[linux-block.git] / io_uring / net.c

// SPDX-License-Identifier: GPL-2.0
#include <linux/kernel.h>
#include <linux/errno.h>
#include <linux/file.h>
#include <linux/slab.h>
#include <linux/net.h>
#include <linux/compat.h>
#include <net/compat.h>
#include <linux/io_uring.h>

#include <uapi/linux/io_uring.h>

#include "io_uring.h"
#include "kbuf.h"
#include "net.h"

#if defined(CONFIG_NET)
struct io_shutdown {
	struct file			*file;
	int				how;
};

struct io_accept {
	struct file			*file;
	struct sockaddr __user		*addr;
	int __user			*addr_len;
	int				flags;
	u32				file_slot;
	unsigned long			nofile;
};

struct io_socket {
	struct file			*file;
	int				domain;
	int				type;
	int				protocol;
	int				flags;
	u32				file_slot;
	unsigned long			nofile;
};

struct io_connect {
	struct file			*file;
	struct sockaddr __user		*addr;
	int				addr_len;
};

struct io_sr_msg {
	struct file			*file;
	union {
		struct compat_msghdr __user	*umsg_compat;
		struct user_msghdr __user	*umsg;
		void __user			*buf;
	};
	int				msg_flags;
	size_t				len;
	size_t				done_io;
	unsigned int			flags;
};

#define IO_APOLL_MULTI_POLLED (REQ_F_APOLL_MULTISHOT | REQ_F_POLLED)

int io_shutdown_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
{
	struct io_shutdown *shutdown = io_kiocb_to_cmd(req);

	if (unlikely(sqe->off || sqe->addr || sqe->rw_flags ||
		     sqe->buf_index || sqe->splice_fd_in))
		return -EINVAL;

	shutdown->how = READ_ONCE(sqe->len);
	return 0;
}

int io_shutdown(struct io_kiocb *req, unsigned int issue_flags)
{
	struct io_shutdown *shutdown = io_kiocb_to_cmd(req);
	struct socket *sock;
	int ret;

	if (issue_flags & IO_URING_F_NONBLOCK)
		return -EAGAIN;

	sock = sock_from_file(req->file);
	if (unlikely(!sock))
		return -ENOTSOCK;

	ret = __sys_shutdown_sock(sock, shutdown->how);
	io_req_set_res(req, ret, 0);
	return IOU_OK;
}

static bool io_net_retry(struct socket *sock, int flags)
{
	if (!(flags & MSG_WAITALL))
		return false;
	return sock->type == SOCK_STREAM || sock->type == SOCK_SEQPACKET;
}

static int io_setup_async_msg(struct io_kiocb *req,
			      struct io_async_msghdr *kmsg)
{
	struct io_async_msghdr *async_msg = req->async_data;

	if (async_msg)
		return -EAGAIN;
	if (io_alloc_async_data(req)) {
		kfree(kmsg->free_iov);
		return -ENOMEM;
	}
	async_msg = req->async_data;
	req->flags |= REQ_F_NEED_CLEANUP;
	memcpy(async_msg, kmsg, sizeof(*kmsg));
	async_msg->msg.msg_name = &async_msg->addr;
	/* if were using fast_iov, set it to the new one */
	if (!async_msg->free_iov)
		async_msg->msg.msg_iter.iov = async_msg->fast_iov;

	return -EAGAIN;
}

static int io_sendmsg_copy_hdr(struct io_kiocb *req,
			       struct io_async_msghdr *iomsg)
{
	struct io_sr_msg *sr = io_kiocb_to_cmd(req);

	iomsg->msg.msg_name = &iomsg->addr;
	iomsg->free_iov = iomsg->fast_iov;
	return sendmsg_copy_msghdr(&iomsg->msg, sr->umsg, sr->msg_flags,
					&iomsg->free_iov);
}

int io_sendmsg_prep_async(struct io_kiocb *req)
{
	int ret;

	ret = io_sendmsg_copy_hdr(req, req->async_data);
	if (!ret)
		req->flags |= REQ_F_NEED_CLEANUP;
	return ret;
}

void io_sendmsg_recvmsg_cleanup(struct io_kiocb *req)
{
	struct io_async_msghdr *io = req->async_data;

	kfree(io->free_iov);
}

int io_sendmsg_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
{
	struct io_sr_msg *sr = io_kiocb_to_cmd(req);

	if (unlikely(sqe->file_index || sqe->addr2))
		return -EINVAL;

	sr->umsg = u64_to_user_ptr(READ_ONCE(sqe->addr));
	sr->len = READ_ONCE(sqe->len);
	sr->flags = READ_ONCE(sqe->ioprio);
	if (sr->flags & ~IORING_RECVSEND_POLL_FIRST)
		return -EINVAL;
	sr->msg_flags = READ_ONCE(sqe->msg_flags) | MSG_NOSIGNAL;
	if (sr->msg_flags & MSG_DONTWAIT)
		req->flags |= REQ_F_NOWAIT;

#ifdef CONFIG_COMPAT
	if (req->ctx->compat)
		sr->msg_flags |= MSG_CMSG_COMPAT;
#endif
	sr->done_io = 0;
	return 0;
}

int io_sendmsg(struct io_kiocb *req, unsigned int issue_flags)
{
	struct io_sr_msg *sr = io_kiocb_to_cmd(req);
	struct io_async_msghdr iomsg, *kmsg;
	struct socket *sock;
	unsigned flags;
	int min_ret = 0;
	int ret;

	sock = sock_from_file(req->file);
	if (unlikely(!sock))
		return -ENOTSOCK;

	if (req_has_async_data(req)) {
		kmsg = req->async_data;
	} else {
		ret = io_sendmsg_copy_hdr(req, &iomsg);
		if (ret)
			return ret;
		kmsg = &iomsg;
	}

	if (!(req->flags & REQ_F_POLLED) &&
	    (sr->flags & IORING_RECVSEND_POLL_FIRST))
		return io_setup_async_msg(req, kmsg);

	flags = sr->msg_flags;
	if (issue_flags & IO_URING_F_NONBLOCK)
		flags |= MSG_DONTWAIT;
	if (flags & MSG_WAITALL)
		min_ret = iov_iter_count(&kmsg->msg.msg_iter);

	ret = __sys_sendmsg_sock(sock, &kmsg->msg, flags);

	if (ret < min_ret) {
		if (ret == -EAGAIN && (issue_flags & IO_URING_F_NONBLOCK))
			return io_setup_async_msg(req, kmsg);
		if (ret == -ERESTARTSYS)
			ret = -EINTR;
		if (ret > 0 && io_net_retry(sock, flags)) {
			sr->done_io += ret;
			req->flags |= REQ_F_PARTIAL_IO;
			return io_setup_async_msg(req, kmsg);
		}
		req_set_fail(req);
	}
	/* fast path, check for non-NULL to avoid function call */
	if (kmsg->free_iov)
		kfree(kmsg->free_iov);
	req->flags &= ~REQ_F_NEED_CLEANUP;
	if (ret >= 0)
		ret += sr->done_io;
	else if (sr->done_io)
		ret = sr->done_io;
	io_req_set_res(req, ret, 0);
	return IOU_OK;
}

