diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 73902acf2b71c8800d81b744a936a7420f33b459..75d50ab7997cdabba047dff05eb0297db651f369 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -481,10 +481,11 @@ static inline void sock_zerocopy_get(struct ubuf_info *uarg)
}
void sock_zerocopy_put(struct ubuf_info *uarg);
-void sock_zerocopy_put_abort(struct ubuf_info *uarg);
+void sock_zerocopy_put_abort(struct ubuf_info *uarg, bool have_uref);
void sock_zerocopy_callback(struct ubuf_info *uarg, bool success);
+int skb_zerocopy_iter_dgram(struct sk_buff *skb, struct msghdr *msg, int len);
int skb_zerocopy_iter_stream(struct sock *sk, struct sk_buff *skb,
struct msghdr *msg, int len,
struct ubuf_info *uarg);
@@ -1325,10 +1326,14 @@ static inline struct ubuf_info *skb_zcopy(struct sk_buff *skb)
return is_zcopy ? skb_uarg(skb) : NULL;
}
-static inline void skb_zcopy_set(struct sk_buff *skb, struct ubuf_info *uarg)
+static inline void skb_zcopy_set(struct sk_buff *skb, struct ubuf_info *uarg,
+ bool *have_ref)
{
if (skb && uarg && !skb_zcopy(skb)) {
- sock_zerocopy_get(uarg);
+ if (unlikely(have_ref && *have_ref))
+ *have_ref = false;
+ else
+ sock_zerocopy_get(uarg);
skb_shinfo(skb)->destructor_arg = uarg;
skb_shinfo(skb)->tx_flags |= SKBTX_ZEROCOPY_FRAG;
}
@@ -1373,7 +1378,7 @@ static inline void skb_zcopy_abort(struct sk_buff *skb)
struct ubuf_info *uarg = skb_zcopy(skb);
if (uarg) {
- sock_zerocopy_put_abort(uarg);
+ sock_zerocopy_put_abort(uarg, false);
skb_shinfo(skb)->tx_flags &= ~SKBTX_ZEROCOPY_FRAG;
}
}
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index 3c814565ed7cd89af02b7169fc92b7224a699c0a..c78ce114537e097d532eb112d8c4eb86c724f6ea 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -1089,7 +1089,7 @@ void sock_zerocopy_put(struct ubuf_info *uarg)
}
EXPORT_SYMBOL_GPL(sock_zerocopy_put);
-void sock_zerocopy_put_abort(struct ubuf_info *uarg)
+void sock_zerocopy_put_abort(struct ubuf_info *uarg, bool have_uref)
{
if (uarg) {
struct sock *sk = skb_from_uarg(uarg)->sk;
@@ -1097,7 +1097,8 @@ void sock_zerocopy_put_abort(struct ubuf_info *uarg)
atomic_dec(&sk->sk_zckey);
uarg->len--;
- sock_zerocopy_put(uarg);
+ if (have_uref)
+ sock_zerocopy_put(uarg);
}
}
EXPORT_SYMBOL_GPL(sock_zerocopy_put_abort);
@@ -1105,6 +1106,12 @@ EXPORT_SYMBOL_GPL(sock_zerocopy_put_abort);
extern int __zerocopy_sg_from_iter(struct sock *sk, struct sk_buff *skb,
struct iov_iter *from, size_t length);
+int skb_zerocopy_iter_dgram(struct sk_buff *skb, struct msghdr *msg, int len)
+{
+ return __zerocopy_sg_from_iter(skb->sk, skb, &msg->msg_iter, len);
+}
+EXPORT_SYMBOL_GPL(skb_zerocopy_iter_dgram);
+
int skb_zerocopy_iter_stream(struct sock *sk, struct sk_buff *skb,
