#include <linux/elevator.h>
#include <linux/rbtree.h>
#include <linux/ioprio.h>
+#include <linux/blktrace_api.h>
/*
* tunables
#define CFQ_MIN_TT (2)
#define CFQ_SLICE_SCALE (5)
+#define CFQ_HW_QUEUE_MIN (5)
#define RQ_CIC(rq) \
((struct cfq_io_context *) (rq)->elevator_private)
-#define RQ_CFQQ(rq) ((rq)->elevator_private2)
+#define RQ_CFQQ(rq) (struct cfq_queue *) ((rq)->elevator_private2)
static struct kmem_cache *cfq_pool;
static struct kmem_cache *cfq_ioc_pool;
static DEFINE_PER_CPU(unsigned long, ioc_count);
static struct completion *ioc_gone;
+static DEFINE_SPINLOCK(ioc_gone_lock);
#define CFQ_PRIO_LISTS IOPRIO_BE_NR
#define cfq_class_idle(cfqq) ((cfqq)->ioprio_class == IOPRIO_CLASS_IDLE)
int rq_in_driver;
int sync_flight;
+
+ /*
+ * queue-depth detection
+ */
+ int rq_queued;
int hw_tag;
+ int hw_tag_samples;
+ int rq_in_driver_peak;
/*
* idle window management
unsigned short ioprio, org_ioprio;
unsigned short ioprio_class, org_ioprio_class;
+ pid_t pid;
};
enum cfqq_state_flags {
CFQ_CFQQ_FNS(sync);
#undef CFQ_CFQQ_FNS
+#define cfq_log_cfqq(cfqd, cfqq, fmt, args...) \
+ blk_add_trace_msg((cfqd)->queue, "cfq%d " fmt, (cfqq)->pid, ##args)
+#define cfq_log(cfqd, fmt, args...) \
+ blk_add_trace_msg((cfqd)->queue, "cfq " fmt, ##args)
+
static void cfq_dispatch_insert(struct request_queue *, struct request *);
static struct cfq_queue *cfq_get_queue(struct cfq_data *, int,
struct io_context *, gfp_t);
*/
static inline void cfq_schedule_dispatch(struct cfq_data *cfqd)
{
- if (cfqd->busy_queues)
- kblockd_schedule_work(&cfqd->unplug_work);
+ if (cfqd->busy_queues) {
+ cfq_log(cfqd, "schedule dispatch");
+ kblockd_schedule_work(cfqd->queue, &cfqd->unplug_work);
+ }
}
static int cfq_queue_empty(struct request_queue *q)
cfq_set_prio_slice(struct cfq_data *cfqd, struct cfq_queue *cfqq)
{
cfqq->slice_end = cfq_prio_to_slice(cfqd, cfqq) + jiffies;
+ cfq_log_cfqq(cfqd, cfqq, "set_slice=%lu", cfqq->slice_end - jiffies);
}
/*
*/
static void cfq_add_cfqq_rr(struct cfq_data *cfqd, struct cfq_queue *cfqq)
{
+ cfq_log_cfqq(cfqd, cfqq, "add_to_rr");
BUG_ON(cfq_cfqq_on_rr(cfqq));
cfq_mark_cfqq_on_rr(cfqq);
cfqd->busy_queues++;
*/
static void cfq_del_cfqq_rr(struct cfq_data *cfqd, struct cfq_queue *cfqq)
{
+ cfq_log_cfqq(cfqd, cfqq, "del_from_rr");
BUG_ON(!cfq_cfqq_on_rr(cfqq));
cfq_clear_cfqq_on_rr(cfqq);
struct cfq_data *cfqd = q->elevator->elevator_data;
cfqd->rq_in_driver++;
-
- /*
- * If the depth is larger 1, it really could be queueing. But lets
- * make the mark a little higher - idling could still be good for
- * low queueing, and a low queueing number could also just indicate
- * a SCSI mid layer like behaviour where limit+1 is often seen.
