[PATCH v12 0/5] io_uring: add napi busy polling support

[PATCH v12 0/5] io_uring: add napi busy polling support

[Stefan Roesch]

This adds the napi busy polling support in io_uring.c. It adds a new
napi_list to the io_ring_ctx structure. This list contains the list of
napi_id's that are currently enabled for busy polling. This list is
used to determine which napi id's enabled busy polling. For faster
access it also adds a hash table.

When a new napi id is added, the hash table is used to locate if
the napi id has already been added. When processing the busy poll
loop the list is used to process the individual elements.

io-uring allows specifying two parameters:
- busy poll timeout and
- prefer busy poll to call of io_napi_busy_loop()
This sets the above parameters for the ring. The settings are passed
with a new structure io_uring_napi.

There is also a corresponding liburing patch series, which enables this
feature. The name of the series is "liburing: add add api for napi busy
poll timeout". It also contains two programs to test the this.

Testing has shown that the round-trip times are reduced to 38us from
55us by enabling napi busy polling with a busy poll timeout of 100us.
More detailled results are part of the commit message of the first
patch.

Changes:
- V12:
  - introduce io_napi_hash_find()
  - use rcu for changes to the hash table
  - use rcu for searching if a napi id is in the napi hash table
  - use rcu hlist functions for adding and removing items from the hash
    table
  - add stale entry detection in __io_napi_do_busy_loop and remove stale
    entries in io_napi_blocking_busy_loop() and io_napi_sqpoll_busy_loop()
  - create io_napi_remove_stale() and __io_napi_remove_stale()
  - __io_napi_do_busy_loop() takes additional loop_end_arg and does stale
    entry detection
  - io_napi_multi_busy_loop is removed. Logic is moved to
    io_napi_blocking_busy_loop()
  - io_napi_free uses rcu function to free
  - io_napi_busy_loop no longer splices
  - io_napi_sqpoll_busy_poll uses rcu
- V11:
  - Fixed long comment lines and whitespace issues
  - Refactor new code io_cqring_wait()
  - Refactor io_napi_adjust_timeout() and remove adjust_timeout
  - Rename io_napi_adjust_timeout to __io_napi_adjust_timeout
  - Add new function io_napi_adjust_timeout
  - Cleanup calls to list_is_singular() in io_napi_multi_busy_loop()
    and io_napi_blocking_busy_loop()
  - Cleanup io_napi_busy_loop_should_end()
  - Rename __io_napi_busy_loop to __io_napi_do_busy_loop() 
- V10:
  - Refreshed to io-uring/for-6.4
  - Repeated performance measurements for 6.4 (same/similar results)
- V9:
  - refreshed to io-uring/for-6.3
  - folded patch 2 and 3 into patch 4
  - fixed commit description for last 2 patches
  - fixed some whitespace issues
  - removed io_napi_busy_loop_on helper
  - removed io_napi_setup_busy helper
  - renamed io_napi_end_busy_loop to io_napi_busy_loop
  - removed NAPI_LIST_HEAD macro
  - split io_napi_blocking_busy_loop into two functions
  - added io_napi function
  - comment for sqpoll check
- V8:
  - added new file napi.c and add napi functions to this file
  - added NAPI_LIST_HEAD function so no ifdef is necessary
  - added io_napi_init and io_napi_free function
  - added io_napi_setup_busy loop helper function
  - added io_napi_adjust_busy_loop helper function
  - added io_napi_end_busy_loop helper function
  - added io_napi_sqpoll_busy_poll helper function
  - some of the definitions in napi.h are macros to avoid ifdef
    definitions in io_uring.c, poll.c and sqpoll.c
  - changed signature of io_napi_add function
  - changed size of hashtable to 16. The number of entries is limited
    by the number of nic queues.
  - Removed ternary in io_napi_blocking_busy_loop
  - Rewrote io_napi_blocking_busy_loop to make it more readable
  - Split off 3 more patches
- V7:
  - allow unregister with NULL value for arg parameter
  - return -EOPNOTSUPP if CONFIG_NET_RX_BUSY_POLL is not enabled
- V6:
  - Add a hash table on top of the list for faster access during the
    add operation. The linked list and the hash table use the same
    data structure
- V5:
  - Refreshed to 6.1-rc6
  - Use copy_from_user instead of memdup/kfree
  - Removed the moving of napi_busy_poll_to
  - Return -EINVAL if any of the reserved or padded fields are not 0.
- V4:
  - Pass structure for napi config, instead of individual parameters
- V3:
  - Refreshed to 6.1-rc5
  - Added a new io-uring api for the prefer napi busy poll api and wire
    it to io_napi_busy_loop().
  - Removed the unregister (implemented as register)
  - Added more performance results to the first commit message.
- V2:
  - Add missing defines if CONFIG_NET_RX_BUSY_POLL is not defined
  - Changes signature of function io_napi_add_list to static inline
    if CONFIG_NET_RX_BUSY_POLL is not defined
  - define some functions as static




Stefan Roesch (5):
  io-uring: move io_wait_queue definition to header file
  io-uring: add napi busy poll support
  io-uring: add sqpoll support for napi busy poll
  io_uring: add register/unregister napi function
  io_uring: add prefer busy poll to register and unregister napi api

 include/linux/io_uring_types.h |  11 ++
 include/uapi/linux/io_uring.h  |  12 ++
 io_uring/Makefile              |   1 +
 io_uring/io_uring.c            |  38 ++--
 io_uring/io_uring.h            |  26 +++
 io_uring/napi.c                | 336 +++++++++++++++++++++++++++++++++
 io_uring/napi.h                |  98 ++++++++++
 io_uring/poll.c                |   2 +
 io_uring/sqpoll.c              |   4 +
 9 files changed, 507 insertions(+), 21 deletions(-)
 create mode 100644 io_uring/napi.c
 create mode 100644 io_uring/napi.h


base-commit: 3c85cc43c8e7855d202da184baf00c7b8eeacf71
-- 
2.39.1

[PATCH v12 1/5] io-uring: move io_wait_queue definition to header file

[Stefan Roesch]

This moves the definition of the io_wait_queue structure to the header
file so it can be also used from other files.