struct msghdr *msg, int len,
struct ubuf_info *uarg)
@@ -1131,7 +1138,7 @@ int skb_zerocopy_iter_stream(struct sock *sk, struct sk_buff *skb,
return err;
}
- skb_zcopy_set(skb, uarg);
+ skb_zcopy_set(skb, uarg, NULL);
return skb->len - orig_len;
}
EXPORT_SYMBOL_GPL(skb_zerocopy_iter_stream);
@@ -1151,7 +1158,7 @@ static int skb_zerocopy_clone(struct sk_buff *nskb, struct sk_buff *orig,
if (skb_copy_ubufs(nskb, GFP_ATOMIC))
return -EIO;
}
- skb_zcopy_set(nskb, skb_uarg(orig));
+ skb_zcopy_set(nskb, skb_uarg(orig), NULL);
}
return 0;
}
diff --git a/net/core/sock.c b/net/core/sock.c
index 6d7e189e3cd9f85f8b5375d79c89b9e13d9daa64..f5bb89785e474951423feeb8307b223b39363807 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -1018,7 +1018,10 @@ int sock_setsockopt(struct socket *sock, int level, int optname,
case SO_ZEROCOPY:
if (sk->sk_family == PF_INET || sk->sk_family == PF_INET6) {
- if (sk->sk_protocol != IPPROTO_TCP)
+ if (!((sk->sk_type == SOCK_STREAM &&
+ sk->sk_protocol == IPPROTO_TCP) ||
+ (sk->sk_type == SOCK_DGRAM &&
+ sk->sk_protocol == IPPROTO_UDP)))
ret = -ENOTSUPP;
} else if (sk->sk_family != PF_RDS) {
ret = -ENOTSUPP;
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index 5dbec21856f4ce458af136bfe5ef1c0c84a4f8a5..78f028bdad30b4c6f2d89a9488f2bd79c79f672c 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -867,6 +867,7 @@ static int __ip_append_data(struct sock *sk,
unsigned int flags)
{
struct inet_sock *inet = inet_sk(sk);
+ struct ubuf_info *uarg = NULL;
struct sk_buff *skb;
struct ip_options *opt = cork->opt;
@@ -880,8 +881,8 @@ static int __ip_append_data(struct sock *sk,
int csummode = CHECKSUM_NONE;
struct rtable *rt = (struct rtable *)cork->dst;
unsigned int wmem_alloc_delta = 0;
+ bool paged, extra_uref;
u32 tskey = 0;
- bool paged;
skb = skb_peek_tail(queue);
@@ -916,6 +917,20 @@ static int __ip_append_data(struct sock *sk,
(!exthdrlen || (rt->dst.dev->features & NETIF_F_HW_ESP_TX_CSUM)))
csummode = CHECKSUM_PARTIAL;
+ if (flags & MSG_ZEROCOPY && length && sock_flag(sk, SOCK_ZEROCOPY)) {
+ uarg = sock_zerocopy_realloc(sk, length, skb_zcopy(skb));
+ if (!uarg)
+ return -ENOBUFS;
+ extra_uref = true;
+ if (rt->dst.dev->features & NETIF_F_SG &&
+ csummode == CHECKSUM_PARTIAL) {
+ paged = true;
+ } else {
+ uarg->zerocopy = 0;
+ skb_zcopy_set(skb, uarg, &extra_uref);
+ }
+ }
+
cork->length += length;
/* So, what's going on in the loop below?
@@ -1001,12 +1016,6 @@ static int __ip_append_data(struct sock *sk,
skb->csum = 0;
skb_reserve(skb, hh_len);
- /* only the initial fragment is time stamped */
- skb_shinfo(skb)->tx_flags = cork->tx_flags;
- cork->tx_flags = 0;
- skb_shinfo(skb)->tskey = tskey;
- tskey = 0;
-
/*
* Find where to start putting bytes.
