Merge branch 'akpm' (patches from Andrew Morton)

Merge fixes from Andrew Morton:
 "Bunch of fixes.

  And a reversion of mhocko's "Soft limit rework" patch series.  This is
  actually your fault for opening the merge window when I was off racing ;)

  I didn't read the email thread before sending everything off.
  Johannes Weiner raised significant issues:

    http://www.spinics.net/lists/cgroups/msg08813.html

  and we agreed to back it all out"

I clearly need to be more aware of Andrew's racing schedule.

* akpm:
  MAINTAINERS: update mach-bcm related email address
  checkpatch: make extern in .h prototypes quieter
  cciss: fix info leak in cciss_ioctl32_passthru()
  cpqarray: fix info leak in ida_locked_ioctl()
  kernel/reboot.c: re-enable the function of variable reboot_default
  audit: fix endless wait in audit_log_start()
  revert "memcg, vmscan: integrate soft reclaim tighter with zone shrinking code"
  revert "memcg: get rid of soft-limit tree infrastructure"
  revert "vmscan, memcg: do softlimit reclaim also for targeted reclaim"
  revert "memcg: enhance memcg iterator to support predicates"
  revert "memcg: track children in soft limit excess to improve soft limit"
  revert "memcg, vmscan: do not attempt soft limit reclaim if it would not scan anything"
  revert "memcg: track all children over limit in the root"
  revert "memcg, vmscan: do not fall into reclaim-all pass too quickly"
  fs/ocfs2/super.c: use a bigger nodestr in ocfs2_dismount_volume
  watchdog: update watchdog_thresh properly
  watchdog: update watchdog attributes atomically

MAINTAINERS

@@ -1812,7 +1812,8 @@ S:	Supported
 F:	drivers/net/ethernet/broadcom/bnx2x/
 
 BROADCOM BCM281XX/BCM11XXX ARM ARCHITECTURE
-M:	Christian Daudt <csd@broadcom.com>
+M:	Christian Daudt <bcm@fixthebug.org>
+L:	bcm-kernel-feedback-list@broadcom.com
 T:	git git://git.github.com/broadcom/bcm11351
 S:	Maintained
 F:	arch/arm/mach-bcm/

drivers/block/cciss.c

@@ -1189,6 +1189,7 @@ static int cciss_ioctl32_passthru(struct block_device *bdev, fmode_t mode,
 	int err;
 	u32 cp;
 
+	memset(&arg64, 0, sizeof(arg64));
 	err = 0;
 	err |=
 	    copy_from_user(&arg64.LUN_info, &arg32->LUN_info,

drivers/block/cpqarray.c

@@ -1193,6 +1193,7 @@ out_passthru:
 		ida_pci_info_struct pciinfo;
 
 		if (!arg) return -EINVAL;
+		memset(&pciinfo, 0, sizeof(pciinfo));
 		pciinfo.bus = host->pci_dev->bus->number;
 		pciinfo.dev_fn = host->pci_dev->devfn;
 		pciinfo.board_id = host->board_id;

fs/ocfs2/super.c

@@ -1924,7 +1924,7 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err)
 {
 	int tmp, hangup_needed = 0;
 	struct ocfs2_super *osb = NULL;
-	char nodestr[8];
+	char nodestr[12];
 
 	trace_ocfs2_dismount_volume(sb);
 

include/linux/memcontrol.h

@@ -53,23 +53,6 @@ struct mem_cgroup_reclaim_cookie {
 	unsigned int generation;
 };
 
-enum mem_cgroup_filter_t {
-	VISIT,		/* visit current node */
-	SKIP,		/* skip the current node and continue traversal */
-	SKIP_TREE,	/* skip the whole subtree and continue traversal */
-};
-
-/*
- * mem_cgroup_filter_t predicate might instruct mem_cgroup_iter_cond how to
- * iterate through the hierarchy tree. Each tree element is checked by the
- * predicate before it is returned by the iterator. If a filter returns
- * SKIP or SKIP_TREE then the iterator code continues traversal (with the
- * next node down the hierarchy or the next node that doesn't belong under the
- * memcg's subtree).
- */
-typedef enum mem_cgroup_filter_t
-(*mem_cgroup_iter_filter)(struct mem_cgroup *memcg, struct mem_cgroup *root);
-
 #ifdef CONFIG_MEMCG
 /*
  * All "charge" functions with gfp_mask should use GFP_KERNEL or
@@ -137,18 +120,9 @@ mem_cgroup_prepare_migration(struct page *page, struct page *newpage,
 extern void mem_cgroup_end_migration(struct mem_cgroup *memcg,
 	struct page *oldpage, struct page *newpage, bool migration_ok);
 
-struct mem_cgroup *mem_cgroup_iter_cond(struct mem_cgroup *root,
-				   struct mem_cgroup *prev,
-				   struct mem_cgroup_reclaim_cookie *reclaim,
-				   mem_cgroup_iter_filter cond);
-
-static inline struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *root,
-				   struct mem_cgroup *prev,
-				   struct mem_cgroup_reclaim_cookie *reclaim)
-{
-	return mem_cgroup_iter_cond(root, prev, reclaim, NULL);
-}
-
+struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *,
+				   struct mem_cgroup *,
+				   struct mem_cgroup_reclaim_cookie *);
 void mem_cgroup_iter_break(struct mem_cgroup *, struct mem_cgroup *);
 
 /*
@@ -260,9 +234,9 @@ static inline void mem_cgroup_dec_page_stat(struct page *page,
 	mem_cgroup_update_page_stat(page, idx, -1);
 }
 
