@@ -19,6 +19,7 @@
#include <net/dst.h>
#include <net/xfrm.h>
#include <net/xdp.h>
+#include <net/udp.h>
#include <linux/veth.h>
#include <linux/module.h>
#include <linux/bpf.h>
@@ -335,9 +336,12 @@ static bool veth_skb_is_eligible_for_gro(const struct net_device *dev,
const struct net_device *rcv,
const struct sk_buff *skb)
{
- return !(dev->features & NETIF_F_ALL_TSO) ||
- (skb->destructor == sock_wfree &&
- rcv->features & (NETIF_F_GRO_FRAGLIST | NETIF_F_GRO_UDP_FWD));
+ if (skb->destructor == sock_wfree ||
+ (IS_ENABLED(CONFIG_INET) && skb->destructor == udp_wfree))
+ return rcv->features & (NETIF_F_GRO_FRAGLIST |
+ NETIF_F_GRO_UDP_FWD);
+
+ return !(dev->features & NETIF_F_ALL_TSO);
}
static netdev_tx_t veth_xmit(struct sk_buff *skb, struct net_device *dev)
@@ -1811,6 +1811,7 @@ static inline bool sock_allow_reclassification(const struct sock *csk)
struct sock *sk_alloc(struct net *net, int family, gfp_t priority,
struct proto *prot, int kern);
void sk_free(struct sock *sk);
+void __sk_free(struct sock *sk);
void sk_destruct(struct sock *sk);
struct sock *sk_clone_lock(const struct sock *sk, const gfp_t priority);
void sk_free_unlock_clone(struct sock *sk);
@@ -288,6 +288,7 @@ int udp_pre_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
int __udp_disconnect(struct sock *sk, int flags);
int udp_disconnect(struct sock *sk, int flags);
__poll_t udp_poll(struct file *file, struct socket *sock, poll_table *wait);
+void udp_wfree(struct sk_buff *skb);
struct sk_buff *skb_udp_tunnel_segment(struct sk_buff *skb,
netdev_features_t features,
bool is_ipv6);
@@ -140,6 +140,7 @@
#include <trace/events/sock.h>
#include <net/tcp.h>
+#include <net/udp.h>
#include <net/busy_poll.h>
#include <net/phonet/phonet.h>
@@ -2221,7 +2222,7 @@ void sk_destruct(struct sock *sk)
__sk_destruct(&sk->sk_rcu);
}
-static void __sk_free(struct sock *sk)
+void __sk_free(struct sock *sk)
{
if (likely(sk->sk_net_refcnt))
sock_inuse_add(sock_net(sk), -1);
@@ -2231,6 +2232,7 @@ static void __sk_free(struct sock *sk)
else
sk_destruct(sk);
}
+EXPORT_SYMBOL(__sk_free);
void sk_free(struct sock *sk)
{
@@ -2531,8 +2533,10 @@ static bool can_skb_orphan_partial(const struct sk_buff *skb)
if (skb->decrypted)
return false;
#endif
- return (skb->destructor == sock_wfree ||
- (IS_ENABLED(CONFIG_INET) && skb->destructor == tcp_wfree));
+ if (skb->destructor == sock_wfree)
+ return true;
+ return IS_ENABLED(CONFIG_INET) &&
+ (skb->destructor == tcp_wfree || skb->destructor == udp_wfree);
}
/* This helper is used by netem, as it can hold packets in its
@@ -796,6 +796,42 @@ int udp_err(struct sk_buff *skb, u32 info)
return __udp4_lib_err(skb, info, dev_net(skb->dev)->ipv4.udp_table);
}
+static inline bool __udp_wfree(struct sk_buff *skb)
+{
+ struct socket_wq *wq;
+ struct sock *sk = skb->sk;
+ bool free;
+
+ free = refcount_sub_and_test(skb->truesize, &sk->sk_wmem_alloc);
+ /* a full barrier is required before waitqueue_active() */
+ smp_mb__after_atomic();
+
+ if (!sock_writeable(sk))
+ goto out;
+
+ wq = rcu_dereference(sk->sk_wq);
+ if (wq && waitqueue_active(&wq->wait))
+ wake_up_interruptible_sync_poll(&wq->wait, EPOLLOUT |
+ EPOLLWRNORM | EPOLLWRBAND);
+ /* Should agree with poll, otherwise some programs break */
+ sk_wake_async(sk, SOCK_WAKE_SPACE, POLL_OUT);
+out:
+ return free;
+}
+
+void udp_wfree(struct sk_buff *skb)
+{
+ bool free;
+
+ rcu_read_lock();
+ free = __udp_wfree(skb);
+ rcu_read_unlock();
+
+ if (unlikely(free))
+ __sk_free(skb->sk);
+}
+EXPORT_SYMBOL_GPL(udp_wfree);
+
/*
* Throw away all pending data and cancel the corking. Socket is locked.
*/
@@ -989,6 +1025,7 @@ int udp_push_pending_frames(struct sock *sk)
if (!skb)
goto out;
+ skb->destructor = udp_wfree;
err = udp_send_skb(skb, fl4, &inet->cork.base);
out:
@@ -1246,8 +1283,10 @@ int udp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
sizeof(struct udphdr), &ipc, &rt,
&cork, msg->msg_flags);
err = PTR_ERR(skb);
- if (!IS_ERR_OR_NULL(skb))
+ if (!IS_ERR_OR_NULL(skb)) {
+ skb->destructor = udp_wfree;
err = udp_send_skb(skb, fl4, &cork);
+ }
goto out;
}
@@ -1309,6 +1309,7 @@ static int udp_v6_push_pending_frames(struct sock *sk)
if (!skb)
goto out;
+ skb->destructor = udp_wfree;
err = udp_v6_send_skb(skb, &inet_sk(sk)->cork.fl.u.ip6,
&inet_sk(sk)->cork.base);
out:
@@ -1576,8 +1577,10 @@ int udpv6_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
(struct rt6_info *)dst,
msg->msg_flags, &cork);
err = PTR_ERR(skb);
- if (!IS_ERR_OR_NULL(skb))
+ if (!IS_ERR_OR_NULL(skb)) {
+ skb->destructor = udp_wfree;
err = udp_v6_send_skb(skb, fl6, &cork.base);
+ }
/* ip6_make_skb steals dst reference */
goto out_no_dst;
}
Introduce a UDP specific skb destructor, udp_wfree(). We'll need it in the next patch, which wires SOCK_NOSPACE support for udp sockets. We can't reuse sock_wfree() there as is because SOCK_NOSPACE also requires support from the poll callback. Signed-off-by: Pavel Begunkov <asml.silence@gmail.com> --- drivers/net/veth.c | 10 +++++++--- include/net/sock.h | 1 + include/net/udp.h | 1 + net/core/sock.c | 10 +++++++--- net/ipv4/udp.c | 41 ++++++++++++++++++++++++++++++++++++++++- net/ipv6/udp.c | 5 ++++- 6 files changed, 60 insertions(+), 8 deletions(-)