*/
@@ -1039,6 +1048,13 @@ static int __ip_append_data(struct sock *sk,
exthdrlen = 0;
csummode = CHECKSUM_NONE;
+ /* only the initial fragment is time stamped */
+ skb_shinfo(skb)->tx_flags = cork->tx_flags;
+ cork->tx_flags = 0;
+ skb_shinfo(skb)->tskey = tskey;
+ tskey = 0;
+ skb_zcopy_set(skb, uarg, &extra_uref);
+
if ((flags & MSG_CONFIRM) && !skb_prev)
skb_set_dst_pending_confirm(skb, 1);
@@ -1068,7 +1084,7 @@ static int __ip_append_data(struct sock *sk,
err = -EFAULT;
goto error;
}
- } else {
+ } else if (!uarg || !uarg->zerocopy) {
int i = skb_shinfo(skb)->nr_frags;
err = -ENOMEM;
@@ -1098,6 +1114,10 @@ static int __ip_append_data(struct sock *sk,
skb->data_len += copy;
skb->truesize += copy;
wmem_alloc_delta += copy;
+ } else {
+ err = skb_zerocopy_iter_dgram(skb, from, copy);
+ if (err < 0)
+ goto error;
}
offset += copy;
length -= copy;
@@ -1110,6 +1130,7 @@ static int __ip_append_data(struct sock *sk,
error_efault:
err = -EFAULT;
error:
+ sock_zerocopy_put_abort(uarg, extra_uref);
cork->length -= length;
IP_INC_STATS(sock_net(sk), IPSTATS_MIB_OUTDISCARDS);
refcount_add(wmem_alloc_delta, &sk->sk_wmem_alloc);
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index 215e4d3b361657c434ff019ebe638437b65dd351..dc68c408bba09af3273882a1c7ffae1ec03ba1d3 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -1423,7 +1423,7 @@ int tcp_sendmsg_locked(struct sock *sk, struct msghdr *msg, size_t size)
if (copied + copied_syn)
goto out;
out_err:
- sock_zerocopy_put_abort(uarg);
+ sock_zerocopy_put_abort(uarg, true);
err = sk_stream_error(sk, flags, err);
/* make sure we wake any epoll edge trigger waiter */
if (unlikely(skb_queue_len(&sk->sk_write_queue) == 0 &&
diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index 827a3f5ff3bbdff2e66e09bb90b5b2195b9ef297..ec8c235ea891a7dfdc978e61fa71aee0d5abcd17 100644
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -1245,6 +1245,7 @@ static int __ip6_append_data(struct sock *sk,
{
struct sk_buff *skb, *skb_prev = NULL;
unsigned int maxfraglen, fragheaderlen, mtu, orig_mtu, pmtu;
+ struct ubuf_info *uarg = NULL;
int exthdrlen = 0;
int dst_exthdrlen = 0;
int hh_len;
@@ -1257,7 +1258,7 @@ static int __ip6_append_data(struct sock *sk,
int csummode = CHECKSUM_NONE;
unsigned int maxnonfragsize, headersize;
unsigned int wmem_alloc_delta = 0;
- bool paged;
+ bool paged, extra_uref;
skb = skb_peek_tail(queue);
if (!skb) {
@@ -1322,6 +1323,20 @@ static int __ip6_append_data(struct sock *sk,
rt->dst.dev->features & (NETIF_F_IPV6_CSUM | NETIF_F_HW_CSUM))
csummode = CHECKSUM_PARTIAL;
+ if (flags & MSG_ZEROCOPY && length && sock_flag(sk, SOCK_ZEROCOPY)) {
+ uarg = sock_zerocopy_realloc(sk, length, skb_zcopy(skb));
+ if (!uarg)
+ return -ENOBUFS;
+ extra_uref = true;
+ if (rt->dst.dev->features & NETIF_F_SG &&
+ csummode == CHECKSUM_PARTIAL) {
+ paged = true;
+ } else {
+ uarg->zerocopy = 0;
+ skb_zcopy_set(skb, uarg, &extra_uref);
+ }
+ }
+
/*
* Let's try using as much space as possible.
* Use MTU if total length of the message fits into the MTU.