int io_send(struct io_kiocb *req, unsigned int issue_flags)
{
	struct io_sr_msg *sr = io_kiocb_to_cmd(req);
	struct msghdr msg;
	struct iovec iov;
	struct socket *sock;
	unsigned flags;
	int min_ret = 0;
	int ret;

	if (!(req->flags & REQ_F_POLLED) &&
	    (sr->flags & IORING_RECVSEND_POLL_FIRST))
		return -EAGAIN;

	sock = sock_from_file(req->file);
	if (unlikely(!sock))
		return -ENOTSOCK;

	ret = import_single_range(WRITE, sr->buf, sr->len, &iov, &msg.msg_iter);
	if (unlikely(ret))
		return ret;

	msg.msg_name = NULL;
	msg.msg_control = NULL;
	msg.msg_controllen = 0;
	msg.msg_namelen = 0;

	flags = sr->msg_flags;
	if (issue_flags & IO_URING_F_NONBLOCK)
		flags |= MSG_DONTWAIT;
	if (flags & MSG_WAITALL)
		min_ret = iov_iter_count(&msg.msg_iter);

	msg.msg_flags = flags;
	ret = sock_sendmsg(sock, &msg);
	if (ret < min_ret) {
		if (ret == -EAGAIN && (issue_flags & IO_URING_F_NONBLOCK))
			return -EAGAIN;
		if (ret == -ERESTARTSYS)
			ret = -EINTR;
		if (ret > 0 && io_net_retry(sock, flags)) {
			sr->len -= ret;
			sr->buf += ret;
			sr->done_io += ret;
			req->flags |= REQ_F_PARTIAL_IO;
			return -EAGAIN;
		}
		req_set_fail(req);
	}
	if (ret >= 0)
		ret += sr->done_io;
	else if (sr->done_io)
		ret = sr->done_io;
	io_req_set_res(req, ret, 0);
	return IOU_OK;
}

static int __io_recvmsg_copy_hdr(struct io_kiocb *req,
				 struct io_async_msghdr *iomsg)
{
	struct io_sr_msg *sr = io_kiocb_to_cmd(req);
	struct iovec __user *uiov;
	size_t iov_len;
	int ret;

	ret = __copy_msghdr_from_user(&iomsg->msg, sr->umsg,
					&iomsg->uaddr, &uiov, &iov_len);
	if (ret)
		return ret;

	if (req->flags & REQ_F_BUFFER_SELECT) {
		if (iov_len > 1)
			return -EINVAL;
		if (copy_from_user(iomsg->fast_iov, uiov, sizeof(*uiov)))
			return -EFAULT;
		sr->len = iomsg->fast_iov[0].iov_len;
		iomsg->free_iov = NULL;
	} else {
		iomsg->free_iov = iomsg->fast_iov;
		ret = __import_iovec(READ, uiov, iov_len, UIO_FASTIOV,
				     &iomsg->free_iov, &iomsg->msg.msg_iter,
				     false);
		if (ret > 0)
			ret = 0;
	}

	return ret;
}

#ifdef CONFIG_COMPAT
static int __io_compat_recvmsg_copy_hdr(struct io_kiocb *req,
					struct io_async_msghdr *iomsg)
{
	struct io_sr_msg *sr = io_kiocb_to_cmd(req);
	struct compat_iovec __user *uiov;
	compat_uptr_t ptr;
	compat_size_t len;
	int ret;

	ret = __get_compat_msghdr(&iomsg->msg, sr->umsg_compat, &iomsg->uaddr,
				  &ptr, &len);
	if (ret)
		return ret;

	uiov = compat_ptr(ptr);
	if (req->flags & REQ_F_BUFFER_SELECT) {
		compat_ssize_t clen;

		if (len > 1)
			return -EINVAL;
		if (!access_ok(uiov, sizeof(*uiov)))
			return -EFAULT;
		if (__get_user(clen, &uiov->iov_len))
			return -EFAULT;
		if (clen < 0)
			return -EINVAL;
		sr->len = clen;
		iomsg->free_iov = NULL;
	} else {
		iomsg->free_iov = iomsg->fast_iov;
		ret = __import_iovec(READ, (struct iovec __user *)uiov, len,
				   UIO_FASTIOV, &iomsg->free_iov,
				   &iomsg->msg.msg_iter, true);
		if (ret < 0)
			return ret;
	}

	return 0;
}
#endif

static int io_recvmsg_copy_hdr(struct io_kiocb *req,
			       struct io_async_msghdr *iomsg)
{
	iomsg->msg.msg_name = &iomsg->addr;

#ifdef CONFIG_COMPAT
	if (req->ctx->compat)
		return __io_compat_recvmsg_copy_hdr(req, iomsg);
#endif

	return __io_recvmsg_copy_hdr(req, iomsg);
}

int io_recvmsg_prep_async(struct io_kiocb *req)
{
	int ret;

	ret = io_recvmsg_copy_hdr(req, req->async_data);
	if (!ret)
		req->flags |= REQ_F_NEED_CLEANUP;
	return ret;
}

int io_recvmsg_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
{
	struct io_sr_msg *sr = io_kiocb_to_cmd(req);

	if (unlikely(sqe->file_index || sqe->addr2))
		return -EINVAL;

	sr->umsg = u64_to_user_ptr(READ_ONCE(sqe->addr));
	sr->len = READ_ONCE(sqe->len);
	sr->flags = READ_ONCE(sqe->ioprio);
	if (sr->flags & ~IORING_RECVSEND_POLL_FIRST)
		return -EINVAL;
	sr->msg_flags = READ_ONCE(sqe->msg_flags) | MSG_NOSIGNAL;
	if (sr->msg_flags & MSG_DONTWAIT)
		req->flags |= REQ_F_NOWAIT;
	if (sr->msg_flags & MSG_ERRQUEUE)
		req->flags |= REQ_F_CLEAR_POLLIN;