- */
- if (!cfqd->hw_tag && cfqd->rq_in_driver > 4)
- cfqd->hw_tag = 1;
+ cfq_log_cfqq(cfqd, RQ_CFQQ(rq), "activate rq, drv=%d",
+ cfqd->rq_in_driver);
cfqd->last_position = rq->hard_sector + rq->hard_nr_sectors;
}
WARN_ON(!cfqd->rq_in_driver);
cfqd->rq_in_driver--;
+ cfq_log_cfqq(cfqd, RQ_CFQQ(rq), "deactivate rq, drv=%d",
+ cfqd->rq_in_driver);
}
static void cfq_remove_request(struct request *rq)
list_del_init(&rq->queuelist);
cfq_del_rq_rb(rq);
+ cfqq->cfqd->rq_queued--;
if (rq_is_meta(rq)) {
WARN_ON(!cfqq->meta_pending);
cfqq->meta_pending--;
struct cfq_queue *cfqq)
{
if (cfqq) {
+ cfq_log_cfqq(cfqd, cfqq, "set_active");
cfqq->slice_end = 0;
cfq_clear_cfqq_must_alloc_slice(cfqq);
cfq_clear_cfqq_fifo_expire(cfqq);
__cfq_slice_expired(struct cfq_data *cfqd, struct cfq_queue *cfqq,
int timed_out)
{
+ cfq_log_cfqq(cfqd, cfqq, "slice expired t=%d", timed_out);
+
if (cfq_cfqq_wait_request(cfqq))
del_timer(&cfqd->idle_slice_timer);
/*
* store what was left of this slice, if the queue idled/timed out
*/
- if (timed_out && !cfq_cfqq_slice_new(cfqq))
+ if (timed_out && !cfq_cfqq_slice_new(cfqq)) {
cfqq->slice_resid = cfqq->slice_end - jiffies;
+ cfq_log_cfqq(cfqd, cfqq, "resid=%ld", cfqq->slice_resid);
+ }
cfq_resort_rr_list(cfqd, cfqq);
struct cfq_io_context *cic;
unsigned long sl;
+ /*
+ * SSD device without seek penalty, disable idling. But only do so
+ * for devices that support queuing, otherwise we still have a problem
+ * with sync vs async workloads.
+ */
+ if (blk_queue_nonrot(cfqd->queue) && cfqd->hw_tag)
+ return;
+
WARN_ON(!RB_EMPTY_ROOT(&cfqq->sort_list));
WARN_ON(cfq_cfqq_slice_new(cfqq));
if (!cfqd->cfq_slice_idle || !cfq_cfqq_idle_window(cfqq))
return;
+ /*
+ * still requests with the driver, don't idle
+ */
+ if (cfqd->rq_in_driver)
+ return;
+
/*
* task has exited, don't wait
*/
sl = min(sl, msecs_to_jiffies(CFQ_MIN_TT));
mod_timer(&cfqd->idle_slice_timer, jiffies + sl);
+ cfq_log(cfqd, "arm_idle: %lu", sl);
}
/*
struct cfq_data *cfqd = q->elevator->elevator_data;
struct cfq_queue *cfqq = RQ_CFQQ(rq);
+ cfq_log_cfqq(cfqd, cfqq, "dispatch_insert");
+
cfq_remove_request(rq);
cfqq->dispatched++;
elv_dispatch_sort(q, rq);
rq = rq_entry_fifo(cfqq->fifo.next);
if (time_before(jiffies, rq->start_time + cfqd->cfq_fifo_expire[fifo]))
- return NULL;
+ rq = NULL;
+ cfq_log_cfqq(cfqd, cfqq, "fifo=%p", rq);
return rq;
}
BUG_ON(cfqd->busy_queues);
+ cfq_log(cfqd, "forced_dispatch=%d\n", dispatched);
return dispatched;
}
dispatched += __cfq_dispatch_requests(cfqd, cfqq, max_dispatch);
}
+ cfq_log(cfqd, "dispatched=%d", dispatched);
return dispatched;
}
if (!atomic_dec_and_test(&cfqq->ref))
return;
+ cfq_log_cfqq(cfqd, cfqq, "put_queue");
BUG_ON(rb_first(&cfqq->sort_list));
BUG_ON(cfqq->allocated[READ] + cfqq->allocated[WRITE]);
BUG_ON(cfq_cfqq_on_rr(cfqq));
kmem_cache_free(cfq_ioc_pool, cic);
elv_ioc_count_dec(ioc_count);
- if (ioc_gone && !elv_ioc_count_read(ioc_count))
- complete(ioc_gone);
+ if (ioc_gone) {
+ /*
+ * CFQ scheduler is exiting, grab exit lock and check
+ * the pending io context count. If it hits zero,
