4 files changed, 108 insertions, 61 deletions

diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
index 2ac1cd8d23ea..db623b3185c8 100644
--- a/io_uring/io_uring.c
+++ b/io_uring/io_uring.c
@@ -1765,17 +1765,12 @@ queue:
 	}
 	spin_unlock(&ctx->completion_lock);
 
-	ret = io_req_prep_async(req);
-	if (ret) {
-fail:
-		io_req_defer_failed(req, ret);
-		return;
-	}
 	io_prep_async_link(req);
 	de = kmalloc(sizeof(*de), GFP_KERNEL);
 	if (!de) {
 		ret = -ENOMEM;
-		goto fail;
+		io_req_defer_failed(req, ret);
+		return;
 	}
 
 	spin_lock(&ctx->completion_lock);
@@ -2048,13 +2043,16 @@ static void io_queue_sqe_fallback(struct io_kiocb *req)
 		req->flags &= ~REQ_F_HARDLINK;
 		req->flags |= REQ_F_LINK;
 		io_req_defer_failed(req, req->cqe.res);
-	} else if (unlikely(req->ctx->drain_active)) {
-		io_drain_req(req);
 	} else {
 		int ret = io_req_prep_async(req);
 
-		if (unlikely(ret))
+		if (unlikely(ret)) {
 			io_req_defer_failed(req, ret);
+			return;
+		}
+
+		if (unlikely(req->ctx->drain_active))
+			io_drain_req(req);
 		else
 			io_queue_iowq(req, NULL);
 	}
@@ -3674,7 +3672,7 @@ static __cold int io_uring_create(unsigned entries, struct io_uring_params *p,
 
 	if (ctx->flags & IORING_SETUP_SINGLE_ISSUER
 	    && !(ctx->flags & IORING_SETUP_R_DISABLED))
-		ctx->submitter_task = get_task_struct(current);
+		WRITE_ONCE(ctx->submitter_task, get_task_struct(current));
 
 	file = io_uring_get_file(ctx);
 	if (IS_ERR(file)) {
@@ -3868,7 +3866,7 @@ static int io_register_enable_rings(struct io_ring_ctx *ctx)
 		return -EBADFD;
 
 	if (ctx->flags & IORING_SETUP_SINGLE_ISSUER && !ctx->submitter_task)
-		ctx->submitter_task = get_task_struct(current);
+		WRITE_ONCE(ctx->submitter_task, get_task_struct(current));
 
 	if (ctx->restrictions.registered)
 		ctx->restricted = 1;
diff --git a/io_uring/msg_ring.c b/io_uring/msg_ring.c
index 2d3cd945a531..15602a136821 100644
--- a/io_uring/msg_ring.c
+++ b/io_uring/msg_ring.c
@@ -25,6 +25,28 @@ struct io_msg {
 	u32 flags;
 };
 
+static void io_double_unlock_ctx(struct io_ring_ctx *octx)
+{
+	mutex_unlock(&octx->uring_lock);
+}
+
+static int io_double_lock_ctx(struct io_ring_ctx *octx,
+			      unsigned int issue_flags)
+{
+	/*
+	 * To ensure proper ordering between the two ctxs, we can only
+	 * attempt a trylock on the target. If that fails and we already have
+	 * the source ctx lock, punt to io-wq.
+	 */
+	if (!(issue_flags & IO_URING_F_UNLOCKED)) {
+		if (!mutex_trylock(&octx->uring_lock))
+			return -EAGAIN;
+		return 0;
+	}
+	mutex_lock(&octx->uring_lock);
+	return 0;
+}
+
 void io_msg_ring_cleanup(struct io_kiocb *req)
 {
 	struct io_msg *msg = io_kiocb_to_cmd(req, struct io_msg);
@@ -36,6 +58,29 @@ void io_msg_ring_cleanup(struct io_kiocb *req)
 	msg->src_file = NULL;
 }
 
