonsoo Kim <iamjoonsoo.kim-Hm3cg6mZ9cc@public.gmane.org>,
	Johannes Weiner <hannes-druUgvl0LCNAfugRpC6u6w@public.gmane.org>,
	Michal Hocko <mhocko-AlSwsSmVLrQ@public.gmane.org>,
	azurIt <azurit-Rm0zKEqwvD4@public.gmane.org>,
	linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org,
	cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	Christian Casteyde
	<casteyde.christian-GANU6spQydw@public.gmane.org>,
	Pekka Enberg <penberg-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
Subject: Re: [patch 2/2] fs: buffer: move allocation failure loop into the allocator
Date: Wed, 4 Dec 2013 15:17:08 +0000	[thread overview]
Message-ID: <00000142be2f1de0-764bb035-adbc-4367-b2b4-bf05498510a6-000000@email.amazonses.com> (raw)
In-Reply-To: <20131203180717.94c013d1.akpm-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org>

On Tue, 3 Dec 2013, Andrew Morton wrote:

> >  	page = alloc_slab_page(alloc_gfp, node, oo);
> >  	if (unlikely(!page)) {
> >  		oo = s->min;
>
> What is the value of s->min?  Please tell me it's zero.

It usually is.

> > @@ -1349,7 +1350,7 @@ static struct page *allocate_slab(struct kmem_cache *s, gfp_t flags, int node)
> >  		&& !(s->flags & (SLAB_NOTRACK | DEBUG_DEFAULT_FLAGS))) {
> >  		int pages = 1 << oo_order(oo);
> >
> > -		kmemcheck_alloc_shadow(page, oo_order(oo), flags, node);
> > +		kmemcheck_alloc_shadow(page, oo_order(oo), alloc_gfp, node);
>
> That seems reasonable, assuming kmemcheck can handle the allocation
> failure.
>
>
> Still I dislike this practice of using unnecessarily large allocations.
> What does it gain us?  Slightly improved object packing density.
> Anything else?

The fastpath for slub works only within the bounds of a single slab page.
Therefore a larger frame increases the number of allocation possible from
the fastpath without having to use the slowpath and also reduces the
management overhead in the partial lists.

There is a kernel parameter that can be used to control the maximum order

	slub_max_order

The default is PAGE_ALLOC_COSTLY_ORDER. See also
Documentation/vm/slub.txt.

Booting with slub_max_order=1 will force order 0/1 pages.

  parent reply	other threads:[~2013-12-04 15:17 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-10-08 20:58 [patch 1/2] mm: memcg: handle non-error OOM situations more gracefully Johannes Weiner
2013-10-08 20:58 ` Johannes Weiner
2013-10-08 20:58 ` Johannes Weiner
2013-10-08 20:58 ` [patch 2/2] fs: buffer: move allocation failure loop into the allocator Johannes Weiner
2013-10-08 20:58   ` Johannes Weiner
2013-10-08 20:58   ` Johannes Weiner
2013-10-11 20:51   ` Andrew Morton
2013-10-11 20:51     ` Andrew Morton
2013-12-04  0:59   ` Andrew Morton
2013-12-04  0:59     ` Andrew Morton
2013-12-04  1:52     ` Joonsoo Kim
2013-12-04  1:52       ` Joonsoo Kim
2013-12-04  1:52       ` Joonsoo Kim
2013-12-04  2:07       ` Andrew Morton
2013-12-04  2:07         ` Andrew Morton
2013-12-04  2:07         ` Andrew Morton
2013-12-04  2:42         ` Joonsoo Kim
2013-12-04  2:42           ` Joonsoo Kim
2013-12-04 15:17         ` Christoph Lameter [this message]
2013-12-04 15:17           ` Christoph Lameter
2013-12-04 15:17           ` Christoph Lameter
2013-12-04 16:02           ` Joonsoo Kim
2013-12-04 16:02             ` Joonsoo Kim
2013-12-04 16:33             ` Christoph Lameter
2013-12-04 16:33               ` Christoph Lameter
2013-12-05  8:44               ` Joonsoo Kim
2013-12-05  8:44                 ` Joonsoo Kim
2013-12-05 18:50                 ` Christoph Lameter
2013-12-05 18:50                   ` Christoph Lameter
2013-12-06  8:57                   ` Joonsoo Kim
2013-12-06  8:57                     ` Joonsoo Kim
2013-12-13  6:58       ` Joonsoo Kim
2013-12-13  6:58         ` Joonsoo Kim
2013-12-13  6:58         ` Joonsoo Kim
2013-12-13 16:40         ` Christoph Lameter
2013-12-13 16:40           ` Christoph Lameter
2013-12-16  8:22           ` Joonsoo Kim
2013-12-16  8:22             ` Joonsoo Kim
2013-12-16  8:22             ` Joonsoo Kim

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=00000142be2f1de0-764bb035-adbc-4367-b2b4-bf05498510a6-000000@email.amazonses.com \
    --to=cl@linux.com \
    --cc=akpm@linux-foundation.org \
    --cc=azurit@pobox.sk \
    --cc=casteyde.christian@free.fr \
    --cc=cgroups@vger.kernel.org \
    --cc=hannes@cmpxchg.org \
    --cc=iamjoonsoo.kim@lge.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mhocko@suse.cz \
    --cc=penberg@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 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.