@@ -1440,12 +1455,6 @@ static int __ip6_append_data(struct sock *sk,
skb_reserve(skb, hh_len + sizeof(struct frag_hdr) +
dst_exthdrlen);
- /* Only the initial fragment is time stamped */
- skb_shinfo(skb)->tx_flags = cork->tx_flags;
- cork->tx_flags = 0;
- skb_shinfo(skb)->tskey = tskey;
- tskey = 0;
-
/*
* Find where to start putting bytes
*/
@@ -1477,6 +1486,13 @@ static int __ip6_append_data(struct sock *sk,
exthdrlen = 0;
dst_exthdrlen = 0;
+ /* Only the initial fragment is time stamped */
+ skb_shinfo(skb)->tx_flags = cork->tx_flags;
+ cork->tx_flags = 0;
+ skb_shinfo(skb)->tskey = tskey;
+ tskey = 0;
+ skb_zcopy_set(skb, uarg, &extra_uref);
+
if ((flags & MSG_CONFIRM) && !skb_prev)
skb_set_dst_pending_confirm(skb, 1);
@@ -1506,7 +1522,7 @@ static int __ip6_append_data(struct sock *sk,
err = -EFAULT;
goto error;
}
- } else {
+ } else if (!uarg || !uarg->zerocopy) {
int i = skb_shinfo(skb)->nr_frags;
err = -ENOMEM;
@@ -1536,6 +1552,10 @@ static int __ip6_append_data(struct sock *sk,
skb->data_len += copy;
skb->truesize += copy;
wmem_alloc_delta += copy;
+ } else {
+ err = skb_zerocopy_iter_dgram(skb, from, copy);
+ if (err < 0)
+ goto error;
}
offset += copy;
length -= copy;
@@ -1548,6 +1568,7 @@ static int __ip6_append_data(struct sock *sk,
error_efault:
err = -EFAULT;
error:
+ sock_zerocopy_put_abort(uarg, extra_uref);
cork->length -= length;
IP6_INC_STATS(sock_net(sk), rt->rt6i_idev, IPSTATS_MIB_OUTDISCARDS);
refcount_add(wmem_alloc_delta, &sk->sk_wmem_alloc);
diff --git a/tools/testing/selftests/net/msg_zerocopy.c b/tools/testing/selftests/net/msg_zerocopy.c
index 406cc70c571dea0c1bfef19165e8cfa251a2cf39..4b02933cab8aa90e9f678739bfe9031a19079341 100644
--- a/tools/testing/selftests/net/msg_zerocopy.c
+++ b/tools/testing/selftests/net/msg_zerocopy.c
@@ -651,12 +651,13 @@ static void do_flush_datagram(int fd, int type)
static void do_rx(int domain, int type, int protocol)
{
+ const int cfg_receiver_wait_ms = 400;
uint64_t tstop;
int fd;
fd = do_setup_rx(domain, type, protocol);
- tstop = gettimeofday_ms() + cfg_runtime_ms;
+ tstop = gettimeofday_ms() + cfg_runtime_ms + cfg_receiver_wait_ms;
do {
if (type == SOCK_STREAM)
do_flush_tcp(fd);
diff --git a/tools/testing/selftests/net/msg_zerocopy.sh b/tools/testing/selftests/net/msg_zerocopy.sh
index c43c6debda06913ab54542e5fd91c33d4f23b297..825ffec85cea3e23cd672541bb331302f6c66aa0 100755
--- a/tools/testing/selftests/net/msg_zerocopy.sh
+++ b/tools/testing/selftests/net/msg_zerocopy.sh
@@ -25,6 +25,8 @@ readonly path_sysctl_mem="net.core.optmem_max"
if [[ "$#" -eq "0" ]]; then
$0 4 tcp -t 1
$0 6 tcp -t 1
+ $0 4 udp -t 1
+ $0 6 udp -t 1
echo "OK. All tests passed"
exit 0
fi
diff --git a/tools/testing/selftests/net/udpgso_bench.sh b/tools/testing/selftests/net/udpgso_bench.sh
index 0f0628613f818d8bad788683e5ab5e4fb0b7277a..5670a9ffd8eb04775ec5dea21815057914e99ebc 100755
--- a/tools/testing/selftests/net/udpgso_bench.sh
+++ b/tools/testing/selftests/net/udpgso_bench.sh
@@ -35,6 +35,9 @@ run_udp() {
echo "udp gso"
run_in_netns ${args} -S 0
+
+ echo "udp gso zerocopy"
+ run_in_netns ${args} -S 0 -z
}
run_tcp() {