From mboxrd@z Thu Jan  1 00:00:00 1970
Subject: Re: slab fragmentation ?
From: Badari Pulavarty <pbadari@us.ibm.com>
In-Reply-To: <29460000.1096555795@[10.10.2.4]>
References: <1096500963.12861.21.camel@dyn318077bld.beaverton.ibm.com>
	 <20040929204143.134154bc.akpm@osdl.org>  <29460000.1096555795@[10.10.2.4]>
Content-Type: text/plain
Message-Id: <1096555693.12861.27.camel@dyn318077bld.beaverton.ibm.com>
Mime-Version: 1.0
Date: 30 Sep 2004 07:48:14 -0700
Content-Transfer-Encoding: 7bit
Sender: owner-linux-mm@kvack.org
Return-Path: <owner-linux-mm@kvack.org>
To: "Martin J. Bligh" <mbligh@aracnet.com>
Cc: Andrew Morton <akpm@osdl.org>, manfred@colorfullife.com, linux-mm@kvack.org
List-ID: <linux-mm.kvack.org>

On Thu, 2004-09-30 at 07:49, Martin J. Bligh wrote:
> --Andrew Morton <akpm@osdl.org> wrote (on Wednesday, September 29, 2004 20:41:43 -0700):
> 
> > Badari Pulavarty <pbadari@us.ibm.com> wrote:
> >> 
> >> # name            <active_objs> <num_objs> <objsize> <objperslab> <pagesperslab> : tunables <batchcount> <limit> <sharedfactor> : slabdata <active_slabs> <num_slabs> <sharedavail>
> >> size-40             2633  11468     64   61    1 : tunables  120   60    8 : slabdata    188    188      0
> >> size-40             2633  11468     64   61    1 : tunables  120   60    8 : slabdata    188    188      0
> >> size-40             2633  11468     64   61    1 : tunables  120   60    8 : slabdata    188    188      0
> >> size-40             2633  11468     64   61    1 : tunables  120   60    8 : slabdata    188    188      0
> >> size-40             4457  27084     64   61    1 : tunables  120   60    8 : slabdata    444    444      0
> >> size-40             7685  59292     64   61    1 : tunables  120   60    8 : slabdata    972    972      0
> >> size-40            10761  89548     64   61    1 : tunables  120   60    8 : slabdata   1468   1468      0
> >> size-40            13589 119316     64   61    1 : tunables  120   60    8 : slabdata   1956   1956      0
> >> size-40            16717 149084     64   61    1 : tunables  120   60    8 : slabdata   2444   2444      0
> > 
> > That looks like plain brokenness rather than fragmentation.  We shouldn't
> > be allocating new pages until active_objs reaches num_objs, should we?
> > 
> > Unless the accouting is broken, or course...
> 
> Doesn't this happen if we allocate 1000 slabs, then free half the elements
> in each of the slabs? Which seemed to be the default action of the slab 
> shrink routines ;-)

Yes. But the next allocations should be satisfied by filling in the
partial slabs, instead of getting a new slab.

As you can see from my tests, we are allocating and freeing few
thousands every second. I can imagine this happening, if we allocated
150K objects and then freed 140K of them randomly.

Thanks,
Badari

--
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:"aart@kvack.org"> aart@kvack.org </a>