/* Scan @scan_size MB every @scan_period after an initial @scan_delay in ms */
unsigned int sysctl_numa_balancing_scan_delay = 1000;
-/*
- * After skipping a page migration on a shared page, skip N more numa page
- * migrations unconditionally. This reduces the number of NUMA migrations
- * in shared memory workloads, and has the effect of pulling tasks towards
- * where their memory lives, over pulling the memory towards the task.
- */
-unsigned int sysctl_numa_balancing_migrate_deferred = 16;
-
static unsigned int task_nr_scan_windows(struct task_struct *p)
{
unsigned long rss = 0;
struct list_head task_list;
struct rcu_head rcu;
+ nodemask_t active_nodes;
unsigned long total_faults;
+ /*
+ * Faults_cpu is used to decide whether memory should move
+ * towards the CPU. As a consequence, these stats are weighted
+ * more by CPU use than by memory faults.
+ */
+ unsigned long *faults_cpu;
unsigned long faults[0];
};
+/* Shared or private faults. */
+#define NR_NUMA_HINT_FAULT_TYPES 2
+
+/* Memory and CPU locality */
+#define NR_NUMA_HINT_FAULT_STATS (NR_NUMA_HINT_FAULT_TYPES * 2)
+
+/* Averaged statistics, and temporary buffers. */
+#define NR_NUMA_HINT_FAULT_BUCKETS (NR_NUMA_HINT_FAULT_STATS * 2)
+
pid_t task_numa_group_id(struct task_struct *p)
{
return p->numa_group ? p->numa_group->gid : 0;
static inline int task_faults_idx(int nid, int priv)
{
- return 2 * nid + priv;
+ return NR_NUMA_HINT_FAULT_TYPES * nid + priv;
}
static inline unsigned long task_faults(struct task_struct *p, int nid)
{
- if (!p->numa_faults)
+ if (!p->numa_faults_memory)
return 0;
- return p->numa_faults[task_faults_idx(nid, 0)] +
- p->numa_faults[task_faults_idx(nid, 1)];
+ return p->numa_faults_memory[task_faults_idx(nid, 0)] +
+ p->numa_faults_memory[task_faults_idx(nid, 1)];
}
static inline unsigned long group_faults(struct task_struct *p, int nid)
p->numa_group->faults[task_faults_idx(nid, 1)];
}
+static inline unsigned long group_faults_cpu(struct numa_group *group, int nid)
+{
+ return group->faults_cpu[task_faults_idx(nid, 0)] +
+ group->faults_cpu[task_faults_idx(nid, 1)];
+}
+
/*
* These return the fraction of accesses done by a particular task, or
* task group, on a particular numa node. The group weight is given a
{
unsigned long total_faults;
- if (!p->numa_faults)
+ if (!p->numa_faults_memory)
return 0;
total_faults = p->total_numa_faults;
return 1000 * group_faults(p, nid) / p->numa_group->total_faults;
}
+bool should_numa_migrate_memory(struct task_struct *p, struct page * page,
+ int src_nid, int dst_cpu)
+{
+ struct numa_group *ng = p->numa_group;
+ int dst_nid = cpu_to_node(dst_cpu);
+ int last_cpupid, this_cpupid;
+
+ this_cpupid = cpu_pid_to_cpupid(dst_cpu, current->pid);
+
+ /*
+ * Multi-stage node selection is used in conjunction with a periodic
+ * migration fault to build a temporal task<->page relation. By using
+ * a two-stage filter we remove short/unlikely relations.
+ *
+ * Using P(p) ~ n_p / n_t as per frequentist probability, we can equate
+ * a task's usage of a particular page (n_p) per total usage of this
+ * page (n_t) (in a given time-span) to a probability.
+ *
+ * Our periodic faults will sample this probability and getting the
+ * same result twice in a row, given these samples are fully
+ * independent, is then given by P(n)^2, provided our sample period
+ * is sufficiently short compared to the usage pattern.
+ *
+ * This quadric squishes small probabilities, making it less likely we
+ * act on an unlikely task<->page relation.