#ifdef CONFIG_COMPAT
	if (req->ctx->compat)
		sr->msg_flags |= MSG_CMSG_COMPAT;
#endif
	sr->done_io = 0;
	return 0;
}

int io_recvmsg(struct io_kiocb *req, unsigned int issue_flags)
{
	struct io_sr_msg *sr = io_kiocb_to_cmd(req);
	struct io_async_msghdr iomsg, *kmsg;
	struct socket *sock;
	unsigned int cflags;
	unsigned flags;
	int ret, min_ret = 0;
	bool force_nonblock = issue_flags & IO_URING_F_NONBLOCK;

	sock = sock_from_file(req->file);
	if (unlikely(!sock))
		return -ENOTSOCK;

	if (req_has_async_data(req)) {
		kmsg = req->async_data;
	} else {
		ret = io_recvmsg_copy_hdr(req, &iomsg);
		if (ret)
			return ret;
		kmsg = &iomsg;
	}

	if (!(req->flags & REQ_F_POLLED) &&
	    (sr->flags & IORING_RECVSEND_POLL_FIRST))
		return io_setup_async_msg(req, kmsg);

	if (io_do_buffer_select(req)) {
		void __user *buf;

		buf = io_buffer_select(req, &sr->len, issue_flags);
		if (!buf)
			return -ENOBUFS;
		kmsg->fast_iov[0].iov_base = buf;
		kmsg->fast_iov[0].iov_len = sr->len;
		iov_iter_init(&kmsg->msg.msg_iter, READ, kmsg->fast_iov, 1,
				sr->len);
	}

	flags = sr->msg_flags;
	if (force_nonblock)
		flags |= MSG_DONTWAIT;
	if (flags & MSG_WAITALL)
		min_ret = iov_iter_count(&kmsg->msg.msg_iter);

	kmsg->msg.msg_get_inq = 1;
	ret = __sys_recvmsg_sock(sock, &kmsg->msg, sr->umsg, kmsg->uaddr, flags);
	if (ret < min_ret) {
		if (ret == -EAGAIN && force_nonblock)
			return io_setup_async_msg(req, kmsg);
		if (ret == -ERESTARTSYS)
			ret = -EINTR;
		if (ret > 0 && io_net_retry(sock, flags)) {
			sr->done_io += ret;
			req->flags |= REQ_F_PARTIAL_IO;
			return io_setup_async_msg(req, kmsg);
		}
		req_set_fail(req);
	} else if ((flags & MSG_WAITALL) && (kmsg->msg.msg_flags & (MSG_TRUNC | MSG_CTRUNC))) {
		req_set_fail(req);
	}

	/* fast path, check for non-NULL to avoid function call */
	if (kmsg->free_iov)
		kfree(kmsg->free_iov);
	req->flags &= ~REQ_F_NEED_CLEANUP;
	if (ret >= 0)
		ret += sr->done_io;
	else if (sr->done_io)
		ret = sr->done_io;
	cflags = io_put_kbuf(req, issue_flags);
	if (kmsg->msg.msg_inq)
		cflags |= IORING_CQE_F_SOCK_NONEMPTY;
	io_req_set_res(req, ret, cflags);
	return IOU_OK;
}

int io_recv(struct io_kiocb *req, unsigned int issue_flags)
{
	struct io_sr_msg *sr = io_kiocb_to_cmd(req);
	struct msghdr msg;
	struct socket *sock;
	struct iovec iov;
	unsigned int cflags;
	unsigned flags;
	int ret, min_ret = 0;
	bool force_nonblock = issue_flags & IO_URING_F_NONBLOCK;

	if (!(req->flags & REQ_F_POLLED) &&
	    (sr->flags & IORING_RECVSEND_POLL_FIRST))
		return -EAGAIN;

	sock = sock_from_file(req->file);
	if (unlikely(!sock))
		return -ENOTSOCK;

	if (io_do_buffer_select(req)) {
		void __user *buf;

		buf = io_buffer_select(req, &sr->len, issue_flags);
		if (!buf)
			return -ENOBUFS;
		sr->buf = buf;
	}

	ret = import_single_range(READ, sr->buf, sr->len, &iov, &msg.msg_iter);
	if (unlikely(ret))
		goto out_free;

	msg.msg_name = NULL;
	msg.msg_namelen = 0;
	msg.msg_control = NULL;
	msg.msg_get_inq = 1;
	msg.msg_flags = 0;
	msg.msg_controllen = 0;
	msg.msg_iocb = NULL;

	flags = sr->msg_flags;
	if (force_nonblock)
		flags |= MSG_DONTWAIT;
	if (flags & MSG_WAITALL)
		min_ret = iov_iter_count(&msg.msg_iter);

	ret = sock_recvmsg(sock, &msg, flags);
	if (ret < min_ret) {
		if (ret == -EAGAIN && force_nonblock)
			return -EAGAIN;
		if (ret == -ERESTARTSYS)
			ret = -EINTR;
		if (ret > 0 && io_net_retry(sock, flags)) {
			sr->len -= ret;
			sr->buf += ret;
			sr->done_io += ret;
			req->flags |= REQ_F_PARTIAL_IO;
			return -EAGAIN;
		}
		req_set_fail(req);
	} else if ((flags & MSG_WAITALL) && (msg.msg_flags & (MSG_TRUNC | MSG_CTRUNC))) {
out_free:
		req_set_fail(req);
	}

	if (ret >= 0)
		ret += sr->done_io;
	else if (sr->done_io)
		ret = sr->done_io;
	cflags = io_put_kbuf(req, issue_flags);
	if (msg.msg_inq)
		cflags |= IORING_CQE_F_SOCK_NONEMPTY;
	io_req_set_res(req, ret, cflags);
	return IOU_OK;
}

int io_accept_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
{
	struct io_accept *accept = io_kiocb_to_cmd(req);
	unsigned flags;

	if (sqe->len || sqe->buf_index)
		return -EINVAL;

	accept->addr = u64_to_user_ptr(READ_ONCE(sqe->addr));
	accept->addr_len = u64_to_user_ptr(READ_ONCE(sqe->addr2));
	accept->flags = READ_ONCE(sqe->accept_flags);
	accept->nofile = rlimit(RLIMIT_NOFILE);
	flags = READ_ONCE(sqe->ioprio);
	if (flags & ~IORING_ACCEPT_MULTISHOT)
		return -EINVAL;

	accept->file_slot = READ_ONCE(sqe->file_index);
	if (accept->file_slot) {
		if (accept->flags & SOCK_CLOEXEC)
			return -EINVAL;
		if (flags & IORING_ACCEPT_MULTISHOT &&
		    accept->file_slot != IORING_FILE_INDEX_ALLOC)
			return -EINVAL;
	}
	if (accept->flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
		return -EINVAL;
	if (SOCK_NONBLOCK != O_NONBLOCK && (accept->flags & SOCK_NONBLOCK))
		accept->flags = (accept->flags & ~SOCK_NONBLOCK) | O_NONBLOCK;
	if (flags & IORING_ACCEPT_MULTISHOT)
		req->flags |= REQ_F_APOLL_MULTISHOT;
	return 0;
}

int io_accept(struct io_kiocb *req, unsigned int issue_flags)
{
	struct io_ring_ctx *ctx = req->ctx;
	struct io_accept *accept = io_kiocb_to_cmd(req);
	bool force_nonblock = issue_flags & IO_URING_F_NONBLOCK;
	unsigned int file_flags = force_nonblock ? O_NONBLOCK : 0;
	bool fixed = !!accept->file_slot;
	struct file *file;
	int ret, fd;