+static inline bool io_msg_need_remote(struct io_ring_ctx *target_ctx)
+{
+	if (!target_ctx->task_complete)
+		return false;
+	return current != target_ctx->submitter_task;
+}
+
+static int io_msg_exec_remote(struct io_kiocb *req, task_work_func_t func)
+{
+	struct io_ring_ctx *ctx = req->file->private_data;
+	struct io_msg *msg = io_kiocb_to_cmd(req, struct io_msg);
+	struct task_struct *task = READ_ONCE(ctx->submitter_task);
+
+	if (unlikely(!task))
+		return -EOWNERDEAD;
+
+	init_task_work(&msg->tw, func);
+	if (task_work_add(ctx->submitter_task, &msg->tw, TWA_SIGNAL))
+		return -EOWNERDEAD;
+
+	return IOU_ISSUE_SKIP_COMPLETE;
+}
+
 static void io_msg_tw_complete(struct callback_head *head)
 {
 	struct io_msg *msg = container_of(head, struct io_msg, tw);
@@ -43,61 +88,54 @@ static void io_msg_tw_complete(struct callback_head *head)
 	struct io_ring_ctx *target_ctx = req->file->private_data;
 	int ret = 0;
 
-	if (current->flags & PF_EXITING)
+	if (current->flags & PF_EXITING) {
 		ret = -EOWNERDEAD;
-	else if (!io_post_aux_cqe(target_ctx, msg->user_data, msg->len, 0))
-		ret = -EOVERFLOW;
+	} else {
+		/*
+		 * If the target ring is using IOPOLL mode, then we need to be
+		 * holding the uring_lock for posting completions. Other ring
+		 * types rely on the regular completion locking, which is
+		 * handled while posting.
+		 */
+		if (target_ctx->flags & IORING_SETUP_IOPOLL)
+			mutex_lock(&target_ctx->uring_lock);
+		if (!io_post_aux_cqe(target_ctx, msg->user_data, msg->len, 0))
+			ret = -EOVERFLOW;
+		if (target_ctx->flags & IORING_SETUP_IOPOLL)
+			mutex_unlock(&target_ctx->uring_lock);
+	}
 
 	if (ret < 0)
 		req_set_fail(req);
 	io_req_queue_tw_complete(req, ret);
 }
 
-static int io_msg_ring_data(struct io_kiocb *req)
+static int io_msg_ring_data(struct io_kiocb *req, unsigned int issue_flags)
 {
 	struct io_ring_ctx *target_ctx = req->file->private_data;
 	struct io_msg *msg = io_kiocb_to_cmd(req, struct io_msg);
+	int ret;
 
 	if (msg->src_fd || msg->dst_fd || msg->flags)
 		return -EINVAL;
+	if (target_ctx->flags & IORING_SETUP_R_DISABLED)
+		return -EBADFD;
 
-	if (target_ctx->task_complete && current != target_ctx->submitter_task) {
-		init_task_work(&msg->tw, io_msg_tw_complete);
-		if (task_work_add(target_ctx->submitter_task, &msg->tw,
-				  TWA_SIGNAL_NO_IPI))
-			return -EOWNERDEAD;
-
-		atomic_or(IORING_SQ_TASKRUN, &target_ctx->rings->sq_flags);
-		return IOU_ISSUE_SKIP_COMPLETE;
-	}
-
-	if (io_post_aux_cqe(target_ctx, msg->user_data, msg->len, 0))
-		return 0;
+	if (io_msg_need_remote(target_ctx))
+		return io_msg_exec_remote(req, io_msg_tw_complete);
 
-	return -EOVERFLOW;
-}
-
-static void io_double_unlock_ctx(struct io_ring_ctx *octx,
-				 unsigned int issue_flags)
-{
-	mutex_unlock(&octx->uring_lock);
-}
-
-static int io_double_lock_ctx(struct io_ring_ctx *octx,
-			      unsigned int issue_flags)
-{
-	/*
-	 * To ensure proper ordering between the two ctxs, we can only
-	 * attempt a trylock on the target. If that fails and we already have
-	 * the source ctx lock, punt to io-wq.
-	 */
-	if (!(issue_flags & IO_URING_F_UNLOCKED)) {
-		if (!mutex_trylock(&octx->uring_lock))
+	ret = -EOVERFLOW;
+	if (target_ctx->flags & IORING_SETUP_IOPOLL) {
+		if (unlikely(io_double_lock_ctx(target_ctx, issue_flags)))
 			return -EAGAIN;
-		return 0;
+		if (io_post_aux_cqe(target_ctx, msg->user_data, msg->len, 0))
+			ret = 0;
+		io_double_unlock_ctx(target_ctx);
+	} else {
+		if (io_post_aux_cqe(target_ctx, msg->user_data, msg->len, 0))
+			ret = 0;
 	}
-	mutex_lock(&octx->uring_lock);
-	return 0;
+	return ret;
 }
 
 static struct file *io_msg_grab_file(struct io_kiocb *req, unsigned int issue_flags)
@@ -148,7 +186,7 @@ static int io_msg_install_complete(struct io_kiocb *req, unsigned int issue_flag
 	if (!io_post_aux_cqe(target_ctx, msg->user_data, msg->len, 0))
 		ret = -EOVERFLOW;
 out_unlock:
-	io_double_unlock_ctx(target_ctx, issue_flags);
+	io_double_unlock_ctx(target_ctx);
 	return ret;
 }
 