+ */
+ last_cpupid = page_cpupid_xchg_last(page, this_cpupid);
+ if (!cpupid_pid_unset(last_cpupid) &&
+ cpupid_to_nid(last_cpupid) != dst_nid)
+ return false;
+
+ /* Always allow migrate on private faults */
+ if (cpupid_match_pid(p, last_cpupid))
+ return true;
+
+ /* A shared fault, but p->numa_group has not been set up yet. */
+ if (!ng)
+ return true;
+
+ /*
+ * Do not migrate if the destination is not a node that
+ * is actively used by this numa group.
+ */
+ if (!node_isset(dst_nid, ng->active_nodes))
+ return false;
+
+ /*
+ * Source is a node that is not actively used by this
+ * numa group, while the destination is. Migrate.
+ */
+ if (!node_isset(src_nid, ng->active_nodes))
+ return true;
+
+ /*
+ * Both source and destination are nodes in active
+ * use by this numa group. Maximize memory bandwidth
+ * by migrating from more heavily used groups, to less
+ * heavily used ones, spreading the load around.
+ * Use a 1/4 hysteresis to avoid spurious page movement.
+ */
+ return group_faults(p, dst_nid) < (group_faults(p, src_nid) * 3 / 4);
+}
+
static unsigned long weighted_cpuload(const int cpu);
static unsigned long source_load(int cpu, int type);
static unsigned long target_load(int cpu, int type);
static void numa_migrate_preferred(struct task_struct *p)
{
/* This task has no NUMA fault statistics yet */
- if (unlikely(p->numa_preferred_nid == -1 || !p->numa_faults))
+ if (unlikely(p->numa_preferred_nid == -1 || !p->numa_faults_memory))
return;
/* Periodically retry migrating the task to the preferred node */
task_numa_migrate(p);
}
+/*
+ * Find the nodes on which the workload is actively running. We do this by
+ * tracking the nodes from which NUMA hinting faults are triggered. This can
+ * be different from the set of nodes where the workload's memory is currently
+ * located.
+ *
+ * The bitmask is used to make smarter decisions on when to do NUMA page
+ * migrations, To prevent flip-flopping, and excessive page migrations, nodes
+ * are added when they cause over 6/16 of the maximum number of faults, but
+ * only removed when they drop below 3/16.
+ */
+static void update_numa_active_node_mask(struct numa_group *numa_group)
+{
+ unsigned long faults, max_faults = 0;
+ int nid;
+
+ for_each_online_node(nid) {
+ faults = group_faults_cpu(numa_group, nid);
+ if (faults > max_faults)
+ max_faults = faults;
+ }
+
+ for_each_online_node(nid) {
+ faults = group_faults_cpu(numa_group, nid);
+ if (!node_isset(nid, numa_group->active_nodes)) {
+ if (faults > max_faults * 6 / 16)
+ node_set(nid, numa_group->active_nodes);
+ } else if (faults < max_faults * 3 / 16)
+ node_clear(nid, numa_group->active_nodes);
+ }
+}
+
/*
* When adapting the scan rate, the period is divided into NUMA_PERIOD_SLOTS
* increments. The more local the fault statistics are, the higher the scan
memset(p->numa_faults_locality, 0, sizeof(p->numa_faults_locality));
}
+/*
+ * Get the fraction of time the task has been running since the last
+ * NUMA placement cycle. The scheduler keeps similar statistics, but
+ * decays those on a 32ms period, which is orders of magnitude off
+ * from the dozens-of-seconds NUMA balancing period. Use the scheduler
+ * stats only if the task is so new there are no NUMA statistics yet.