-enum mem_cgroup_filter_t
-mem_cgroup_soft_reclaim_eligible(struct mem_cgroup *memcg,
-		struct mem_cgroup *root);
+unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order,
+						gfp_t gfp_mask,
+						unsigned long *total_scanned);
 
 void __mem_cgroup_count_vm_event(struct mm_struct *mm, enum vm_event_item idx);
 static inline void mem_cgroup_count_vm_event(struct mm_struct *mm,
@@ -376,15 +350,6 @@ static inline void mem_cgroup_end_migration(struct mem_cgroup *memcg,
 		struct page *oldpage, struct page *newpage, bool migration_ok)
 {
 }
-static inline struct mem_cgroup *
-mem_cgroup_iter_cond(struct mem_cgroup *root,
-		struct mem_cgroup *prev,
-		struct mem_cgroup_reclaim_cookie *reclaim,
-		mem_cgroup_iter_filter cond)
-{
-	/* first call must return non-NULL, second return NULL */
-	return (struct mem_cgroup *)(unsigned long)!prev;
-}
 
 static inline struct mem_cgroup *
 mem_cgroup_iter(struct mem_cgroup *root,
@@ -471,11 +436,11 @@ static inline void mem_cgroup_dec_page_stat(struct page *page,
 }
 
 static inline
-enum mem_cgroup_filter_t
-mem_cgroup_soft_reclaim_eligible(struct mem_cgroup *memcg,
-		struct mem_cgroup *root)
+unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order,
+					    gfp_t gfp_mask,
+					    unsigned long *total_scanned)
 {
-	return VISIT;
+	return 0;
 }
 
 static inline void mem_cgroup_split_huge_fixup(struct page *head)

include/linux/smp.h

@@ -155,6 +155,12 @@ smp_call_function_any(const struct cpumask *mask, smp_call_func_t func,
 
 static inline void kick_all_cpus_sync(void) {  }
 
+static inline void __smp_call_function_single(int cpuid,
+		struct call_single_data *data, int wait)
+{
+	on_each_cpu(data->func, data->info, wait);
+}
+
 #endif /* !SMP */
 
 /*

kernel/audit.c

@@ -1117,9 +1117,10 @@ struct audit_buffer *audit_log_start(struct audit_context *ctx, gfp_t gfp_mask,
 
 			sleep_time = timeout_start + audit_backlog_wait_time -
 					jiffies;
-			if ((long)sleep_time > 0)
+			if ((long)sleep_time > 0) {
 				wait_for_auditd(sleep_time);
-			continue;
+				continue;
+			}
 		}
 		if (audit_rate_check() && printk_ratelimit())
 			printk(KERN_WARNING

kernel/reboot.c

@@ -32,7 +32,14 @@ EXPORT_SYMBOL(cad_pid);
 #endif
 enum reboot_mode reboot_mode DEFAULT_REBOOT_MODE;
 
-int reboot_default;
+/*
+ * This variable is used privately to keep track of whether or not
+ * reboot_type is still set to its default value (i.e., reboot= hasn't
+ * been set on the command line).  This is needed so that we can
+ * suppress DMI scanning for reboot quirks.  Without it, it's
+ * impossible to override a faulty reboot quirk without recompiling.
+ */
+int reboot_default = 1;
 int reboot_cpu;
 enum reboot_type reboot_type = BOOT_ACPI;
 int reboot_force;

kernel/watchdog.c

@@ -486,7 +486,52 @@ static struct smp_hotplug_thread watchdog_threads = {
 	.unpark			= watchdog_enable,
 };
 
-static int watchdog_enable_all_cpus(void)
+static void restart_watchdog_hrtimer(void *info)
+{
+	struct hrtimer *hrtimer = &__raw_get_cpu_var(watchdog_hrtimer);
+	int ret;
+
+	/*
+	 * No need to cancel and restart hrtimer if it is currently executing
+	 * because it will reprogram itself with the new period now.
+	 * We should never see it unqueued here because we are running per-cpu
+	 * with interrupts disabled.
+	 */
+	ret = hrtimer_try_to_cancel(hrtimer);
+	if (ret == 1)
+		hrtimer_start(hrtimer, ns_to_ktime(sample_period),
+				HRTIMER_MODE_REL_PINNED);
+}
+
+static void update_timers(int cpu)
+{
+	struct call_single_data data = {.func = restart_watchdog_hrtimer};
+	/*
+	 * Make sure that perf event counter will adopt to a new
+	 * sampling period. Updating the sampling period directly would
+	 * be much nicer but we do not have an API for that now so
+	 * let's use a big hammer.
+	 * Hrtimer will adopt the new period on the next tick but this
+	 * might be late already so we have to restart the timer as well.
+	 */
+	watchdog_nmi_disable(cpu);
+	__smp_call_function_single(cpu, &data, 1);
+	watchdog_nmi_enable(cpu);
+}
+
+static void update_timers_all_cpus(void)
+{
+	int cpu;
+
+	get_online_cpus();
+	preempt_disable();
+	for_each_online_cpu(cpu)
+		update_timers(cpu);
+	preempt_enable();
+	put_online_cpus();
+}
+
+static int watchdog_enable_all_cpus(bool sample_period_changed)
 {
 	int err = 0;
 
@@ -496,6 +541,8 @@ static int watchdog_enable_all_cpus(void)
 			pr_err("Failed to create watchdog threads, disabled\n");
 		else
 			watchdog_running = 1;
+	} else if (sample_period_changed) {
+		update_timers_all_cpus();
 	}
 
 	return err;
@@ -520,13 +567,15 @@ int proc_dowatchdog(struct ctl_table *table, int write,
 		    void __user *buffer, size_t *lenp, loff_t *ppos)
 {
 	int err, old_thresh, old_enabled;
+	static DEFINE_MUTEX(watchdog_proc_mutex);
 
+	mutex_lock(&watchdog_proc_mutex);
 	old_thresh = ACCESS_ONCE(watchdog_thresh);
 	old_enabled = ACCESS_ONCE(watchdog_user_enabled);
 
 	err = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
 	if (err || !write)
-		return err;
+		goto out;
 
 	set_sample_period();
 	/*
@@ -535,7 +584,7 @@ int proc_dowatchdog(struct ctl_table *table, int write,
 	 * watchdog_*_all_cpus() function takes care of this.
 	 */
 	if (watchdog_user_enabled && watchdog_thresh)
-		err = watchdog_enable_all_cpus();
+		err = watchdog_enable_all_cpus(old_thresh != watchdog_thresh);
 	else
 		watchdog_disable_all_cpus();
 
