From: Mel Gorman <mgorman@techsingularity.net>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Vlastimil Babka <vbabka@suse.cz>, Linux-MM <linux-mm@kvack.org>,
LKML <linux-kernel@vger.kernel.org>,
Mel Gorman <mgorman@techsingularity.net>
Subject: [PATCH 24/24] mm, page_alloc: Do not lookup pcp migratetype during bulk free
Date: Tue, 12 Apr 2016 11:12:25 +0100 [thread overview]
Message-ID: <1460455945-29644-25-git-send-email-mgorman@techsingularity.net> (raw)
In-Reply-To: <1460455945-29644-1-git-send-email-mgorman@techsingularity.net>
During bulk free, the pcp type of the page is known as it was removed
from a specific list. It only needs to be rechecked if an isolated
pageblock exists. This patch removes an unnecessary variable in
the process. The impact is that the round-robin freeing of PCP
lists is distorted when an isolated pageblock is encountered but
that is a rare and harmless corner-case.
The impact on the page allocator microbench of the bulk free patches is
visible for higher batch counts when the bulk free paths are hit.
pagealloc
4.6.0-rc3 4.6.0-rc3
cpuset-v2r2 micro-v2r2
Min free-odr0-1 191.00 ( 0.00%) 195.00 ( -2.09%)
Min free-odr0-2 136.00 ( 0.00%) 136.00 ( 0.00%)
Min free-odr0-4 107.00 ( 0.00%) 107.00 ( 0.00%)
Min free-odr0-8 95.00 ( 0.00%) 95.00 ( 0.00%)
Min free-odr0-16 87.00 ( 0.00%) 87.00 ( 0.00%)
Min free-odr0-32 82.00 ( 0.00%) 82.00 ( 0.00%)
Min free-odr0-64 80.00 ( 0.00%) 80.00 ( 0.00%)
Min free-odr0-128 79.00 ( 0.00%) 79.00 ( 0.00%)
Min free-odr0-256 94.00 ( 0.00%) 97.00 ( -3.19%)
Min free-odr0-512 112.00 ( 0.00%) 109.00 ( 2.68%)
Min free-odr0-1024 118.00 ( 0.00%) 118.00 ( 0.00%)
Min free-odr0-2048 123.00 ( 0.00%) 121.00 ( 1.63%)
Min free-odr0-4096 127.00 ( 0.00%) 125.00 ( 1.57%)
Min free-odr0-8192 129.00 ( 0.00%) 127.00 ( 1.55%)
Min free-odr0-16384 128.00 ( 0.00%) 127.00 ( 0.78%)
It's tiny but the patches are trivial.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
---
mm/page_alloc.c | 13 ++++---------
1 file changed, 4 insertions(+), 9 deletions(-)
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 1b1553c1156c..4d4079309760 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -876,7 +876,7 @@ static void free_pcppages_bulk(struct zone *zone, int count,
*/
do {
batch_free++;
- if (++migratetype == MIGRATE_PCPTYPES)
+ if (++migratetype >= MIGRATE_PCPTYPES)
migratetype = 0;
list = &pcp->lists[migratetype];
} while (list_empty(list));
@@ -886,21 +886,16 @@ static void free_pcppages_bulk(struct zone *zone, int count,
batch_free = count;
do {
- int mt; /* migratetype of the to-be-freed page */
-
page = list_last_entry(list, struct page, lru);
/* must delete as __free_one_page list manipulates */
list_del(&page->lru);
- mt = get_pcppage_migratetype(page);
- /* MIGRATE_ISOLATE page should not go to pcplists */
- VM_BUG_ON_PAGE(is_migrate_isolate(mt), page);
/* Pageblock could have been isolated meanwhile */
if (unlikely(isolated_pageblocks))
- mt = get_pageblock_migratetype(page);
+ migratetype = get_pageblock_migratetype(page);
- __free_one_page(page, page_to_pfn(page), zone, 0, mt);
- trace_mm_page_pcpu_drain(page, 0, mt);
+ __free_one_page(page, page_to_pfn(page), zone, 0, migratetype);