+ */
+static u64 numa_get_avg_runtime(struct task_struct *p, u64 *period)
+{
+ u64 runtime, delta, now;
+ /* Use the start of this time slice to avoid calculations. */
+ now = p->se.exec_start;
+ runtime = p->se.sum_exec_runtime;
+
+ if (p->last_task_numa_placement) {
+ delta = runtime - p->last_sum_exec_runtime;
+ *period = now - p->last_task_numa_placement;
+ } else {
+ delta = p->se.avg.runnable_avg_sum;
+ *period = p->se.avg.runnable_avg_period;
+ }
+
+ p->last_sum_exec_runtime = runtime;
+ p->last_task_numa_placement = now;
+
+ return delta;
+}
+
static void task_numa_placement(struct task_struct *p)
{
int seq, nid, max_nid = -1, max_group_nid = -1;
unsigned long max_faults = 0, max_group_faults = 0;
unsigned long fault_types[2] = { 0, 0 };
+ unsigned long total_faults;
+ u64 runtime, period;
spinlock_t *group_lock = NULL;
seq = ACCESS_ONCE(p->mm->numa_scan_seq);
p->numa_scan_seq = seq;
p->numa_scan_period_max = task_scan_max(p);
+ total_faults = p->numa_faults_locality[0] +
+ p->numa_faults_locality[1];
+ runtime = numa_get_avg_runtime(p, &period);
+
/* If the task is part of a group prevent parallel updates to group stats */
if (p->numa_group) {
group_lock = &p->numa_group->lock;
unsigned long faults = 0, group_faults = 0;
int priv, i;
- for (priv = 0; priv < 2; priv++) {
- long diff;
+ for (priv = 0; priv < NR_NUMA_HINT_FAULT_TYPES; priv++) {
+ long diff, f_diff, f_weight;
i = task_faults_idx(nid, priv);
- diff = -p->numa_faults[i];
/* Decay existing window, copy faults since last scan */
- p->numa_faults[i] >>= 1;
- p->numa_faults[i] += p->numa_faults_buffer[i];
- fault_types[priv] += p->numa_faults_buffer[i];
- p->numa_faults_buffer[i] = 0;
+ diff = p->numa_faults_buffer_memory[i] - p->numa_faults_memory[i] / 2;
+ fault_types[priv] += p->numa_faults_buffer_memory[i];
+ p->numa_faults_buffer_memory[i] = 0;
- faults += p->numa_faults[i];
- diff += p->numa_faults[i];
+ /*
+ * Normalize the faults_from, so all tasks in a group
+ * count according to CPU use, instead of by the raw
+ * number of faults. Tasks with little runtime have
+ * little over-all impact on throughput, and thus their
+ * faults are less important.
+ */
+ f_weight = div64_u64(runtime << 16, period + 1);
+ f_weight = (f_weight * p->numa_faults_buffer_cpu[i]) /
+ (total_faults + 1);
+ f_diff = f_weight - p->numa_faults_cpu[i] / 2;
+ p->numa_faults_buffer_cpu[i] = 0;
+
+ p->numa_faults_memory[i] += diff;
+ p->numa_faults_cpu[i] += f_diff;
+ faults += p->numa_faults_memory[i];
p->total_numa_faults += diff;
if (p->numa_group) {
/* safe because we can only change our own group */
p->numa_group->faults[i] += diff;
+ p->numa_group->faults_cpu[i] += f_diff;
p->numa_group->total_faults += diff;
group_faults += p->numa_group->faults[i];
}
update_task_scan_period(p, fault_types[0], fault_types[1]);
if (p->numa_group) {
+ update_numa_active_node_mask(p->numa_group);
/*
* If the preferred task and group nids are different,
* iterate over the nodes again to find the best place.
if (unlikely(!p->numa_group)) {
unsigned int size = sizeof(struct numa_group) +
- 2*nr_node_ids*sizeof(unsigned long);
+ 4*nr_node_ids*sizeof(unsigned long);
grp = kzalloc(size, GFP_KERNEL | __GFP_NOWARN);
if (!grp)
spin_lock_init(&grp->lock);
INIT_LIST_HEAD(&grp->task_list);
grp->gid = p->pid;
+ /* Second half of the array tracks nids where faults happen */
+ grp->faults_cpu = grp->faults + NR_NUMA_HINT_FAULT_TYPES *
+ nr_node_ids;
+
+ node_set(task_node(current), grp->active_nodes);
- for (i = 0; i < 2*nr_node_ids; i++)
- grp->faults[i] = p->numa_faults[i];
+ for (i = 0; i < NR_NUMA_HINT_FAULT_STATS * nr_node_ids; i++)
+ grp->faults[i] = p->numa_faults_memory[i];
grp->total_faults = p->total_numa_faults;
double_lock(&my_grp->lock, &grp->lock);
- for (i = 0; i < 2*nr_node_ids; i++) {
- my_grp->faults[i] -= p->numa_faults[i];
- grp->faults[i] += p->numa_faults[i];
+ for (i = 0; i < NR_NUMA_HINT_FAULT_STATS * nr_node_ids; i++) {
+ my_grp->faults[i] -= p->numa_faults_memory[i];
+ grp->faults[i] += p->numa_faults_memory[i];
}
my_grp->total_faults -= p->total_numa_faults;
grp->total_faults += p->total_numa_faults;
{
struct numa_group *grp = p->numa_group;
int i;
- void *numa_faults = p->numa_faults;
+ void *numa_faults = p->numa_faults_memory;
if (grp) {
spin_lock(&grp->lock);
- for (i = 0; i < 2*nr_node_ids; i++)
- grp->faults[i] -= p->numa_faults[i];
+ for (i = 0; i < NR_NUMA_HINT_FAULT_STATS * nr_node_ids; i++)
+ grp->faults[i] -= p->numa_faults_memory[i];
grp->total_faults -= p->total_numa_faults;
list_del(&p->numa_entry);
put_numa_group(grp);
}
- p->numa_faults = NULL;
- p->numa_faults_buffer = NULL;
+ p->numa_faults_memory = NULL;
+ p->numa_faults_buffer_memory = NULL;
+ p->numa_faults_cpu= NULL;
+ p->numa_faults_buffer_cpu = NULL;
kfree(numa_faults);
}
/*
* Got a PROT_NONE fault for a page on @node.