retry:
	if (!fixed) {
		fd = __get_unused_fd_flags(accept->flags, accept->nofile);
		if (unlikely(fd < 0))
			return fd;
	}
	file = do_accept(req->file, file_flags, accept->addr, accept->addr_len,
			 accept->flags);
	if (IS_ERR(file)) {
		if (!fixed)
			put_unused_fd(fd);
		ret = PTR_ERR(file);
		if (ret == -EAGAIN && force_nonblock) {
			/*
			 * if it's multishot and polled, we don't need to
			 * return EAGAIN to arm the poll infra since it
			 * has already been done
			 */
			if ((req->flags & IO_APOLL_MULTI_POLLED) ==
			    IO_APOLL_MULTI_POLLED)
				ret = IOU_ISSUE_SKIP_COMPLETE;
			return ret;
		}
		if (ret == -ERESTARTSYS)
			ret = -EINTR;
		req_set_fail(req);
	} else if (!fixed) {
		fd_install(fd, file);
		ret = fd;
	} else {
		ret = io_fixed_fd_install(req, issue_flags, file,
						accept->file_slot);
	}

	if (!(req->flags & REQ_F_APOLL_MULTISHOT)) {
		io_req_set_res(req, ret, 0);
		return IOU_OK;
	}

	if (ret < 0)
		return ret;
	if (io_post_aux_cqe(ctx, req->cqe.user_data, ret, IORING_CQE_F_MORE))
		goto retry;
	return -ECANCELED;
}

int io_socket_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
{
	struct io_socket *sock = io_kiocb_to_cmd(req);

	if (sqe->addr || sqe->rw_flags || sqe->buf_index)
		return -EINVAL;

	sock->domain = READ_ONCE(sqe->fd);
	sock->type = READ_ONCE(sqe->off);
	sock->protocol = READ_ONCE(sqe->len);
	sock->file_slot = READ_ONCE(sqe->file_index);
	sock->nofile = rlimit(RLIMIT_NOFILE);

	sock->flags = sock->type & ~SOCK_TYPE_MASK;
	if (sock->file_slot && (sock->flags & SOCK_CLOEXEC))
		return -EINVAL;
	if (sock->flags & ~(SOCK_CLOEXEC | SOCK_NONBLOCK))
		return -EINVAL;
	return 0;
}

int io_socket(struct io_kiocb *req, unsigned int issue_flags)
{
	struct io_socket *sock = io_kiocb_to_cmd(req);
	bool fixed = !!sock->file_slot;
	struct file *file;
	int ret, fd;

	if (!fixed) {
		fd = __get_unused_fd_flags(sock->flags, sock->nofile);
		if (unlikely(fd < 0))
			return fd;
	}
	file = __sys_socket_file(sock->domain, sock->type, sock->protocol);
	if (IS_ERR(file)) {
		if (!fixed)
			put_unused_fd(fd);
		ret = PTR_ERR(file);
		if (ret == -EAGAIN && (issue_flags & IO_URING_F_NONBLOCK))
			return -EAGAIN;
		if (ret == -ERESTARTSYS)
			ret = -EINTR;
		req_set_fail(req);
	} else if (!fixed) {
		fd_install(fd, file);
		ret = fd;
	} else {
		ret = io_fixed_fd_install(req, issue_flags, file,
					    sock->file_slot);
	}
	io_req_set_res(req, ret, 0);
	return IOU_OK;
}

int io_connect_prep_async(struct io_kiocb *req)
{
	struct io_async_connect *io = req->async_data;
	struct io_connect *conn = io_kiocb_to_cmd(req);

	return move_addr_to_kernel(conn->addr, conn->addr_len, &io->address);
}

int io_connect_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
{
	struct io_connect *conn = io_kiocb_to_cmd(req);

	if (sqe->len || sqe->buf_index || sqe->rw_flags || sqe->splice_fd_in)
		return -EINVAL;

	conn->addr = u64_to_user_ptr(READ_ONCE(sqe->addr));
	conn->addr_len =  READ_ONCE(sqe->addr2);
	return 0;
}

int io_connect(struct io_kiocb *req, unsigned int issue_flags)
{
	struct io_connect *connect = io_kiocb_to_cmd(req);
	struct io_async_connect __io, *io;
	unsigned file_flags;
	int ret;
	bool force_nonblock = issue_flags & IO_URING_F_NONBLOCK;

	if (req_has_async_data(req)) {
		io = req->async_data;
	} else {
		ret = move_addr_to_kernel(connect->addr,
						connect->addr_len,
						&__io.address);
		if (ret)
			goto out;
		io = &__io;
	}

	file_flags = force_nonblock ? O_NONBLOCK : 0;

