Linux-mm Archive on lore.kernel.org
 help / color / Atom feed
From: Mel Gorman <mgorman@suse.de>
To: Peter Zijlstra <a.p.zijlstra@chello.nl>, Rik van Riel <riel@redhat.com>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>,
	Ingo Molnar <mingo@kernel.org>,
	Andrea Arcangeli <aarcange@redhat.com>,
	Johannes Weiner <hannes@cmpxchg.org>,
	Linux-MM <linux-mm@kvack.org>,
	LKML <linux-kernel@vger.kernel.org>, Mel Gorman <mgorman@suse.de>
Subject: [PATCH 53/63] sched: numa: Decide whether to favour task or group weights based on swap candidate relationships
Date: Mon,  7 Oct 2013 11:29:31 +0100
Message-ID: <1381141781-10992-54-git-send-email-mgorman@suse.de> (raw)
In-Reply-To: <1381141781-10992-1-git-send-email-mgorman@suse.de>

From: Rik van Riel <riel@redhat.com>

This patch separately considers task and group affinities when searching
for swap candidates during task NUMA placement. If tasks are not part of
a group or the same group then the task weights are considered.
Otherwise the group weights are compared.

Signed-off-by: Rik van Riel <riel@redhat.com>
---
 kernel/sched/fair.c | 59 ++++++++++++++++++++++++++++++++---------------------
 1 file changed, 36 insertions(+), 23 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 58d1070..e7da6f2 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -1039,13 +1039,15 @@ static void task_numa_assign(struct task_numa_env *env,
  * into account that it might be best if task running on the dst_cpu should
  * be exchanged with the source task
  */
-static void task_numa_compare(struct task_numa_env *env, long imp)
+static void task_numa_compare(struct task_numa_env *env,
+			      long taskimp, long groupimp)
 {
 	struct rq *src_rq = cpu_rq(env->src_cpu);
 	struct rq *dst_rq = cpu_rq(env->dst_cpu);
 	struct task_struct *cur;
 	long dst_load, src_load;
 	long load;
+	long imp = (groupimp > 0) ? groupimp : taskimp;
 
 	rcu_read_lock();
 	cur = ACCESS_ONCE(dst_rq->curr);
@@ -1064,10 +1066,19 @@ static void task_numa_compare(struct task_numa_env *env, long imp)
 		if (!cpumask_test_cpu(env->src_cpu, tsk_cpus_allowed(cur)))
 			goto unlock;
 
-		imp += task_weight(cur, env->src_nid) +
-		       group_weight(cur, env->src_nid) -
-		       task_weight(cur, env->dst_nid) -
-		       group_weight(cur, env->dst_nid);
+		/*
+		 * If dst and source tasks are in the same NUMA group, or not
+		 * in any group then look only at task weights otherwise give
+		 * priority to the group weights.
+		 */
+		if (!cur->numa_group || !env->p->numa_group ||
+		    cur->numa_group == env->p->numa_group) {
+			imp = taskimp + task_weight(cur, env->src_nid) -
+			      task_weight(cur, env->dst_nid);
+		} else {
+			imp = groupimp + group_weight(cur, env->src_nid) -
+			       group_weight(cur, env->dst_nid);
+		}
 	}
 
 	if (imp < env->best_imp)
@@ -1117,7 +1128,8 @@ unlock:
 	rcu_read_unlock();
 }
 
-static void task_numa_find_cpu(struct task_numa_env *env, long imp)
+static void task_numa_find_cpu(struct task_numa_env *env,
+				long taskimp, long groupimp)
 {
 	int cpu;
 
@@ -1127,7 +1139,7 @@ static void task_numa_find_cpu(struct task_numa_env *env, long imp)
 			continue;
 
 		env->dst_cpu = cpu;
-		task_numa_compare(env, imp);
+		task_numa_compare(env, taskimp, groupimp);
 	}
 }
 
@@ -1146,9 +1158,9 @@ static int task_numa_migrate(struct task_struct *p)
 		.best_cpu = -1
 	};
 	struct sched_domain *sd;
-	unsigned long weight;
+	unsigned long taskweight, groupweight;
 	int nid, ret;
-	long imp;
+	long taskimp, groupimp;
 
 	/*
 	 * Pick the lowest SD_NUMA domain, as that would have the smallest
@@ -1163,15 +1175,17 @@ static int task_numa_migrate(struct task_struct *p)
 	env.imbalance_pct = 100 + (sd->imbalance_pct - 100) / 2;
 	rcu_read_unlock();
 