Signed-off-by: Stefan Roesch <shr@devkernel.io>
---
 io_uring/io_uring.c | 21 ---------------------
 io_uring/io_uring.h | 22 ++++++++++++++++++++++
 2 files changed, 22 insertions(+), 21 deletions(-)

diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
index 3d43df8f1e4e..efbd6c9c56e5 100644
--- a/io_uring/io_uring.c
+++ b/io_uring/io_uring.c
@@ -2516,33 +2516,12 @@ int io_submit_sqes(struct io_ring_ctx *ctx, unsigned int nr)
 	return ret;
 }
 
-struct io_wait_queue {
-	struct wait_queue_entry wq;
-	struct io_ring_ctx *ctx;
-	unsigned cq_tail;
-	unsigned nr_timeouts;
-	ktime_t timeout;
-};
-
 static inline bool io_has_work(struct io_ring_ctx *ctx)
 {
 	return test_bit(IO_CHECK_CQ_OVERFLOW_BIT, &ctx->check_cq) ||
 	       !llist_empty(&ctx->work_llist);
 }
 
-static inline bool io_should_wake(struct io_wait_queue *iowq)
-{
-	struct io_ring_ctx *ctx = iowq->ctx;
-	int dist = READ_ONCE(ctx->rings->cq.tail) - (int) iowq->cq_tail;
-
-	/*
-	 * Wake up if we have enough events, or if a timeout occurred since we
-	 * started waiting. For timeouts, we always want to return to userspace,
-	 * regardless of event count.
-	 */
-	return dist >= 0 || atomic_read(&ctx->cq_timeouts) != iowq->nr_timeouts;
-}
-
 static int io_wake_function(struct wait_queue_entry *curr, unsigned int mode,
 			    int wake_flags, void *key)
 {
diff --git a/io_uring/io_uring.h b/io_uring/io_uring.h
index 25515d69d205..5f04bd47562a 100644
--- a/io_uring/io_uring.h
+++ b/io_uring/io_uring.h
@@ -41,6 +41,28 @@ enum {
 	IOU_STOP_MULTISHOT	= -ECANCELED,
 };
 
+struct io_wait_queue {
+	struct wait_queue_entry wq;
+	struct io_ring_ctx *ctx;
+	unsigned cq_tail;
+	unsigned nr_timeouts;
+	ktime_t timeout;
+
+};
+
+static inline bool io_should_wake(struct io_wait_queue *iowq)
+{
+	struct io_ring_ctx *ctx = iowq->ctx;
+	int dist = READ_ONCE(ctx->rings->cq.tail) - (int) iowq->cq_tail;
+
+	/*
+	 * Wake up if we have enough events, or if a timeout occurred since we
+	 * started waiting. For timeouts, we always want to return to userspace,
+	 * regardless of event count.
+	 */
+	return dist >= 0 || atomic_read(&ctx->cq_timeouts) != iowq->nr_timeouts;
+}
+
 struct io_uring_cqe *__io_get_cqe(struct io_ring_ctx *ctx, bool overflow);
 bool io_req_cqe_overflow(struct io_kiocb *req);
 int io_run_task_work_sig(struct io_ring_ctx *ctx);
-- 
2.39.1

[PATCH v12 2/5] io-uring: add napi busy poll support

[Stefan Roesch]

This adds the napi busy polling support in io_uring.c. It adds a new
napi_list to the io_ring_ctx structure. This list contains the list of
napi_id's that are currently enabled for busy polling. The list is
synchronized by the new napi_lock spin lock. The current default napi
busy polling time is stored in napi_busy_poll_to. If napi busy polling
is not enabled, the value is 0.

In addition there is also a hash table. The hash table store the napi
id ond the pointer to the above list nodes. The hash table is used to
speed up the lookup to the list elements. The hash table is synchronized
with rcu.

The NAPI_TIMEOUT is stored as a timeout to make sure that the time a
napi entry is stored in the napi list is limited.

The busy poll timeout is also stored as part of the io_wait_queue. This
is necessary as for sq polling the poll interval needs to be adjusted
and the napi callback allows only to pass in one value.

This has been tested with two simple programs from the liburing library
repository: the napi client and the napi server program. The client
sends a request, which has a timestamp in its payload and the server
replies with the same payload. The client calculates the roundtrip time
and stores it to calcualte the results.

