linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Andrew Morton <akpm@linux-foundation.org>
To: Jesper Dangaard Brouer <brouer@redhat.com>
Cc: linux-mm@kvack.org, Christoph Lameter <cl@linux.com>,
	netdev@vger.kernel.org,
	Alexander Duyck <alexander.duyck@gmail.com>
Subject: Re: [PATCH 2/7] slub bulk alloc: extract objects from the per cpu slab
Date: Tue, 16 Jun 2015 14:48:40 -0700	[thread overview]
Message-ID: <20150616144840.1b669e149d937365a4b54c1c@linux-foundation.org> (raw)
In-Reply-To: <20150615155207.18824.8674.stgit@devil>

On Mon, 15 Jun 2015 17:52:07 +0200 Jesper Dangaard Brouer <brouer@redhat.com> wrote:

> From: Christoph Lameter <cl@linux.com>
> 
> [NOTICE: Already in AKPM's quilt-queue]
> 
> First piece: acceleration of retrieval of per cpu objects
> 
> If we are allocating lots of objects then it is advantageous to disable
> interrupts and avoid the this_cpu_cmpxchg() operation to get these objects
> faster.
> 
> Note that we cannot do the fast operation if debugging is enabled, because
> we would have to add extra code to do all the debugging checks.  And it
> would not be fast anyway.
> 
> Note also that the requirement of having interrupts disabled
> avoids having to do processor flag operations.
> 
> Allocate as many objects as possible in the fast way and then fall back to
> the generic implementation for the rest of the objects.
> 
> ...
>
> --- a/mm/slub.c
> +++ b/mm/slub.c
> @@ -2759,7 +2759,32 @@ EXPORT_SYMBOL(kmem_cache_free_bulk);
>  bool kmem_cache_alloc_bulk(struct kmem_cache *s, gfp_t flags, size_t size,
>  								void **p)
>  {
> -	return kmem_cache_alloc_bulk(s, flags, size, p);
> +	if (!kmem_cache_debug(s)) {
> +		struct kmem_cache_cpu *c;
> +
> +		/* Drain objects in the per cpu slab */
> +		local_irq_disable();
> +		c = this_cpu_ptr(s->cpu_slab);
> +
> +		while (size) {
> +			void *object = c->freelist;
> +
> +			if (!object)
> +				break;
> +
> +			c->freelist = get_freepointer(s, object);
> +			*p++ = object;
> +			size--;
> +
> +			if (unlikely(flags & __GFP_ZERO))
> +				memset(object, 0, s->object_size);
> +		}
> +		c->tid = next_tid(c->tid);
> +
> +		local_irq_enable();

It might be worth adding

		if (!size)
			return true;

here.  To avoid the pointless call to __kmem_cache_alloc_bulk().

It depends on the typical success rate of this allocation loop.  Do you
know what this is?

> +	}
> +
> +	return __kmem_cache_alloc_bulk(s, flags, size, p);
>  }
>  EXPORT_SYMBOL(kmem_cache_alloc_bulk);

--
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:[~2015-06-16 21:48 UTC|newest]

Thread overview: 35+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-06-15 15:51 [PATCH 0/7] slub: bulk alloc and free for slub allocator Jesper Dangaard Brouer
2015-06-15 15:51 ` [PATCH 1/7] slab: infrastructure for bulk object allocation and freeing Jesper Dangaard Brouer
2015-06-15 16:45   ` Alexander Duyck
2015-06-15 16:50     ` Christoph Lameter
2015-06-16 21:44   ` Andrew Morton
2015-06-15 15:52 ` [PATCH 2/7] slub bulk alloc: extract objects from the per cpu slab Jesper Dangaard Brouer
2015-06-16  7:21   ` Joonsoo Kim
2015-06-16 15:05     ` Christoph Lameter
2015-06-16 21:48   ` Andrew Morton [this message]
2015-06-17  6:24     ` Jesper Dangaard Brouer
2015-06-15 15:52 ` [PATCH 3/7] slub: reduce indention level in kmem_cache_alloc_bulk() Jesper Dangaard Brouer
2015-06-15 15:52 ` [PATCH 4/7] slub: fix error path bug in kmem_cache_alloc_bulk Jesper Dangaard Brouer
2015-06-16 21:51   ` Andrew Morton
2015-06-17  6:25     ` Jesper Dangaard Brouer
2015-06-15 15:52 ` [PATCH 5/7] slub: kmem_cache_alloc_bulk() move clearing outside IRQ disabled section Jesper Dangaard Brouer
2015-06-15 15:52 ` [PATCH 6/7] slub: improve bulk alloc strategy Jesper Dangaard Brouer
2015-06-15 16:36   ` Christoph Lameter
2015-06-16 21:53   ` Andrew Morton
2015-06-17  6:29     ` Jesper Dangaard Brouer
2015-06-15 15:52 ` [PATCH 7/7] slub: initial bulk free implementation Jesper Dangaard Brouer
2015-06-15 16:34   ` Christoph Lameter
2015-06-16  8:04     ` Jesper Dangaard Brouer
2015-06-15 17:04   ` Alexander Duyck
2015-06-16  7:23   ` Joonsoo Kim
2015-06-16  9:20     ` Jesper Dangaard Brouer
2015-06-16 12:00       ` Joonsoo Kim
2015-06-16 13:58         ` Jesper Dangaard Brouer
2015-06-16 15:06         ` Christoph Lameter
2015-06-16  7:28   ` Joonsoo Kim
2015-06-16  8:21     ` Jesper Dangaard Brouer
2015-06-16  8:57       ` Jesper Dangaard Brouer
2015-06-16 12:05         ` Joonsoo Kim
2015-06-16 15:10           ` Christoph Lameter
2015-06-16 15:52             ` Jesper Dangaard Brouer
2015-06-16 16:04               ` Christoph Lameter

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=20150616144840.1b669e149d937365a4b54c1c@linux-foundation.org \
    --to=akpm@linux-foundation.org \
    --cc=alexander.duyck@gmail.com \
    --cc=brouer@redhat.com \
    --cc=cl@linux.com \
    --cc=linux-mm@kvack.org \
    --cc=netdev@vger.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox