1 files changed, 138 insertions, 79 deletions

diff --git a/io_uring/msg_ring.c b/io_uring/msg_ring.c
index 15602a136821..7063ea7964e7 100644
--- a/io_uring/msg_ring.c
+++ b/io_uring/msg_ring.c
@@ -13,6 +13,10 @@
 #include "filetable.h"
 #include "msg_ring.h"
 
+/* All valid masks for MSG_RING */
+#define IORING_MSG_RING_MASK		(IORING_MSG_RING_CQE_SKIP | \
+					IORING_MSG_RING_FLAGS_PASS)
+
 struct io_msg {
 	struct file			*file;
 	struct file			*src_file;
@@ -21,7 +25,10 @@ struct io_msg {
 	u32 len;
 	u32 cmd;
 	u32 src_fd;
-	u32 dst_fd;
+	union {
+		u32 dst_fd;
+		u32 cqe_flags;
+	};
 	u32 flags;
 };
 
@@ -30,8 +37,8 @@ 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)
+static int io_lock_external_ctx(struct io_ring_ctx *octx,
+				unsigned int issue_flags)
 {
 	/*
 	 * To ensure proper ordering between the two ctxs, we can only
@@ -60,102 +67,111 @@ void io_msg_ring_cleanup(struct io_kiocb *req)
 
 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;
+	return target_ctx->task_complete;
 }
 
-static int io_msg_exec_remote(struct io_kiocb *req, task_work_func_t func)
+static void io_msg_tw_complete(struct io_tw_req tw_req, io_tw_token_t tw)
 {
-	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);
+	struct io_kiocb *req = tw_req.req;
+	struct io_ring_ctx *ctx = req->ctx;
 
-	if (unlikely(!task))
-		return -EOWNERDEAD;
+	io_add_aux_cqe(ctx, req->cqe.user_data, req->cqe.res, req->cqe.flags);
+	kfree_rcu(req, rcu_head);
+	percpu_ref_put(&ctx->refs);
+}
 
-	init_task_work(&msg->tw, func);
-	if (task_work_add(ctx->submitter_task, &msg->tw, TWA_SIGNAL))
+static int io_msg_remote_post(struct io_ring_ctx *ctx, struct io_kiocb *req,
+			      int res, u32 cflags, u64 user_data)
+{
+	if (!READ_ONCE(ctx->submitter_task)) {
+		kfree_rcu(req, rcu_head);
 		return -EOWNERDEAD;
-
-	return IOU_ISSUE_SKIP_COMPLETE;
+	}
+	req->opcode = IORING_OP_NOP;
+	req->cqe.user_data = user_data;
+	io_req_set_res(req, res, cflags);
+	percpu_ref_get(&ctx->refs);
+	req->ctx = ctx;
+	req->tctx = NULL;
+	req->io_task_work.func = io_msg_tw_complete;
+	io_req_task_work_add_remote(req, IOU_F_TWQ_LAZY_WAKE);
+	return 0;
 }
 
-static void io_msg_tw_complete(struct callback_head *head)
+static int io_msg_data_remote(struct io_ring_ctx *target_ctx,
+			      struct io_msg *msg)
 {
-	struct io_msg *msg = container_of(head, struct io_msg, tw);
-	struct io_kiocb *req = cmd_to_io_kiocb(msg);
-	struct io_ring_ctx *target_ctx = req->file->private_data;
-	int ret = 0;
-
-	if (current->flags & PF_EXITING) {
-		ret = -EOWNERDEAD;
-	} 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);
-	}
+	struct io_kiocb *target;
+	u32 flags = 0;
 
-	if (ret < 0)
-		req_set_fail(req);
-	io_req_queue_tw_complete(req, ret);
+	target = kmem_cache_alloc(req_cachep, GFP_KERNEL | __GFP_NOWARN | __GFP_ZERO)  ;
+	if (unlikely(!target))
+		return -ENOMEM;
+
+	if (msg->flags & IORING_MSG_RING_FLAGS_PASS)
+		flags = msg->cqe_flags;
+
+	return io_msg_remote_post(target_ctx, target, msg->len, flags,
+					msg->user_data);
 }
 
