RE: [PATCH 15/17] switch kernel_sendmsg() and kernel_recvmsg() to iov_iter_kvec()

From
David Laight <hidden>
Date
2015-04-14 16:22:57
Reply to
Al Viro, 2015-04-11
From: Al Viro
quoted hunk ↗ jump to hunk
Sent: 11 April 2015 22:18
From: Al Viro <viro@zeniv.linux.org.uk>

For kernel_sendmsg() that eliminates the need to play with setfs();
for kernel_recvmsg() it does *not* - a couple of callers are using
it with non-NULL ->msg_control, which would be treated as userland
address on recvmsg side of things.

In all cases we are really setting a kvec-backed iov_iter, though.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 net/socket.c | 20 +++-----------------
 1 file changed, 3 insertions(+), 17 deletions(-)
diff --git a/net/socket.c b/net/socket.c
index e5669ce..b6ceeda 100644
--- a/net/socket.c
+++ b/net/socket.c
@@ -627,18 +627,8 @@ EXPORT_SYMBOL(sock_sendmsg);
 int kernel_sendmsg(struct socket *sock, struct msghdr *msg,
 		   struct kvec *vec, size_t num, size_t size)
 {
-	mm_segment_t oldfs = get_fs();
-	int result;
-
-	set_fs(KERNEL_DS);
-	/*
-	 * the following is safe, since for compiler definitions of kvec and
-	 * iovec are identical, yielding the same in-core layout and alignment
-	 */
-	iov_iter_init(&msg->msg_iter, WRITE, (struct iovec *)vec, num, size);
-	result = sock_sendmsg(sock, msg, size);
-	set_fs(oldfs);
-	return result;
+	iov_iter_kvec(&msg->msg_iter, WRITE | ITER_KVEC, vec, num, size);
+	return sock_sendmsg(sock, msg, size);
Massive NAK.
This breaks any code that is using msg_control to set SCTP parameters
when sending data.

	David
Thread (36 messages) 36 messages, 6 authors, last reply 2015-04-15
Keyboard shortcuts
hback out one level
jnext message in thread / next series-diff position
kprevious message in thread / previous series-diff position
ldrill in
Escclose help / fold thread tree
?toggle this help