io_uring: cache task struct refs

tctx in submission part is always synchronised because is executed from
the task's context, so we can batch allocate tctx/task references and
store them across syscall boundaries. It avoids enough of operations,
including an atomic for getting task ref and a percpu_counter_add()
function call, which still fallback to spinlock for large batching
cases (around >=32). Should be good for SQPOLL submitting in small
portions and coming at some moment bpf submissions.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/14b327b973410a3eec1f702ecf650e100513aca9.1623634181.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>

1 files changed, 28 insertions, 9 deletions

diff --git a/fs/io_uring.c b/fs/io_uring.c
index bc4d03b15aa6..0ca0282e96b0 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -110,6 +110,8 @@
 				IOSQE_IO_HARDLINK | IOSQE_ASYNC | \
 				IOSQE_BUFFER_SELECT)
 
+#define IO_TCTX_REFS_CACHE_NR	(1U << 10)
+
 struct io_uring {
 	u32 head ____cacheline_aligned_in_smp;
 	u32 tail ____cacheline_aligned_in_smp;
@@ -472,6 +474,7 @@ struct io_ring_ctx {
 
 struct io_uring_task {
 	/* submission side */
+	int			cached_refs;
 	struct xarray		xa;
 	struct wait_queue_head	wait;
 	const struct io_ring_ctx *last;
@@ -6707,16 +6710,23 @@ static const struct io_uring_sqe *io_get_sqe(struct io_ring_ctx *ctx)
 
 static int io_submit_sqes(struct io_ring_ctx *ctx, unsigned int nr)
 {
+	struct io_uring_task *tctx;
 	int submitted = 0;
 
 	/* make sure SQ entry isn't read before tail */
 	nr = min3(nr, ctx->sq_entries, io_sqring_entries(ctx));
-
 	if (!percpu_ref_tryget_many(&ctx->refs, nr))
 		return -EAGAIN;
 
-	percpu_counter_add(&current->io_uring->inflight, nr);
-	refcount_add(nr, &current->usage);
+	tctx = current->io_uring;
+	tctx->cached_refs -= nr;
+	if (unlikely(tctx->cached_refs < 0)) {
+		unsigned int refill = -tctx->cached_refs + IO_TCTX_REFS_CACHE_NR;
+
+		percpu_counter_add(&tctx->inflight, refill);
+		refcount_add(refill, &current->usage);
+		tctx->cached_refs += refill;
+	}
 	io_submit_state_start(&ctx->submit_state, nr);
 
 	while (submitted < nr) {
@@ -6742,12 +6752,10 @@ static int io_submit_sqes(struct io_ring_ctx *ctx, unsigned int nr)
 
 	if (unlikely(submitted != nr)) {
 		int ref_used = (submitted == -EAGAIN) ? 0 : submitted;
-		struct io_uring_task *tctx = current->io_uring;
 		int unused = nr - ref_used;
 
+		current->io_uring->cached_refs += unused;
 		percpu_ref_put_many(&ctx->refs, unused);
-		percpu_counter_sub(&tctx->inflight, unused);
-		put_task_struct_many(current, unused);
 	}
 
 	io_submit_state_end(&ctx->submit_state, ctx);
@@ -7929,7 +7937,7 @@ static int io_uring_alloc_task_context(struct task_struct *task,
 	struct io_uring_task *tctx;
 	int ret;
 
-	tctx = kmalloc(sizeof(*tctx), GFP_KERNEL);
+	tctx = kzalloc(sizeof(*tctx), GFP_KERNEL);
 	if (unlikely(!tctx))
 		return -ENOMEM;
 
@@ -7949,13 +7957,11 @@ static int io_uring_alloc_task_context(struct task_struct *task,
 
 	xa_init(&tctx->xa);
 	init_waitqueue_head(&tctx->wait);
-	tctx->last = NULL;
 	atomic_set(&tctx->in_idle, 0);
 	atomic_set(&tctx->inflight_tracked, 0);
 	task->io_uring = tctx;
 	spin_lock_init(&tctx->task_lock);
 	INIT_WQ_LIST(&tctx->task_list);
-	tctx->task_state = 0;
 	init_task_work(&tctx->task_work, tctx_task_work);
 	return 0;
 }
@@ -7966,6 +7972,7 @@ void __io_uring_free(struct task_struct *tsk)
 
 	WARN_ON_ONCE(!xa_empty(&tctx->xa));
 	WARN_ON_ONCE(tctx->io_wq);
+	WARN_ON_ONCE(tctx->cached_refs);
 
 	percpu_counter_destroy(&tctx->inflight);
 	kfree(tctx);
@@ -9110,6 +9117,16 @@ static void io_uring_try_cancel(bool cancel_all)
 	}
 }
 
+static void io_uring_drop_tctx_refs(struct task_struct *task)
+{
+	struct io_uring_task *tctx = task->io_uring;
+	unsigned int refs = tctx->cached_refs;
+
+	tctx->cached_refs = 0;
+	percpu_counter_sub(&tctx->inflight, refs);
+	put_task_struct_many(task, refs);
+}
+
 /* should only be called by SQPOLL task */
 static void io_uring_cancel_sqpoll(struct io_sq_data *sqd)
 {
@@ -9125,6 +9142,7 @@ static void io_uring_cancel_sqpoll(struct io_sq_data *sqd)
 
 	WARN_ON_ONCE(!sqd || sqd->thread != current);
 
+	io_uring_drop_tctx_refs(current);
 	atomic_inc(&tctx->in_idle);
 	do {
 		/* read completions before cancelations */
@@ -9162,6 +9180,7 @@ void __io_uring_cancel(struct files_struct *files)
 		io_wq_exit_start(tctx->io_wq);
 
 	/* make sure overflow events are dropped */
+	io_uring_drop_tctx_refs(current);
 	atomic_inc(&tctx->in_idle);
 	do {
 		/* read completions before cancelations */