The patch titled
mm: batch activate_page() to reduce lock contention
has been added to the -mm tree. Its filename is
mm-batch-activate_page-to-reduce-lock-contention.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/SubmitChecklist when testing your code ***
See http://userweb.kernel.org/~akpm/stuff/added-to-mm.txt to find
out what to do about this
The current -mm tree may be found at http://userweb.kernel.org/~akpm/mmotm/
------------------------------------------------------
Subject: mm: batch activate_page() to reduce lock contention
From: Shaohua Li <shaohua.li@xxxxxxxxx>
The zone->lru_lock is heavily contented in workload where activate_page()
is frequently used. We could do batch activate_page() to reduce the lock
contention. The batched pages will be added into zone list when the pool
is full or page reclaim is trying to drain them.
For example, in a 4 socket 64 CPU system, create a sparse file and 64
processes, processes shared map to the file. Each process read access the
whole file and then exit. The process exit will do unmap_vmas() and cause
a lot of activate_page() call. In such workload, we saw about 58% total
time reduction with below patch. Other workloads with a lot of
activate_page also benefits a lot too.
I tested some microbenchmarks:
case-anon-cow-rand-mt 0.58%
case-anon-cow-rand -3.30%
case-anon-cow-seq-mt -0.51%
case-anon-cow-seq -5.68%
case-anon-r-rand-mt 0.23%
case-anon-r-rand 0.81%
case-anon-r-seq-mt -0.71%
case-anon-r-seq -1.99%
case-anon-rx-rand-mt 2.11%
case-anon-rx-seq-mt 3.46%
case-anon-w-rand-mt -0.03%
case-anon-w-rand -0.50%
case-anon-w-seq-mt -1.08%
case-anon-w-seq -0.12%
case-anon-wx-rand-mt -5.02%
case-anon-wx-seq-mt -1.43%
case-fork 1.65%
case-fork-sleep -0.07%
case-fork-withmem 1.39%
case-hugetlb -0.59%
case-lru-file-mmap-read-mt -0.54%
case-lru-file-mmap-read 0.61%
case-lru-file-mmap-read-rand -2.24%
case-lru-file-readonce -0.64%
case-lru-file-readtwice -11.69%
case-lru-memcg -1.35%
case-mmap-pread-rand-mt 1.88%
case-mmap-pread-rand -15.26%
case-mmap-pread-seq-mt 0.89%
case-mmap-pread-seq -69.72%
case-mmap-xread-rand-mt 0.71%
case-mmap-xread-seq-mt 0.38%
The most significent are:
case-lru-file-readtwice -11.69%
case-mmap-pread-rand -15.26%
case-mmap-pread-seq -69.72%
which use activate_page a lot. others are basically variations because
each run has slightly difference.
Signed-off-by: Shaohua Li <shaohua.li@xxxxxxxxx>
Cc: Andi Kleen <andi@xxxxxxxxxxxxxx>
Cc: Minchan Kim <minchan.kim@xxxxxxxxx>
Cc: KOSAKI Motohiro <kosaki.motohiro@xxxxxxxxxxxxxx>
Cc: Rik van Riel <riel@xxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
mm/internal.h | 9 ++++
mm/swap.c | 90 ++++++++++++++++++++++++++++++++++++++++++------
mm/vmscan.c | 6 ++-
3 files changed, 92 insertions(+), 13 deletions(-)
diff -puN mm/internal.h~mm-batch-activate_page-to-reduce-lock-contention mm/internal.h
--- a/mm/internal.h~mm-batch-activate_page-to-reduce-lock-contention
+++ a/mm/internal.h
@@ -39,6 +39,15 @@ static inline void __put_page(struct pag
extern unsigned long highest_memmap_pfn;
+#ifdef CONFIG_SMP
+extern int putback_active_lru_page(struct zone *zone, struct page *page);
+#else
+static inline int putback_active_lru_page(struct zone *zone, struct page *page)
+{
+ return 0;
+}
+#endif
+
/*
* in mm/vmscan.c:
*/
diff -puN mm/swap.c~mm-batch-activate_page-to-reduce-lock-contention mm/swap.c
--- a/mm/swap.c~mm-batch-activate_page-to-reduce-lock-contention
+++ a/mm/swap.c
@@ -273,27 +273,94 @@ static void update_page_reclaim_stat(str
}
/*
- * FIXME: speed this up?
+ * A page will go to active list either by activate_page or putback_lru_page.
+ * In the activate_page case, the page hasn't active bit set. The page might
+ * not in LRU list because it's isolated before it gets a chance to be moved to
+ * active list. The window is small because pagevec just stores several pages.
+ * For such case, we do nothing for such page.
