sock: Make sock->sk_stamp thread-safe

Al Viro mentioned (Message-ID
<20170626041334.GZ10672@ZenIV.linux.org.uk>)
that there is probably a race condition
lurking in accesses of sk_stamp on 32-bit machines.

sock->sk_stamp is of type ktime_t which is always an s64.
On a 32 bit architecture, we might run into situations of
unsafe access as the access to the field becomes non atomic.

Use seqlocks for synchronization.
This allows us to avoid using spinlocks for readers as
readers do not need mutual exclusion.

Another approach to solve this is to require sk_lock for all
modifications of the timestamps. The current approach allows
for timestamps to have their own lock: sk_stamp_lock.
This allows for the patch to not compete with already
existing critical sections, and side effects are limited
to the paths in the patch.

The addition of the new field maintains the data locality
optimizations from
commit 9115e8cd2a0c ("net: reorganize struct sock for better data
locality")

Note that all the instances of the sk_stamp accesses
are either through the ioctl or the syscall recvmsg.

Signed-off-by: Deepa Dinamani <deepa.kernel@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

3 files changed, 20 insertions, 12 deletions

diff --git a/net/compat.c b/net/compat.c
index 47a614b370cd..d1f3a8a0b3ef 100644
--- a/net/compat.c
+++ b/net/compat.c
@@ -467,12 +467,14 @@ int compat_sock_get_timestamp(struct sock *sk, struct timeval __user *userstamp)
 	ctv = (struct compat_timeval __user *) userstamp;
 	err = -ENOENT;
 	sock_enable_timestamp(sk, SOCK_TIMESTAMP);
-	tv = ktime_to_timeval(sk->sk_stamp);
+	tv = ktime_to_timeval(sock_read_timestamp(sk));
+
 	if (tv.tv_sec == -1)
 		return err;
 	if (tv.tv_sec == 0) {
-		sk->sk_stamp = ktime_get_real();
-		tv = ktime_to_timeval(sk->sk_stamp);
+		ktime_t kt = ktime_get_real();
+		sock_write_timestamp(sk, kt);
+		tv = ktime_to_timeval(kt);
 	}
 	err = 0;
 	if (put_user(tv.tv_sec, &ctv->tv_sec) ||
@@ -494,12 +496,13 @@ int compat_sock_get_timestampns(struct sock *sk, struct timespec __user *usersta
 	ctv = (struct compat_timespec __user *) userstamp;
 	err = -ENOENT;
 	sock_enable_timestamp(sk, SOCK_TIMESTAMP);
-	ts = ktime_to_timespec(sk->sk_stamp);
+	ts = ktime_to_timespec(sock_read_timestamp(sk));
 	if (ts.tv_sec == -1)
 		return err;
 	if (ts.tv_sec == 0) {
-		sk->sk_stamp = ktime_get_real();
-		ts = ktime_to_timespec(sk->sk_stamp);
+		ktime_t kt = ktime_get_real();
+		sock_write_timestamp(sk, kt);
+		ts = ktime_to_timespec(kt);
 	}
 	err = 0;
 	if (put_user(ts.tv_sec, &ctv->tv_sec) ||
diff --git a/net/core/sock.c b/net/core/sock.c
index f00902c532cc..6aa2e7e0b4fb 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -2751,6 +2751,9 @@ void sock_init_data(struct socket *sock, struct sock *sk)
 	sk->sk_sndtimeo		=	MAX_SCHEDULE_TIMEOUT;
 
 	sk->sk_stamp = SK_DEFAULT_STAMP;
+#if BITS_PER_LONG==32
+	seqlock_init(&sk->sk_stamp_seq);
+#endif
 	atomic_set(&sk->sk_zckey, 0);
 
 #ifdef CONFIG_NET_RX_BUSY_POLL
@@ -2850,12 +2853,13 @@ int sock_get_timestamp(struct sock *sk, struct timeval __user *userstamp)
 	struct timeval tv;
 
 	sock_enable_timestamp(sk, SOCK_TIMESTAMP);
-	tv = ktime_to_timeval(sk->sk_stamp);
+	tv = ktime_to_timeval(sock_read_timestamp(sk));
 	if (tv.tv_sec == -1)
 		return -ENOENT;
 	if (tv.tv_sec == 0) {
-		sk->sk_stamp = ktime_get_real();
-		tv = ktime_to_timeval(sk->sk_stamp);
+		ktime_t kt = ktime_get_real();
+		sock_write_timestamp(sk, kt);
+		tv = ktime_to_timeval(kt);
 	}
 	return copy_to_user(userstamp, &tv, sizeof(tv)) ? -EFAULT : 0;
 }
@@ -2866,11 +2870,12 @@ int sock_get_timestampns(struct sock *sk, struct timespec __user *userstamp)
 	struct timespec ts;
 
 	sock_enable_timestamp(sk, SOCK_TIMESTAMP);
-	ts = ktime_to_timespec(sk->sk_stamp);
+	ts = ktime_to_timespec(sock_read_timestamp(sk));
 	if (ts.tv_sec == -1)
 		return -ENOENT;
 	if (ts.tv_sec == 0) {
-		sk->sk_stamp = ktime_get_real();
+		ktime_t kt = ktime_get_real();
+		sock_write_timestamp(sk, kt);
 		ts = ktime_to_timespec(sk->sk_stamp);
 	}
 	return copy_to_user(userstamp, &ts, sizeof(ts)) ? -EFAULT : 0;
diff --git a/net/sunrpc/svcsock.c b/net/sunrpc/svcsock.c
index 986f3ed7d1a2..b7e67310ec37 100644
--- a/net/sunrpc/svcsock.c
+++ b/net/sunrpc/svcsock.c
@@ -549,7 +549,7 @@ static int svc_udp_recvfrom(struct svc_rqst *rqstp)
 		/* Don't enable netstamp, sunrpc doesn't
 		   need that much accuracy */
 	}
-	svsk->sk_sk->sk_stamp = skb->tstamp;
+	sock_write_timestamp(svsk->sk_sk, skb->tstamp);
 	set_bit(XPT_DATA, &svsk->sk_xprt.xpt_flags); /* there may be more data... */
 
 	len  = skb->len;