Skip to content

Commit

Permalink
userspace: Respect tso/gso segment size.
Browse files Browse the repository at this point in the history 
Currently OVS will calculate the segment size based on the
MTU of the egress port. That usually happens to be correct
when the ports share the same MTU, but that is not always true.

Therefore, if the segment size is provided, then use that and
make sure the over sized packets are dropped.

Signed-off-by: Flavio Leitner <[email protected]>
Co-authored-by: Mike Pattrick <[email protected]>
Signed-off-by: Mike Pattrick <[email protected]>
Acked-by: Simon Horman <[email protected]>
Signed-off-by: Ilya Maximets <[email protected]>
  • Loading branch information
@fleitner @mkp-rh @igsilya
2 people authored and igsilya committed Dec 1, 2023
1 parent 6c59c19 commit e005601
Show file tree
Hide file tree
Showing 4 changed files with 95 additions and 23 deletions.
3 changes: 3 additions & 0 deletions lib/dp-packet.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -34,6 +34,7 @@ dp_packet_init__(struct dp_packet *b, size_t allocated, enum dp_packet_source so
pkt_metadata_init(&b->md, 0);
dp_packet_reset_cutlen(b);
dp_packet_reset_offload(b);
dp_packet_set_tso_segsz(b, 0);
/* Initialize implementation-specific fields of dp_packet. */
dp_packet_init_specific(b);
/* By default assume the packet type to be Ethernet. */
Expand Down Expand Up @@ -203,6 +204,8 @@ dp_packet_clone_with_headroom(const struct dp_packet *buffer, size_t headroom)
*dp_packet_ol_flags_ptr(new_buffer) = *dp_packet_ol_flags_ptr(buffer);
*dp_packet_ol_flags_ptr(new_buffer) &= DP_PACKET_OL_SUPPORTED_MASK;

dp_packet_set_tso_segsz(new_buffer, dp_packet_get_tso_segsz(buffer));

if (dp_packet_rss_valid(buffer)) {
dp_packet_set_rss_hash(new_buffer, dp_packet_get_rss_hash(buffer));
}
Expand Down
26 changes: 26 additions & 0 deletions lib/dp-packet.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -126,6 +126,7 @@ struct dp_packet {
uint32_t ol_flags; /* Offloading flags. */
uint32_t rss_hash; /* Packet hash. */
uint32_t flow_mark; /* Packet flow mark. */
uint16_t tso_segsz; /* TCP segment size. */
#endif
enum dp_packet_source source; /* Source of memory allocated as 'base'. */

Expand Down Expand Up @@ -166,6 +167,9 @@ static inline void dp_packet_set_size(struct dp_packet *, uint32_t);
static inline uint16_t dp_packet_get_allocated(const struct dp_packet *);
static inline void dp_packet_set_allocated(struct dp_packet *, uint16_t);

static inline uint16_t dp_packet_get_tso_segsz(const struct dp_packet *);
static inline void dp_packet_set_tso_segsz(struct dp_packet *, uint16_t);

void *dp_packet_resize_l2(struct dp_packet *, int increment);
void *dp_packet_resize_l2_5(struct dp_packet *, int increment);
static inline void *dp_packet_eth(const struct dp_packet *);
Expand Down Expand Up @@ -644,6 +648,17 @@ dp_packet_set_allocated(struct dp_packet *b, uint16_t s)
b->mbuf.buf_len = s;
}

static inline uint16_t
dp_packet_get_tso_segsz(const struct dp_packet *p)
{
return p->mbuf.tso_segsz;
}

static inline void
dp_packet_set_tso_segsz(struct dp_packet *p, uint16_t s)
{
p->mbuf.tso_segsz = s;
}
#else /* DPDK_NETDEV */

static inline void
Expand Down Expand Up @@ -700,6 +715,17 @@ dp_packet_set_allocated(struct dp_packet *b, uint16_t s)
b->allocated_ = s;
}

static inline uint16_t
dp_packet_get_tso_segsz(const struct dp_packet *p)
{
return p->tso_segsz;
}

static inline void
dp_packet_set_tso_segsz(struct dp_packet *p, uint16_t s)
{
p->tso_segsz = s;
}
#endif /* DPDK_NETDEV */

static inline void
Expand Down
12 changes: 11 additions & 1 deletion lib/netdev-dpdk.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2486,6 +2486,7 @@ netdev_dpdk_prep_hwol_packet(struct netdev_dpdk *dev, struct rte_mbuf *mbuf)

if (mbuf->ol_flags & RTE_MBUF_F_TX_TCP_SEG) {
struct tcp_header *th = dp_packet_l4(pkt);
int hdr_len;