The client is running on host1 and the server is running on host 2 (in
the same rack). The measured times below are roundtrip times. They are
average times over 5 runs each. Each run measures 1 million roundtrips.

                   no rx coal          rx coal: frames=88,usecs=33
Default              57us                    56us

client_poll=100us    47us                    46us

server_poll=100us    51us                    46us

client_poll=100us+   40us                    40us
server_poll=100us

client_poll=100us+   41us                    39us
server_poll=100us+
prefer napi busy poll on client

client_poll=100us+   41us                    39us
server_poll=100us+
prefer napi busy poll on server

client_poll=100us+   41us                    39us
server_poll=100us+
prefer napi busy poll on client + server

Signed-off-by: Stefan Roesch <shr@devkernel.io>
Suggested-by: Olivier Langlois <olivier@trillion01.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
---
 include/linux/io_uring_types.h |  11 ++
 io_uring/Makefile              |   1 +
 io_uring/io_uring.c            |   8 +
 io_uring/io_uring.h            |   4 +
 io_uring/napi.c                | 260 +++++++++++++++++++++++++++++++++
 io_uring/napi.h                |  83 +++++++++++
 io_uring/poll.c                |   2 +
 7 files changed, 369 insertions(+)
 create mode 100644 io_uring/napi.c
 create mode 100644 io_uring/napi.h

diff --git a/include/linux/io_uring_types.h b/include/linux/io_uring_types.h
index 1b2a20a42413..9def1de4eaf4 100644
--- a/include/linux/io_uring_types.h
+++ b/include/linux/io_uring_types.h
@@ -2,6 +2,7 @@
 #define IO_URING_TYPES_H
 
 #include <linux/blkdev.h>
+#include <linux/hashtable.h>
 #include <linux/task_work.h>
 #include <linux/bitmap.h>
 #include <linux/llist.h>
@@ -277,6 +278,16 @@ struct io_ring_ctx {
 	struct xarray		personalities;
 	u32			pers_next;
 
+#ifdef CONFIG_NET_RX_BUSY_POLL
+	struct list_head	napi_list;	/* track busy poll napi_id */
+	spinlock_t		napi_lock;	/* napi_list lock */
+
+	DECLARE_HASHTABLE(napi_ht, 4);
+	/* napi busy poll default timeout */
+	unsigned int		napi_busy_poll_to;
+	bool			napi_prefer_busy_poll;
+#endif
+
 	struct {
 		/*
 		 * We cache a range of free CQEs we can use, once exhausted it
diff --git a/io_uring/Makefile b/io_uring/Makefile
index 8cc8e5387a75..2efe7c5f07ba 100644
--- a/io_uring/Makefile
+++ b/io_uring/Makefile
@@ -9,3 +9,4 @@ obj-$(CONFIG_IO_URING)		+= io_uring.o xattr.o nop.o fs.o splice.o \
 					sqpoll.o fdinfo.o tctx.o poll.o \
 					cancel.o kbuf.o rsrc.o rw.o opdef.o notif.o
 obj-$(CONFIG_IO_WQ)		+= io-wq.o
+obj-$(CONFIG_NET_RX_BUSY_POLL) += napi.o
diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
index efbd6c9c56e5..f06175b36b41 100644
--- a/io_uring/io_uring.c
+++ b/io_uring/io_uring.c
@@ -91,6 +91,7 @@
 #include "rsrc.h"
 #include "cancel.h"
 #include "net.h"
+#include "napi.h"
 #include "notif.h"
 
 #include "timeout.h"
@@ -337,6 +338,8 @@ static __cold struct io_ring_ctx *io_ring_ctx_alloc(struct io_uring_params *p)
 	INIT_WQ_LIST(&ctx->locked_free_list);
 	INIT_DELAYED_WORK(&ctx->fallback_work, io_fallback_req_func);
 	INIT_WQ_LIST(&ctx->submit_state.compl_reqs);
+	io_napi_init(ctx);
+
 	return ctx;
 err:
 	kfree(ctx->dummy_ubuf);
@@ -2619,9 +2622,13 @@ 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());
+		io_napi_adjust_timeout(ctx, &iowq, &ts);
 	}
 
+	io_napi_busy_loop(ctx, &iowq);
+
 	trace_io_uring_cqring_wait(ctx, min_events);
 	do {
 		unsigned long check_cq;
@@ -2856,6 +2863,7 @@ static __cold void io_ring_ctx_free(struct io_ring_ctx *ctx)
 	io_req_caches_free(ctx);
 	if (ctx->hash_map)
 		io_wq_put_hash(ctx->hash_map);
+	io_napi_free(ctx);
 	kfree(ctx->cancel_table.hbs);
 	kfree(ctx->cancel_table_locked.hbs);
 	kfree(ctx->dummy_ubuf);
diff --git a/io_uring/io_uring.h b/io_uring/io_uring.h
index 5f04bd47562a..d669e06f54f0 100644
--- a/io_uring/io_uring.h
+++ b/io_uring/io_uring.h
@@ -48,6 +48,10 @@ struct io_wait_queue {
 	unsigned nr_timeouts;
 	ktime_t timeout;
 
+#ifdef CONFIG_NET_RX_BUSY_POLL
+	unsigned int napi_busy_poll_to;
+	bool napi_prefer_busy_poll;
+#endif
 };
 