-	weight = task_weight(p, env.src_nid) + group_weight(p, env.src_nid);
+	taskweight = task_weight(p, env.src_nid);
+	groupweight = group_weight(p, env.src_nid);
 	update_numa_stats(&env.src_stats, env.src_nid);
 	env.dst_nid = p->numa_preferred_nid;
-	imp = task_weight(p, env.dst_nid) + group_weight(p, env.dst_nid) - weight;
+	taskimp = task_weight(p, env.dst_nid) - taskweight;
+	groupimp = group_weight(p, env.dst_nid) - groupweight;
 	update_numa_stats(&env.dst_stats, env.dst_nid);
 
 	/* If the preferred nid has capacity, try to use it. */
 	if (env.dst_stats.has_capacity)
-		task_numa_find_cpu(&env, imp);
+		task_numa_find_cpu(&env, taskimp, groupimp);
 
 	/* No space available on the preferred nid. Look elsewhere. */
 	if (env.best_cpu == -1) {
@@ -1180,13 +1194,14 @@ static int task_numa_migrate(struct task_struct *p)
 				continue;
 
 			/* Only consider nodes where both task and groups benefit */
-			imp = task_weight(p, nid) + group_weight(p, nid) - weight;
-			if (imp < 0)
+			taskimp = task_weight(p, nid) - taskweight;
+			groupimp = group_weight(p, nid) - groupweight;
+			if (taskimp < 0 && groupimp < 0)
 				continue;
 
 			env.dst_nid = nid;
 			update_numa_stats(&env.dst_stats, env.dst_nid);
-			task_numa_find_cpu(&env, imp);
+			task_numa_find_cpu(&env, taskimp, groupimp);
 		}
 	}
 
@@ -4678,10 +4693,9 @@ static bool migrate_improves_locality(struct task_struct *p, struct lb_env *env)
 	if (dst_nid == p->numa_preferred_nid)
 		return true;
 
-	/* After the task has settled, check if the new node is better. */
-	if (p->numa_migrate_seq >= sysctl_numa_balancing_settle_count &&
-			task_weight(p, dst_nid) + group_weight(p, dst_nid) >
-			task_weight(p, src_nid) + group_weight(p, src_nid))
+	/* If both task and group weight improve, this move is a winner. */
+	if (task_weight(p, dst_nid) > task_weight(p, src_nid) &&
+	    group_weight(p, dst_nid) > group_weight(p, src_nid))
 		return true;
 
 	return false;
@@ -4708,10 +4722,9 @@ static bool migrate_degrades_locality(struct task_struct *p, struct lb_env *env)
 	if (src_nid == p->numa_preferred_nid)
 		return true;
 
-	/* After the task has settled, check if the new node is worse. */
-	if (p->numa_migrate_seq >= sysctl_numa_balancing_settle_count &&
-			task_weight(p, dst_nid) + group_weight(p, dst_nid) <
-			task_weight(p, src_nid) + group_weight(p, src_nid))
+	/* If either task or group weight get worse, don't do it. */
+	if (task_weight(p, dst_nid) < task_weight(p, src_nid) ||
+	    group_weight(p, dst_nid) < group_weight(p, src_nid))
 		return true;
 
