X-Git-Url: https://err.no/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=kernel%2Fcpuset.c;h=d5ab79cf516d7edf77f8e45afe15c04cdae2cda4;hb=73909f7a665991013dcff42a815fda76d3a7300a;hp=7326d51eefe13169254a14edd4e05e77bb108fa1;hpb=0b2f630a28d53b5a2082a5275bc3334b10373508;p=linux-2.6

diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index 7326d51eef..d5ab79cf51 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -54,7 +54,6 @@
 #include <asm/uaccess.h>
 #include <asm/atomic.h>
 #include <linux/mutex.h>
-#include <linux/kfifo.h>
 #include <linux/workqueue.h>
 #include <linux/cgroup.h>
 
@@ -365,7 +364,7 @@ void cpuset_update_task_memory_state(void)
 		my_cpusets_mem_gen = top_cpuset.mems_generation;
 	} else {
 		rcu_read_lock();
-		my_cpusets_mem_gen = task_cs(current)->mems_generation;
+		my_cpusets_mem_gen = task_cs(tsk)->mems_generation;
 		rcu_read_unlock();
 	}
 
@@ -486,21 +485,51 @@ static int cpusets_overlap(struct cpuset *a, struct cpuset *b)
 static void
 update_domain_attr(struct sched_domain_attr *dattr, struct cpuset *c)
 {
-	if (!dattr)
-		return;
 	if (dattr->relax_domain_level < c->relax_domain_level)
 		dattr->relax_domain_level = c->relax_domain_level;
 	return;
 }
 
+static void
+update_domain_attr_tree(struct sched_domain_attr *dattr, struct cpuset *c)
+{
+	LIST_HEAD(q);
+
+	list_add(&c->stack_list, &q);
+	while (!list_empty(&q)) {
+		struct cpuset *cp;
+		struct cgroup *cont;
+		struct cpuset *child;
+
+		cp = list_first_entry(&q, struct cpuset, stack_list);
+		list_del(q.next);
+
+		if (cpus_empty(cp->cpus_allowed))
+			continue;
+
+		if (is_sched_load_balance(cp))
+			update_domain_attr(dattr, cp);
+
+		list_for_each_entry(cont, &cp->css.cgroup->children, sibling) {
+			child = cgroup_cs(cont);
+			list_add_tail(&child->stack_list, &q);
+		}
+	}
+}
+
 /*
  * rebuild_sched_domains()
  *
- * If the flag 'sched_load_balance' of any cpuset with non-empty
- * 'cpus' changes, or if the 'cpus' allowed changes in any cpuset
- * which has that flag enabled, or if any cpuset with a non-empty
- * 'cpus' is removed, then call this routine to rebuild the
- * scheduler's dynamic sched domains.
+ * This routine will be called to rebuild the scheduler's dynamic
+ * sched domains:
+ * - if the flag 'sched_load_balance' of any cpuset with non-empty
+ *   'cpus' changes,
+ * - or if the 'cpus' allowed changes in any cpuset which has that
+ *   flag enabled,
+ * - or if the 'sched_relax_domain_level' of any cpuset which has
+ *   that flag enabled and with non-empty 'cpus' changes,
+ * - or if any cpuset with non-empty 'cpus' is removed,
+ * - or if a cpu gets offlined.
  *
  * This routine builds a partial partition of the systems CPUs
  * (the set of non-overlappping cpumask_t's in the array 'part'
@@ -527,7 +556,7 @@ update_domain_attr(struct sched_domain_attr *dattr, struct cpuset *c)
  * So the reverse nesting would risk an ABBA deadlock.
  *
  * The three key local variables below are:
- *    q  - a kfifo queue of cpuset pointers, used to implement a
+ *    q  - a linked-list queue of cpuset pointers, used to implement a
  *	   top-down scan of all cpusets.  This scan loads a pointer
  *	   to each cpuset marked is_sched_load_balance into the
  *	   array 'csa'.  For our purposes, rebuilding the schedulers
@@ -562,7 +591,7 @@ update_domain_attr(struct sched_domain_attr *dattr, struct cpuset *c)
 
