io_uring: clean cqe filling functions

Split io_cqring_fill_event() into a couple of more targeted functions.
The first on is io_fill_cqe_aux() for completions that are not
associated with request completions and doing the ->cq_extra accounting.
Examples are additional CQEs from multishot poll and rsrc notifications.

The second is io_fill_cqe_req(), should be called when it's a normal
request completion. Nothing more to it at the moment, will be used in
later patches.

The last one is inlined __io_fill_cqe() for a finer grained control,
should be used with caution and in hottest places.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/59a9117a4a44fc9efcf04b3afa51e0d080f5943c.1636559119.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>

fs/io_uring.c

@@ -1108,8 +1108,8 @@ static void io_uring_try_cancel_requests(struct io_ring_ctx *ctx,
 					 bool cancel_all);
 static void io_uring_cancel_generic(bool cancel_all, struct io_sq_data *sqd);
 
-static bool io_cqring_fill_event(struct io_ring_ctx *ctx, u64 user_data,
+static void io_fill_cqe_req(struct io_kiocb *req, s32 res, u32 cflags);
-				 s32 res, u32 cflags);
+
 static void io_put_req(struct io_kiocb *req);
 static void io_put_req_deferred(struct io_kiocb *req);
 static void io_dismantle_req(struct io_kiocb *req);
@@ -1560,7 +1560,7 @@ static void io_kill_timeout(struct io_kiocb *req, int status)
 		atomic_set(&req->ctx->cq_timeouts,
 			atomic_read(&req->ctx->cq_timeouts) + 1);
 		list_del_init(&req->timeout.list);
-		io_cqring_fill_event(req->ctx, req->user_data, status, 0);
+		io_fill_cqe_req(req, status, 0);
 		io_put_req_deferred(req);
 	}
 }
@@ -1819,8 +1819,8 @@ static bool io_cqring_event_overflow(struct io_ring_ctx *ctx, u64 user_data,
 	return true;
 }
 
-static inline bool __io_cqring_fill_event(struct io_ring_ctx *ctx, u64 user_data,
+static inline bool __io_fill_cqe(struct io_ring_ctx *ctx, u64 user_data,
-					  s32 res, u32 cflags)
+				 s32 res, u32 cflags)
 {
 	struct io_uring_cqe *cqe;
 
@@ -1841,11 +1841,16 @@ static inline bool __io_cqring_fill_event(struct io_ring_ctx *ctx, u64 user_data
 	return io_cqring_event_overflow(ctx, user_data, res, cflags);
 }
 
-/* not as hot to bloat with inlining */
+static noinline void io_fill_cqe_req(struct io_kiocb *req, s32 res, u32 cflags)
-static noinline bool io_cqring_fill_event(struct io_ring_ctx *ctx, u64 user_data,
-					  s32 res, u32 cflags)
 {
-	return __io_cqring_fill_event(ctx, user_data, res, cflags);
+	__io_fill_cqe(req->ctx, req->user_data, res, cflags);
+}
+
+static noinline bool io_fill_cqe_aux(struct io_ring_ctx *ctx, u64 user_data,
+				     s32 res, u32 cflags)
+{
+	ctx->cq_extra++;
+	return __io_fill_cqe(ctx, user_data, res, cflags);
 }
 
 static void io_req_complete_post(struct io_kiocb *req, s32 res,
@@ -1854,7 +1859,7 @@ static void io_req_complete_post(struct io_kiocb *req, s32 res,
 	struct io_ring_ctx *ctx = req->ctx;
 
 	spin_lock(&ctx->completion_lock);
-	__io_cqring_fill_event(ctx, req->user_data, res, cflags);
+	__io_fill_cqe(ctx, req->user_data, res, cflags);
 	/*
 	 * If we're the last reference to this request, add to our locked
 	 * free_list cache.
@@ -2062,8 +2067,7 @@ static bool io_kill_linked_timeout(struct io_kiocb *req)
 		link->timeout.head = NULL;
 		if (hrtimer_try_to_cancel(&io->timer) != -1) {
 			list_del(&link->timeout.list);
-			io_cqring_fill_event(link->ctx, link->user_data,
+			io_fill_cqe_req(link, -ECANCELED, 0);
-					     -ECANCELED, 0);
 			io_put_req_deferred(link);
 			return true;
 		}
@@ -2087,7 +2091,7 @@ static void io_fail_links(struct io_kiocb *req)
 		link->link = NULL;
 