*/
-void task_numa_fault(int last_cpupid, int node, int pages, int flags)
+void task_numa_fault(int last_cpupid, int mem_node, int pages, int flags)
{
struct task_struct *p = current;
bool migrated = flags & TNF_MIGRATED;
+ int cpu_node = task_node(current);
int priv;
if (!numabalancing_enabled)
return;
/* Allocate buffer to track faults on a per-node basis */
- if (unlikely(!p->numa_faults)) {
- int size = sizeof(*p->numa_faults) * 2 * nr_node_ids;
+ if (unlikely(!p->numa_faults_memory)) {
+ int size = sizeof(*p->numa_faults_memory) *
+ NR_NUMA_HINT_FAULT_BUCKETS * nr_node_ids;
- /* numa_faults and numa_faults_buffer share the allocation */
- p->numa_faults = kzalloc(size * 2, GFP_KERNEL|__GFP_NOWARN);
- if (!p->numa_faults)
+ p->numa_faults_memory = kzalloc(size, GFP_KERNEL|__GFP_NOWARN);
+ if (!p->numa_faults_memory)
return;
- BUG_ON(p->numa_faults_buffer);
- p->numa_faults_buffer = p->numa_faults + (2 * nr_node_ids);
+ BUG_ON(p->numa_faults_buffer_memory);
+ /*
+ * The averaged statistics, shared & private, memory & cpu,
+ * occupy the first half of the array. The second half of the
+ * array is for current counters, which are averaged into the
+ * first set by task_numa_placement.
+ */
+ p->numa_faults_cpu = p->numa_faults_memory + (2 * nr_node_ids);
+ p->numa_faults_buffer_memory = p->numa_faults_memory + (4 * nr_node_ids);
+ p->numa_faults_buffer_cpu = p->numa_faults_memory + (6 * nr_node_ids);
p->total_numa_faults = 0;
memset(p->numa_faults_locality, 0, sizeof(p->numa_faults_locality));
}
if (migrated)
p->numa_pages_migrated += pages;
- p->numa_faults_buffer[task_faults_idx(node, priv)] += pages;
+ p->numa_faults_buffer_memory[task_faults_idx(mem_node, priv)] += pages;
+ p->numa_faults_buffer_cpu[task_faults_idx(cpu_node, priv)] += pages;
p->numa_faults_locality[!!(flags & TNF_FAULT_LOCAL)] += pages;
}
{
int src_nid, dst_nid;
- if (!sched_feat(NUMA_FAVOUR_HIGHER) || !p->numa_faults ||
+ if (!sched_feat(NUMA_FAVOUR_HIGHER) || !p->numa_faults_memory ||
!(env->sd->flags & SD_NUMA)) {
return false;
}
if (!sched_feat(NUMA) || !sched_feat(NUMA_RESIST_LOWER))
return false;
- if (!p->numa_faults || !(env->sd->flags & SD_NUMA))
+ if (!p->numa_faults_memory || !(env->sd->flags & SD_NUMA))
return false;
src_nid = cpu_to_node(env->src_cpu);
projects / cascardo / linux.git / commitdiff