 void rebuild_sched_domains(void)
 {
-	struct kfifo *q;	/* queue of cpusets to be scanned */
+	LIST_HEAD(q);		/* queue of cpusets to be scanned*/
 	struct cpuset *cp;	/* scans q */
 	struct cpuset **csa;	/* array of all cpuset ptrs */
 	int csn;		/* how many cpuset ptrs in csa so far */
@@ -572,7 +601,6 @@ void rebuild_sched_domains(void)
 	int ndoms;		/* number of sched domains in result */
 	int nslot;		/* next empty doms[] cpumask_t slot */
 
-	q = NULL;
 	csa = NULL;
 	doms = NULL;
 	dattr = NULL;
@@ -586,30 +614,42 @@ void rebuild_sched_domains(void)
 		dattr = kmalloc(sizeof(struct sched_domain_attr), GFP_KERNEL);
 		if (dattr) {
 			*dattr = SD_ATTR_INIT;
-			update_domain_attr(dattr, &top_cpuset);
+			update_domain_attr_tree(dattr, &top_cpuset);
 		}
 		*doms = top_cpuset.cpus_allowed;
 		goto rebuild;
 	}
 
-	q = kfifo_alloc(number_of_cpusets * sizeof(cp), GFP_KERNEL, NULL);
-	if (IS_ERR(q))
-		goto done;
 	csa = kmalloc(number_of_cpusets * sizeof(cp), GFP_KERNEL);
 	if (!csa)
 		goto done;
 	csn = 0;
 
-	cp = &top_cpuset;
-	__kfifo_put(q, (void *)&cp, sizeof(cp));
-	while (__kfifo_get(q, (void *)&cp, sizeof(cp))) {
+	list_add(&top_cpuset.stack_list, &q);
+	while (!list_empty(&q)) {
 		struct cgroup *cont;
 		struct cpuset *child;   /* scans child cpusets of cp */
-		if (is_sched_load_balance(cp))
+
+		cp = list_first_entry(&q, struct cpuset, stack_list);
+		list_del(q.next);
+
+		if (cpus_empty(cp->cpus_allowed))
+			continue;
+
+		/*
+		 * All child cpusets contain a subset of the parent's cpus, so
+		 * just skip them, and then we call update_domain_attr_tree()
+		 * to calc relax_domain_level of the corresponding sched
+		 * domain.
+		 */
+		if (is_sched_load_balance(cp)) {
 			csa[csn++] = cp;
+			continue;
+		}
+
 		list_for_each_entry(cont, &cp->css.cgroup->children, sibling) {
 			child = cgroup_cs(cont);
-			__kfifo_put(q, (void *)&child, sizeof(cp));
+			list_add_tail(&child->stack_list, &q);
 		}
   	}
 
@@ -676,7 +716,7 @@ restart:
 					cpus_or(*dp, *dp, b->cpus_allowed);
 					b->pn = -1;
 					if (dattr)
-						update_domain_attr(dattr
+						update_domain_attr_tree(dattr
 								   + nslot, b);
 				}
 			}
@@ -692,43 +732,11 @@ rebuild:
 	put_online_cpus();
 
 done:
-	if (q && !IS_ERR(q))
-		kfifo_free(q);
 	kfree(csa);
 	/* Don't kfree(doms) -- partition_sched_domains() does that. */
 	/* Don't kfree(dattr) -- partition_sched_domains() does that. */
 }
 