@@ -544,7 +593,8 @@ int proc_dowatchdog(struct ctl_table *table, int write,
 		watchdog_thresh = old_thresh;
 		watchdog_user_enabled = old_enabled;
 	}
-
+out:
+	mutex_unlock(&watchdog_proc_mutex);
 	return err;
 }
 #endif /* CONFIG_SYSCTL */
@@ -554,5 +604,5 @@ void __init lockup_detector_init(void)
 	set_sample_period();
 
 	if (watchdog_user_enabled)
-		watchdog_enable_all_cpus();
+		watchdog_enable_all_cpus(false);
 }

mm/memcontrol.c

@@ -39,6 +39,7 @@
 #include <linux/limits.h>
 #include <linux/export.h>
 #include <linux/mutex.h>
+#include <linux/rbtree.h>
 #include <linux/slab.h>
 #include <linux/swap.h>
 #include <linux/swapops.h>
@@ -160,6 +161,10 @@ struct mem_cgroup_per_zone {
 
 	struct mem_cgroup_reclaim_iter reclaim_iter[DEF_PRIORITY + 1];
 
+	struct rb_node		tree_node;	/* RB tree node */
+	unsigned long long	usage_in_excess;/* Set to the value by which */
+						/* the soft limit is exceeded*/
+	bool			on_tree;
 	struct mem_cgroup	*memcg;		/* Back pointer, we cannot */
 						/* use container_of	   */
 };
@@ -168,6 +173,26 @@ struct mem_cgroup_per_node {
 	struct mem_cgroup_per_zone zoneinfo[MAX_NR_ZONES];
 };
 
+/*
+ * Cgroups above their limits are maintained in a RB-Tree, independent of
+ * their hierarchy representation
+ */
+
+struct mem_cgroup_tree_per_zone {
+	struct rb_root rb_root;
+	spinlock_t lock;
+};
+
+struct mem_cgroup_tree_per_node {
+	struct mem_cgroup_tree_per_zone rb_tree_per_zone[MAX_NR_ZONES];
+};
+
+struct mem_cgroup_tree {
+	struct mem_cgroup_tree_per_node *rb_tree_per_node[MAX_NUMNODES];
+};
+
+static struct mem_cgroup_tree soft_limit_tree __read_mostly;
+
 struct mem_cgroup_threshold {
 	struct eventfd_ctx *eventfd;
 	u64 threshold;
@@ -303,22 +328,6 @@ struct mem_cgroup {
 	atomic_t	numainfo_events;
 	atomic_t	numainfo_updating;
 #endif
-	/*
-	 * Protects soft_contributed transitions.
-	 * See mem_cgroup_update_soft_limit
-	 */
-	spinlock_t soft_lock;
-
-	/*
-	 * If true then this group has increased parents' children_in_excess
-	 * when it got over the soft limit.
-	 * When a group falls bellow the soft limit, parents' children_in_excess
-	 * is decreased and soft_contributed changed to false.
-	 */
-	bool soft_contributed;
-
-	/* Number of children that are in soft limit excess */
-	atomic_t children_in_excess;
 
 	struct mem_cgroup_per_node *nodeinfo[0];
 	/* WARNING: nodeinfo must be the last member here */
@@ -422,6 +431,7 @@ static bool move_file(void)
  * limit reclaim to prevent infinite loops, if they ever occur.
  */
 #define	MEM_CGROUP_MAX_RECLAIM_LOOPS		100
+#define	MEM_CGROUP_MAX_SOFT_LIMIT_RECLAIM_LOOPS	2
 
 enum charge_type {
 	MEM_CGROUP_CHARGE_TYPE_CACHE = 0,
@@ -648,6 +658,164 @@ page_cgroup_zoneinfo(struct mem_cgroup *memcg, struct page *page)
 	return mem_cgroup_zoneinfo(memcg, nid, zid);
 }
 