	ret = __sys_connect_file(req->file, &io->address,
					connect->addr_len, file_flags);
	if ((ret == -EAGAIN || ret == -EINPROGRESS) && force_nonblock) {
		if (req_has_async_data(req))
			return -EAGAIN;
		if (io_alloc_async_data(req)) {
			ret = -ENOMEM;
			goto out;
		}
		memcpy(req->async_data, &__io, sizeof(__io));
		return -EAGAIN;
	}
	if (ret == -ERESTARTSYS)
		ret = -EINTR;
out:
	if (ret < 0)
		req_set_fail(req);
	io_req_set_res(req, ret, 0);
	return IOU_OK;
}
#endif
Commit	Line	Data
f9ead18c JA	1	// SPDX-License-Identifier: GPL-2.0
	2	#include <linux/kernel.h>
	3	#include <linux/errno.h>
	4	#include <linux/file.h>
	5	#include <linux/slab.h>
	6	#include <linux/net.h>
	7	#include <linux/compat.h>
	8	#include <net/compat.h>
	9	#include <linux/io_uring.h>
	10
	11	#include <uapi/linux/io_uring.h>
	12
f9ead18c	13	#include "io_uring.h"
3b77495a	14	#include "kbuf.h"
f9ead18c JA	15	#include "net.h"
	16
	17	#if defined(CONFIG_NET)
	18	struct io_shutdown {
	19	struct file *file;
	20	int how;
	21	};
	22
	23	struct io_accept {
	24	struct file *file;
	25	struct sockaddr __user *addr;
	26	int __user *addr_len;
	27	int flags;
	28	u32 file_slot;
	29	unsigned long nofile;
	30	};
	31
	32	struct io_socket {
	33	struct file *file;
	34	int domain;
	35	int type;
	36	int protocol;
	37	int flags;
	38	u32 file_slot;
	39	unsigned long nofile;
	40	};
	41
	42	struct io_connect {
	43	struct file *file;
	44	struct sockaddr __user *addr;
	45	int addr_len;
	46	};
	47
	48	struct io_sr_msg {
	49	struct file *file;
	50	union {
	51	struct compat_msghdr __user *umsg_compat;
	52	struct user_msghdr __user *umsg;
	53	void __user *buf;
	54	};
	55	int msg_flags;
	56	size_t len;
	57	size_t done_io;
	58	unsigned int flags;
	59	};
	60
	61	#define IO_APOLL_MULTI_POLLED (REQ_F_APOLL_MULTISHOT \| REQ_F_POLLED)
	62
	63	int io_shutdown_prep(struct io_kiocb req, const struct io_uring_sqe sqe)
	64	{
	65	struct io_shutdown *shutdown = io_kiocb_to_cmd(req);
	66
	67	if (unlikely(sqe->off \|\| sqe->addr \|\| sqe->rw_flags \|\|
	68	sqe->buf_index \|\| sqe->splice_fd_in))
	69	return -EINVAL;
	70
	71	shutdown->how = READ_ONCE(sqe->len);
	72	return 0;
	73	}
	74
	75	int io_shutdown(struct io_kiocb *req, unsigned int issue_flags)
	76	{
	77	struct io_shutdown *shutdown = io_kiocb_to_cmd(req);
	78	struct socket *sock;
79	int ret;
80
81	if (issue_flags & IO_URING_F_NONBLOCK)
82	return -EAGAIN;
83
84	sock = sock_from_file(req->file);
85	if (unlikely(!sock))
86	return -ENOTSOCK;
87
88	ret = __sys_shutdown_sock(sock, shutdown->how);
89	io_req_set_res(req, ret, 0);
90	return IOU_OK;
91	}
92
93	static bool io_net_retry(struct socket *sock, int flags)
94	{
95	if (!(flags & MSG_WAITALL))
96	return false;
97	return sock->type == SOCK_STREAM \|\| sock->type == SOCK_SEQPACKET;
98	}
99
100	static int io_setup_async_msg(struct io_kiocb *req,
101	struct io_async_msghdr *kmsg)
102	{
103	struct io_async_msghdr *async_msg = req->async_data;
104
105	if (async_msg)
106	return -EAGAIN;
107	if (io_alloc_async_data(req)) {
108	kfree(kmsg->free_iov);
109	return -ENOMEM;
110	}
111	async_msg = req->async_data;
112	req->flags \|= REQ_F_NEED_CLEANUP;
113	memcpy(async_msg, kmsg, sizeof(*kmsg));
114	async_msg->msg.msg_name = &async_msg->addr;
115	/* if were using fast_iov, set it to the new one */
116	if (!async_msg->free_iov)
117	async_msg->msg.msg_iter.iov = async_msg->fast_iov;
118
119	return -EAGAIN;
120	}
121
122	static int io_sendmsg_copy_hdr(struct io_kiocb *req,
123	struct io_async_msghdr *iomsg)
124	{
125	struct io_sr_msg *sr = io_kiocb_to_cmd(req);
126
127	iomsg->msg.msg_name = &iomsg->addr;
128	iomsg->free_iov = iomsg->fast_iov;
129	return sendmsg_copy_msghdr(&iomsg->msg, sr->umsg, sr->msg_flags,
130	&iomsg->free_iov);
131	}
132
133	int io_sendmsg_prep_async(struct io_kiocb *req)
134	{
135	int ret;
136
137	ret = io_sendmsg_copy_hdr(req, req->async_data);
138	if (!ret)
139	req->flags \|= REQ_F_NEED_CLEANUP;
140	return ret;
141	}
142
143	void io_sendmsg_recvmsg_cleanup(struct io_kiocb *req)
144	{
145	struct io_async_msghdr *io = req->async_data;
146
147	kfree(io->free_iov);
148	}
149
150	int io_sendmsg_prep(struct io_kiocb req, const struct io_uring_sqe sqe)
151	{
152	struct io_sr_msg *sr = io_kiocb_to_cmd(req);
153
154	if (unlikely(sqe->file_index \|\| sqe->addr2))
155	return -EINVAL;
156
157	sr->umsg = u64_to_user_ptr(READ_ONCE(sqe->addr));
158	sr->len = READ_ONCE(sqe->len);
159	sr->flags = READ_ONCE(sqe->ioprio);
160	if (sr->flags & ~IORING_RECVSEND_POLL_FIRST)
161	return -EINVAL;
162	sr->msg_flags = READ_ONCE(sqe->msg_flags) \| MSG_NOSIGNAL;
163	if (sr->msg_flags & MSG_DONTWAIT)
164	req->flags \|= REQ_F_NOWAIT;
165
166	#ifdef CONFIG_COMPAT
167	if (req->ctx->compat)
168	sr->msg_flags \|= MSG_CMSG_COMPAT;
169	#endif
170	sr->done_io = 0;
171	return 0;
172	}
173
174	int io_sendmsg(struct io_kiocb *req, unsigned int issue_flags)
175	{
176	struct io_sr_msg *sr = io_kiocb_to_cmd(req);
177	struct io_async_msghdr iomsg, *kmsg;
178	struct socket *sock;
179	unsigned flags;
180	int min_ret = 0;
181	int ret;
182
183	sock = sock_from_file(req->file);
184	if (unlikely(!sock))
185	return -ENOTSOCK;
186
187	if (req_has_async_data(req)) {
188	kmsg = req->async_data;
189	} else {
190	ret = io_sendmsg_copy_hdr(req, &iomsg);
191	if (ret)
