diff mbox series

[v6,7/8] io_uring/cmd: Introduce SOCKET_URING_OP_SETSOCKOPT

Message ID 20230913152744.2333228-8-leitao@debian.org (mailing list archive)
State New
Headers show
Series io_uring: Initial support for {s,g}etsockopt commands | expand

Commit Message

Breno Leitao Sept. 13, 2023, 3:27 p.m. UTC 
Add support for SOCKET_URING_OP_SETSOCKOPT. This new command is similar
to setsockopt(2). This implementation leverages the function
do_sock_setsockopt(), which is shared with the setsockopt() system call
path.

Important to say that userspace needs to keep the pointer's memory alive
until the operation is completed. I.e, the memory could not be
deallocated before the CQE is returned to userspace.

Reviewed-by: Gabriel Krisman Bertazi <krisman@suse.de>
Signed-off-by: Breno Leitao <leitao@debian.org>
---
 include/uapi/linux/io_uring.h |  1 +
 io_uring/uring_cmd.c          | 16 ++++++++++++++++
 2 files changed, 17 insertions(+)

Comments

Jens Axboe Sept. 13, 2023, 7:39 p.m. UTC | #1 
On 9/13/23 9:27 AM, Breno Leitao wrote:
> Add support for SOCKET_URING_OP_SETSOCKOPT. This new command is similar
> to setsockopt(2). This implementation leverages the function
> do_sock_setsockopt(), which is shared with the setsockopt() system call
> path.
> 
> Important to say that userspace needs to keep the pointer's memory alive
> until the operation is completed. I.e, the memory could not be
> deallocated before the CQE is returned to userspace.

This is different than other commands that write data. Since
IORING_FEAT_SUBMIT_STABLE was introduced, any command that writes data
should ensure that this data is stable. Eg it follows the life time of
the SQE, and doesn't need to be available until a CQE has been posted
for it. This is _generally_ true, even if we do have a few exceptions.

The problem is that then you cannot use user pointers, obviously, you'd
need to be able to pass in the value directly to do_sock_setsockopt()...
diff mbox series

Patch

diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
index 1c789ee6462d..99cdb2c1e240 100644
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -743,6 +743,7 @@  enum {
 	SOCKET_URING_OP_SIOCINQ		= 0,
 	SOCKET_URING_OP_SIOCOUTQ,
 	SOCKET_URING_OP_GETSOCKOPT,
+	SOCKET_URING_OP_SETSOCKOPT,
 };
 
 #ifdef __cplusplus
diff --git a/io_uring/uring_cmd.c b/io_uring/uring_cmd.c
index a2a6ac0c503b..2266c829d1c4 100644
--- a/io_uring/uring_cmd.c
+++ b/io_uring/uring_cmd.c
@@ -180,6 +180,20 @@  static inline int io_uring_cmd_getsockopt(struct socket *sock,
 	return do_sock_getsockopt(sock, compat, level, optname, optval, optlen);
 }
 
+static inline int io_uring_cmd_setsockopt(struct socket *sock,
+					  struct io_uring_cmd *cmd,
+					  unsigned int issue_flags)
+{
+	void __user *optval = u64_to_user_ptr(READ_ONCE(cmd->sqe->optval));
+	int optlen = READ_ONCE(cmd->sqe->optlen);
+	bool compat = !!(issue_flags & IO_URING_F_COMPAT);
+	int optname = READ_ONCE(cmd->sqe->optname);
+	int level = READ_ONCE(cmd->sqe->level);
+
+	return do_sock_setsockopt(sock, compat, level, optname, optval,
+				  optlen);
+}
+
 #if defined(CONFIG_NET)
 int io_uring_cmd_sock(struct io_uring_cmd *cmd, unsigned int issue_flags)
 {
@@ -204,6 +218,8 @@  int io_uring_cmd_sock(struct io_uring_cmd *cmd, unsigned int issue_flags)
 		return arg;
 	case SOCKET_URING_OP_GETSOCKOPT:
 		return io_uring_cmd_getsockopt(sock, cmd, issue_flags);
+	case SOCKET_URING_OP_SETSOCKOPT:
+		return io_uring_cmd_setsockopt(sock, cmd, issue_flags);
 	default:
 		return -EOPNOTSUPP;
 	}