Merge tag 'for-6.2/io_uring-2022-12-08' of git://git.kernel.dk/linux

Pull io_uring updates from Jens Axboe:

 - Always ensure proper ordering in case of CQ ring overflow, which then
   means we can remove some work-arounds for that (Dylan)

 - Support completion batching for multishot, greatly increasing the
   efficiency for those (Dylan)

 - Flag epoll/eventfd wakeups done from io_uring, so that we can easily
   tell if we're recursing into io_uring again.

   Previously, this would have resulted in repeated multishot
   notifications if we had a dependency there. That could happen if an
   eventfd was registered as the ring eventfd, and we multishot polled
   for events on it. Or if an io_uring fd was added to epoll, and
   io_uring had a multishot request for the epoll fd.

   Test cases here:
	https://git.kernel.dk/cgit/liburing/commit/?id=919755a7d0096fda08fb6d65ac54ad8d0fe027cd

   Previously these got terminated when the CQ ring eventually
   overflowed, now it's handled gracefully (me).

 - Tightening of the IOPOLL based completions (Pavel)

 - Optimizations of the networking zero-copy paths (Pavel)

 - Various tweaks and fixes (Dylan, Pavel)

* tag 'for-6.2/io_uring-2022-12-08' of git://git.kernel.dk/linux: (41 commits)
  io_uring: keep unlock_post inlined in hot path
  io_uring: don't use complete_post in kbuf
  io_uring: spelling fix
  io_uring: remove io_req_complete_post_tw
  io_uring: allow multishot polled reqs to defer completion
  io_uring: remove overflow param from io_post_aux_cqe
  io_uring: add lockdep assertion in io_fill_cqe_aux
  io_uring: make io_fill_cqe_aux static
  io_uring: add io_aux_cqe which allows deferred completion
  io_uring: allow defer completion for aux posted cqes
  io_uring: defer all io_req_complete_failed
  io_uring: always lock in io_apoll_task_func
  io_uring: remove iopoll spinlock
  io_uring: iopoll protect complete_post
  io_uring: inline __io_req_complete_put()
  io_uring: remove io_req_tw_post_queue
  io_uring: use io_req_task_complete() in timeout
  io_uring: hold locks for io_req_complete_failed
  io_uring: add completion locking for iopoll
  io_uring: kill io_cqring_ev_posted() and __io_cq_unlock_post()
  ...

Author: torvalds

fs/eventfd.c

@@ -43,21 +43,7 @@ struct eventfd_ctx {
 	int id;
 };
 
-/**
- * eventfd_signal - Adds @n to the eventfd counter.
- * @ctx: [in] Pointer to the eventfd context.
- * @n: [in] Value of the counter to be added to the eventfd internal counter.
- *          The value cannot be negative.
- *
- * This function is supposed to be called by the kernel in paths that do not
- * allow sleeping. In this function we allow the counter to reach the ULLONG_MAX
- * value, and we signal this as overflow condition by returning a EPOLLERR
- * to poll(2).
- *
- * Returns the amount by which the counter was incremented.  This will be less
- * than @n if the counter has overflowed.
- */
-__u64 eventfd_signal(struct eventfd_ctx *ctx, __u64 n)
+__u64 eventfd_signal_mask(struct eventfd_ctx *ctx, __u64 n, unsigned mask)
 {
 	unsigned long flags;
 
@@ -78,12 +64,31 @@ __u64 eventfd_signal(struct eventfd_ctx *ctx, __u64 n)
 		n = ULLONG_MAX - ctx->count;
 	ctx->count += n;
 	if (waitqueue_active(&ctx->wqh))
-		wake_up_locked_poll(&ctx->wqh, EPOLLIN);
+		wake_up_locked_poll(&ctx->wqh, EPOLLIN | mask);
 	current->in_eventfd = 0;
 	spin_unlock_irqrestore(&ctx->wqh.lock, flags);
 
 	return n;
 }
+
+/**
+ * eventfd_signal - Adds @n to the eventfd counter.
+ * @ctx: [in] Pointer to the eventfd context.
+ * @n: [in] Value of the counter to be added to the eventfd internal counter.
+ *          The value cannot be negative.
+ *
+ * This function is supposed to be called by the kernel in paths that do not
+ * allow sleeping. In this function we allow the counter to reach the ULLONG_MAX
+ * value, and we signal this as overflow condition by returning a EPOLLERR
+ * to poll(2).
+ *
+ * Returns the amount by which the counter was incremented.  This will be less
+ * than @n if the counter has overflowed.
+ */
+__u64 eventfd_signal(struct eventfd_ctx *ctx, __u64 n)
+{
+	return eventfd_signal_mask(ctx, n, 0);
+}
 EXPORT_SYMBOL_GPL(eventfd_signal);
 
 static void eventfd_free_ctx(struct eventfd_ctx *ctx)

fs/eventpoll.c

@@ -491,7 +491,8 @@ static inline void ep_set_busy_poll_napi_id(struct epitem *epi)
  */
 #ifdef CONFIG_DEBUG_LOCK_ALLOC
 
-static void ep_poll_safewake(struct eventpoll *ep, struct epitem *epi)
+static void ep_poll_safewake(struct eventpoll *ep, struct epitem *epi,
+			     unsigned pollflags)
 {
 	struct eventpoll *ep_src;
 	unsigned long flags;
@@ -522,16 +523,17 @@ static void ep_poll_safewake(struct eventpoll *ep, struct epitem *epi)
 	}
 	spin_lock_irqsave_nested(&ep->poll_wait.lock, flags, nests);
 	ep->nests = nests + 1;
-	wake_up_locked_poll(&ep->poll_wait, EPOLLIN);
+	wake_up_locked_poll(&ep->poll_wait, EPOLLIN | pollflags);
 	ep->nests = 0;
 	spin_unlock_irqrestore(&ep->poll_wait.lock, flags);
 }
 