+ * complete ioc_gone and set it back to NULL
+ */
+ spin_lock(&ioc_gone_lock);
+ if (ioc_gone && !elv_ioc_count_read(ioc_count)) {
+ complete(ioc_gone);
+ ioc_gone = NULL;
+ }
+ spin_unlock(&ioc_gone_lock);
+ }
}
static void cfq_cic_free(struct cfq_io_context *cic)
cfq_mark_cfqq_idle_window(cfqq);
cfq_mark_cfqq_sync(cfqq);
}
+ cfqq->pid = current->pid;
+ cfq_log_cfqq(cfqd, cfqq, "alloced");
}
if (new_cfqq)
cfq_update_idle_window(struct cfq_data *cfqd, struct cfq_queue *cfqq,
struct cfq_io_context *cic)
{
- int enable_idle;
+ int old_idle, enable_idle;
/*
* Don't idle for async or idle io prio class
if (!cfq_cfqq_sync(cfqq) || cfq_class_idle(cfqq))
return;
- enable_idle = cfq_cfqq_idle_window(cfqq);
+ enable_idle = old_idle = cfq_cfqq_idle_window(cfqq);
if (!atomic_read(&cic->ioc->nr_tasks) || !cfqd->cfq_slice_idle ||
(cfqd->hw_tag && CIC_SEEKY(cic)))
enable_idle = 1;
}
- if (enable_idle)
- cfq_mark_cfqq_idle_window(cfqq);
- else
- cfq_clear_cfqq_idle_window(cfqq);
+ if (old_idle != enable_idle) {
+ cfq_log_cfqq(cfqd, cfqq, "idle=%d", enable_idle);
+ if (enable_idle)
+ cfq_mark_cfqq_idle_window(cfqq);
+ else
+ cfq_clear_cfqq_idle_window(cfqq);
+ }
}
/*
*/
static void cfq_preempt_queue(struct cfq_data *cfqd, struct cfq_queue *cfqq)
{
+ cfq_log_cfqq(cfqd, cfqq, "preempt");
cfq_slice_expired(cfqd, 1);
/*
{
struct cfq_io_context *cic = RQ_CIC(rq);
+ cfqd->rq_queued++;
if (rq_is_meta(rq))
cfqq->meta_pending++;
struct cfq_data *cfqd = q->elevator->elevator_data;
struct cfq_queue *cfqq = RQ_CFQQ(rq);
+ cfq_log_cfqq(cfqd, cfqq, "insert_request");
cfq_init_prio_data(cfqq, RQ_CIC(rq)->ioc);
cfq_add_rq_rb(rq);
cfq_rq_enqueued(cfqd, cfqq, rq);
}
+/*
+ * Update hw_tag based on peak queue depth over 50 samples under
+ * sufficient load.
+ */
+static void cfq_update_hw_tag(struct cfq_data *cfqd)
+{
+ if (cfqd->rq_in_driver > cfqd->rq_in_driver_peak)
+ cfqd->rq_in_driver_peak = cfqd->rq_in_driver;
+
+ if (cfqd->rq_queued <= CFQ_HW_QUEUE_MIN &&
+ cfqd->rq_in_driver <= CFQ_HW_QUEUE_MIN)
+ return;
+
+ if (cfqd->hw_tag_samples++ < 50)
+ return;
+
+ if (cfqd->rq_in_driver_peak >= CFQ_HW_QUEUE_MIN)
+ cfqd->hw_tag = 1;
+ else
+ cfqd->hw_tag = 0;
+
+ cfqd->hw_tag_samples = 0;
+ cfqd->rq_in_driver_peak = 0;
+}
+
static void cfq_completed_request(struct request_queue *q, struct request *rq)
{
struct cfq_queue *cfqq = RQ_CFQQ(rq);
unsigned long now;
now = jiffies;
+ cfq_log_cfqq(cfqd, cfqq, "complete");
+
+ cfq_update_hw_tag(cfqd);
WARN_ON(!cfqd->rq_in_driver);
WARN_ON(!cfqq->dispatched);
cfq_schedule_dispatch(cfqd);
spin_unlock_irqrestore(q->queue_lock, flags);
+ cfq_log(cfqd, "set_request fail");
return 1;
}
unsigned long flags;
int timed_out = 1;
+ cfq_log(cfqd, "idle timer fired");
+
spin_lock_irqsave(cfqd->queue->queue_lock, flags);
cfqq = cfqd->active_queue;
cfqd->cfq_slice[1] = cfq_slice_sync;
cfqd->cfq_slice_async_rq = cfq_slice_async_rq;
cfqd->cfq_slice_idle = cfq_slice_idle;
+ cfqd->hw_tag = 1;
return cfqd;
}
* pending RCU callbacks
*/
if (elv_ioc_count_read(ioc_count))
- wait_for_completion(ioc_gone);
+ wait_for_completion(&all_gone);
cfq_slab_kill();
}