@@ -2529,12 +2529,65 @@ static enum hrtimer_restart io_cqring_timer_wakeup(struct hrtimer *timer)
return HRTIMER_NORESTART;
}
-static int io_cqring_schedule_timeout(struct io_wait_queue *iowq)
+/*
+ * Doing min_timeout portion. If we saw any timeouts, events, or have work,
+ * wake up. If not, and we have a normal timeout, switch to that and keep
+ * sleeping.
+ */
+static enum hrtimer_restart io_cqring_min_timer_wakeup(struct hrtimer *timer)
{
+ struct io_wait_queue *iowq = container_of(timer, struct io_wait_queue, t);
+ struct io_ring_ctx *ctx = iowq->ctx;
+ ktime_t timeout;
+
+ /* no general timeout, or shorter, we are done */
+ if (iowq->timeout == KTIME_MAX ||
+ ktime_after(iowq->min_timeout, iowq->timeout))
+ goto out_wake;
+ /* work we may need to run, wake function will see if we need to wake */
+ if (io_has_work(ctx))
+ goto out_wake;
+ /* got events since we started waiting, min timeout is done */
+ if (iowq->cq_min_tail != READ_ONCE(ctx->rings->cq.tail))
+ goto out_wake;
+ /* if we have any events and min timeout expired, we're done */
+ if (io_cqring_events(ctx))
+ goto out_wake;
+
+ /*
+ * If using deferred task_work running and application is waiting on
+ * more than one request, ensure we reset it now where we are switching
+ * to normal sleeps. Any request completion post min_wait should wake
+ * the task and return.
+ */
+ if (ctx->flags & IORING_SETUP_DEFER_TASKRUN)
+ atomic_set(&ctx->cq_wait_nr, 1);
+
+ timeout = ktime_sub_ns(iowq->timeout, iowq->min_timeout);
+ iowq->t.function = io_cqring_timer_wakeup;
+ hrtimer_set_expires(timer, ktime_add_ns(timeout, ktime_get_ns()));
+ return HRTIMER_RESTART;
+out_wake:
+ return io_cqring_timer_wakeup(timer);
+}
+
+static int io_cqring_schedule_timeout(struct io_wait_queue *iowq,
+ ktime_t start_time)
+{
+ ktime_t timeout;
+
iowq->hit_timeout = 0;
hrtimer_init_on_stack(&iowq->t, CLOCK_MONOTONIC, HRTIMER_MODE_ABS);
- iowq->t.function = io_cqring_timer_wakeup;
- hrtimer_set_expires_range_ns(&iowq->t, iowq->timeout, 0);
+
+ if (iowq->min_timeout != KTIME_MAX) {
+ timeout = ktime_add_ns(iowq->min_timeout, start_time);
+ iowq->t.function = io_cqring_min_timer_wakeup;
+ } else {
+ timeout = ktime_add_ns(iowq->timeout, start_time);
+ iowq->t.function = io_cqring_timer_wakeup;
+ }
+
+ hrtimer_set_expires_range_ns(&iowq->t, timeout, 0);
hrtimer_start_expires(&iowq->t, HRTIMER_MODE_ABS);
if (!READ_ONCE(iowq->hit_timeout))
@@ -2548,7 +2601,8 @@ static int io_cqring_schedule_timeout(struct io_wait_queue *iowq)
}
static int __io_cqring_wait_schedule(struct io_ring_ctx *ctx,
- struct io_wait_queue *iowq)
+ struct io_wait_queue *iowq,
+ ktime_t start_time)
{
int io_wait, ret = 0;
@@ -2560,8 +2614,8 @@ static int __io_cqring_wait_schedule(struct io_ring_ctx *ctx,
io_wait = current->in_iowait;
if (current_pending_io())
current->in_iowait = 1;
- if (iowq->timeout != KTIME_MAX)
- ret = io_cqring_schedule_timeout(iowq);
+ if (iowq->timeout != KTIME_MAX || iowq->min_timeout != KTIME_MAX)
+ ret = io_cqring_schedule_timeout(iowq, start_time);
else
schedule();
current->in_iowait = io_wait;
@@ -2570,7 +2624,8 @@ static int __io_cqring_wait_schedule(struct io_ring_ctx *ctx,
/* when returns >0, the caller should retry */
static inline int io_cqring_wait_schedule(struct io_ring_ctx *ctx,
- struct io_wait_queue *iowq)