 	return false;
-- 
1.8.4

--
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>

  parent reply index

Thread overview: 136+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-10-07 10:28 [PATCH 0/63] Basic scheduler support for automatic NUMA balancing V9 Mel Gorman
2013-10-07 10:28 ` [PATCH 01/63] hotplug: Optimize {get,put}_online_cpus() Mel Gorman
2013-10-07 10:28 ` [PATCH 02/63] mm: numa: Document automatic NUMA balancing sysctls Mel Gorman
2013-10-07 12:46   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 03/63] sched, numa: Comment fixlets Mel Gorman
2013-10-07 12:46   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 04/63] mm: numa: Do not account for a hinting fault if we raced Mel Gorman
2013-10-07 12:47   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 05/63] mm: Wait for THP migrations to complete during NUMA hinting faults Mel Gorman
2013-10-07 13:55   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 06/63] mm: Prevent parallel splits during THP migration Mel Gorman
2013-10-07 14:01   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 07/63] mm: numa: Sanitize task_numa_fault() callsites Mel Gorman
2013-10-07 14:02   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 08/63] mm: Close races between THP migration and PMD numa clearing Mel Gorman
2013-10-07 14:02   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 09/63] mm: Account for a THP NUMA hinting update as one PTE update Mel Gorman
2013-10-07 14:02   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 10/63] mm: Do not flush TLB during protection change if !pte_present && !migration_entry Mel Gorman
2013-10-07 15:12   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 11/63] mm: Only flush TLBs if a transhuge PMD is modified for NUMA pte scanning Mel Gorman
2013-10-07 10:28 ` [PATCH 12/63] mm: numa: Do not migrate or account for hinting faults on the zero page Mel Gorman
2013-10-07 17:10   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 13/63] sched: numa: Mitigate chance that same task always updates PTEs Mel Gorman
2013-10-07 17:24   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 14/63] sched: numa: Continue PTE scanning even if migrate rate limited Mel Gorman
2013-10-07 17:24   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 15/63] Revert "mm: sched: numa: Delay PTE scanning until a task is scheduled on a new node" Mel Gorman
2013-10-07 17:42   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 16/63] sched: numa: Initialise numa_next_scan properly Mel Gorman
2013-10-07 17:44   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 17/63] sched: Set the scan rate proportional to the memory usage of the task being scanned Mel Gorman
2013-10-07 17:44   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 18/63] sched: numa: Slow scan rate if no NUMA hinting faults are being recorded Mel Gorman
2013-10-07 18:02   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 19/63] sched: Track NUMA hinting faults on per-node basis Mel Gorman
2013-10-07 18:02   ` Rik van Riel
2013-12-04  5:32   ` Wanpeng Li
2013-12-04  5:37     ` Wanpeng Li
2013-10-07 10:28 ` [PATCH 20/63] sched: Select a preferred node with the most numa hinting faults Mel Gorman
2013-10-07 18:04   ` Rik van Riel
2013-10-07 10:28 ` [PATCH 21/63] sched: Update NUMA hinting faults once per scan Mel Gorman
2013-10-07 18:39   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 22/63] sched: Favour moving tasks towards the preferred node Mel Gorman
2013-10-07 18:39   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 23/63] sched: Resist moving tasks towards nodes with fewer hinting faults Mel Gorman
2013-10-07 18:40   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 24/63] sched: Reschedule task on preferred NUMA node once selected Mel Gorman
2013-10-07 18:40   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 25/63] sched: Add infrastructure for split shared/private accounting of NUMA hinting faults Mel Gorman
2013-10-07 18:41   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 26/63] sched: Check current->mm before allocating NUMA faults Mel Gorman
2013-10-07 18:41   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 27/63] mm: numa: Scan pages with elevated page_mapcount Mel Gorman
2013-10-07 18:43   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 28/63] sched: Remove check that skips small VMAs Mel Gorman
2013-10-07 18:44   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 29/63] sched: Set preferred NUMA node based on number of private faults Mel Gorman
2013-10-07 18:45   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 30/63] sched: Do not migrate memory immediately after switching node Mel Gorman
2013-10-07 10:29 ` [PATCH 31/63] mm: numa: only unmap migrate-on-fault VMAs Mel Gorman
2013-10-07 10:29 ` [PATCH 32/63] sched: Avoid overloading CPUs on a preferred NUMA node Mel Gorman
2013-10-07 18:58   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 33/63] sched: Retry migration of tasks to CPU on a preferred node Mel Gorman
2013-10-07 18:58   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 34/63] sched: numa: increment numa_migrate_seq when task runs in correct location Mel Gorman
2013-10-07 10:29 ` [PATCH 35/63] sched: numa: Do not trap hinting faults for shared libraries Mel Gorman
2013-10-07 19:04   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 36/63] mm: numa: Only trap pmd hinting faults if we would otherwise trap PTE faults Mel Gorman