 static inline bool io_should_wake(struct io_wait_queue *iowq)
diff --git a/io_uring/napi.c b/io_uring/napi.c
new file mode 100644
index 000000000000..76bcabeb7ef0
--- /dev/null
+++ b/io_uring/napi.c
@@ -0,0 +1,260 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include "io_uring.h"
+#include "napi.h"
+
+#ifdef CONFIG_NET_RX_BUSY_POLL
+
+/* Timeout for cleanout of stale entries. */
+#define NAPI_TIMEOUT		(60 * SEC_CONVERSION)
+
+struct io_napi_entry {
+	unsigned int		napi_id;
+	struct list_head	list;
+
+	unsigned long		timeout;
+	struct hlist_node	node;
+
+	struct rcu_head		rcu;
+};
+
+static struct io_napi_entry *io_napi_hash_find(struct hlist_head *hash_list,
+					       unsigned int napi_id)
+{
+	struct io_napi_entry *e;
+
+	hlist_for_each_entry_rcu(e, hash_list, node) {
+		if (e->napi_id != napi_id)
+			continue;
+		e->timeout = jiffies + NAPI_TIMEOUT;
+		return e;
+	}
+
+	return NULL;
+}
+
+void __io_napi_add(struct io_ring_ctx *ctx, struct file *file)
+{
+	struct hlist_head *hash_list;
+	unsigned int napi_id;
+	struct socket *sock;
+	struct sock *sk;
+	struct io_napi_entry *e;
+
+	sock = sock_from_file(file);
+	if (!sock)
+		return;
+
+	sk = sock->sk;
+	if (!sk)
+		return;
+
+	napi_id = READ_ONCE(sk->sk_napi_id);
+
+	/* Non-NAPI IDs can be rejected. */
+	if (napi_id < MIN_NAPI_ID)
+		return;
+
+	hash_list = &ctx->napi_ht[hash_min(napi_id, HASH_BITS(ctx->napi_ht))];
+
+	rcu_read_lock();
+	e = io_napi_hash_find(hash_list, napi_id);
+	if (e) {
+		e->timeout = jiffies + NAPI_TIMEOUT;
+		rcu_read_unlock();
+		return;
+	}
+	rcu_read_unlock();
+
+	e = kmalloc(sizeof(*e), GFP_NOWAIT);
+	if (!e)
+		return;
+
+	e->napi_id = napi_id;
+	e->timeout = jiffies + NAPI_TIMEOUT;
+
+	spin_lock(&ctx->napi_lock);
+	if (unlikely(io_napi_hash_find(hash_list, napi_id))) {
+		spin_unlock(&ctx->napi_lock);
+		kfree(e);
+		return;
+	}
+
+	hlist_add_tail_rcu(&e->node, hash_list);
+	list_add_tail(&e->list, &ctx->napi_list);
+	spin_unlock(&ctx->napi_lock);
+}
+
+static void __io_napi_remove_stale(struct io_ring_ctx *ctx)
+{
+	struct io_napi_entry *e;
+	unsigned int i;
+
+	spin_lock(&ctx->napi_lock);
+	hash_for_each(ctx->napi_ht, i, e, node) {
+		if (time_after(jiffies, e->timeout)) {
+			list_del(&e->list);
+			hash_del_rcu(&e->node);
+			kfree_rcu(e, rcu);
+		}
+	}
+	spin_unlock(&ctx->napi_lock);
+}
+
+static inline void io_napi_remove_stale(struct io_ring_ctx *ctx, bool is_stale)
+{
+	if (is_stale)
+		__io_napi_remove_stale(ctx);
+}
+
+static inline bool io_napi_busy_loop_timeout(unsigned long start_time,
+					     unsigned long bp_usec)
+{
+	if (bp_usec) {
+		unsigned long end_time = start_time + bp_usec;
+		unsigned long now = busy_loop_current_time();
+
+		return time_after(now, end_time);
+	}
+
+	return true;
+}
+
+static bool io_napi_busy_loop_should_end(void *data,
+					 unsigned long start_time)
+{
+	struct io_wait_queue *iowq = data;
+
+	if (signal_pending(current))
+		return true;
+	if (io_should_wake(iowq))
+		return true;
+	if (io_napi_busy_loop_timeout(start_time, iowq->napi_busy_poll_to))
+		return true;
+
+	return false;
+}
+
+static bool __io_napi_do_busy_loop(struct io_ring_ctx *ctx,
+				   void *loop_end_arg)
+{
+	struct io_napi_entry *e;
+	bool (*loop_end)(void *, unsigned long) = NULL;
+	bool is_stale = false;
+
+	if (loop_end_arg)
+		loop_end = io_napi_busy_loop_should_end;
+
+	list_for_each_entry_rcu(e, &ctx->napi_list, list) {
+		napi_busy_loop(e->napi_id, loop_end, loop_end_arg,
+			       ctx->napi_prefer_busy_poll, BUSY_POLL_BUDGET);
+
+		if (time_after(jiffies, e->timeout))
+			is_stale = true;
+	}
+
+	return is_stale;
+}
+
+static void io_napi_blocking_busy_loop(struct io_ring_ctx *ctx,
+				       struct io_wait_queue *iowq)
+{
+	unsigned long start_time = busy_loop_current_time();
+	void *loop_end_arg = NULL;
+	bool is_stale = false;
+
+	/* Singular lists use a different napi loop end check function and are
+	 * only executed once.
+	 */
+	if (list_is_singular(&ctx->napi_list))
+		loop_end_arg = iowq;
+
+	rcu_read_lock();
+	do {
+		is_stale = __io_napi_do_busy_loop(ctx, loop_end_arg);
+	} while (!io_napi_busy_loop_should_end(iowq, start_time) && !loop_end_arg);
+	rcu_read_unlock();
+
+	io_napi_remove_stale(ctx, is_stale);
+}
+
+/*