 #else
 
-static void ep_poll_safewake(struct eventpoll *ep, struct epitem *epi)
+static void ep_poll_safewake(struct eventpoll *ep, struct epitem *epi,
+			     unsigned pollflags)
 {
-	wake_up_poll(&ep->poll_wait, EPOLLIN);
+	wake_up_poll(&ep->poll_wait, EPOLLIN | pollflags);
 }
 
 #endif
@@ -742,7 +744,7 @@ static void ep_free(struct eventpoll *ep)
 
 	/* We need to release all tasks waiting for these file */
 	if (waitqueue_active(&ep->poll_wait))
-		ep_poll_safewake(ep, NULL);
+		ep_poll_safewake(ep, NULL, 0);
 
 	/*
 	 * We need to lock this because we could be hit by
@@ -1208,7 +1210,7 @@ static int ep_poll_callback(wait_queue_entry_t *wait, unsigned mode, int sync, v
 
 	/* We have to call this outside the lock */
 	if (pwake)
-		ep_poll_safewake(ep, epi);
+		ep_poll_safewake(ep, epi, pollflags & EPOLL_URING_WAKE);
 
 	if (!(epi->event.events & EPOLLEXCLUSIVE))
 		ewake = 1;
@@ -1553,7 +1555,7 @@ static int ep_insert(struct eventpoll *ep, const struct epoll_event *event,
 
 	/* We have to call this outside the lock */
 	if (pwake)
-		ep_poll_safewake(ep, NULL);
+		ep_poll_safewake(ep, NULL, 0);
 
 	return 0;
 }
@@ -1629,7 +1631,7 @@ static int ep_modify(struct eventpoll *ep, struct epitem *epi,
 
 	/* We have to call this outside the lock */
 	if (pwake)
-		ep_poll_safewake(ep, NULL);
+		ep_poll_safewake(ep, NULL, 0);
 
 	return 0;
 }

include/linux/eventfd.h

@@ -40,6 +40,7 @@ struct file *eventfd_fget(int fd);
 struct eventfd_ctx *eventfd_ctx_fdget(int fd);
 struct eventfd_ctx *eventfd_ctx_fileget(struct file *file);
 __u64 eventfd_signal(struct eventfd_ctx *ctx, __u64 n);
+__u64 eventfd_signal_mask(struct eventfd_ctx *ctx, __u64 n, unsigned mask);
 int eventfd_ctx_remove_wait_queue(struct eventfd_ctx *ctx, wait_queue_entry_t *wait,
 				  __u64 *cnt);
 void eventfd_ctx_do_read(struct eventfd_ctx *ctx, __u64 *cnt);
@@ -66,6 +67,12 @@ static inline int eventfd_signal(struct eventfd_ctx *ctx, __u64 n)
 	return -ENOSYS;
 }
 
+static inline int eventfd_signal_mask(struct eventfd_ctx *ctx, __u64 n,
+				      unsigned mask)
+{
+	return -ENOSYS;
+}
+
 static inline void eventfd_ctx_put(struct eventfd_ctx *ctx)
 {
 

include/linux/io_uring_types.h

@@ -174,7 +174,9 @@ struct io_submit_state {
 	bool			plug_started;
 	bool			need_plug;
 	unsigned short		submit_nr;
+	unsigned int		cqes_count;
 	struct blk_plug		plug;
+	struct io_uring_cqe	cqes[16];
 };
 
 struct io_ev_fd {

include/uapi/linux/eventpoll.h

@@ -41,6 +41,12 @@
 #define EPOLLMSG	(__force __poll_t)0x00000400
 #define EPOLLRDHUP	(__force __poll_t)0x00002000
 
+/*
+ * Internal flag - wakeup generated by io_uring, used to detect recursion back
+ * into the io_uring poll handler.
+ */
+#define EPOLL_URING_WAKE	((__force __poll_t)(1U << 27))
+
 /* Set exclusive wakeup mode for the target file descriptor */
 #define EPOLLEXCLUSIVE	((__force __poll_t)(1U << 28))
 

include/uapi/linux/io_uring.h

@@ -296,10 +296,28 @@ enum io_uring_op {
  *
  * IORING_RECVSEND_FIXED_BUF	Use registered buffers, the index is stored in
  *				the buf_index field.
+ *
+ * IORING_SEND_ZC_REPORT_USAGE
+ *				If set, SEND[MSG]_ZC should report
+ *				the zerocopy usage in cqe.res
+ *				for the IORING_CQE_F_NOTIF cqe.
+ *				0 is reported if zerocopy was actually possible.
+ *				IORING_NOTIF_USAGE_ZC_COPIED if data was copied
+ *				(at least partially).
  */
 #define IORING_RECVSEND_POLL_FIRST	(1U << 0)
 #define IORING_RECV_MULTISHOT		(1U << 1)
 #define IORING_RECVSEND_FIXED_BUF	(1U << 2)
+#define IORING_SEND_ZC_REPORT_USAGE	(1U << 3)
+
+/*
+ * cqe.res for IORING_CQE_F_NOTIF if
+ * IORING_SEND_ZC_REPORT_USAGE was requested
+ *
+ * It should be treated as a flag, all other
+ * bits of cqe.res should be treated as reserved!
+ */
+#define IORING_NOTIF_USAGE_ZC_COPIED    (1U << 31)
 
 /*
  * accept flags stored in sqe->ioprio