@@ -174,6 +212,8 @@ static int io_msg_send_fd(struct io_kiocb *req, unsigned int issue_flags)
 
 	if (target_ctx == ctx)
 		return -EINVAL;
+	if (target_ctx->flags & IORING_SETUP_R_DISABLED)
+		return -EBADFD;
 	if (!src_file) {
 		src_file = io_msg_grab_file(req, issue_flags);
 		if (!src_file)
@@ -182,14 +222,8 @@ static int io_msg_send_fd(struct io_kiocb *req, unsigned int issue_flags)
 		req->flags |= REQ_F_NEED_CLEANUP;
 	}
 
-	if (target_ctx->task_complete && current != target_ctx->submitter_task) {
-		init_task_work(&msg->tw, io_msg_tw_fd_complete);
-		if (task_work_add(target_ctx->submitter_task, &msg->tw,
-				  TWA_SIGNAL))
-			return -EOWNERDEAD;
-
-		return IOU_ISSUE_SKIP_COMPLETE;
-	}
+	if (io_msg_need_remote(target_ctx))
+		return io_msg_exec_remote(req, io_msg_tw_fd_complete);
 	return io_msg_install_complete(req, issue_flags);
 }
 
@@ -224,7 +258,7 @@ int io_msg_ring(struct io_kiocb *req, unsigned int issue_flags)
 
 	switch (msg->cmd) {
 	case IORING_MSG_DATA:
-		ret = io_msg_ring_data(req);
+		ret = io_msg_ring_data(req, issue_flags);
 		break;
 	case IORING_MSG_SEND_FD:
 		ret = io_msg_send_fd(req, issue_flags);
diff --git a/io_uring/net.c b/io_uring/net.c
index fbc34a7c2743..90326b279965 100644
--- a/io_uring/net.c
+++ b/io_uring/net.c
@@ -62,6 +62,7 @@ struct io_sr_msg {
 	u16				flags;
 	/* initialised and used only by !msg send variants */
 	u16				addr_len;
+	u16				buf_group;
 	void __user			*addr;
 	/* used only for send zerocopy */
 	struct io_kiocb 		*notif;
@@ -580,6 +581,15 @@ int io_recvmsg_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 		if (req->opcode == IORING_OP_RECV && sr->len)
 			return -EINVAL;
 		req->flags |= REQ_F_APOLL_MULTISHOT;
+		/*
+		 * Store the buffer group for this multishot receive separately,
+		 * as if we end up doing an io-wq based issue that selects a
+		 * buffer, it has to be committed immediately and that will
+		 * clear ->buf_list. This means we lose the link to the buffer
+		 * list, and the eventual buffer put on completion then cannot
+		 * restore it.
+		 */
+		sr->buf_group = req->buf_index;
 	}
 
 #ifdef CONFIG_COMPAT
@@ -596,6 +606,7 @@ static inline void io_recv_prep_retry(struct io_kiocb *req)
 
 	sr->done_io = 0;
 	sr->len = 0; /* get from the provided buffer */
+	req->buf_index = sr->buf_group;
 }
 
 /*
diff --git a/io_uring/poll.c b/io_uring/poll.c
index 32e5fc8365e6..2ac1366adbd7 100644
--- a/io_uring/poll.c
+++ b/io_uring/poll.c
@@ -283,8 +283,12 @@ static int io_poll_check_events(struct io_kiocb *req, bool *locked)
 			 * to the waitqueue, so if we get nothing back, we
 			 * should be safe and attempt a reissue.
 			 */
-			if (unlikely(!req->cqe.res))
+			if (unlikely(!req->cqe.res)) {
+				/* Multishot armed need not reissue */
+				if (!(req->apoll_events & EPOLLONESHOT))
+					continue;
 				return IOU_POLL_REISSUE;
+			}
 		}
 		if (req->apoll_events & EPOLLONESHOT)
 			return IOU_POLL_DONE;