From: Mel Gorman <mgorman@suse.de>
To: Christoph Lameter <cl@linux.com>
Cc: Michal Hocko <mhocko@kernel.org>,
Joonsoo Kim <iamjoonsoo.kim@lge.com>,
Aruna Ramakrishna <aruna.ramakrishna@oracle.com>,
linux-mm@kvack.org, linux-kernel@vger.kernel.org,
Mike Kravetz <mike.kravetz@oracle.com>,
Pekka Enberg <penberg@kernel.org>,
David Rientjes <rientjes@google.com>,
Andrew Morton <akpm@linux-foundation.org>,
Jiri Slaby <jslaby@suse.cz>
Subject: Re: what is the purpose of SLAB and SLUB (was: Re: [PATCH v3] mm/slab: Improve performance of gathering slabinfo) stats
Date: Thu, 25 Aug 2016 11:07:07 +0100 [thread overview]
Message-ID: <20160825100707.GU2693@suse.de> (raw)
In-Reply-To: <alpine.DEB.2.20.1608242240460.1837@east.gentwo.org>
On Wed, Aug 24, 2016 at 11:01:43PM -0500, Christoph Lameter wrote:
> On Wed, 24 Aug 2016, Mel Gorman wrote:
> > If/when I get back to the page allocator, the priority would be a bulk
> > API for faster allocs of batches of order-0 pages instead of allocating
> > a large page and splitting.
> >
>
> OMG. Do we really want to continue this? There are billions of Linux
> devices out there that require a reboot at least once a week. This is now
> standard with certain Android phones. In our company we reboot all
> machines every week because fragmentation degrades performance
> significantly. We need to finally face up to it and deal with the issue
> instead of continuing to produce more half ass-ed solutions.
>
Flipping the lid aside, there will always be a need for fast management
of 4K pages. The primary use case is networking that sometimes uses
high-order pages to avoid allocator overhead and amortise DMA setup.
Userspace-mapped pages will always be 4K although fault-around may benefit
from bulk allocating the pages. That is relatively low hanging fruit that
would take a few weeks given a free schedule.
Dirty tracking of pages on a 4K boundary will always be required to avoid IO
multiplier effects that cannot be side-stepped by increasing the fundamental
unit of allocation.
Batching of tree_lock during reclaim for large files and swapping is also
relatively low hanging fruit that also is doable in a week or two.
A high-order per-cpu cache for SLUB to reduce zone->lock contention is
also relatively low hanging fruit with the caveat it makes per_cpu_pages
larger than a cache line.
If you want to rework the VM to use a larger fundamental unit, track
sub-units where required and deal with the internal fragmentation issues
then by all means go ahead and deal with it.
--
Mel Gorman
SUSE Labs
--
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: Mel Gorman <mgorman@suse.de>
To: Christoph Lameter <cl@linux.com>
Cc: Michal Hocko <mhocko@kernel.org>,
Joonsoo Kim <iamjoonsoo.kim@lge.com>,
Aruna Ramakrishna <aruna.ramakrishna@oracle.com>,
linux-mm@kvack.org, linux-kernel@vger.kernel.org,
Mike Kravetz <mike.kravetz@oracle.com>,
Pekka Enberg <penberg@kernel.org>,
David Rientjes <rientjes@google.com>,
Andrew Morton <akpm@linux-foundation.org>,
Jiri Slaby <jslaby@suse.cz>
Subject: Re: what is the purpose of SLAB and SLUB (was: Re: [PATCH v3] mm/slab: Improve performance of gathering slabinfo) stats
Date: Thu, 25 Aug 2016 11:07:07 +0100 [thread overview]
Message-ID: <20160825100707.GU2693@suse.de> (raw)
In-Reply-To: <alpine.DEB.2.20.1608242240460.1837@east.gentwo.org>
On Wed, Aug 24, 2016 at 11:01:43PM -0500, Christoph Lameter wrote:
> On Wed, 24 Aug 2016, Mel Gorman wrote:
> > If/when I get back to the page allocator, the priority would be a bulk
> > API for faster allocs of batches of order-0 pages instead of allocating
> > a large page and splitting.
> >
>
> OMG. Do we really want to continue this? There are billions of Linux
> devices out there that require a reboot at least once a week. This is now
> standard with certain Android phones. In our company we reboot all
> machines every week because fragmentation degrades performance
> significantly. We need to finally face up to it and deal with the issue
> instead of continuing to produce more half ass-ed solutions.