-static int io_msg_ring_data(struct io_kiocb *req, unsigned int issue_flags)
+static int __io_msg_ring_data(struct io_ring_ctx *target_ctx,
+			      struct io_msg *msg, 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);
+	u32 flags = 0;
 	int ret;
 
-	if (msg->src_fd || msg->dst_fd || msg->flags)
+	if (msg->src_fd || msg->flags & ~IORING_MSG_RING_FLAGS_PASS)
+		return -EINVAL;
+	if (!(msg->flags & IORING_MSG_RING_FLAGS_PASS) && msg->dst_fd)
 		return -EINVAL;
 	if (target_ctx->flags & IORING_SETUP_R_DISABLED)
 		return -EBADFD;
 
 	if (io_msg_need_remote(target_ctx))
-		return io_msg_exec_remote(req, io_msg_tw_complete);
+		return io_msg_data_remote(target_ctx, msg);
+
+	if (msg->flags & IORING_MSG_RING_FLAGS_PASS)
+		flags = msg->cqe_flags;
 
 	ret = -EOVERFLOW;
 	if (target_ctx->flags & IORING_SETUP_IOPOLL) {
-		if (unlikely(io_double_lock_ctx(target_ctx, issue_flags)))
+		if (unlikely(io_lock_external_ctx(target_ctx, issue_flags)))
 			return -EAGAIN;
-		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;
 	}
+	if (io_post_aux_cqe(target_ctx, msg->user_data, msg->len, flags))
+		ret = 0;
+	if (target_ctx->flags & IORING_SETUP_IOPOLL)
+		io_double_unlock_ctx(target_ctx);
 	return ret;
 }
 
-static struct file *io_msg_grab_file(struct io_kiocb *req, unsigned int issue_flags)
+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);
+
+	return __io_msg_ring_data(target_ctx, msg, issue_flags);
+}
+
+static int io_msg_grab_file(struct io_kiocb *req, unsigned int issue_flags)
 {
 	struct io_msg *msg = io_kiocb_to_cmd(req, struct io_msg);
 	struct io_ring_ctx *ctx = req->ctx;
-	struct file *file = NULL;
-	unsigned long file_ptr;
-	int idx = msg->src_fd;
+	struct io_rsrc_node *node;
+	int ret = -EBADF;
 
 	io_ring_submit_lock(ctx, issue_flags);
-	if (likely(idx < ctx->nr_user_files)) {
-		idx = array_index_nospec(idx, ctx->nr_user_files);
-		file_ptr = io_fixed_file_slot(&ctx->file_table, idx)->file_ptr;
-		file = (struct file *) (file_ptr & FFS_MASK);
-		if (file)
-			get_file(file);
+	node = io_rsrc_node_lookup(&ctx->file_table.data, msg->src_fd);
+	if (node) {
+		msg->src_file = io_slot_file(node);
+		if (msg->src_file)
+			get_file(msg->src_file);
+		req->flags |= REQ_F_NEED_CLEANUP;
+		ret = 0;
 	}
 	io_ring_submit_unlock(ctx, issue_flags);
-	return file;
+	return ret;
 }
 
 static int io_msg_install_complete(struct io_kiocb *req, unsigned int issue_flags)
@@ -165,7 +181,7 @@ static int io_msg_install_complete(struct io_kiocb *req, unsigned int issue_flag
 	struct file *src_file = msg->src_file;
 	int ret;
 