+ * io_napi_init() - Init napi settings
+ * @ctx: pointer to io-uring context structure
+ *
+ * Init napi settings in the io-uring context.
+ */
+void io_napi_init(struct io_ring_ctx *ctx)
+{
+	INIT_LIST_HEAD(&ctx->napi_list);
+	spin_lock_init(&ctx->napi_lock);
+	ctx->napi_prefer_busy_poll = false;
+	ctx->napi_busy_poll_to = READ_ONCE(sysctl_net_busy_poll);
+}
+
+/*
+ * io_napi_free() - Deallocate napi
+ * @ctx: pointer to io-uring context structure
+ *
+ * Free the napi list and the hash table in the io-uring context.
+ */
+void io_napi_free(struct io_ring_ctx *ctx)
+{
+	struct io_napi_entry *e;
+	LIST_HEAD(napi_list);
+	unsigned int i;
+
+	spin_lock(&ctx->napi_lock);
+	hash_for_each(ctx->napi_ht, i, e, node) {
+		hash_del_rcu(&e->node);
+		kfree_rcu(e, rcu);
+	}
+	spin_unlock(&ctx->napi_lock);
+}
+
+/*
+ * __io_napi_adjust_timeout() - Add napi id to the busy poll list
+ * @ctx: pointer to io-uring context structure
+ * @iowq: pointer to io wait queue
+ * @ts: pointer to timespec or NULL
+ *
+ * Adjust the busy loop timeout according to timespec and busy poll timeout.
+ */
+void __io_napi_adjust_timeout(struct io_ring_ctx *ctx, struct io_wait_queue *iowq,
+			      struct timespec64 *ts)
+{
+	unsigned int poll_to = READ_ONCE(ctx->napi_busy_poll_to);
+
+	if (ts) {
+		struct timespec64 poll_to_ts = ns_to_timespec64(1000 * (s64)poll_to);
+
+		if (timespec64_compare(ts, &poll_to_ts) > 0) {
+			*ts = timespec64_sub(*ts, poll_to_ts);
+		} else {
+			u64 to = timespec64_to_ns(ts);
+
+			do_div(to, 1000);
+			ts->tv_sec = 0;
+			ts->tv_nsec = 0;
+		}
+	}
+
+	iowq->napi_busy_poll_to = poll_to;
+}
+
+/*
+ * __io_napi_busy_loop() - execute busy poll loop
+ * @ctx: pointer to io-uring context structure
+ * @iowq: pointer to io wait queue
+ *
+ * Execute the busy poll loop and merge the spliced off list.
+ */
+void __io_napi_busy_loop(struct io_ring_ctx *ctx, struct io_wait_queue *iowq)
+{
+	iowq->napi_prefer_busy_poll = READ_ONCE(ctx->napi_prefer_busy_poll);
+
+	if (!(ctx->flags & IORING_SETUP_SQPOLL) && iowq->napi_busy_poll_to)
+		io_napi_blocking_busy_loop(ctx, iowq);
+}
+
+#endif
diff --git a/io_uring/napi.h b/io_uring/napi.h
new file mode 100644
index 000000000000..1bdf8442081f
--- /dev/null
+++ b/io_uring/napi.h
@@ -0,0 +1,83 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+
+#ifndef IOU_NAPI_H
+#define IOU_NAPI_H
+
+#include <linux/kernel.h>
+#include <linux/io_uring.h>
+#include <net/busy_poll.h>
+
+#ifdef CONFIG_NET_RX_BUSY_POLL
+
+void io_napi_init(struct io_ring_ctx *ctx);
+void io_napi_free(struct io_ring_ctx *ctx);
+
+void __io_napi_add(struct io_ring_ctx *ctx, struct file *file);
+
+void __io_napi_adjust_timeout(struct io_ring_ctx *ctx,
+		struct io_wait_queue *iowq, struct timespec64 *ts);
+void __io_napi_busy_loop(struct io_ring_ctx *ctx, struct io_wait_queue *iowq);
+
+static inline bool io_napi(struct io_ring_ctx *ctx)
+{
+	return !list_empty(&ctx->napi_list);
+}
+
+static inline void io_napi_adjust_timeout(struct io_ring_ctx *ctx,
+					  struct io_wait_queue *iowq,
+					  struct timespec64 *ts)
+{
+	if (!io_napi(ctx))
+		return;
+	__io_napi_adjust_timeout(ctx, iowq, ts);
+}
+
+static inline void io_napi_busy_loop(struct io_ring_ctx *ctx,
+				     struct io_wait_queue *iowq)
+{
+	if (!io_napi(ctx))
+		return;
+	__io_napi_busy_loop(ctx, iowq);
+}
+
+/*
+ * io_napi_add() - Add napi id to the busy poll list
+ * @req: pointer to io_kiocb request
+ *
+ * Add the napi id of the socket to the napi busy poll list and hash table.
+ */
+static inline void io_napi_add(struct io_kiocb *req)
+{
+	struct io_ring_ctx *ctx = req->ctx;
+
+	if (!READ_ONCE(ctx->napi_busy_poll_to))
+		return;
+
+	__io_napi_add(ctx, req->file);
+}
+
+#else
+
+static inline void io_napi_init(struct io_ring_ctx *ctx)
+{
+}
+
+static inline void io_napi_free(struct io_ring_ctx *ctx)
+{
+}
+
+static inline bool io_napi(struct io_ring_ctx *ctx)
+{
+	return false;
+}
+
+static inline void io_napi_add(struct io_kiocb *req)
+{
+}
+
+#define io_napi_adjust_timeout(ctx, iowq, ts) do {} while (0)
+#define io_napi_busy_loop(ctx, iowq) do {} while (0)
+
+#endif
+
+#endif
diff --git a/io_uring/poll.c b/io_uring/poll.c
index c90e47dc1e29..0284849793bb 100644
--- a/io_uring/poll.c
+++ b/io_uring/poll.c
@@ -15,6 +15,7 @@
 