>
Flipping the lid aside, there will always be a need for fast management
of 4K pages. The primary use case is networking that sometimes uses
high-order pages to avoid allocator overhead and amortise DMA setup.
Userspace-mapped pages will always be 4K although fault-around may benefit
from bulk allocating the pages. That is relatively low hanging fruit that
would take a few weeks given a free schedule.
Dirty tracking of pages on a 4K boundary will always be required to avoid IO
multiplier effects that cannot be side-stepped by increasing the fundamental
unit of allocation.
Batching of tree_lock during reclaim for large files and swapping is also
relatively low hanging fruit that also is doable in a week or two.
A high-order per-cpu cache for SLUB to reduce zone->lock contention is
also relatively low hanging fruit with the caveat it makes per_cpu_pages
larger than a cache line.
If you want to rework the VM to use a larger fundamental unit, track
sub-units where required and deal with the internal fragmentation issues
then by all means go ahead and deal with it.
--
Mel Gorman
SUSE Labs
next prev parent reply other threads:[~2016-08-25 10:07 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-08-17 18:20 [PATCH v3] mm/slab: Improve performance of gathering slabinfo stats Aruna Ramakrishna
2016-08-17 18:20 ` Aruna Ramakrishna
2016-08-17 19:03 ` Eric Dumazet
2016-08-17 19:03 ` Eric Dumazet
2016-08-17 19:25 ` Aruna Ramakrishna
2016-08-17 19:25 ` Aruna Ramakrishna
2016-08-18 11:52 ` Michal Hocko
2016-08-18 11:52 ` Michal Hocko
2016-08-19 5:47 ` aruna.ramakrishna
2016-08-19 5:47 ` aruna.ramakrishna
2016-08-23 2:13 ` Joonsoo Kim
2016-08-23 2:13 ` Joonsoo Kim
2016-08-23 15:38 ` what is the purpose of SLAB and SLUB (was: Re: [PATCH v3] mm/slab: Improve performance of gathering slabinfo) stats Michal Hocko
2016-08-23 15:38 ` Michal Hocko
2016-08-23 15:54 ` what is the purpose of SLAB and SLUB Andi Kleen
2016-08-23 15:54 ` Andi Kleen
2016-08-25 4:10 ` Christoph Lameter
2016-08-25 4:10 ` Christoph Lameter
2016-08-25 7:32 ` Michal Hocko
2016-08-25 7:32 ` Michal Hocko
2016-08-25 19:49 ` Christoph Lameter
2016-08-25 19:49 ` Christoph Lameter
2016-08-24 1:15 ` what is the purpose of SLAB and SLUB (was: Re: [PATCH v3] mm/slab: Improve performance of gathering slabinfo) stats Joonsoo Kim
2016-08-24 1:15 ` Joonsoo Kim
2016-08-24 8:05 ` Michal Hocko
2016-08-24 8:05 ` Michal Hocko
2016-08-24 8:20 ` Mel Gorman
2016-08-24 8:20 ` Mel Gorman
2016-08-25 4:01 ` Christoph Lameter
2016-08-25 4:01 ` Christoph Lameter
2016-08-25 10:07 ` Mel Gorman [this message]
2016-08-25 10:07 ` Mel Gorman
2016-08-25 19:55 ` Christoph Lameter
2016-08-25 19:55 ` Christoph Lameter
2016-08-26 20:47 ` what is the purpose of SLAB and SLUB Andi Kleen
2016-08-26 20:47 ` Andi Kleen
2016-08-29 13:44 ` Michal Hocko
2016-08-29 13:44 ` Michal Hocko
2016-08-29 14:49 ` Christoph Lameter
2016-08-29 14:49 ` Christoph Lameter
2016-08-30 9:39 ` what is the purpose of SLAB and SLUB (was: Re: [PATCH v3] mm/slab: Improve performance of gathering slabinfo) stats Mel Gorman
2016-08-30 9:39 ` Mel Gorman
2016-08-30 19:32 ` Christoph Lameter
2016-08-30 19:32 ` 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=20160825100707.GU2693@suse.de \
--to=mgorman@suse.de \
--cc=akpm@linux-foundation.org \
--cc=aruna.ramakrishna@oracle.com \
--cc=cl@linux.com \
--cc=iamjoonsoo.kim@lge.com \
--cc=jslaby@suse.cz \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mhocko@kernel.org \
--cc=mike.kravetz@oracle.com \
--cc=penberg@kernel.org \
--cc=rientjes@google.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 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.