-	if (unlikely(io_double_lock_ctx(target_ctx, issue_flags)))
+	if (unlikely(io_lock_external_ctx(target_ctx, issue_flags)))
 		return -EAGAIN;
 
 	ret = __io_fixed_fd_install(target_ctx, src_file, msg->dst_fd);
@@ -183,7 +199,7 @@ static int io_msg_install_complete(struct io_kiocb *req, unsigned int issue_flag
 	 * completes with -EOVERFLOW, then the sender must ensure that a
 	 * later IORING_OP_MSG_RING delivers the message.
 	 */
-	if (!io_post_aux_cqe(target_ctx, msg->user_data, msg->len, 0))
+	if (!io_post_aux_cqe(target_ctx, msg->user_data, ret, 0))
 		ret = -EOVERFLOW;
 out_unlock:
 	io_double_unlock_ctx(target_ctx);
@@ -203,34 +219,47 @@ static void io_msg_tw_fd_complete(struct callback_head *head)
 	io_req_queue_tw_complete(req, ret);
 }
 
+static int io_msg_fd_remote(struct io_kiocb *req)
+{
+	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, io_msg_tw_fd_complete);
+	if (task_work_add(task, &msg->tw, TWA_SIGNAL))
+		return -EOWNERDEAD;
+
+	return IOU_ISSUE_SKIP_COMPLETE;
+}
+
 static int io_msg_send_fd(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);
 	struct io_ring_ctx *ctx = req->ctx;
-	struct file *src_file = msg->src_file;
 
+	if (msg->len)
+		return -EINVAL;
 	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)
-			return -EBADF;
-		msg->src_file = src_file;
-		req->flags |= REQ_F_NEED_CLEANUP;
+	if (!msg->src_file) {
+		int ret = io_msg_grab_file(req, issue_flags);
+		if (unlikely(ret))
+			return ret;
 	}
 
 	if (io_msg_need_remote(target_ctx))
-		return io_msg_exec_remote(req, io_msg_tw_fd_complete);
+		return io_msg_fd_remote(req);
 	return io_msg_install_complete(req, issue_flags);
 }
 
-int io_msg_ring_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
+static int __io_msg_ring_prep(struct io_msg *msg, const struct io_uring_sqe *sqe)
 {
-	struct io_msg *msg = io_kiocb_to_cmd(req, struct io_msg);
-
 	if (unlikely(sqe->buf_index || sqe->personality))
 		return -EINVAL;
 
@@ -241,12 +270,17 @@ int io_msg_ring_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 	msg->src_fd = READ_ONCE(sqe->addr3);
 	msg->dst_fd = READ_ONCE(sqe->file_index);
 	msg->flags = READ_ONCE(sqe->msg_ring_flags);
-	if (msg->flags & ~IORING_MSG_RING_CQE_SKIP)
+	if (msg->flags & ~IORING_MSG_RING_MASK)
 		return -EINVAL;
 
 	return 0;
 }
 
+int io_msg_ring_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
+{
+	return __io_msg_ring_prep(io_kiocb_to_cmd(req, struct io_msg), sqe);
+}
+
 int io_msg_ring(struct io_kiocb *req, unsigned int issue_flags)
 {
 	struct io_msg *msg = io_kiocb_to_cmd(req, struct io_msg);
@@ -275,5 +309,30 @@ done:
 		req_set_fail(req);
 	}
 	io_req_set_res(req, ret, 0);
-	return IOU_OK;
+	return IOU_COMPLETE;
+}
+
+int io_uring_sync_msg_ring(struct io_uring_sqe *sqe)
+{
+	struct io_msg io_msg = { };
+	int ret;
+
+	ret = __io_msg_ring_prep(&io_msg, sqe);
+	if (unlikely(ret))
+		return ret;
+
+	/*
+	 * Only data sending supported, not IORING_MSG_SEND_FD as that one
+	 * doesn't make sense without a source ring to send files from.
+	 */
+	if (io_msg.cmd != IORING_MSG_DATA)
+		return -EINVAL;
+
+	CLASS(fd, f)(sqe->fd);
+	if (fd_empty(f))
+		return -EBADF;
+	if (!io_is_uring_fops(fd_file(f)))
+		return -EBADFD;
+	return  __io_msg_ring_data(fd_file(f)->private_data,
+				   &io_msg, IO_URING_F_UNLOCKED);
 }