+ struct io_wait_queue *iowq,
+ ktime_t start_time)
{
if (unlikely(READ_ONCE(ctx->check_cq)))
return 1;
@@ -2583,7 +2638,7 @@ static inline int io_cqring_wait_schedule(struct io_ring_ctx *ctx,
if (unlikely(io_should_wake(iowq)))
return 0;
- return __io_cqring_wait_schedule(ctx, iowq);
+ return __io_cqring_wait_schedule(ctx, iowq, start_time);
}
/*
@@ -2596,6 +2651,7 @@ static int io_cqring_wait(struct io_ring_ctx *ctx, int min_events,
{
struct io_wait_queue iowq;
struct io_rings *rings = ctx->rings;
+ ktime_t start_time;
int ret;
if (!io_allowed_run_tw(ctx))
@@ -2626,8 +2682,11 @@ static int io_cqring_wait(struct io_ring_ctx *ctx, int min_events,
INIT_LIST_HEAD(&iowq.wq.entry);
iowq.ctx = ctx;
iowq.nr_timeouts = atomic_read(&ctx->cq_timeouts);
+ iowq.cq_min_tail = READ_ONCE(ctx->rings->cq.tail);
iowq.cq_tail = READ_ONCE(ctx->rings->cq.head) + min_events;
+ iowq.min_timeout = KTIME_MAX;
iowq.timeout = KTIME_MAX;
+ start_time = ktime_get_ns();
if (uts) {
struct timespec64 ts;
@@ -2635,7 +2694,7 @@ static int io_cqring_wait(struct io_ring_ctx *ctx, int min_events,
if (get_timespec64(&ts, uts))
return -EFAULT;
- iowq.timeout = ktime_add_ns(timespec64_to_ktime(ts), ktime_get_ns());
+ iowq.timeout = timespec64_to_ktime(ts);
io_napi_adjust_timeout(ctx, &iowq, &ts);
}
@@ -2654,7 +2713,7 @@ static int io_cqring_wait(struct io_ring_ctx *ctx, int min_events,
TASK_INTERRUPTIBLE);
}
- ret = io_cqring_wait_schedule(ctx, &iowq);
+ ret = io_cqring_wait_schedule(ctx, &iowq, start_time);
__set_current_state(TASK_RUNNING);
atomic_set(&ctx->cq_wait_nr, IO_CQ_WAKE_INIT);
@@ -39,8 +39,10 @@ struct io_wait_queue {
struct wait_queue_entry wq;
struct io_ring_ctx *ctx;
unsigned cq_tail;
+ unsigned cq_min_tail;
unsigned nr_timeouts;
int hit_timeout;
+ ktime_t min_timeout;
ktime_t timeout;
struct hrtimer t;
Waiting for events with io_uring has two knobs that can be set: 1) The number of events to wake for 2) The timeout associated with the event Waiting will abort when either of those conditions are met, as expected. This adds support for a third event, which is associated with the number of events to wait for. Applications generally like to handle batches of completions, and right now they'd set a number of events to wait for and the timeout for that. If no events have been received but the timeout triggers, control is returned to the application and it can wait again. However, if the application doesn't have anything to do until events are reaped, then it's possible to make this waiting more efficient. For example, the application may have a latency time of 50 usecs and wanting to handle a batch of 8 requests at the time. If it uses 50 usecs as the timeout, then it'll be doing 20K context switches per second even if nothing is happening. This introduces the notion of min batch wait time. If the min batch wait time expires, then we'll return to userspace if we have any events at all. If none are available, the general wait time is applied. Any request arriving after the min batch wait time will cause waiting to stop and return control to the application. Signed-off-by: Jens Axboe <axboe@kernel.dk> --- io_uring/io_uring.c | 79 +++++++++++++++++++++++++++++++++++++++------ io_uring/io_uring.h | 2 ++ 2 files changed, 71 insertions(+), 10 deletions(-)