 #include "io_uring.h"
 #include "refs.h"
+#include "napi.h"
 #include "opdef.h"
 #include "kbuf.h"
 #include "poll.h"
@@ -631,6 +632,7 @@ static int __io_arm_poll_handler(struct io_kiocb *req,
 		__io_poll_execute(req, mask);
 		return 0;
 	}
+	io_napi_add(req);
 
 	if (ipt->owning) {
 		/*
-- 
2.39.1

[PATCH v12 3/5] io-uring: add sqpoll support for napi busy poll

[Stefan Roesch]

This adds the sqpoll support to the io-uring napi.

Signed-off-by: Stefan Roesch <shr@devkernel.io>
Suggested-by: Olivier Langlois <olivier@trillion01.com>
---
 io_uring/napi.c   | 24 ++++++++++++++++++++++++
 io_uring/napi.h   |  2 ++
 io_uring/sqpoll.c |  4 ++++
 3 files changed, 30 insertions(+)

diff --git a/io_uring/napi.c b/io_uring/napi.c
index 76bcabeb7ef0..491d028e8146 100644
--- a/io_uring/napi.c
+++ b/io_uring/napi.c
@@ -257,4 +257,28 @@ void __io_napi_busy_loop(struct io_ring_ctx *ctx, struct io_wait_queue *iowq)
 		io_napi_blocking_busy_loop(ctx, iowq);
 }
 
+/*
+ * io_napi_sqpoll_busy_poll() - busy poll loop for sqpoll
+ * @ctx: pointer to io-uring context structure
+ *
+ * Splice of the napi list and execute the napi busy poll loop.
+ */
+int io_napi_sqpoll_busy_poll(struct io_ring_ctx *ctx)
+{
+	LIST_HEAD(napi_list);
+	bool is_stale = false;
+
+	if (!READ_ONCE(ctx->napi_busy_poll_to))
+		return 0;
+	if (list_empty_careful(&ctx->napi_list))
+		return 0;
+
+	rcu_read_lock();
+	is_stale = __io_napi_do_busy_loop(ctx, NULL);
+	rcu_read_unlock();
+
+	io_napi_remove_stale(ctx, is_stale);
+	return 1;
+}
+
 #endif
diff --git a/io_uring/napi.h b/io_uring/napi.h
index 1bdf8442081f..0484dfd9a396 100644
--- a/io_uring/napi.h
+++ b/io_uring/napi.h
@@ -17,6 +17,7 @@ void __io_napi_add(struct io_ring_ctx *ctx, struct file *file);
 void __io_napi_adjust_timeout(struct io_ring_ctx *ctx,
 		struct io_wait_queue *iowq, struct timespec64 *ts);
 void __io_napi_busy_loop(struct io_ring_ctx *ctx, struct io_wait_queue *iowq);
+int io_napi_sqpoll_busy_poll(struct io_ring_ctx *ctx);
 
 static inline bool io_napi(struct io_ring_ctx *ctx)
 {
@@ -77,6 +78,7 @@ static inline void io_napi_add(struct io_kiocb *req)
 
 #define io_napi_adjust_timeout(ctx, iowq, ts) do {} while (0)
 #define io_napi_busy_loop(ctx, iowq) do {} while (0)
+#define io_napi_sqpoll_busy_poll(ctx) (0)
 
 #endif
 
diff --git a/io_uring/sqpoll.c b/io_uring/sqpoll.c
index 9db4bc1f521a..0c8d53ef134a 100644
--- a/io_uring/sqpoll.c
+++ b/io_uring/sqpoll.c
@@ -15,6 +15,7 @@
 #include <uapi/linux/io_uring.h>
 
 #include "io_uring.h"
+#include "napi.h"
 #include "sqpoll.h"
 
 #define IORING_SQPOLL_CAP_ENTRIES_VALUE 8
@@ -193,6 +194,9 @@ static int __io_sq_thread(struct io_ring_ctx *ctx, bool cap_entries)
 			ret = io_submit_sqes(ctx, to_submit);
 		mutex_unlock(&ctx->uring_lock);
 
+		if (io_napi(ctx))
+			ret += io_napi_sqpoll_busy_poll(ctx);
+
 		if (to_submit && wq_has_sleeper(&ctx->sqo_sq_wait))
 			wake_up(&ctx->sqo_sq_wait);
 		if (creds)
-- 
2.39.1

[PATCH v12 4/5] io_uring: add register/unregister napi function

[Stefan Roesch]

This adds an api to register and unregister the napi for io-uring. If
the arg value is specified when unregistering, the current napi setting
for the busy poll timeout is copied into the user structure. If this is
not required, NULL can be passed as the arg value.

Signed-off-by: Stefan Roesch <shr@devkernel.io>
Acked-by: Jakub Kicinski <kuba@kernel.org>
---
 include/uapi/linux/io_uring.h | 11 ++++++++
 io_uring/io_uring.c           |  9 +++++++
 io_uring/napi.c               | 48 +++++++++++++++++++++++++++++++++++
 io_uring/napi.h               | 13 ++++++++++
 4 files changed, 81 insertions(+)

diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
index 0716cb17e436..278c1a9de78c 100644
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -523,6 +523,10 @@ enum {
 	/* register a range of fixed file slots for automatic slot allocation */
 	IORING_REGISTER_FILE_ALLOC_RANGE	= 25,
 
+	/* set/clear busy poll settings */
+	IORING_REGISTER_NAPI			= 26,
+	IORING_UNREGISTER_NAPI			= 27,
+
 	/* this goes last */
 	IORING_REGISTER_LAST,
 
@@ -649,6 +653,13 @@ struct io_uring_buf_reg {
 	__u64	resv[3];
 };
 
