Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757015Ab3HGPvf (ORCPT ); Wed, 7 Aug 2013 11:51:35 -0400 Received: from usmamail.tilera.com ([12.216.194.151]:63734 "EHLO USMAMAIL.TILERA.COM" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753806Ab3HGPve (ORCPT ); Wed, 7 Aug 2013 11:51:34 -0400 Message-ID: <201308071551.r77FpWTf022475@farm-0012.internal.tilera.com> In-Reply-To: <201308071458.r77EwuJV013106@farm-0012.internal.tilera.com> References: <201308071458.r77EwuJV013106@farm-0012.internal.tilera.com> From: Chris Metcalf Date: Tue, 6 Aug 2013 16:22:39 -0400 Subject: [PATCH v2] mm: make lru_add_drain_all() selective To: , , Tejun Heo , Thomas Gleixner , Frederic Weisbecker MIME-Version: 1.0 Content-Type: text/plain Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 5557 Lines: 182 This change makes lru_add_drain_all() only selectively interrupt the cpus that have per-cpu free pages that can be drained. This is important in nohz mode where calling mlockall(), for example, otherwise will interrupt every core unnecessarily. Signed-off-by: Chris Metcalf --- Oops! In the previous version of this change I had just blindly patched it forward from a slightly older version of mm/swap.c. This version is now properly against a version of mm/swap.c that includes all the latest changes to lru_add_drain_all(). include/linux/workqueue.h | 3 +++ kernel/workqueue.c | 35 ++++++++++++++++++++++++++--------- mm/swap.c | 37 ++++++++++++++++++++++++++++++++++++- 3 files changed, 65 insertions(+), 10 deletions(-) diff --git a/include/linux/workqueue.h b/include/linux/workqueue.h index a0ed78a..71a3fe7 100644 --- a/include/linux/workqueue.h +++ b/include/linux/workqueue.h @@ -13,6 +13,8 @@ #include #include +struct cpumask; + struct workqueue_struct; struct work_struct; @@ -470,6 +472,7 @@ extern void flush_workqueue(struct workqueue_struct *wq); extern void drain_workqueue(struct workqueue_struct *wq); extern void flush_scheduled_work(void); +extern int schedule_on_cpu_mask(work_func_t func, const struct cpumask *mask); extern int schedule_on_each_cpu(work_func_t func); int execute_in_process_context(work_func_t fn, struct execute_work *); diff --git a/kernel/workqueue.c b/kernel/workqueue.c index f02c4a4..a6d1809 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -2962,17 +2962,18 @@ bool cancel_delayed_work_sync(struct delayed_work *dwork) EXPORT_SYMBOL(cancel_delayed_work_sync); /** - * schedule_on_each_cpu - execute a function synchronously on each online CPU + * schedule_on_cpu_mask - execute a function synchronously on each listed CPU * @func: the function to call + * @mask: the cpumask to invoke the function on * - * schedule_on_each_cpu() executes @func on each online CPU using the + * schedule_on_cpu_mask() executes @func on each listed CPU using the * system workqueue and blocks until all CPUs have completed. - * schedule_on_each_cpu() is very slow. + * schedule_on_cpu_mask() is very slow. * * RETURNS: * 0 on success, -errno on failure. */ -int schedule_on_each_cpu(work_func_t func) +int schedule_on_cpu_mask(work_func_t func, const struct cpumask *mask) { int cpu; struct work_struct __percpu *works; @@ -2981,24 +2982,40 @@ int schedule_on_each_cpu(work_func_t func) if (!works) return -ENOMEM; - get_online_cpus(); - - for_each_online_cpu(cpu) { + for_each_cpu(cpu, mask) { struct work_struct *work = per_cpu_ptr(works, cpu); INIT_WORK(work, func); schedule_work_on(cpu, work); } - for_each_online_cpu(cpu) + for_each_cpu(cpu, mask) flush_work(per_cpu_ptr(works, cpu)); - put_online_cpus(); free_percpu(works); return 0; } /** + * schedule_on_each_cpu - execute a function synchronously on each online CPU + * @func: the function to call + * + * schedule_on_each_cpu() executes @func on each online CPU using the + * system workqueue and blocks until all CPUs have completed. + * schedule_on_each_cpu() is very slow. + * + * RETURNS: + * 0 on success, -errno on failure. + */ +int schedule_on_each_cpu(work_func_t func) +{ + get_online_cpus(); + schedule_on_cpu_mask(func, cpu_online_mask); + put_online_cpus(); + return 0; +} + +/** * flush_scheduled_work - ensure that any scheduled work has run to completion. * * Forces execution of the kernel-global workqueue and blocks until its diff --git a/mm/swap.c b/mm/swap.c index 4a1d0d2..d4a862b 100644 --- a/mm/swap.c +++ b/mm/swap.c @@ -405,6 +405,11 @@ static void activate_page_drain(int cpu) pagevec_lru_move_fn(pvec, __activate_page, NULL); } +static bool need_activate_page_drain(int cpu) +{ + return pagevec_count(&per_cpu(activate_page_pvecs, cpu)) != 0; +} + void activate_page(struct page *page) { if (PageLRU(page) && !PageActive(page) && !PageUnevictable(page)) { @@ -422,6 +427,11 @@ static inline void activate_page_drain(int cpu) { } +static bool need_activate_page_drain(int cpu) +{ + return false; +} + void activate_page(struct page *page) { struct zone *zone = page_zone(page); @@ -683,7 +693,32 @@ static void lru_add_drain_per_cpu(struct work_struct *dummy) */ int lru_add_drain_all(void) { - return schedule_on_each_cpu(lru_add_drain_per_cpu); + cpumask_var_t mask; + int cpu, rc; + + if (!alloc_cpumask_var(&mask, GFP_KERNEL)) + return -ENOMEM; + cpumask_clear(mask); + + /* + * Figure out which cpus need flushing. It's OK if we race + * with changes to the per-cpu lru pvecs, since it's no worse + * than if we flushed all cpus, since a cpu could still end + * up putting pages back on its pvec before we returned. + * And this avoids interrupting other cpus unnecessarily. + */ + for_each_online_cpu(cpu) { + if (pagevec_count(&per_cpu(lru_add_pvec, cpu)) || + pagevec_count(&per_cpu(lru_rotate_pvecs, cpu)) || + pagevec_count(&per_cpu(lru_deactivate_pvecs, cpu)) || + need_activate_page_drain(cpu)) + cpumask_set_cpu(cpu, mask); + } + + rc = schedule_on_cpu_mask(lru_add_drain_per_cpu, mask); + + free_cpumask_var(mask); + return rc; } /* -- 1.8.3.1 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/