/* Throttling is performed over 100ms slice and after that slice is renewed */
static unsigned long throtl_slice = HZ/10; /* 100 ms */
+/* A workqueue to queue throttle related work */
+static struct workqueue_struct *kthrotld_workqueue;
+static void throtl_schedule_delayed_work(struct throtl_data *td,
+ unsigned long delay);
+
struct throtl_rb_root {
struct rb_root rb;
struct rb_node *left;
unsigned long slice_end[2];
/* Some throttle limits got updated for the group */
- bool limits_changed;
+ int limits_changed;
};
struct throtl_data
/* Work for dispatching throttled bios */
struct delayed_work throtl_work;
- atomic_t limits_changed;
+ int limits_changed;
};
enum tg_state_flags {
RB_CLEAR_NODE(&tg->rb_node);
bio_list_init(&tg->bio_lists[0]);
bio_list_init(&tg->bio_lists[1]);
+ td->limits_changed = false;
/*
* Take the initial reference that will be released on destroy
update_min_dispatch_time(st);
if (time_before_eq(st->min_disptime, jiffies))
- throtl_schedule_delayed_work(td->queue, 0);
+ throtl_schedule_delayed_work(td, 0);
else
- throtl_schedule_delayed_work(td->queue,
- (st->min_disptime - jiffies));
+ throtl_schedule_delayed_work(td, (st->min_disptime - jiffies));
}
static inline void
struct throtl_grp *tg;
struct hlist_node *pos, *n;
- if (!atomic_read(&td->limits_changed))
+ if (!td->limits_changed)
return;
- throtl_log(td, "limit changed =%d", atomic_read(&td->limits_changed));
+ xchg(&td->limits_changed, false);
- /*
- * Make sure updates from throtl_update_blkio_group_read_bps() group
- * of functions to tg->limits_changed are visible. We do not
- * want update td->limits_changed to be visible but update to
- * tg->limits_changed not being visible yet on this cpu. Hence
- * the read barrier.
- */
- smp_rmb();
+ throtl_log(td, "limits changed");
hlist_for_each_entry_safe(tg, pos, n, &td->tg_list, tg_node) {
- if (throtl_tg_on_rr(tg) && tg->limits_changed) {
- throtl_log_tg(td, tg, "limit change rbps=%llu wbps=%llu"
- " riops=%u wiops=%u", tg->bps[READ],
- tg->bps[WRITE], tg->iops[READ],
- tg->iops[WRITE]);
+ if (!tg->limits_changed)
+ continue;
+
+ if (!xchg(&tg->limits_changed, false))
+ continue;
+
+ throtl_log_tg(td, tg, "limit change rbps=%llu wbps=%llu"
+ " riops=%u wiops=%u", tg->bps[READ], tg->bps[WRITE],
+ tg->iops[READ], tg->iops[WRITE]);
+
+ /*
+ * Restart the slices for both READ and WRITES. It
+ * might happen that a group's limit are dropped
+ * suddenly and we don't want to account recently
+ * dispatched IO with new low rate
+ */
+ throtl_start_new_slice(td, tg, 0);
+ throtl_start_new_slice(td, tg, 1);
+
+ if (throtl_tg_on_rr(tg))
tg_update_disptime(td, tg);
- tg->limits_changed = false;
- }
}
-
- smp_mb__before_atomic_dec();
- atomic_dec(&td->limits_changed);
- smp_mb__after_atomic_dec();
}
/* Dispatch throttled bios. Should be called without queue lock held. */
unsigned int nr_disp = 0;
struct bio_list bio_list_on_stack;
struct bio *bio;
+ struct blk_plug plug;
spin_lock_irq(q->queue_lock);
* immediate dispatch
*/
if (nr_disp) {
+ blk_start_plug(&plug);
while((bio = bio_list_pop(&bio_list_on_stack)))
generic_make_request(bio);
- blk_unplug(q);
+ blk_finish_plug(&plug);
}
return nr_disp;
}
}
/* Call with queue lock held */
-void throtl_schedule_delayed_work(struct request_queue *q, unsigned long delay)
+static void
+throtl_schedule_delayed_work(struct throtl_data *td, unsigned long delay)
{
- struct throtl_data *td = q->td;
struct delayed_work *dwork = &td->throtl_work;
- if (total_nr_queued(td) > 0) {
+ /* schedule work if limits changed even if no bio is queued */
+ if (total_nr_queued(td) > 0 || td->limits_changed) {
/*
* We might have a work scheduled to be executed in future.
* Cancel that and schedule a new one.