+static struct mem_cgroup_tree_per_zone *
+soft_limit_tree_node_zone(int nid, int zid)
+{
+	return &soft_limit_tree.rb_tree_per_node[nid]->rb_tree_per_zone[zid];
+}
+
+static struct mem_cgroup_tree_per_zone *
+soft_limit_tree_from_page(struct page *page)
+{
+	int nid = page_to_nid(page);
+	int zid = page_zonenum(page);
+
+	return &soft_limit_tree.rb_tree_per_node[nid]->rb_tree_per_zone[zid];
+}
+
+static void
+__mem_cgroup_insert_exceeded(struct mem_cgroup *memcg,
+				struct mem_cgroup_per_zone *mz,
+				struct mem_cgroup_tree_per_zone *mctz,
+				unsigned long long new_usage_in_excess)
+{
+	struct rb_node **p = &mctz->rb_root.rb_node;
+	struct rb_node *parent = NULL;
+	struct mem_cgroup_per_zone *mz_node;
+
+	if (mz->on_tree)
+		return;
+
+	mz->usage_in_excess = new_usage_in_excess;
+	if (!mz->usage_in_excess)
+		return;
+	while (*p) {
+		parent = *p;
+		mz_node = rb_entry(parent, struct mem_cgroup_per_zone,
+					tree_node);
+		if (mz->usage_in_excess < mz_node->usage_in_excess)
+			p = &(*p)->rb_left;
+		/*
+		 * We can't avoid mem cgroups that are over their soft
+		 * limit by the same amount
+		 */
+		else if (mz->usage_in_excess >= mz_node->usage_in_excess)
+			p = &(*p)->rb_right;
+	}
+	rb_link_node(&mz->tree_node, parent, p);
+	rb_insert_color(&mz->tree_node, &mctz->rb_root);
+	mz->on_tree = true;
+}
+
+static void
+__mem_cgroup_remove_exceeded(struct mem_cgroup *memcg,
+				struct mem_cgroup_per_zone *mz,
+				struct mem_cgroup_tree_per_zone *mctz)
+{
+	if (!mz->on_tree)
+		return;
+	rb_erase(&mz->tree_node, &mctz->rb_root);
+	mz->on_tree = false;
+}
+
+static void
+mem_cgroup_remove_exceeded(struct mem_cgroup *memcg,
+				struct mem_cgroup_per_zone *mz,
+				struct mem_cgroup_tree_per_zone *mctz)
+{
+	spin_lock(&mctz->lock);
+	__mem_cgroup_remove_exceeded(memcg, mz, mctz);
+	spin_unlock(&mctz->lock);
+}
+
+
+static void mem_cgroup_update_tree(struct mem_cgroup *memcg, struct page *page)
+{
+	unsigned long long excess;
+	struct mem_cgroup_per_zone *mz;
+	struct mem_cgroup_tree_per_zone *mctz;
+	int nid = page_to_nid(page);
+	int zid = page_zonenum(page);
+	mctz = soft_limit_tree_from_page(page);
+
+	/*
+	 * Necessary to update all ancestors when hierarchy is used.
+	 * because their event counter is not touched.
+	 */
+	for (; memcg; memcg = parent_mem_cgroup(memcg)) {
+		mz = mem_cgroup_zoneinfo(memcg, nid, zid);
+		excess = res_counter_soft_limit_excess(&memcg->res);
+		/*
+		 * We have to update the tree if mz is on RB-tree or
+		 * mem is over its softlimit.
+		 */
+		if (excess || mz->on_tree) {
+			spin_lock(&mctz->lock);
+			/* if on-tree, remove it */
+			if (mz->on_tree)
+				__mem_cgroup_remove_exceeded(memcg, mz, mctz);
+			/*
+			 * Insert again. mz->usage_in_excess will be updated.
+			 * If excess is 0, no tree ops.
+			 */
+			__mem_cgroup_insert_exceeded(memcg, mz, mctz, excess);
+			spin_unlock(&mctz->lock);
+		}
+	}
+}
+
+static void mem_cgroup_remove_from_trees(struct mem_cgroup *memcg)
+{
+	int node, zone;
+	struct mem_cgroup_per_zone *mz;
+	struct mem_cgroup_tree_per_zone *mctz;
+
+	for_each_node(node) {
+		for (zone = 0; zone < MAX_NR_ZONES; zone++) {
+			mz = mem_cgroup_zoneinfo(memcg, node, zone);
+			mctz = soft_limit_tree_node_zone(node, zone);
+			mem_cgroup_remove_exceeded(memcg, mz, mctz);
+		}
+	}
+}
+
+static struct mem_cgroup_per_zone *
+__mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_zone *mctz)
+{
+	struct rb_node *rightmost = NULL;
+	struct mem_cgroup_per_zone *mz;
+
+retry:
+	mz = NULL;
+	rightmost = rb_last(&mctz->rb_root);
+	if (!rightmost)
+		goto done;		/* Nothing to reclaim from */
+
+	mz = rb_entry(rightmost, struct mem_cgroup_per_zone, tree_node);
+	/*
+	 * Remove the node now but someone else can add it back,
+	 * we will to add it back at the end of reclaim to its correct
+	 * position in the tree.
+	 */
+	__mem_cgroup_remove_exceeded(mz->memcg, mz, mctz);
+	if (!res_counter_soft_limit_excess(&mz->memcg->res) ||
+		!css_tryget(&mz->memcg->css))
+		goto retry;
+done:
+	return mz;
+}
+
+static struct mem_cgroup_per_zone *
+mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_zone *mctz)
+{
+	struct mem_cgroup_per_zone *mz;
+
+	spin_lock(&mctz->lock);
+	mz = __mem_cgroup_largest_soft_limit_node(mctz);
+	spin_unlock(&mctz->lock);
+	return mz;
+}
+
 /*
  * Implementation Note: reading percpu statistics for memcg.
  *
@@ -821,48 +989,6 @@ static bool mem_cgroup_event_ratelimit(struct mem_cgroup *memcg,
 	return false;
 }
 
-/*
- * Called from rate-limited memcg_check_events when enough
- * MEM_CGROUP_TARGET_SOFTLIMIT events are accumulated and it makes sure
- * that all the parents up the hierarchy will be notified that this group
- * is in excess or that it is not in excess anymore. mmecg->soft_contributed
- * makes the transition a single action whenever the state flips from one to
- * the other.
- */
-static void mem_cgroup_update_soft_limit(struct mem_cgroup *memcg)
-{
-	unsigned long long excess = res_counter_soft_limit_excess(&memcg->res);
-	struct mem_cgroup *parent = memcg;
-	int delta = 0;
-
-	spin_lock(&memcg->soft_lock);
-	if (excess) {
-		if (!memcg->soft_contributed) {
-			delta = 1;
-			memcg->soft_contributed = true;
-		}
-	} else {
-		if (memcg->soft_contributed) {
-			delta = -1;
-			memcg->soft_contributed = false;
-		}
-	}
-
-	/*
-	 * Necessary to update all ancestors when hierarchy is used
-	 * because their event counter is not touched.
-	 * We track children even outside the hierarchy for the root
-	 * cgroup because tree walk starting at root should visit
-	 * all cgroups and we want to prevent from pointless tree
-	 * walk if no children is below the limit.
-	 */
-	while (delta && (parent = parent_mem_cgroup(parent)))
-		atomic_add(delta, &parent->children_in_excess);
-	if (memcg != root_mem_cgroup && !root_mem_cgroup->use_hierarchy)
-		atomic_add(delta, &root_mem_cgroup->children_in_excess);
-	spin_unlock(&memcg->soft_lock);
-}
-
 /*
  * Check events in order.
  *
@@ -886,7 +1012,7 @@ static void memcg_check_events(struct mem_cgroup *memcg, struct page *page)
 
 		mem_cgroup_threshold(memcg);
 		if (unlikely(do_softlimit))
-			mem_cgroup_update_soft_limit(memcg);
+			mem_cgroup_update_tree(memcg, page);
 #if MAX_NUMNODES > 1
 		if (unlikely(do_numainfo))
 			atomic_inc(&memcg->numainfo_events);
@@ -929,15 +1055,6 @@ struct mem_cgroup *try_get_mem_cgroup_from_mm(struct mm_struct *mm)
 	return memcg;
 }
 
