From: Andi Kleen <andi@firstfloor.org>
To: Mel Gorman <mgorman@suse.de>
Cc: Linux-MM <linux-mm@kvack.org>,
Johannes Weiner <hannes@cmpxchg.org>, Dave Hansen <dave@sr71.net>,
Christoph Lameter <cl@linux.com>,
LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH 18/22] mm: page allocator: Split magazine lock in two to reduce contention
Date: Wed, 15 May 2013 12:44:31 -0700 [thread overview]
Message-ID: <m24ne43u5c.fsf@firstfloor.org> (raw)
In-Reply-To: <1368028987-8369-19-git-send-email-mgorman@suse.de> (Mel Gorman's message of "Wed, 8 May 2013 17:03:03 +0100")
Mel Gorman <mgorman@suse.de> writes:
>
> -static inline struct free_area_magazine *find_lock_filled_magazine(struct zone *zone)
> +static inline struct free_magazine *find_lock_magazine(struct zone *zone)
> {
> - struct free_area_magazine *area = &zone->_noirq_magazine;
> - if (!area->nr_free)
> + int i = (raw_smp_processor_id() >> 1) & (NR_MAGAZINES-1);
> + int start = i;
> +
> + do {
> + if (spin_trylock(&zone->noirq_magazine[i].lock))
> + goto out;
I'm not sure doing it this way is great. It optimizes for lock
contention vs the initial cost of just fetching the cache line.
Doing the try lock already has to fetch the cache line, even
if the lock is contended.
Page allocation should be limited more by the cache line bouncing
than long contention
So you may be paying the fetch cost multiple times without actually
amortizing it.
If you want to do it this way I would read the lock only. That can
be much cheaper because it doesn't have to take the cache line
exclusive. It may still need to transfer it though (because another
CPU just took it exclusive), which may be already somewhat expensive.
So overall I'm not sure it's a good idea.
-Andi
--
ak@linux.intel.com -- Speaking for myself only
--
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>
next prev parent reply other threads:[~2013-05-15 19:44 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-05-08 16:02 [RFC PATCH 00/22] Per-cpu page allocator replacement prototype Mel Gorman
2013-05-08 16:02 ` [PATCH 01/22] mm: page allocator: Lookup pageblock migratetype with IRQs enabled during free Mel Gorman
2013-05-08 16:02 ` [PATCH 02/22] mm: page allocator: Push down where IRQs are disabled during page free Mel Gorman
2013-05-08 16:02 ` [PATCH 03/22] mm: page allocator: Use unsigned int for order in more places Mel Gorman
2013-05-08 16:02 ` [PATCH 04/22] mm: page allocator: Only check migratetype of pages being drained while CMA active Mel Gorman
2013-05-08 16:02 ` [PATCH 05/22] oom: Use number of online nodes when deciding whether to suppress messages Mel Gorman
2013-05-08 16:02 ` [PATCH 06/22] mm: page allocator: Convert hot/cold parameter and immediate callers to bool Mel Gorman
2013-05-08 16:02 ` [PATCH 07/22] mm: page allocator: Do not lookup the pageblock migratetype during allocation Mel Gorman
2013-05-08 16:02 ` [PATCH 08/22] mm: page allocator: Remove the per-cpu page allocator Mel Gorman
2013-05-08 16:02 ` [PATCH 09/22] mm: page allocator: Allocate/free order-0 pages from a per-zone magazine Mel Gorman
2013-05-08 18:41 ` Christoph Lameter
2013-05-09 15:23 ` Mel Gorman
2013-05-09 16:21 ` Christoph Lameter
2013-05-09 17:27 ` Mel Gorman
2013-05-09 18:08 ` Christoph Lameter
2013-05-08 16:02 ` [PATCH 10/22] mm: page allocator: Allocate and free pages from magazine in batches Mel Gorman
2013-05-08 16:02 ` [PATCH 11/22] mm: page allocator: Shrink the magazine to the migratetypes in use Mel Gorman
2013-05-08 16:02 ` [PATCH 12/22] mm: page allocator: Remove knowledge of hot/cold from page allocator Mel Gorman
2013-05-08 16:02 ` [PATCH 13/22] mm: page allocator: Use list_splice to refill the magazine Mel Gorman
2013-05-08 16:02 ` [PATCH 14/22] mm: page allocator: Do not disable IRQs just to update stats Mel Gorman
2013-05-08 16:03 ` [PATCH 15/22] mm: page allocator: Check if interrupts are enabled only once per allocation attempt Mel Gorman
2013-05-08 16:03 ` [PATCH 16/22] mm: page allocator: Remove coalescing improvement heuristic during page free Mel Gorman
2013-05-08 16:03 ` [PATCH 17/22] mm: page allocator: Move magazine access behind accessors Mel Gorman
2013-05-08 16:03 ` [PATCH 18/22] mm: page allocator: Split magazine lock in two to reduce contention Mel Gorman
2013-05-09 15:21 ` Dave Hansen
2013-05-15 19:44 ` Andi Kleen [this message]
2013-05-08 16:03 ` [PATCH 19/22] mm: page allocator: Watch for magazine and zone lock contention Mel Gorman
2013-05-08 16:03 ` [PATCH 20/22] mm: page allocator: Hold magazine lock for a batch of pages Mel Gorman
2013-05-08 16:03 ` [PATCH 21/22] mm: compaction: Release free page list under a batched magazine lock Mel Gorman
2013-05-08 16:03 ` [PATCH 22/22] mm: page allocator: Drain magazines for direct compact failures Mel Gorman
2013-05-09 15:41 ` [RFC PATCH 00/22] Per-cpu page allocator replacement prototype Dave Hansen
2013-05-09 16:25 ` Christoph Lameter
2013-05-09 17:33 ` Mel Gorman
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=m24ne43u5c.fsf@firstfloor.org \
--to=andi@firstfloor.org \
--cc=cl@linux.com \
--cc=dave@sr71.net \
--cc=hannes@cmpxchg.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mgorman@suse.de \
/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