+ * In the putback_lru_page case, the page isn't in lru list but has active
+ * bit set
*/
-void activate_page(struct page *page)
+static void __activate_page(struct page *page, void *arg)
{
struct zone *zone = page_zone(page);
+ int file = page_is_file_cache(page);
+ int lru = page_lru_base_type(page);
+ bool putback = !PageLRU(page);
+
+ /* The page is isolated before it's moved to active list */
+ if (!PageLRU(page) && !PageActive(page))
+ return;
+ if ((PageLRU(page) && PageActive(page)) || PageUnevictable(page))
+ return;
- spin_lock_irq(&zone->lru_lock);
- if (PageLRU(page) && !PageActive(page) && !PageUnevictable(page)) {
- int file = page_is_file_cache(page);
- int lru = page_lru_base_type(page);
+ if (!putback)
del_page_from_lru_list(zone, page, lru);
+ else
+ SetPageLRU(page);
- SetPageActive(page);
- lru += LRU_ACTIVE;
- add_page_to_lru_list(zone, page, lru);
- __count_vm_event(PGACTIVATE);
+ SetPageActive(page);
+ lru += LRU_ACTIVE;
+ add_page_to_lru_list(zone, page, lru);
+
+ if (putback)
+ return;
+ __count_vm_event(PGACTIVATE);
+ update_page_reclaim_stat(zone, page, file, 1);
+}
+
+#ifdef CONFIG_SMP
+static DEFINE_PER_CPU(struct pagevec, activate_page_pvecs);
+
+static void activate_page_drain(int cpu)
+{
+ struct pagevec *pvec = &per_cpu(activate_page_pvecs, cpu);
+
+ if (pagevec_count(pvec))
+ pagevec_lru_move_fn(pvec, __activate_page, NULL);
+}
+
+void activate_page(struct page *page)
+{
+ if (PageLRU(page) && !PageActive(page) && !PageUnevictable(page)) {
+ struct pagevec *pvec = &get_cpu_var(activate_page_pvecs);
- update_page_reclaim_stat(zone, page, file, 1);
+ page_cache_get(page);
+ if (!pagevec_add(pvec, page))
+ pagevec_lru_move_fn(pvec, __activate_page, NULL);
+ put_cpu_var(activate_page_pvecs);
}
+}
+
+/* Caller should hold zone->lru_lock */
+int putback_active_lru_page(struct zone *zone, struct page *page)
+{
+ struct pagevec *pvec = &get_cpu_var(activate_page_pvecs);
+
+ if (!pagevec_add(pvec, page)) {
+ spin_unlock_irq(&zone->lru_lock);
+ pagevec_lru_move_fn(pvec, __activate_page, NULL);
+ spin_lock_irq(&zone->lru_lock);
+ }
+ put_cpu_var(activate_page_pvecs);
+ return 1;
+}
+
+#else
+static void inline activate_page_drain(int cpu)
+{
+}
+
+void activate_page(struct page *page)
+{
+ struct zone *zone = page_zone(page);
+
+ spin_lock_irq(&zone->lru_lock);
+ if (PageLRU(page) && !PageActive(page) && !PageUnevictable(page))
+ __activate_page(page, NULL);
spin_unlock_irq(&zone->lru_lock);
}
+#endif
/*
* Mark a page as having seen activity.
@@ -435,6 +502,7 @@ static void drain_cpu_pagevecs(int cpu)
pvec = &per_cpu(lru_deactivate_pvecs, cpu);
if (pagevec_count(pvec))
__pagevec_lru_deactive(pvec);
+ activate_page_drain(cpu);
}
/*
diff -puN mm/vmscan.c~mm-batch-activate_page-to-reduce-lock-contention mm/vmscan.c
--- a/mm/vmscan.c~mm-batch-activate_page-to-reduce-lock-contention
+++ a/mm/vmscan.c
@@ -1271,14 +1271,16 @@ putback_lru_pages(struct zone *zone, str
spin_lock_irq(&zone->lru_lock);
continue;
}
- SetPageLRU(page);
lru = page_lru(page);
- add_page_to_lru_list(zone, page, lru);
if (is_active_lru(lru)) {
int file = is_file_lru(lru);
int numpages = hpage_nr_pages(page);
reclaim_stat->recent_rotated[file] += numpages;
+ if (putback_active_lru_page(zone, page))
+ continue;
}
+ SetPageLRU(page);
+ add_page_to_lru_list(zone, page, lru);
if (!pagevec_add(&pvec, page)) {
spin_unlock_irq(&zone->lru_lock);
__pagevec_release(&pvec);
_
Patches currently in -mm which might be from shaohua.li@xxxxxxxxx are
linux-next.patch
intel_idle-open-broadcast-clock-event-resend.patch
mm-page-allocator-adjust-the-per-cpu-counter-threshold-when-memory-is-low.patch
writeback-io-less-balance_dirty_pages.patch
writeback-consolidate-variable-names-in-balance_dirty_pages.patch
writeback-per-task-rate-limit-on-balance_dirty_pages.patch
writeback-per-task-rate-limit-on-balance_dirty_pages-fix.patch
writeback-prevent-duplicate-balance_dirty_pages_ratelimited-calls.patch
writeback-account-per-bdi-accumulated-written-pages.patch
writeback-bdi-write-bandwidth-estimation.patch
writeback-bdi-write-bandwidth-estimation-fix.patch
writeback-show-bdi-write-bandwidth-in-debugfs.patch
writeback-quit-throttling-when-bdi-dirty-pages-dropped-low.patch
writeback-reduce-per-bdi-dirty-threshold-ramp-up-time.patch
writeback-make-reasonable-gap-between-the-dirty-background-thresholds.patch
writeback-scale-down-max-throttle-bandwidth-on-concurrent-dirtiers.patch
writeback-add-trace-event-for-balance_dirty_pages.patch
writeback-make-nr_to_write-a-per-file-limit.patch
writeback-make-nr_to_write-a-per-file-limit-fix.patch
mm-kswapd-stop-high-order-balancing-when-any-suitable-zone-is-balanced.patch
mm-kswapd-keep-kswapd-awake-for-high-order-allocations-until-a-percentage-of-the-node-is-balanced.patch
mm-kswapd-use-the-order-that-kswapd-was-reclaiming-at-for-sleeping_prematurely.patch
mm-kswapd-reset-kswapd_max_order-and-classzone_idx-after-reading.patch
mm-kswapd-treat-zone-all_unreclaimable-in-sleeping_prematurely-similar-to-balance_pgdat.patch
mm-kswapd-use-the-classzone-idx-that-kswapd-was-using-for-sleeping_prematurely.patch
mm-simplify-code-of-swapc.patch
mm-batch-activate_page-to-reduce-lock-contention.patch
include-asm-generic-vmlinuxldsh-make-readmostly-section-correctly-align.patch
--
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