if (!th) {
VLOG_WARN_RL(&rl, "%s: TCP Segmentation without L4 header"
Expand All @@ -2495,7 +2496,15 @@ netdev_dpdk_prep_hwol_packet(struct netdev_dpdk *dev, struct rte_mbuf *mbuf)

mbuf->l4_len = TCP_OFFSET(th->tcp_ctl) * 4;
mbuf->ol_flags |= RTE_MBUF_F_TX_TCP_CKSUM;
hdr_len = mbuf->l2_len + mbuf->l3_len + mbuf->l4_len;
mbuf->tso_segsz = dev->mtu - mbuf->l3_len - mbuf->l4_len;
if (OVS_UNLIKELY((hdr_len + mbuf->tso_segsz) > dev->max_packet_len)) {
VLOG_WARN_RL(&rl, "%s: Oversized TSO packet. "
"hdr: %"PRIu32", gso: %"PRIu32", max len: %"PRIu32"",
dev->up.name, hdr_len, mbuf->tso_segsz,
dev->max_packet_len);
return false;
}

if (mbuf->ol_flags & RTE_MBUF_F_TX_IPV4) {
mbuf->ol_flags |= RTE_MBUF_F_TX_IP_CKSUM;
Expand Down Expand Up @@ -2770,7 +2779,8 @@ netdev_dpdk_filter_packet_len(struct netdev_dpdk *dev, struct rte_mbuf **pkts,
int cnt = 0;
struct rte_mbuf *pkt;

/* Filter oversized packets, unless are marked for TSO. */
/* Filter oversized packets. The TSO packets are filtered out
* during the offloading preparation for performance reasons. */
for (i = 0; i < pkt_cnt; i++) {
pkt = pkts[i];
if (OVS_UNLIKELY((pkt->pkt_len > dev->max_packet_len)
Expand Down
77 changes: 55 additions & 22 deletions lib/netdev-linux.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -539,7 +539,7 @@ static atomic_count miimon_cnt = ATOMIC_COUNT_INIT(0);
static bool tap_supports_vnet_hdr = true;

static int netdev_linux_parse_vnet_hdr(struct dp_packet *b);
static void netdev_linux_prepend_vnet_hdr(struct dp_packet *b, int mtu);
static int netdev_linux_prepend_vnet_hdr(struct dp_packet *b, int mtu);
static int netdev_linux_do_ethtool(const char *name, struct ethtool_cmd *,
int cmd, const char *cmd_name);
static int get_flags(const struct netdev *, unsigned int *flags);
Expand Down Expand Up @@ -1597,9 +1597,10 @@ netdev_linux_rxq_drain(struct netdev_rxq *rxq_)
}

static int
netdev_linux_sock_batch_send(int sock, int ifindex, bool tso, int mtu,
struct dp_packet_batch *batch)
netdev_linux_sock_batch_send(struct netdev *netdev_, int sock, int ifindex,
bool tso, int mtu, struct dp_packet_batch *batch)
{
struct netdev_linux *netdev = netdev_linux_cast(netdev_);
const size_t size = dp_packet_batch_size(batch);
/* We don't bother setting most fields in sockaddr_ll because the
* kernel ignores them for SOCK_RAW. */
Expand All @@ -1608,26 +1609,36 @@ netdev_linux_sock_batch_send(int sock, int ifindex, bool tso, int mtu,

struct mmsghdr *mmsg = xmalloc(sizeof(*mmsg) * size);
struct iovec *iov = xmalloc(sizeof(*iov) * size);

struct dp_packet *packet;
int cnt = 0;

DP_PACKET_BATCH_FOR_EACH (i, packet, batch) {
if (tso) {
netdev_linux_prepend_vnet_hdr(packet, mtu);
}
int ret = netdev_linux_prepend_vnet_hdr(packet, mtu);

if (OVS_UNLIKELY(ret)) {
netdev->tx_dropped += 1;
VLOG_WARN_RL(&rl, "%s: Prepend vnet hdr failed, packet "
"dropped. %s", netdev_get_name(netdev_),
ovs_strerror(ret));
continue;
}
}

iov[i].iov_base = dp_packet_data(packet);
iov[i].iov_len = dp_packet_size(packet);
mmsg[i].msg_hdr = (struct msghdr) { .msg_name = &sll,
.msg_namelen = sizeof sll,
.msg_iov = &iov[i],
.msg_iovlen = 1 };
iov[cnt].iov_base = dp_packet_data(packet);
iov[cnt].iov_len = dp_packet_size(packet);
mmsg[cnt].msg_hdr = (struct msghdr) { .msg_name = &sll,
.msg_namelen = sizeof sll,
.msg_iov = &iov[cnt],
.msg_iovlen = 1 };
cnt++;
}

int error = 0;
for (uint32_t ofs = 0; ofs < size; ) {
for (uint32_t ofs = 0; ofs < cnt;) {
ssize_t retval;
do {
retval = sendmmsg(sock, mmsg + ofs, size - ofs, 0);
retval = sendmmsg(sock, mmsg + ofs, cnt - ofs, 0);
error = retval < 0 ? errno : 0;
} while (error == EINTR);
if (error) {
Expand Down Expand Up @@ -1669,7 +1680,14 @@ netdev_linux_tap_batch_send(struct netdev *netdev_, int mtu,
int error;

if (OVS_LIKELY(tap_supports_vnet_hdr)) {
netdev_linux_prepend_vnet_hdr(packet, mtu);
error = netdev_linux_prepend_vnet_hdr(packet, mtu);
if (OVS_UNLIKELY(error)) {
netdev->tx_dropped++;
VLOG_WARN_RL(&rl, "%s: Prepend vnet hdr failed, packet "
"dropped. %s", netdev_get_name(netdev_),
ovs_strerror(error));
continue;
}
}

size = dp_packet_size(packet);
Expand Down Expand Up @@ -1799,7 +1817,8 @@ netdev_linux_send(struct netdev *netdev_, int qid OVS_UNUSED,
goto free_batch;
}

error = netdev_linux_sock_batch_send(sock, ifindex, tso, mtu, batch);
error = netdev_linux_sock_batch_send(netdev_, sock, ifindex, tso, mtu,
batch);
} else {
error = netdev_linux_tap_batch_send(netdev_, mtu, batch);
}
Expand Down Expand Up @@ -7092,8 +7111,7 @@ netdev_linux_parse_vnet_hdr(struct dp_packet *b)
switch (vnet->gso_type) {
case VIRTIO_NET_HDR_GSO_TCPV4:
case VIRTIO_NET_HDR_GSO_TCPV6:
/* FIXME: The packet has offloaded TCP segmentation. The gso_size
* is given and needs to be respected. */
dp_packet_set_tso_segsz(b, (OVS_FORCE uint16_t) vnet->gso_size);
dp_packet_hwol_set_tcp_seg(b);
break;

Expand All @@ -7115,18 +7133,32 @@ netdev_linux_parse_vnet_hdr(struct dp_packet *b)
return ret;
}

static void
/* Prepends struct virtio_net_hdr to packet 'b'.
* Returns 0 if successful, otherwise a positive errno value.
* Returns EMSGSIZE if the packet 'b' cannot be sent over MTU 'mtu'. */
static int
netdev_linux_prepend_vnet_hdr(struct dp_packet *b, int mtu)
{
struct virtio_net_hdr v;
struct virtio_net_hdr *vnet = &v;

if (dp_packet_hwol_is_tso(b)) {
uint16_t hdr_len = ((char *)dp_packet_l4(b) - (char *)dp_packet_eth(b))
+ TCP_HEADER_LEN;
uint16_t tso_segsz = dp_packet_get_tso_segsz(b);
struct tcp_header *tcp = dp_packet_l4(b);
int tcp_hdr_len = TCP_OFFSET(tcp->tcp_ctl) * 4;
int hdr_len = ((char *) dp_packet_l4(b) - (char *) dp_packet_eth(b))
+ tcp_hdr_len;
int max_packet_len = mtu + ETH_HEADER_LEN + VLAN_HEADER_LEN;

if (OVS_UNLIKELY((hdr_len + tso_segsz) > max_packet_len)) {
VLOG_WARN_RL(&rl, "Oversized TSO packet. hdr_len: %"PRIu32", "
"gso: %"PRIu16", max length: %"PRIu32".", hdr_len,
tso_segsz, max_packet_len);
return EMSGSIZE;
}

vnet->hdr_len = (OVS_FORCE __virtio16)hdr_len;
vnet->gso_size = (OVS_FORCE __virtio16)(mtu - hdr_len);
vnet->gso_size = (OVS_FORCE __virtio16)(tso_segsz);
if (dp_packet_hwol_is_ipv4(b)) {
vnet->gso_type = VIRTIO_NET_HDR_GSO_TCPV4;
} else if (dp_packet_hwol_tx_ipv6(b)) {
Expand Down Expand Up @@ -7216,4 +7248,5 @@ netdev_linux_prepend_vnet_hdr(struct dp_packet *b, int mtu)
}

dp_packet_push(b, vnet, sizeof *vnet);
return 0;
}

0 comments on commit e005601

Please sign in to comment.