-static enum mem_cgroup_filter_t
-mem_cgroup_filter(struct mem_cgroup *memcg, struct mem_cgroup *root,
-		mem_cgroup_iter_filter cond)
-{
-	if (!cond)
-		return VISIT;
-	return cond(memcg, root);
-}
-
 /*
  * Returns a next (in a pre-order walk) alive memcg (with elevated css
  * ref. count) or NULL if the whole root's subtree has been visited.
@@ -945,7 +1062,7 @@ mem_cgroup_filter(struct mem_cgroup *memcg, struct mem_cgroup *root,
  * helper function to be used by mem_cgroup_iter
  */
 static struct mem_cgroup *__mem_cgroup_iter_next(struct mem_cgroup *root,
-		struct mem_cgroup *last_visited, mem_cgroup_iter_filter cond)
+		struct mem_cgroup *last_visited)
 {
 	struct cgroup_subsys_state *prev_css, *next_css;
 
@@ -963,31 +1080,11 @@ skip_node:
 	if (next_css) {
 		struct mem_cgroup *mem = mem_cgroup_from_css(next_css);
 
-		switch (mem_cgroup_filter(mem, root, cond)) {
-		case SKIP:
+		if (css_tryget(&mem->css))
+			return mem;
+		else {
 			prev_css = next_css;
 			goto skip_node;
-		case SKIP_TREE:
-			if (mem == root)
-				return NULL;
-			/*
-			 * css_rightmost_descendant is not an optimal way to
-			 * skip through a subtree (especially for imbalanced
-			 * trees leaning to right) but that's what we have right
-			 * now. More effective solution would be traversing
-			 * right-up for first non-NULL without calling
-			 * css_next_descendant_pre afterwards.
-			 */
-			prev_css = css_rightmost_descendant(next_css);
-			goto skip_node;
-		case VISIT:
-			if (css_tryget(&mem->css))
-				return mem;
-			else {
-				prev_css = next_css;
-				goto skip_node;
-			}
-			break;
 		}
 	}
 
@@ -1051,7 +1148,6 @@ static void mem_cgroup_iter_update(struct mem_cgroup_reclaim_iter *iter,
  * @root: hierarchy root
  * @prev: previously returned memcg, NULL on first invocation
  * @reclaim: cookie for shared reclaim walks, NULL for full walks
- * @cond: filter for visited nodes, NULL for no filter
  *
  * Returns references to children of the hierarchy below @root, or
  * @root itself, or %NULL after a full round-trip.
@@ -1064,18 +1160,15 @@ static void mem_cgroup_iter_update(struct mem_cgroup_reclaim_iter *iter,
  * divide up the memcgs in the hierarchy among all concurrent
  * reclaimers operating on the same zone and priority.
  */
-struct mem_cgroup *mem_cgroup_iter_cond(struct mem_cgroup *root,
+struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *root,
 				   struct mem_cgroup *prev,
-				   struct mem_cgroup_reclaim_cookie *reclaim,
-				   mem_cgroup_iter_filter cond)
+				   struct mem_cgroup_reclaim_cookie *reclaim)
 {
 	struct mem_cgroup *memcg = NULL;
 	struct mem_cgroup *last_visited = NULL;
 
-	if (mem_cgroup_disabled()) {
-		/* first call must return non-NULL, second return NULL */
-		return (struct mem_cgroup *)(unsigned long)!prev;
-	}
+	if (mem_cgroup_disabled())
+		return NULL;
 
 	if (!root)
 		root = root_mem_cgroup;
@@ -1086,9 +1179,7 @@ struct mem_cgroup *mem_cgroup_iter_cond(struct mem_cgroup *root,
 	if (!root->use_hierarchy && root != root_mem_cgroup) {
 		if (prev)
 			goto out_css_put;
-		if (mem_cgroup_filter(root, root, cond) == VISIT)
-			return root;
-		return NULL;
+		return root;
 	}
 
 	rcu_read_lock();
@@ -1111,7 +1202,7 @@ struct mem_cgroup *mem_cgroup_iter_cond(struct mem_cgroup *root,
 			last_visited = mem_cgroup_iter_load(iter, root, &seq);
 		}
 
-		memcg = __mem_cgroup_iter_next(root, last_visited, cond);
+		memcg = __mem_cgroup_iter_next(root, last_visited);
 
 		if (reclaim) {
 			mem_cgroup_iter_update(iter, last_visited, memcg, seq);
@@ -1122,11 +1213,7 @@ struct mem_cgroup *mem_cgroup_iter_cond(struct mem_cgroup *root,
 				reclaim->generation = iter->generation;
 		}
 
-		/*
-		 * We have finished the whole tree walk or no group has been
-		 * visited because filter told us to skip the root node.
-		 */
-		if (!memcg && (prev || (cond && !last_visited)))
+		if (prev && !memcg)
 			goto out_unlock;
 	}
 out_unlock:
@@ -1767,7 +1854,6 @@ static unsigned long mem_cgroup_reclaim(struct mem_cgroup *memcg,
 	return total;
 }
 
-#if MAX_NUMNODES > 1
 /**
  * test_mem_cgroup_node_reclaimable
  * @memcg: the target memcg
@@ -1790,6 +1876,7 @@ static bool test_mem_cgroup_node_reclaimable(struct mem_cgroup *memcg,
 	return false;
 
 }
+#if MAX_NUMNODES > 1
 
 /*
  * Always updating the nodemask is not very good - even if we have an empty
@@ -1857,50 +1944,104 @@ int mem_cgroup_select_victim_node(struct mem_cgroup *memcg)
 	return node;
 }
 
+/*
+ * Check all nodes whether it contains reclaimable pages or not.
+ * For quick scan, we make use of scan_nodes. This will allow us to skip
+ * unused nodes. But scan_nodes is lazily updated and may not cotain
+ * enough new information. We need to do double check.
+ */
+static bool mem_cgroup_reclaimable(struct mem_cgroup *memcg, bool noswap)
+{
+	int nid;
+
+	/*
+	 * quick check...making use of scan_node.
+	 * We can skip unused nodes.
+	 */
+	if (!nodes_empty(memcg->scan_nodes)) {
+		for (nid = first_node(memcg->scan_nodes);
+		     nid < MAX_NUMNODES;
+		     nid = next_node(nid, memcg->scan_nodes)) {
+
+			if (test_mem_cgroup_node_reclaimable(memcg, nid, noswap))
+				return true;
+		}
+	}
+	/*
+	 * Check rest of nodes.
+	 */
+	for_each_node_state(nid, N_MEMORY) {
+		if (node_isset(nid, memcg->scan_nodes))
+			continue;
+		if (test_mem_cgroup_node_reclaimable(memcg, nid, noswap))
+			return true;
+	}
+	return false;
+}
+
 #else
 int mem_cgroup_select_victim_node(struct mem_cgroup *memcg)
 {
 	return 0;
 }
 