192	return ret;
193	kmsg = &iomsg;
194	}
195
196	if (!(req->flags & REQ_F_POLLED) &&
197	(sr->flags & IORING_RECVSEND_POLL_FIRST))
198	return io_setup_async_msg(req, kmsg);
199
200	flags = sr->msg_flags;
201	if (issue_flags & IO_URING_F_NONBLOCK)
202	flags \|= MSG_DONTWAIT;
203	if (flags & MSG_WAITALL)
204	min_ret = iov_iter_count(&kmsg->msg.msg_iter);
205
206	ret = __sys_sendmsg_sock(sock, &kmsg->msg, flags);
207
208	if (ret < min_ret) {
209	if (ret == -EAGAIN && (issue_flags & IO_URING_F_NONBLOCK))
210	return io_setup_async_msg(req, kmsg);
211	if (ret == -ERESTARTSYS)
212	ret = -EINTR;
213	if (ret > 0 && io_net_retry(sock, flags)) {
214	sr->done_io += ret;
215	req->flags \|= REQ_F_PARTIAL_IO;
216	return io_setup_async_msg(req, kmsg);
217	}
218	req_set_fail(req);
219	}
220	/* fast path, check for non-NULL to avoid function call */
221	if (kmsg->free_iov)
222	kfree(kmsg->free_iov);
223	req->flags &= ~REQ_F_NEED_CLEANUP;
224	if (ret >= 0)
225	ret += sr->done_io;
226	else if (sr->done_io)
227	ret = sr->done_io;
228	io_req_set_res(req, ret, 0);
229	return IOU_OK;
230	}
231
232	int io_send(struct io_kiocb *req, unsigned int issue_flags)
233	{
234	struct io_sr_msg *sr = io_kiocb_to_cmd(req);
235	struct msghdr msg;
236	struct iovec iov;
237	struct socket *sock;
238	unsigned flags;
239	int min_ret = 0;
240	int ret;
241
242	if (!(req->flags & REQ_F_POLLED) &&
243	(sr->flags & IORING_RECVSEND_POLL_FIRST))
244	return -EAGAIN;
245
246	sock = sock_from_file(req->file);
247	if (unlikely(!sock))
248	return -ENOTSOCK;
249
250	ret = import_single_range(WRITE, sr->buf, sr->len, &iov, &msg.msg_iter);
251	if (unlikely(ret))
252	return ret;
253
254	msg.msg_name = NULL;
255	msg.msg_control = NULL;
256	msg.msg_controllen = 0;
257	msg.msg_namelen = 0;
258
259	flags = sr->msg_flags;
260	if (issue_flags & IO_URING_F_NONBLOCK)
261	flags \|= MSG_DONTWAIT;
262	if (flags & MSG_WAITALL)
263	min_ret = iov_iter_count(&msg.msg_iter);
264
265	msg.msg_flags = flags;
266	ret = sock_sendmsg(sock, &msg);
267	if (ret < min_ret) {
268	if (ret == -EAGAIN && (issue_flags & IO_URING_F_NONBLOCK))
269	return -EAGAIN;
270	if (ret == -ERESTARTSYS)
271	ret = -EINTR;
272	if (ret > 0 && io_net_retry(sock, flags)) {
273	sr->len -= ret;
274	sr->buf += ret;
275	sr->done_io += ret;
276	req->flags \|= REQ_F_PARTIAL_IO;
277	return -EAGAIN;
278	}
279	req_set_fail(req);
280	}
281	if (ret >= 0)
282	ret += sr->done_io;
283	else if (sr->done_io)
284	ret = sr->done_io;
285	io_req_set_res(req, ret, 0);
286	return IOU_OK;
287	}
288
289	static int __io_recvmsg_copy_hdr(struct io_kiocb *req,
290	struct io_async_msghdr *iomsg)
291	{
292	struct io_sr_msg *sr = io_kiocb_to_cmd(req);
293	struct iovec __user *uiov;
294	size_t iov_len;
295	int ret;
296
297	ret = __copy_msghdr_from_user(&iomsg->msg, sr->umsg,
298	&iomsg->uaddr, &uiov, &iov_len);
299	if (ret)
300	return ret;
301
302	if (req->flags & REQ_F_BUFFER_SELECT) {
303	if (iov_len > 1)
304	return -EINVAL;
305	if (copy_from_user(iomsg->fast_iov, uiov, sizeof(*uiov)))
306	return -EFAULT;
307	sr->len = iomsg->fast_iov[0].iov_len;
308	iomsg->free_iov = NULL;
309	} else {
310	iomsg->free_iov = iomsg->fast_iov;
311	ret = __import_iovec(READ, uiov, iov_len, UIO_FASTIOV,
312	&iomsg->free_iov, &iomsg->msg.msg_iter,
313	false);
314	if (ret > 0)
315	ret = 0;
316	}
317
318	return ret;
319	}
320
321	#ifdef CONFIG_COMPAT
322	static int __io_compat_recvmsg_copy_hdr(struct io_kiocb *req,
323	struct io_async_msghdr *iomsg)
324	{
325	struct io_sr_msg *sr = io_kiocb_to_cmd(req);
326	struct compat_iovec __user *uiov;
327	compat_uptr_t ptr;
328	compat_size_t len;
329	int ret;
330
331	ret = __get_compat_msghdr(&iomsg->msg, sr->umsg_compat, &iomsg->uaddr,
332	&ptr, &len);
333	if (ret)
334	return ret;
335
336	uiov = compat_ptr(ptr);
337	if (req->flags & REQ_F_BUFFER_SELECT) {
338	compat_ssize_t clen;
339
340	if (len > 1)
341	return -EINVAL;
342	if (!access_ok(uiov, sizeof(*uiov)))
343	return -EFAULT;
344	if (__get_user(clen, &uiov->iov_len))
345	return -EFAULT;
346	if (clen < 0)
347	return -EINVAL;
348	sr->len = clen;
349	iomsg->free_iov = NULL;
350	} else {
351	iomsg->free_iov = iomsg->fast_iov;
352	ret = __import_iovec(READ, (struct iovec __user *)uiov, len,
353	UIO_FASTIOV, &iomsg->free_iov,
354	&iomsg->msg.msg_iter, true);
355	if (ret < 0)
356	return ret;
357	}
358
359	return 0;
360	}
361	#endif
362
363	static int io_recvmsg_copy_hdr(struct io_kiocb *req,
364	struct io_async_msghdr *iomsg)
365	{
366	iomsg->msg.msg_name = &iomsg->addr;
367
368	#ifdef CONFIG_COMPAT
369	if (req->ctx->compat)
370	return __io_compat_recvmsg_copy_hdr(req, iomsg);
371	#endif
372
373	return __io_recvmsg_copy_hdr(req, iomsg);
374	}
375
376	int io_recvmsg_prep_async(struct io_kiocb *req)
377	{
378	int ret;
379
380	ret = io_recvmsg_copy_hdr(req, req->async_data);
381	if (!ret)
382	req->flags \|= REQ_F_NEED_CLEANUP;
383	return ret;
384	}
385
386	int io_recvmsg_prep(struct io_kiocb req, const struct io_uring_sqe sqe)
387	{
388	struct io_sr_msg *sr = io_kiocb_to_cmd(req);
389
390	if (unlikely(sqe->file_index \|\| sqe->addr2))
391	return -EINVAL;
392
393	sr->umsg = u64_to_user_ptr(READ_ONCE(sqe->addr));
394	sr->len = READ_ONCE(sqe->len);
395	sr->flags = READ_ONCE(sqe->ioprio);
396	if (sr->flags & ~IORING_RECVSEND_POLL_FIRST)