+/* argument for IORING_(UN)REGISTER_NAPI */
+struct io_uring_napi {
+	__u32	busy_poll_to;
+	__u32	pad;
+	__u64	resv;
+};
+
 /*
  * io_uring_restriction->opcode values
  */
diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
index f06175b36b41..66e4591fbe2b 100644
--- a/io_uring/io_uring.c
+++ b/io_uring/io_uring.c
@@ -4405,6 +4405,15 @@ static int __io_uring_register(struct io_ring_ctx *ctx, unsigned opcode,
 			break;
 		ret = io_register_file_alloc_range(ctx, arg);
 		break;
+	case IORING_REGISTER_NAPI:
+		ret = -EINVAL;
+		if (!arg)
+			break;
+		ret = io_register_napi(ctx, arg);
+		break;
+	case IORING_UNREGISTER_NAPI:
+		ret = io_unregister_napi(ctx, arg);
+		break;
 	default:
 		ret = -EINVAL;
 		break;
diff --git a/io_uring/napi.c b/io_uring/napi.c
index 491d028e8146..fa531949fc6f 100644
--- a/io_uring/napi.c
+++ b/io_uring/napi.c
@@ -212,6 +212,54 @@ void io_napi_free(struct io_ring_ctx *ctx)
 	spin_unlock(&ctx->napi_lock);
 }
 
+/*
+ * io_napi_register() - Register napi with io-uring
+ * @ctx: pointer to io-uring context structure
+ * @arg: pointer to io_uring_napi structure
+ *
+ * Register napi in the io-uring context.
+ */
+int io_register_napi(struct io_ring_ctx *ctx, void __user *arg)
+{
+	const struct io_uring_napi curr = {
+		.busy_poll_to = ctx->napi_busy_poll_to,
+	};
+	struct io_uring_napi napi;
+
+	if (copy_from_user(&napi, arg, sizeof(napi)))
+		return -EFAULT;
+	if (napi.pad || napi.resv)
+		return -EINVAL;
+
+	WRITE_ONCE(ctx->napi_busy_poll_to, napi.busy_poll_to);
+
+	if (copy_to_user(arg, &curr, sizeof(curr)))
+		return -EFAULT;
+
+	return 0;
+}
+
+/*
+ * io_napi_unregister() - Unregister napi with io-uring
+ * @ctx: pointer to io-uring context structure
+ * @arg: pointer to io_uring_napi structure
+ *
+ * Unregister napi. If arg has been specified copy the busy poll timeout and
+ * prefer busy poll setting to the passed in structure.
+ */
+int io_unregister_napi(struct io_ring_ctx *ctx, void __user *arg)
+{
+	const struct io_uring_napi curr = {
+		.busy_poll_to = ctx->napi_busy_poll_to,
+	};
+
+	if (arg && copy_to_user(arg, &curr, sizeof(curr)))
+		return -EFAULT;
+
+	WRITE_ONCE(ctx->napi_busy_poll_to, 0);
+	return 0;
+}
+
 /*
  * __io_napi_adjust_timeout() - Add napi id to the busy poll list
  * @ctx: pointer to io-uring context structure
diff --git a/io_uring/napi.h b/io_uring/napi.h
index 0484dfd9a396..b5e93b3777c0 100644
--- a/io_uring/napi.h
+++ b/io_uring/napi.h
@@ -12,6 +12,9 @@
 void io_napi_init(struct io_ring_ctx *ctx);
 void io_napi_free(struct io_ring_ctx *ctx);
 
+int io_register_napi(struct io_ring_ctx *ctx, void __user *arg);
+int io_unregister_napi(struct io_ring_ctx *ctx, void __user *arg);
+
 void __io_napi_add(struct io_ring_ctx *ctx, struct file *file);
 
 void __io_napi_adjust_timeout(struct io_ring_ctx *ctx,
@@ -67,6 +70,16 @@ static inline void io_napi_free(struct io_ring_ctx *ctx)
 {
 }
 
+static inline int io_register_napi(struct io_ring_ctx *ctx, void __user *arg)
+{
+	return -EOPNOTSUPP;
+}
+
+static inline int io_unregister_napi(struct io_ring_ctx *ctx, void __user *arg)
+{
+	return -EOPNOTSUPP;
+}
+
 static inline bool io_napi(struct io_ring_ctx *ctx)
 {
 	return false;
-- 
2.39.1

[PATCH v12 5/5] io_uring: add prefer busy poll to register and unregister napi api

[Stefan Roesch]

This adds the napi prefer busy poll setting to the register and
unregister napi api. When napi is unregistered and arg is specified,
both napi settings: busy poll timeout and the prefer busy poll setting
are copied into the user structure.

Signed-off-by: Stefan Roesch <shr@devkernel.io>
Acked-by: Jakub Kicinski <kuba@kernel.org>
---
 include/uapi/linux/io_uring.h |  3 ++-
 io_uring/napi.c               | 10 +++++++---
 2 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
index 278c1a9de78c..16d17d6ab7f7 100644
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -656,7 +656,8 @@ struct io_uring_buf_reg {
 /* argument for IORING_(UN)REGISTER_NAPI */
 struct io_uring_napi {
 	__u32	busy_poll_to;
-	__u32	pad;
+	__u8	prefer_busy_poll;
+	__u8	pad[3];
 	__u64	resv;
 };
 
diff --git a/io_uring/napi.c b/io_uring/napi.c
index fa531949fc6f..8b5df89fbb2c 100644
--- a/io_uring/napi.c
+++ b/io_uring/napi.c
@@ -222,16 +222,18 @@ void io_napi_free(struct io_ring_ctx *ctx)
 int io_register_napi(struct io_ring_ctx *ctx, void __user *arg)
 {
 	const struct io_uring_napi curr = {
-		.busy_poll_to = ctx->napi_busy_poll_to,
+		.busy_poll_to 	  = ctx->napi_busy_poll_to,
+		.prefer_busy_poll = ctx->napi_prefer_busy_poll
 	};
 	struct io_uring_napi napi;
 
 	if (copy_from_user(&napi, arg, sizeof(napi)))
 		return -EFAULT;
-	if (napi.pad || napi.resv)
+	if (napi.pad[0] || napi.pad[1] || napi.pad[2] || napi.resv)
 		return -EINVAL;
 