 		trace_io_uring_fail_link(req, link);
-		io_cqring_fill_event(link->ctx, link->user_data, res, 0);
+		io_fill_cqe_req(link, res, 0);
 		io_put_req_deferred(link);
 		link = nxt;
 	}
@@ -2104,8 +2108,7 @@ static bool io_disarm_next(struct io_kiocb *req)
 		req->flags &= ~REQ_F_ARM_LTIMEOUT;
 		if (link && link->opcode == IORING_OP_LINK_TIMEOUT) {
 			io_remove_next_linked(req);
-			io_cqring_fill_event(link->ctx, link->user_data,
+			io_fill_cqe_req(link, -ECANCELED, 0);
-					     -ECANCELED, 0);
 			io_put_req_deferred(link);
 			posted = true;
 		}
@@ -2369,8 +2372,8 @@ static void __io_submit_flush_completions(struct io_ring_ctx *ctx)
 		struct io_kiocb *req = container_of(node, struct io_kiocb,
 						    comp_list);
 
-		__io_cqring_fill_event(ctx, req->user_data, req->result,
+		__io_fill_cqe(ctx, req->user_data, req->result,
-					req->cflags);
+			      req->cflags);
 	}
 	io_commit_cqring(ctx);
 	spin_unlock(&ctx->completion_lock);
@@ -2504,8 +2507,8 @@ static int io_do_iopoll(struct io_ring_ctx *ctx, bool force_nonspin)
 		/* order with io_complete_rw_iopoll(), e.g. ->result updates */
 		if (!smp_load_acquire(&req->iopoll_completed))
 			break;
-		__io_cqring_fill_event(ctx, req->user_data, req->result,
+		__io_fill_cqe(ctx, req->user_data, req->result,
-					io_put_rw_kbuf(req));
+			      io_put_rw_kbuf(req));
 		nr_events++;
 	}
 
@@ -5360,13 +5363,13 @@ static bool __io_poll_complete(struct io_kiocb *req, __poll_t mask)
 	}
 	if (req->poll.events & EPOLLONESHOT)
 		flags = 0;
-	if (!io_cqring_fill_event(ctx, req->user_data, error, flags)) {
+
+	if (!(flags & IORING_CQE_F_MORE)) {
+		io_fill_cqe_req(req, error, flags);
+	} else if (!io_fill_cqe_aux(ctx, req->user_data, error, flags)) {
 		req->poll.events |= EPOLLONESHOT;
 		flags = 0;
 	}
-	if (flags & IORING_CQE_F_MORE)
-		ctx->cq_extra++;
-
 	return !(flags & IORING_CQE_F_MORE);
 }
 
@@ -5684,9 +5687,9 @@ static bool io_poll_remove_one(struct io_kiocb *req)
 	do_complete = __io_poll_remove_one(req, io_poll_get_single(req), true);
 
 	if (do_complete) {
-		io_cqring_fill_event(req->ctx, req->user_data, -ECANCELED, 0);
-		io_commit_cqring(req->ctx);
 		req_set_fail(req);
+		io_fill_cqe_req(req, -ECANCELED, 0);
+		io_commit_cqring(req->ctx);
 		io_put_req_deferred(req);
 	}
 	return do_complete;
@@ -5986,7 +5989,7 @@ static int io_timeout_cancel(struct io_ring_ctx *ctx, __u64 user_data)
 		return PTR_ERR(req);
 
 	req_set_fail(req);
-	io_cqring_fill_event(ctx, req->user_data, -ECANCELED, 0);
+	io_fill_cqe_req(req, -ECANCELED, 0);
 	io_put_req_deferred(req);
 	return 0;
 }
@@ -8219,8 +8222,7 @@ static void __io_rsrc_put_work(struct io_rsrc_node *ref_node)
 
 			io_ring_submit_lock(ctx, lock_ring);
 			spin_lock(&ctx->completion_lock);
-			io_cqring_fill_event(ctx, prsrc->tag, 0, 0);
+			io_fill_cqe_aux(ctx, prsrc->tag, 0, 0);
-			ctx->cq_extra++;
 			io_commit_cqring(ctx);
 			spin_unlock(&ctx->completion_lock);
 			io_cqring_ev_posted(ctx);