2013-10-07 19:06   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 37/63] stop_machine: Introduce stop_two_cpus() Mel Gorman
2013-10-07 10:29 ` [PATCH 38/63] sched: Introduce migrate_swap() Mel Gorman
2013-10-07 19:06   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 39/63] sched: numa: Use a system-wide search to find swap/migration candidates Mel Gorman
2013-10-07 19:07   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 40/63] sched: numa: Favor placing a task on the preferred node Mel Gorman
2013-10-07 19:07   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 41/63] sched: numa: fix placement of workloads spread across multiple nodes Mel Gorman
2013-10-07 10:29 ` [PATCH 42/63] mm: numa: Change page last {nid,pid} into {cpu,pid} Mel Gorman
2013-10-07 19:08   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 43/63] sched: numa: Use {cpu, pid} to create task groups for shared faults Mel Gorman
2013-10-07 19:09   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 44/63] sched: numa: Report a NUMA task group ID Mel Gorman
2013-10-07 19:09   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 45/63] mm: numa: copy cpupid on page migration Mel Gorman
2013-10-07 10:29 ` [PATCH 46/63] mm: numa: Do not group on RO pages Mel Gorman
2013-10-07 19:10   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 47/63] mm: numa: Do not batch handle PMD pages Mel Gorman
2013-10-07 19:11   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 48/63] sched: numa: stay on the same node if CLONE_VM Mel Gorman
2013-10-07 10:29 ` [PATCH 49/63] sched: numa: use group fault statistics in numa placement Mel Gorman
2013-10-07 10:29 ` [PATCH 50/63] sched: numa: call task_numa_free from do_execve Mel Gorman
2013-10-07 10:29 ` [PATCH 51/63] sched: numa: Prevent parallel updates to group stats during placement Mel Gorman
2013-10-07 19:13   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 52/63] sched: numa: add debugging Mel Gorman
2013-10-07 19:13   ` Rik van Riel
2013-10-07 10:29 ` Mel Gorman [this message]
2013-10-07 10:29 ` [PATCH 54/63] sched: numa: fix task or group comparison Mel Gorman
2013-10-07 10:29 ` [PATCH 55/63] sched: numa: Avoid migrating tasks that are placed on their preferred node Mel Gorman
2013-10-07 19:14   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 56/63] sched: numa: be more careful about joining numa groups Mel Gorman
2013-10-07 10:29 ` [PATCH 57/63] sched: numa: Take false sharing into account when adapting scan rate Mel Gorman
2013-10-07 19:14   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 58/63] sched: numa: adjust scan rate in task_numa_placement Mel Gorman
2013-10-07 10:29 ` [PATCH 59/63] sched: numa: Remove the numa_balancing_scan_period_reset sysctl Mel Gorman
2013-10-07 19:14   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 60/63] mm: numa: revert temporarily disabling of NUMA migration Mel Gorman
2013-10-07 10:29 ` [PATCH 61/63] sched: numa: skip some page migrations after a shared fault Mel Gorman
2013-10-07 10:29 ` [PATCH 62/63] sched: numa: use unsigned longs for numa group fault stats Mel Gorman
2013-10-07 19:15   ` Rik van Riel
2013-10-07 10:29 ` [PATCH 63/63] sched: numa: periodically retry task_numa_migrate Mel Gorman
2013-10-09 11:03 ` [PATCH 0/63] Basic scheduler support for automatic NUMA balancing V9 Ingo Molnar
2013-10-09 11:11   ` Ingo Molnar
2013-10-09 11:13     ` Ingo Molnar
2013-10-09 12:05   ` Peter Zijlstra
2013-10-09 12:48     ` Ingo Molnar
2013-10-10  7:05   ` Mel Gorman
2013-10-09 16:28 ` Ingo Molnar
2013-10-09 16:29   ` Ingo Molnar
2013-10-09 16:57     ` Ingo Molnar
2013-10-09 17:09       ` Ingo Molnar
2013-10-09 17:11         ` Peter Zijlstra
2013-10-09 17:08   ` Peter Zijlstra
2013-10-09 17:15     ` Ingo Molnar
2013-10-09 17:18       ` Peter Zijlstra
2013-10-24 12:26 ` Automatic NUMA balancing patches for tip-urgent/stable Mel Gorman
2013-10-26 12:11   ` Ingo Molnar
2013-10-29  9:42     ` Mel Gorman
2013-10-29  9:48       ` Ingo Molnar
2013-10-29 10:24         ` Mel Gorman
2013-10-29 10:41           ` Ingo Molnar
2013-10-29 12:48             ` Mel Gorman
2013-10-31  9:51   ` [RFC GIT PULL] NUMA-balancing memory corruption fixes Ingo Molnar
2013-10-31 22:25     ` Linus Torvalds
2013-11-01  7:36       ` Ingo Molnar
  -- strict thread matches above, loose matches on Subject: below --
2013-09-27 13:26 [PATCH 0/63] Basic scheduler support for automatic NUMA balancing V8 Mel Gorman
2013-09-27 13:27 ` [PATCH 53/63] sched: numa: Decide whether to favour task or group weights based on swap candidate relationships 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=1381141781-10992-54-git-send-email-mgorman@suse.de \
    --to=mgorman@suse.de \
    --cc=a.p.zijlstra@chello.nl \
    --cc=aarcange@redhat.com \
    --cc=hannes@cmpxchg.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mingo@kernel.org \
    --cc=riel@redhat.com \
    --cc=srikar@linux.vnet.ibm.com \
    /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

Linux-mm Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-mm/0 linux-mm/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-mm linux-mm/ https://lore.kernel.org/linux-mm \
		linux-mm@kvack.org
	public-inbox-index linux-mm

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kvack.linux-mm


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git