lkml.org 
[lkml]   [2020]   [Dec]   [1]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
From
Subject[PATCH 2/3] mm/swap.c: bail out early for no memcg and no numa
Date
If a system has memcg disabled and no numa node, like a embedded system,
there is no needs to do the pagevec sort, since only just one lruvec in
system. In this situation, we could skip the pagevec sorting.

Signed-off-by: Alex Shi <alex.shi@linux.alibaba.com>
Cc: Konstantin Khlebnikov <koct9i@gmail.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Yu Zhao <yuzhao@google.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: linux-mm@kvack.org
Cc: linux-kernel@vger.kernel.org
---
mm/swap.c | 19 ++++++++++++-------
1 file changed, 12 insertions(+), 7 deletions(-)

diff --git a/mm/swap.c b/mm/swap.c
index 17d8990e5ca7..814809845700 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -264,12 +264,17 @@ static void pagevec_lru_move_fn(struct pagevec *pvec,
unsigned long flags = 0;
unsigned long lvaddr[PAGEVEC_SIZE];
struct pagevec isopv;
+ struct pagevec *pv;

- pagevec_init(&isopv);
-
- sort_isopv(pvec, &isopv, lvaddr);
+ if (!mem_cgroup_disabled() || num_online_nodes() > 1) {
+ pagevec_init(&isopv);
+ sort_isopv(pvec, &isopv, lvaddr);
+ pv = &isopv;
+ } else {
+ pv = pvec;
+ }

- n = pagevec_count(&isopv);
+ n = pagevec_count(pv);
if (!n)
return;

@@ -284,12 +289,12 @@ static void pagevec_lru_move_fn(struct pagevec *pvec,
spin_lock_irqsave(&lruvec->lru_lock, flags);
}

- (*move_fn)(isopv.pages[i], lruvec);
+ (*move_fn)(pv->pages[i], lruvec);

- SetPageLRU(isopv.pages[i]);
+ SetPageLRU(pv->pages[i]);
}
spin_unlock_irqrestore(&lruvec->lru_lock, flags);
- release_pages(isopv.pages, isopv.nr);
+ release_pages(pv->pages, pv->nr);
}

static void pagevec_move_tail_fn(struct page *page, struct lruvec *lruvec)
--
2.29.GIT
\
 
 \ /
  Last update: 2020-12-01 09:04    [W:0.095 / U:0.148 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site