-static inline int started_after_time(struct task_struct *t1,
-				     struct timespec *time,
-				     struct task_struct *t2)
-{
-	int start_diff = timespec_compare(&t1->start_time, time);
-	if (start_diff > 0) {
-		return 1;
-	} else if (start_diff < 0) {
-		return 0;
-	} else {
-		/*
-		 * Arbitrarily, if two processes started at the same
-		 * time, we'll say that the lower pointer value
-		 * started first. Note that t2 may have exited by now
-		 * so this may not be a valid pointer any longer, but
-		 * that's fine - it still serves to distinguish
-		 * between two tasks started (effectively)
-		 * simultaneously.
-		 */
-		return t1 > t2;
-	}
-}
-
-static inline int started_after(void *p1, void *p2)
-{
-	struct task_struct *t1 = p1;
-	struct task_struct *t2 = p2;
-	return started_after_time(t1, &t2->start_time, t2);
-}
-
 /**
  * cpuset_test_cpumask - test a task's cpus_allowed versus its cpuset's
  * @tsk: task to test
@@ -780,7 +788,12 @@ static int update_tasks_cpumask(struct cpuset *cs)
 	struct ptr_heap heap;
 	int retval;
 
-	retval = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, &started_after);
+	/*
+	 * cgroup_scan_tasks() will initialize heap->gt for us.
+	 * heap_init() is still needed here for we should not change
+	 * cs->cpus_allowed when heap_init() fails.
+	 */
+	retval = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, NULL);
 	if (retval)
 		return retval;
 
@@ -1076,7 +1089,8 @@ static int update_relax_domain_level(struct cpuset *cs, s64 val)
 
 	if (val != cs->relax_domain_level) {
 		cs->relax_domain_level = val;
-		rebuild_sched_domains();
+		if (!cpus_empty(cs->cpus_allowed) && is_sched_load_balance(cs))
+			rebuild_sched_domains();
 	}
 
 	return 0;
@@ -1791,7 +1805,7 @@ static void move_member_tasks_to_cpuset(struct cpuset *from, struct cpuset *to)
 	scan.scan.heap = NULL;
 	scan.to = to->css.cgroup;
 
-	if (cgroup_scan_tasks((struct cgroup_scanner *)&scan))
+	if (cgroup_scan_tasks(&scan.scan))
 		printk(KERN_ERR "move_member_tasks_to_cpuset: "
 				"cgroup_scan_tasks failed\n");
 }
@@ -1847,29 +1861,29 @@ static void remove_tasks_in_empty_cpuset(struct cpuset *cs)
  */
 static void scan_for_empty_cpusets(const struct cpuset *root)
 {
+	LIST_HEAD(queue);
 	struct cpuset *cp;	/* scans cpusets being updated */
 	struct cpuset *child;	/* scans child cpusets of cp */
-	struct list_head queue;
 	struct cgroup *cont;
-
-	INIT_LIST_HEAD(&queue);
+	nodemask_t oldmems;
 
 	list_add_tail((struct list_head *)&root->stack_list, &queue);
 
 	while (!list_empty(&queue)) {
-		cp = container_of(queue.next, struct cpuset, stack_list);
+		cp = list_first_entry(&queue, struct cpuset, stack_list);
 		list_del(queue.next);
 		list_for_each_entry(cont, &cp->css.cgroup->children, sibling) {
 			child = cgroup_cs(cont);
 			list_add_tail(&child->stack_list, &queue);
 		}
-		cont = cp->css.cgroup;
 
 		/* Continue past cpusets with all cpus, mems online */
 		if (cpus_subset(cp->cpus_allowed, cpu_online_map) &&
 		    nodes_subset(cp->mems_allowed, node_states[N_HIGH_MEMORY]))
 			continue;
 
+		oldmems = cp->mems_allowed;
+
 		/* Remove offline cpus and mems from this cpuset. */
 		mutex_lock(&callback_mutex);
 		cpus_and(cp->cpus_allowed, cp->cpus_allowed, cpu_online_map);
@@ -1881,6 +1895,10 @@ static void scan_for_empty_cpusets(const struct cpuset *root)
 		if (cpus_empty(cp->cpus_allowed) ||
 		     nodes_empty(cp->mems_allowed))
 			remove_tasks_in_empty_cpuset(cp);
+		else {
+			update_tasks_cpumask(cp);
+			update_tasks_nodemask(cp, &oldmems);
+		}
 	}
 }