-#endif
-
-/*
- * A group is eligible for the soft limit reclaim under the given root
- * hierarchy if
- *	a) it is over its soft limit
- *	b) any parent up the hierarchy is over its soft limit
- *
- * If the given group doesn't have any children over the limit then it
- * doesn't make any sense to iterate its subtree.
- */
-enum mem_cgroup_filter_t
-mem_cgroup_soft_reclaim_eligible(struct mem_cgroup *memcg,
-		struct mem_cgroup *root)
+static bool mem_cgroup_reclaimable(struct mem_cgroup *memcg, bool noswap)
 {
-	struct mem_cgroup *parent;
-
-	if (!memcg)
-		memcg = root_mem_cgroup;
-	parent = memcg;
-
-	if (res_counter_soft_limit_excess(&memcg->res))
-		return VISIT;
+	return test_mem_cgroup_node_reclaimable(memcg, 0, noswap);
+}
+#endif
 
-	/*
-	 * If any parent up to the root in the hierarchy is over its soft limit
-	 * then we have to obey and reclaim from this group as well.
-	 */
-	while ((parent = parent_mem_cgroup(parent))) {
-		if (res_counter_soft_limit_excess(&parent->res))
-			return VISIT;
-		if (parent == root)
+static int mem_cgroup_soft_reclaim(struct mem_cgroup *root_memcg,
+				   struct zone *zone,
+				   gfp_t gfp_mask,
+				   unsigned long *total_scanned)
+{
+	struct mem_cgroup *victim = NULL;
+	int total = 0;
+	int loop = 0;
+	unsigned long excess;
+	unsigned long nr_scanned;
+	struct mem_cgroup_reclaim_cookie reclaim = {
+		.zone = zone,
+		.priority = 0,
+	};
+
+	excess = res_counter_soft_limit_excess(&root_memcg->res) >> PAGE_SHIFT;
+
+	while (1) {
+		victim = mem_cgroup_iter(root_memcg, victim, &reclaim);
+		if (!victim) {
+			loop++;
+			if (loop >= 2) {
+				/*
+				 * If we have not been able to reclaim
+				 * anything, it might because there are
+				 * no reclaimable pages under this hierarchy
+				 */
+				if (!total)
+					break;
+				/*
+				 * We want to do more targeted reclaim.
+				 * excess >> 2 is not to excessive so as to
+				 * reclaim too much, nor too less that we keep
+				 * coming back to reclaim from this cgroup
+				 */
+				if (total >= (excess >> 2) ||
+					(loop > MEM_CGROUP_MAX_RECLAIM_LOOPS))
+					break;
+			}
+			continue;
+		}
+		if (!mem_cgroup_reclaimable(victim, false))
+			continue;
+		total += mem_cgroup_shrink_node_zone(victim, gfp_mask, false,
+						     zone, &nr_scanned);
+		*total_scanned += nr_scanned;
+		if (!res_counter_soft_limit_excess(&root_memcg->res))
 			break;
 	}
-
-	if (!atomic_read(&memcg->children_in_excess))
-		return SKIP_TREE;
-	return SKIP;
+	mem_cgroup_iter_break(root_memcg, victim);
+	return total;
 }
 
 static DEFINE_SPINLOCK(memcg_oom_lock);
@@ -2812,7 +2953,9 @@ static void __mem_cgroup_commit_charge(struct mem_cgroup *memcg,
 	unlock_page_cgroup(pc);
 
 	/*
-	 * "charge_statistics" updated event counter.
+	 * "charge_statistics" updated event counter. Then, check it.
+	 * Insert ancestor (and ancestor's ancestors), to softlimit RB-tree.
+	 * if they exceeds softlimit.
 	 */
 	memcg_check_events(memcg, page);
 }
@@ -4647,6 +4790,98 @@ static int mem_cgroup_resize_memsw_limit(struct mem_cgroup *memcg,
 	return ret;
 }
 
+unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order,
+					    gfp_t gfp_mask,
+					    unsigned long *total_scanned)
+{
+	unsigned long nr_reclaimed = 0;
+	struct mem_cgroup_per_zone *mz, *next_mz = NULL;
+	unsigned long reclaimed;
+	int loop = 0;
+	struct mem_cgroup_tree_per_zone *mctz;
+	unsigned long long excess;
+	unsigned long nr_scanned;
+
+	if (order > 0)
+		return 0;
+
+	mctz = soft_limit_tree_node_zone(zone_to_nid(zone), zone_idx(zone));
+	/*
+	 * This loop can run a while, specially if mem_cgroup's continuously
+	 * keep exceeding their soft limit and putting the system under
+	 * pressure
+	 */
+	do {
+		if (next_mz)
+			mz = next_mz;
+		else
+			mz = mem_cgroup_largest_soft_limit_node(mctz);
+		if (!mz)
+			break;
+
+		nr_scanned = 0;
+		reclaimed = mem_cgroup_soft_reclaim(mz->memcg, zone,
+						    gfp_mask, &nr_scanned);
+		nr_reclaimed += reclaimed;
+		*total_scanned += nr_scanned;
+		spin_lock(&mctz->lock);
+
+		/*
+		 * If we failed to reclaim anything from this memory cgroup
+		 * it is time to move on to the next cgroup
+		 */
+		next_mz = NULL;
+		if (!reclaimed) {
+			do {
+				/*
+				 * Loop until we find yet another one.
+				 *
+				 * By the time we get the soft_limit lock
+				 * again, someone might have aded the
+				 * group back on the RB tree. Iterate to
+				 * make sure we get a different mem.
+				 * mem_cgroup_largest_soft_limit_node returns
+				 * NULL if no other cgroup is present on
+				 * the tree
+				 */
+				next_mz =
+				__mem_cgroup_largest_soft_limit_node(mctz);
+				if (next_mz == mz)
+					css_put(&next_mz->memcg->css);
+				else /* next_mz == NULL or other memcg */
+					break;
+			} while (1);
+		}
+		__mem_cgroup_remove_exceeded(mz->memcg, mz, mctz);
+		excess = res_counter_soft_limit_excess(&mz->memcg->res);
+		/*
+		 * One school of thought says that we should not add
+		 * back the node to the tree if reclaim returns 0.
+		 * But our reclaim could return 0, simply because due
+		 * to priority we are exposing a smaller subset of
+		 * memory to reclaim from. Consider this as a longer
+		 * term TODO.
+		 */
+		/* If excess == 0, no tree ops */
+		__mem_cgroup_insert_exceeded(mz->memcg, mz, mctz, excess);
+		spin_unlock(&mctz->lock);
+		css_put(&mz->memcg->css);
+		loop++;
+		/*
+		 * Could not reclaim anything and there are no more
+		 * mem cgroups to try or we seem to be looping without
+		 * reclaiming anything.
+		 */
+		if (!nr_reclaimed &&
+			(next_mz == NULL ||
+			loop > MEM_CGROUP_MAX_SOFT_LIMIT_RECLAIM_LOOPS))
+			break;
+	} while (!nr_reclaimed);
+	if (next_mz)
+		css_put(&next_mz->memcg->css);
+	return nr_reclaimed;
+}
+
 /**
  * mem_cgroup_force_empty_list - clears LRU of a group
  * @memcg: group to clear
@@ -5911,6 +6146,8 @@ static int alloc_mem_cgroup_per_zone_info(struct mem_cgroup *memcg, int node)
 	for (zone = 0; zone < MAX_NR_ZONES; zone++) {
 		mz = &pn->zoneinfo[zone];
 		lruvec_init(&mz->lruvec);
+		mz->usage_in_excess = 0;
+		mz->on_tree = false;
 		mz->memcg = memcg;
 	}
 	memcg->nodeinfo[node] = pn;
@@ -5966,6 +6203,7 @@ static void __mem_cgroup_free(struct mem_cgroup *memcg)
 	int node;
 	size_t size = memcg_size();
 
+	mem_cgroup_remove_from_trees(memcg);
 	free_css_id(&mem_cgroup_subsys, &memcg->css);
 
 	for_each_node(node)
@@ -6002,6 +6240,29 @@ struct mem_cgroup *parent_mem_cgroup(struct mem_cgroup *memcg)
 }
 EXPORT_SYMBOL(parent_mem_cgroup);
 
+static void __init mem_cgroup_soft_limit_tree_init(void)
+{
+	struct mem_cgroup_tree_per_node *rtpn;
+	struct mem_cgroup_tree_per_zone *rtpz;
+	int tmp, node, zone;
+
+	for_each_node(node) {
+		tmp = node;
+		if (!node_state(node, N_NORMAL_MEMORY))
+			tmp = -1;
+		rtpn = kzalloc_node(sizeof(*rtpn), GFP_KERNEL, tmp);
+		BUG_ON(!rtpn);
+
+		soft_limit_tree.rb_tree_per_node[node] = rtpn;
+
+		for (zone = 0; zone < MAX_NR_ZONES; zone++) {
+			rtpz = &rtpn->rb_tree_per_zone[zone];
+			rtpz->rb_root = RB_ROOT;
+			spin_lock_init(&rtpz->lock);
+		}
+	}
+}
+
 static struct cgroup_subsys_state * __ref
 mem_cgroup_css_alloc(struct cgroup_subsys_state *parent_css)
 {
@@ -6031,7 +6292,6 @@ mem_cgroup_css_alloc(struct cgroup_subsys_state *parent_css)
 	mutex_init(&memcg->thresholds_lock);
 	spin_lock_init(&memcg->move_lock);
 	vmpressure_init(&memcg->vmpressure);
-	spin_lock_init(&memcg->soft_lock);
 
 	return &memcg->css;
 
@@ -6109,13 +6369,6 @@ static void mem_cgroup_css_offline(struct cgroup_subsys_state *css)
 
 	mem_cgroup_invalidate_reclaim_iterators(memcg);
 	mem_cgroup_reparent_charges(memcg);
-	if (memcg->soft_contributed) {
-		while ((memcg = parent_mem_cgroup(memcg)))
-			atomic_dec(&memcg->children_in_excess);
-
-		if (memcg != root_mem_cgroup && !root_mem_cgroup->use_hierarchy)
-			atomic_dec(&root_mem_cgroup->children_in_excess);
-	}
 	mem_cgroup_destroy_all_caches(memcg);
 	vmpressure_cleanup(&memcg->vmpressure);
 }
@@ -6790,6 +7043,7 @@ static int __init mem_cgroup_init(void)
 {
 	hotcpu_notifier(memcg_cpu_hotplug_callback, 0);
 	enable_swap_cgroup();
+	mem_cgroup_soft_limit_tree_init();
 	memcg_stock_init();
 	return 0;
 }

mm/vmscan.c

@@ -139,23 +139,11 @@ static bool global_reclaim(struct scan_control *sc)
 {
 	return !sc->target_mem_cgroup;
 }
-
-static bool mem_cgroup_should_soft_reclaim(struct scan_control *sc)
-{
-	struct mem_cgroup *root = sc->target_mem_cgroup;
-	return !mem_cgroup_disabled() &&
-		mem_cgroup_soft_reclaim_eligible(root, root) != SKIP_TREE;
-}
 #else
 static bool global_reclaim(struct scan_control *sc)
 {
 	return true;
 }
-
-static bool mem_cgroup_should_soft_reclaim(struct scan_control *sc)
-{
-	return false;
-}
 #endif
 
 unsigned long zone_reclaimable_pages(struct zone *zone)
@@ -2176,11 +2164,9 @@ static inline bool should_continue_reclaim(struct zone *zone,
 	}
 }
 