397	return -EINVAL;
398	sr->msg_flags = READ_ONCE(sqe->msg_flags) \| MSG_NOSIGNAL;
399	if (sr->msg_flags & MSG_DONTWAIT)
400	req->flags \|= REQ_F_NOWAIT;
401	if (sr->msg_flags & MSG_ERRQUEUE)
402	req->flags \|= REQ_F_CLEAR_POLLIN;
403
404	#ifdef CONFIG_COMPAT
405	if (req->ctx->compat)
406	sr->msg_flags \|= MSG_CMSG_COMPAT;
407	#endif
408	sr->done_io = 0;
409	return 0;
410	}
411
412	int io_recvmsg(struct io_kiocb *req, unsigned int issue_flags)
413	{
414	struct io_sr_msg *sr = io_kiocb_to_cmd(req);
415	struct io_async_msghdr iomsg, *kmsg;
416	struct socket *sock;
417	unsigned int cflags;
418	unsigned flags;
419	int ret, min_ret = 0;
420	bool force_nonblock = issue_flags & IO_URING_F_NONBLOCK;
421
422	sock = sock_from_file(req->file);
423	if (unlikely(!sock))
424	return -ENOTSOCK;
425
426	if (req_has_async_data(req)) {
427	kmsg = req->async_data;
428	} else {
429	ret = io_recvmsg_copy_hdr(req, &iomsg);
430	if (ret)
431	return ret;
432	kmsg = &iomsg;
433	}
434
435	if (!(req->flags & REQ_F_POLLED) &&
436	(sr->flags & IORING_RECVSEND_POLL_FIRST))
437	return io_setup_async_msg(req, kmsg);
438
439	if (io_do_buffer_select(req)) {
440	void __user *buf;
441
442	buf = io_buffer_select(req, &sr->len, issue_flags);
443	if (!buf)
444	return -ENOBUFS;
445	kmsg->fast_iov[0].iov_base = buf;
446	kmsg->fast_iov[0].iov_len = sr->len;
447	iov_iter_init(&kmsg->msg.msg_iter, READ, kmsg->fast_iov, 1,
448	sr->len);
449	}
450
451	flags = sr->msg_flags;
452	if (force_nonblock)
453	flags \|= MSG_DONTWAIT;
454	if (flags & MSG_WAITALL)
455	min_ret = iov_iter_count(&kmsg->msg.msg_iter);
456
457	kmsg->msg.msg_get_inq = 1;
458	ret = __sys_recvmsg_sock(sock, &kmsg->msg, sr->umsg, kmsg->uaddr, flags);
459	if (ret < min_ret) {
460	if (ret == -EAGAIN && force_nonblock)
461	return io_setup_async_msg(req, kmsg);
462	if (ret == -ERESTARTSYS)
463	ret = -EINTR;
464	if (ret > 0 && io_net_retry(sock, flags)) {
465	sr->done_io += ret;
466	req->flags \|= REQ_F_PARTIAL_IO;
467	return io_setup_async_msg(req, kmsg);
468	}
469	req_set_fail(req);
470	} else if ((flags & MSG_WAITALL) && (kmsg->msg.msg_flags & (MSG_TRUNC \| MSG_CTRUNC))) {
471	req_set_fail(req);
472	}
473
474	/* fast path, check for non-NULL to avoid function call */
475	if (kmsg->free_iov)
476	kfree(kmsg->free_iov);
477	req->flags &= ~REQ_F_NEED_CLEANUP;
478	if (ret >= 0)
479	ret += sr->done_io;
480	else if (sr->done_io)
481	ret = sr->done_io;
482	cflags = io_put_kbuf(req, issue_flags);
483	if (kmsg->msg.msg_inq)
484	cflags \|= IORING_CQE_F_SOCK_NONEMPTY;
485	io_req_set_res(req, ret, cflags);
486	return IOU_OK;
487	}
488
489	int io_recv(struct io_kiocb *req, unsigned int issue_flags)
490	{
491	struct io_sr_msg *sr = io_kiocb_to_cmd(req);
492	struct msghdr msg;
493	struct socket *sock;
494	struct iovec iov;
495	unsigned int cflags;
496	unsigned flags;
497	int ret, min_ret = 0;
498	bool force_nonblock = issue_flags & IO_URING_F_NONBLOCK;
499
500	if (!(req->flags & REQ_F_POLLED) &&
501	(sr->flags & IORING_RECVSEND_POLL_FIRST))
502	return -EAGAIN;
503
504	sock = sock_from_file(req->file);
505	if (unlikely(!sock))
506	return -ENOTSOCK;
507
508	if (io_do_buffer_select(req)) {
509	void __user *buf;
510
511	buf = io_buffer_select(req, &sr->len, issue_flags);
512	if (!buf)
513	return -ENOBUFS;
514	sr->buf = buf;
515	}
516
517	ret = import_single_range(READ, sr->buf, sr->len, &iov, &msg.msg_iter);
518	if (unlikely(ret))
519	goto out_free;
520
521	msg.msg_name = NULL;
522	msg.msg_namelen = 0;
523	msg.msg_control = NULL;
524	msg.msg_get_inq = 1;
525	msg.msg_flags = 0;
526	msg.msg_controllen = 0;
527	msg.msg_iocb = NULL;
528
529	flags = sr->msg_flags;
530	if (force_nonblock)
531	flags \|= MSG_DONTWAIT;
532	if (flags & MSG_WAITALL)
533	min_ret = iov_iter_count(&msg.msg_iter);
534
535	ret = sock_recvmsg(sock, &msg, flags);
536	if (ret < min_ret) {
537	if (ret == -EAGAIN && force_nonblock)
538	return -EAGAIN;
539	if (ret == -ERESTARTSYS)
540	ret = -EINTR;
541	if (ret > 0 && io_net_retry(sock, flags)) {
542	sr->len -= ret;
543	sr->buf += ret;
544	sr->done_io += ret;
545	req->flags \|= REQ_F_PARTIAL_IO;
546	return -EAGAIN;
547	}
548	req_set_fail(req);
549	} else if ((flags & MSG_WAITALL) && (msg.msg_flags & (MSG_TRUNC \| MSG_CTRUNC))) {
550	out_free:
551	req_set_fail(req);
552	}
553
554	if (ret >= 0)
555	ret += sr->done_io;
556	else if (sr->done_io)
557	ret = sr->done_io;
558	cflags = io_put_kbuf(req, issue_flags);
559	if (msg.msg_inq)
560	cflags \|= IORING_CQE_F_SOCK_NONEMPTY;
561	io_req_set_res(req, ret, cflags);
562	return IOU_OK;
563	}
564
565	int io_accept_prep(struct io_kiocb req, const struct io_uring_sqe sqe)
566	{
567	struct io_accept *accept = io_kiocb_to_cmd(req);
568	unsigned flags;
569
570	if (sqe->len \|\| sqe->buf_index)
571	return -EINVAL;
572
573	accept->addr = u64_to_user_ptr(READ_ONCE(sqe->addr));
574	accept->addr_len = u64_to_user_ptr(READ_ONCE(sqe->addr2));
575	accept->flags = READ_ONCE(sqe->accept_flags);
576	accept->nofile = rlimit(RLIMIT_NOFILE);
577	flags = READ_ONCE(sqe->ioprio);
578	if (flags & ~IORING_ACCEPT_MULTISHOT)
579	return -EINVAL;
580
581	accept->file_slot = READ_ONCE(sqe->file_index);
582	if (accept->file_slot) {
583	if (accept->flags & SOCK_CLOEXEC)
584	return -EINVAL;
585	if (flags & IORING_ACCEPT_MULTISHOT &&