+ trace_mm_page_pcpu_drain(page, 0, migratetype);
} while (--count && --batch_free && !list_empty(list));
}
spin_unlock(&zone->lock);
--
2.6.4
--
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>
prev parent reply other threads:[~2016-04-12 10:14 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-04-12 10:12 [PATCH 00/24] Optimise page alloc/free fast paths v2 Mel Gorman
2016-04-12 10:12 ` [PATCH 01/24] mm, page_alloc: Only check PageCompound for high-order pages Mel Gorman
2016-04-12 10:12 ` [PATCH 02/24] mm, page_alloc: Use new PageAnonHead helper in the free page fast path Mel Gorman
2016-04-12 10:12 ` [PATCH 03/24] mm, page_alloc: Reduce branches in zone_statistics Mel Gorman
2016-04-12 10:12 ` [PATCH 04/24] mm, page_alloc: Inline zone_statistics Mel Gorman
2016-04-12 10:12 ` [PATCH 05/24] mm, page_alloc: Inline the fast path of the zonelist iterator Mel Gorman
2016-04-12 10:12 ` [PATCH 06/24] mm, page_alloc: Use __dec_zone_state for order-0 page allocation Mel Gorman
2016-04-12 10:12 ` [PATCH 07/24] mm, page_alloc: Avoid unnecessary zone lookups during pageblock operations Mel Gorman
2016-04-12 10:12 ` [PATCH 08/24] mm, page_alloc: Convert alloc_flags to unsigned Mel Gorman
2016-04-12 10:12 ` [PATCH 09/24] mm, page_alloc: Convert nr_fair_skipped to bool Mel Gorman
2016-04-12 10:12 ` [PATCH 10/24] mm, page_alloc: Remove unnecessary local variable in get_page_from_freelist Mel Gorman
2016-04-12 10:12 ` [PATCH 11/24] mm, page_alloc: Remove unnecessary initialisation " Mel Gorman
2016-04-12 10:12 ` [PATCH 12/24] mm, page_alloc: Remove unnecessary initialisation from __alloc_pages_nodemask() Mel Gorman
2016-04-12 10:12 ` [PATCH 13/24] mm, page_alloc: Remove redundant check for empty zonelist Mel Gorman
2016-04-12 10:12 ` [PATCH 14/24] mm, page_alloc: Simplify last cpupid reset Mel Gorman
2016-04-12 10:12 ` [PATCH 15/24] mm, page_alloc: Move might_sleep_if check to the allocator slowpath Mel Gorman
2016-04-12 10:12 ` [PATCH 16/24] mm, page_alloc: Move __GFP_HARDWALL modifications out of the fastpath Mel Gorman
2016-04-12 10:12 ` [PATCH 17/24] mm, page_alloc: Reduce cost of fair zone allocation policy retry Mel Gorman
2016-04-12 10:12 ` [PATCH 18/24] mm, page_alloc: Shortcut watermark checks for order-0 pages Mel Gorman
2016-04-12 10:12 ` [PATCH 19/24] mm, page_alloc: Avoid looking up the first zone in a zonelist twice Mel Gorman
2016-04-12 10:12 ` [PATCH 20/24] mm, page_alloc: Check multiple page fields with a single branch Mel Gorman
2016-04-12 10:12 ` [PATCH 21/24] cpuset: use static key better and convert to new API Mel Gorman
2016-04-12 10:12 ` [PATCH 22/24] mm, page_alloc: Check once if a zone has isolated pageblocks Mel Gorman
2016-04-12 10:12 ` [PATCH 23/24] mm, page_alloc: Remove unnecessary variable from free_pcppages_bulk Mel Gorman
2016-04-12 10:12 ` Mel Gorman [this message]
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=1460455945-29644-25-git-send-email-mgorman@techsingularity.net \
--to=mgorman@techsingularity.net \
--cc=akpm@linux-foundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=vbabka@suse.cz \
/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;
as well as URLs for NNTP newsgroup(s).