 	WRITE_ONCE(ctx->napi_busy_poll_to, napi.busy_poll_to);
+	WRITE_ONCE(ctx->napi_prefer_busy_poll, !!napi.prefer_busy_poll);
 
 	if (copy_to_user(arg, &curr, sizeof(curr)))
 		return -EFAULT;
@@ -250,13 +252,15 @@ int io_register_napi(struct io_ring_ctx *ctx, void __user *arg)
 int io_unregister_napi(struct io_ring_ctx *ctx, void __user *arg)
 {
 	const struct io_uring_napi curr = {
-		.busy_poll_to = ctx->napi_busy_poll_to,
+		.busy_poll_to 	  = ctx->napi_busy_poll_to,
+		.prefer_busy_poll = ctx->napi_prefer_busy_poll
 	};
 
 	if (arg && copy_to_user(arg, &curr, sizeof(curr)))
 		return -EFAULT;
 
 	WRITE_ONCE(ctx->napi_busy_poll_to, 0);
+	WRITE_ONCE(ctx->napi_prefer_busy_poll, false);
 	return 0;
 }
 
-- 
2.39.1

Re: [PATCH v12 2/5] io-uring: add napi busy poll support

[Jens Axboe]

On 5/2/23 10:53?AM, Stefan Roesch wrote:
> diff --git a/io_uring/napi.c b/io_uring/napi.c
> new file mode 100644
> index 000000000000..76bcabeb7ef0
> --- /dev/null
> +++ b/io_uring/napi.c
> +static bool io_napi_busy_loop_should_end(void *data,
> +					 unsigned long start_time)
> +{
> +	struct io_wait_queue *iowq = data;
> +
> +	if (signal_pending(current))
> +		return true;
> +	if (io_should_wake(iowq))
> +		return true;
> +	if (io_napi_busy_loop_timeout(start_time, iowq->napi_busy_poll_to))
> +		return true;
> +
> +	return false;
> +}
> +
> +static bool __io_napi_do_busy_loop(struct io_ring_ctx *ctx,
> +				   void *loop_end_arg)
> +{
> +	struct io_napi_entry *e;
> +	bool (*loop_end)(void *, unsigned long) = NULL;
> +	bool is_stale = false;
> +
> +	if (loop_end_arg)
> +		loop_end = io_napi_busy_loop_should_end;
> +
> +	list_for_each_entry_rcu(e, &ctx->napi_list, list) {
> +		napi_busy_loop(e->napi_id, loop_end, loop_end_arg,
> +			       ctx->napi_prefer_busy_poll, BUSY_POLL_BUDGET);
> +
> +		if (time_after(jiffies, e->timeout))
> +			is_stale = true;
> +	}
> +
> +	return is_stale;
> +}
> +
> +static void io_napi_blocking_busy_loop(struct io_ring_ctx *ctx,
> +				       struct io_wait_queue *iowq)
> +{
> +	unsigned long start_time = busy_loop_current_time();
> +	void *loop_end_arg = NULL;
> +	bool is_stale = false;
> +
> +	/* Singular lists use a different napi loop end check function and are
> +	 * only executed once.
> +	 */
> +	if (list_is_singular(&ctx->napi_list))
> +		loop_end_arg = iowq;
> +
> +	rcu_read_lock();
> +	do {
> +		is_stale = __io_napi_do_busy_loop(ctx, loop_end_arg);
> +	} while (!io_napi_busy_loop_should_end(iowq, start_time) && !loop_end_arg);
> +	rcu_read_unlock();
> +
> +	io_napi_remove_stale(ctx, is_stale);
> +}

I mentioned this in our out-of-band discussions on this patch set, and
we cannot call napi_busy_loop() under rcu_read_lock() if loop_end and
loop_end_arg is set AND loop_end() doesn't always return true. Because
otherwise we can end up with napi_busy_loop() doing:

	if (unlikely(need_resched())) {
		if (napi_poll)
			busy_poll_stop(napi, have_poll_lock, prefer_busy_poll, budget);
		preempt_enable();
		rcu_read_unlock();
		cond_resched();
		if (loop_end(loop_end_arg, start_time))
			return;
		goto restart;
	}

and hence we're now scheduling with rcu read locking disabled. So we
need to handle that case appropriately as well.

-- 
Jens Axboe

Re: [PATCH v12 2/5] io-uring: add napi busy poll support

[Stefan Roesch]

Jens Axboe <axboe@kernel.dk> writes:

>
> I mentioned this in our out-of-band discussions on this patch set, and
> we cannot call napi_busy_loop() under rcu_read_lock() if loop_end and
> loop_end_arg is set AND loop_end() doesn't always return true. Because
> otherwise we can end up with napi_busy_loop() doing:
>
> 	if (unlikely(need_resched())) {
> 		if (napi_poll)
> 			busy_poll_stop(napi, have_poll_lock, prefer_busy_poll, budget);
> 		preempt_enable();
> 		rcu_read_unlock();
> 		cond_resched();
> 		if (loop_end(loop_end_arg, start_time))
> 			return;
> 		goto restart;
> 	}
>
> and hence we're now scheduling with rcu read locking disabled. So we
> need to handle that case appropriately as well.

I'll have a look.