-static int
-__shrink_zone(struct zone *zone, struct scan_control *sc, bool soft_reclaim)
+static void shrink_zone(struct zone *zone, struct scan_control *sc)
 {
 	unsigned long nr_reclaimed, nr_scanned;
-	int groups_scanned = 0;
 
 	do {
 		struct mem_cgroup *root = sc->target_mem_cgroup;
@@ -2188,17 +2174,15 @@ __shrink_zone(struct zone *zone, struct scan_control *sc, bool soft_reclaim)
 			.zone = zone,
 			.priority = sc->priority,
 		};
-		struct mem_cgroup *memcg = NULL;
-		mem_cgroup_iter_filter filter = (soft_reclaim) ?
-			mem_cgroup_soft_reclaim_eligible : NULL;
+		struct mem_cgroup *memcg;
 
 		nr_reclaimed = sc->nr_reclaimed;
 		nr_scanned = sc->nr_scanned;
 
-		while ((memcg = mem_cgroup_iter_cond(root, memcg, &reclaim, filter))) {
+		memcg = mem_cgroup_iter(root, NULL, &reclaim);
+		do {
 			struct lruvec *lruvec;
 
-			groups_scanned++;
 			lruvec = mem_cgroup_zone_lruvec(zone, memcg);
 
 			shrink_lruvec(lruvec, sc);
@@ -2218,7 +2202,8 @@ __shrink_zone(struct zone *zone, struct scan_control *sc, bool soft_reclaim)
 				mem_cgroup_iter_break(root, memcg);
 				break;
 			}
-		}
+			memcg = mem_cgroup_iter(root, memcg, &reclaim);
+		} while (memcg);
 
 		vmpressure(sc->gfp_mask, sc->target_mem_cgroup,
 			   sc->nr_scanned - nr_scanned,
@@ -2226,37 +2211,6 @@ __shrink_zone(struct zone *zone, struct scan_control *sc, bool soft_reclaim)
 
 	} while (should_continue_reclaim(zone, sc->nr_reclaimed - nr_reclaimed,
 					 sc->nr_scanned - nr_scanned, sc));
-
-	return groups_scanned;
-}
-
-
-static void shrink_zone(struct zone *zone, struct scan_control *sc)
-{
-	bool do_soft_reclaim = mem_cgroup_should_soft_reclaim(sc);
-	unsigned long nr_scanned = sc->nr_scanned;
-	int scanned_groups;
-
-	scanned_groups = __shrink_zone(zone, sc, do_soft_reclaim);
-	/*
-	 * memcg iterator might race with other reclaimer or start from
-	 * a incomplete tree walk so the tree walk in __shrink_zone
-	 * might have missed groups that are above the soft limit. Try
-	 * another loop to catch up with others. Do it just once to
-	 * prevent from reclaim latencies when other reclaimers always
-	 * preempt this one.
-	 */
-	if (do_soft_reclaim && !scanned_groups)
-		__shrink_zone(zone, sc, do_soft_reclaim);
-
-	/*
-	 * No group is over the soft limit or those that are do not have
-	 * pages in the zone we are reclaiming so we have to reclaim everybody
-	 */
-	if (do_soft_reclaim && (sc->nr_scanned == nr_scanned)) {
-		__shrink_zone(zone, sc, false);
-		return;
-	}
 }
 
 /* Returns true if compaction should go ahead for a high-order request */
@@ -2320,6 +2274,8 @@ static bool shrink_zones(struct zonelist *zonelist, struct scan_control *sc)
 {
 	struct zoneref *z;
 	struct zone *zone;
+	unsigned long nr_soft_reclaimed;
+	unsigned long nr_soft_scanned;
 	bool aborted_reclaim = false;
 
 	/*
@@ -2359,6 +2315,18 @@ static bool shrink_zones(struct zonelist *zonelist, struct scan_control *sc)
 					continue;
 				}
 			}
+			/*
+			 * This steals pages from memory cgroups over softlimit
+			 * and returns the number of reclaimed pages and
+			 * scanned pages. This works for global memory pressure
+			 * and balancing, not for a memcg's limit.
+			 */
+			nr_soft_scanned = 0;
+			nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone,
+						sc->order, sc->gfp_mask,
+						&nr_soft_scanned);
+			sc->nr_reclaimed += nr_soft_reclaimed;
+			sc->nr_scanned += nr_soft_scanned;
 			/* need some check for avoid more shrink_zone() */
 		}
 
@@ -2952,6 +2920,8 @@ static unsigned long balance_pgdat(pg_data_t *pgdat, int order,
 {
 	int i;
 	int end_zone = 0;	/* Inclusive.  0 = ZONE_DMA */
+	unsigned long nr_soft_reclaimed;
+	unsigned long nr_soft_scanned;
 	struct scan_control sc = {
 		.gfp_mask = GFP_KERNEL,
 		.priority = DEF_PRIORITY,
@@ -3066,6 +3036,15 @@ static unsigned long balance_pgdat(pg_data_t *pgdat, int order,
 
 			sc.nr_scanned = 0;
 
+			nr_soft_scanned = 0;
+			/*
+			 * Call soft limit reclaim before calling shrink_zone.
+			 */
+			nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone,
+							order, sc.gfp_mask,
+							&nr_soft_scanned);
+			sc.nr_reclaimed += nr_soft_reclaimed;
+
 			/*
 			 * There should be no need to raise the scanning
 			 * priority if enough pages are already being scanned

scripts/checkpatch.pl

@@ -3975,8 +3975,8 @@ sub string_find_replace {
 # check for new externs in .h files.
 		if ($realfile =~ /\.h$/ &&
 		    $line =~ /^\+\s*(extern\s+)$Type\s*$Ident\s*\(/s) {
-			if (WARN("AVOID_EXTERNS",
-				 "extern prototypes should be avoided in .h files\n" . $herecurr) &&
+			if (CHK("AVOID_EXTERNS",
+				"extern prototypes should be avoided in .h files\n" . $herecurr) &&
 			    $fix) {
 				$fixed[$linenr - 1] =~ s/(.*)\bextern\b\s*(.*)/$1$2/;
 			}