*/
__cancel_delayed_work(dwork);
- kblockd_schedule_delayed_work(q, dwork, delay);
+ queue_delayed_work(kthrotld_workqueue, dwork, delay);
throtl_log(td, "schedule work. delay=%lu jiffies=%lu",
delay, jiffies);
}
}
-EXPORT_SYMBOL(throtl_schedule_delayed_work);
static void
throtl_destroy_tg(struct throtl_data *td, struct throtl_grp *tg)
spin_unlock_irqrestore(td->queue->queue_lock, flags);
}
+static void throtl_update_blkio_group_common(struct throtl_data *td,
+ struct throtl_grp *tg)
+{
+ xchg(&tg->limits_changed, true);
+ xchg(&td->limits_changed, true);
+ /* Schedule a work now to process the limit change */
+ throtl_schedule_delayed_work(td, 0);
+}
+
/*
* For all update functions, key should be a valid pointer because these
* update functions are called under blkcg_lock, that means, blkg is
- * valid and in turn key is valid. queue exit path can not race becuase
+ * valid and in turn key is valid. queue exit path can not race because
* of blkcg_lock
*
* Can not take queue lock in update functions as queue lock under blkcg_lock
struct blkio_group *blkg, u64 read_bps)
{
struct throtl_data *td = key;
+ struct throtl_grp *tg = tg_of_blkg(blkg);
- tg_of_blkg(blkg)->bps[READ] = read_bps;
- /* Make sure read_bps is updated before setting limits_changed */
- smp_wmb();
- tg_of_blkg(blkg)->limits_changed = true;
-
- /* Make sure tg->limits_changed is updated before td->limits_changed */
- smp_mb__before_atomic_inc();
- atomic_inc(&td->limits_changed);
- smp_mb__after_atomic_inc();
-
- /* Schedule a work now to process the limit change */
- throtl_schedule_delayed_work(td->queue, 0);
+ tg->bps[READ] = read_bps;
+ throtl_update_blkio_group_common(td, tg);
}
static void throtl_update_blkio_group_write_bps(void *key,
struct blkio_group *blkg, u64 write_bps)
{
struct throtl_data *td = key;
+ struct throtl_grp *tg = tg_of_blkg(blkg);
- tg_of_blkg(blkg)->bps[WRITE] = write_bps;
- smp_wmb();
- tg_of_blkg(blkg)->limits_changed = true;
- smp_mb__before_atomic_inc();
- atomic_inc(&td->limits_changed);
- smp_mb__after_atomic_inc();
- throtl_schedule_delayed_work(td->queue, 0);
+ tg->bps[WRITE] = write_bps;
+ throtl_update_blkio_group_common(td, tg);
}
static void throtl_update_blkio_group_read_iops(void *key,
struct blkio_group *blkg, unsigned int read_iops)
{
struct throtl_data *td = key;
+ struct throtl_grp *tg = tg_of_blkg(blkg);
- tg_of_blkg(blkg)->iops[READ] = read_iops;
- smp_wmb();
- tg_of_blkg(blkg)->limits_changed = true;
- smp_mb__before_atomic_inc();
- atomic_inc(&td->limits_changed);
- smp_mb__after_atomic_inc();
- throtl_schedule_delayed_work(td->queue, 0);
+ tg->iops[READ] = read_iops;
+ throtl_update_blkio_group_common(td, tg);
}
static void throtl_update_blkio_group_write_iops(void *key,
struct blkio_group *blkg, unsigned int write_iops)
{
struct throtl_data *td = key;
+ struct throtl_grp *tg = tg_of_blkg(blkg);
- tg_of_blkg(blkg)->iops[WRITE] = write_iops;
- smp_wmb();
- tg_of_blkg(blkg)->limits_changed = true;
- smp_mb__before_atomic_inc();
- atomic_inc(&td->limits_changed);
- smp_mb__after_atomic_inc();
- throtl_schedule_delayed_work(td->queue, 0);
+ tg->iops[WRITE] = write_iops;
+ throtl_update_blkio_group_common(td, tg);
}
-void throtl_shutdown_timer_wq(struct request_queue *q)
+static void throtl_shutdown_wq(struct request_queue *q)
{
struct throtl_data *td = q->td;
/*
* There is already another bio queued in same dir. No
* need to update dispatch time.
- * Still update the disptime if rate limits on this group
- * were changed.
*/
- if (!tg->limits_changed)
- update_disptime = false;
- else
- tg->limits_changed = false;
-
+ update_disptime = false;
goto queue_bio;
+
}
/* Bio is with-in rate limit of group */
if (tg_may_dispatch(td, tg, bio, NULL)) {
throtl_charge_bio(tg, bio);
+
+ /*
+ * We need to trim slice even when bios are not being queued
+ * otherwise it might happen that a bio is not queued for
+ * a long time and slice keeps on extending and trim is not
+ * called for a long time. Now if limits are reduced suddenly
+ * we take into account all the IO dispatched so far at new
+ * low rate and * newly queued IO gets a really long dispatch
+ * time.
+ *
+ * So keep on trimming slice even if bio is not queued.
+ */
+ throtl_trim_slice(td, tg, rw);
goto out;
}
INIT_HLIST_HEAD(&td->tg_list);
td->tg_service_tree = THROTL_RB_ROOT;
- atomic_set(&td->limits_changed, 0);
+ td->limits_changed = false;
/* Init root group */
tg = &td->root_tg;
/* Practically unlimited BW */
tg->bps[0] = tg->bps[1] = -1;
tg->iops[0] = tg->iops[1] = -1;
+ td->limits_changed = false;
/*
* Set root group reference to 2. One reference will be dropped when
BUG_ON(!td);
- throtl_shutdown_timer_wq(q);
+ throtl_shutdown_wq(q);
spin_lock_irq(q->queue_lock);
throtl_release_tgs(td);
* update limits through cgroup and another work got queued, cancel
* it.
*/
- throtl_shutdown_timer_wq(q);
+ throtl_shutdown_wq(q);
throtl_td_free(td);
}
static int __init throtl_init(void)
{
+ kthrotld_workqueue = alloc_workqueue("kthrotld", WQ_MEM_RECLAIM, 0);
+ if (!kthrotld_workqueue)
+ panic("Failed to create kthrotld\n");
+
blkio_policy_register(&blkio_policy_throtl);
return 0;
}