586	accept->file_slot != IORING_FILE_INDEX_ALLOC)
587	return -EINVAL;
588	}
589	if (accept->flags & ~(SOCK_CLOEXEC \| SOCK_NONBLOCK))
590	return -EINVAL;
591	if (SOCK_NONBLOCK != O_NONBLOCK && (accept->flags & SOCK_NONBLOCK))
592	accept->flags = (accept->flags & ~SOCK_NONBLOCK) \| O_NONBLOCK;
593	if (flags & IORING_ACCEPT_MULTISHOT)
594	req->flags \|= REQ_F_APOLL_MULTISHOT;
595	return 0;
596	}
597
598	int io_accept(struct io_kiocb *req, unsigned int issue_flags)
599	{
600	struct io_ring_ctx *ctx = req->ctx;
601	struct io_accept *accept = io_kiocb_to_cmd(req);
602	bool force_nonblock = issue_flags & IO_URING_F_NONBLOCK;
603	unsigned int file_flags = force_nonblock ? O_NONBLOCK : 0;
604	bool fixed = !!accept->file_slot;
605	struct file *file;
606	int ret, fd;
607
608	retry:
609	if (!fixed) {
610	fd = __get_unused_fd_flags(accept->flags, accept->nofile);
611	if (unlikely(fd < 0))
612	return fd;
613	}
614	file = do_accept(req->file, file_flags, accept->addr, accept->addr_len,
615	accept->flags);
616	if (IS_ERR(file)) {
617	if (!fixed)
618	put_unused_fd(fd);
619	ret = PTR_ERR(file);
620	if (ret == -EAGAIN && force_nonblock) {
621	/*
622	* if it's multishot and polled, we don't need to
623	* return EAGAIN to arm the poll infra since it
624	* has already been done
625	*/
626	if ((req->flags & IO_APOLL_MULTI_POLLED) ==
627	IO_APOLL_MULTI_POLLED)
628	ret = IOU_ISSUE_SKIP_COMPLETE;
629	return ret;
630	}
631	if (ret == -ERESTARTSYS)
632	ret = -EINTR;
633	req_set_fail(req);
634	} else if (!fixed) {
635	fd_install(fd, file);
636	ret = fd;
637	} else {
638	ret = io_fixed_fd_install(req, issue_flags, file,
639	accept->file_slot);
640	}
641
642	if (!(req->flags & REQ_F_APOLL_MULTISHOT)) {
643	io_req_set_res(req, ret, 0);
644	return IOU_OK;
645	}
f9ead18c	646
d245bca6 PB	647	if (ret < 0)
	648	return ret;
	649	if (io_post_aux_cqe(ctx, req->cqe.user_data, ret, IORING_CQE_F_MORE))
	650	goto retry;
	651	return -ECANCELED;
f9ead18c JA	652	}
	653
	654	int io_socket_prep(struct io_kiocb req, const struct io_uring_sqe sqe)
	655	{
	656	struct io_socket *sock = io_kiocb_to_cmd(req);
	657
	658	if (sqe->addr \|\| sqe->rw_flags \|\| sqe->buf_index)
	659	return -EINVAL;
	660
	661	sock->domain = READ_ONCE(sqe->fd);
	662	sock->type = READ_ONCE(sqe->off);
	663	sock->protocol = READ_ONCE(sqe->len);
	664	sock->file_slot = READ_ONCE(sqe->file_index);
	665	sock->nofile = rlimit(RLIMIT_NOFILE);
	666
	667	sock->flags = sock->type & ~SOCK_TYPE_MASK;
	668	if (sock->file_slot && (sock->flags & SOCK_CLOEXEC))
	669	return -EINVAL;
	670	if (sock->flags & ~(SOCK_CLOEXEC \| SOCK_NONBLOCK))
	671	return -EINVAL;
	672	return 0;
	673	}
	674
	675	int io_socket(struct io_kiocb *req, unsigned int issue_flags)
	676	{
	677	struct io_socket *sock = io_kiocb_to_cmd(req);
	678	bool fixed = !!sock->file_slot;
	679	struct file *file;
	680	int ret, fd;
	681
	682	if (!fixed) {
	683	fd = __get_unused_fd_flags(sock->flags, sock->nofile);
	684	if (unlikely(fd < 0))
	685	return fd;
	686	}
	687	file = __sys_socket_file(sock->domain, sock->type, sock->protocol);
	688	if (IS_ERR(file)) {
	689	if (!fixed)
	690	put_unused_fd(fd);
	691	ret = PTR_ERR(file);
	692	if (ret == -EAGAIN && (issue_flags & IO_URING_F_NONBLOCK))
	693	return -EAGAIN;
	694	if (ret == -ERESTARTSYS)
	695	ret = -EINTR;
	696	req_set_fail(req);
	697	} else if (!fixed) {
	698	fd_install(fd, file);
	699	ret = fd;
	700	} else {
	701	ret = io_fixed_fd_install(req, issue_flags, file,
	702	sock->file_slot);
	703	}
	704	io_req_set_res(req, ret, 0);
	705	return IOU_OK;
	706	}
	707
	708	int io_connect_prep_async(struct io_kiocb *req)
	709	{
	710	struct io_async_connect *io = req->async_data;
	711	struct io_connect *conn = io_kiocb_to_cmd(req);
	712
	713	return move_addr_to_kernel(conn->addr, conn->addr_len, &io->address);
	714	}
	715
716	int io_connect_prep(struct io_kiocb req, const struct io_uring_sqe sqe)
717	{
718	struct io_connect *conn = io_kiocb_to_cmd(req);
719
720	if (sqe->len \|\| sqe->buf_index \|\| sqe->rw_flags \|\| sqe->splice_fd_in)
721	return -EINVAL;
722
723	conn->addr = u64_to_user_ptr(READ_ONCE(sqe->addr));
724	conn->addr_len = READ_ONCE(sqe->addr2);
725	return 0;
726	}
727
728	int io_connect(struct io_kiocb *req, unsigned int issue_flags)
729	{
730	struct io_connect *connect = io_kiocb_to_cmd(req);
731	struct io_async_connect __io, *io;
732	unsigned file_flags;
733	int ret;
734	bool force_nonblock = issue_flags & IO_URING_F_NONBLOCK;
735
736	if (req_has_async_data(req)) {
737	io = req->async_data;
738	} else {
739	ret = move_addr_to_kernel(connect->addr,
740	connect->addr_len,
741	&__io.address);
742	if (ret)
743	goto out;
744	io = &__io;
745	}
746
747	file_flags = force_nonblock ? O_NONBLOCK : 0;
748
749	ret = __sys_connect_file(req->file, &io->address,
750	connect->addr_len, file_flags);
751	if ((ret == -EAGAIN \|\| ret == -EINPROGRESS) && force_nonblock) {
752	if (req_has_async_data(req))
753	return -EAGAIN;
754	if (io_alloc_async_data(req)) {
755	ret = -ENOMEM;
756	goto out;
757	}
758	memcpy(req->async_data, &__io, sizeof(__io));
759	return -EAGAIN;
760	}
761	if (ret == -ERESTARTSYS)
762	ret = -EINTR;
763	out:
764	if (ret < 0)
765	req_set_fail(req);
766	io_req_set_res(req, ret, 0);
767	return IOU_OK;
768	}
769	#endif