Currently, if a non-SYN TCP packet arrives which doesn't match any existing
connection, we simply ignore it. However RFC 9293, section 3.10.7.1 says
we should respond with an RST to a non-SYN, non-RST packet that's for a
CLOSED (i.e. non-existent) connection.
This can arise in practice with migration, in cases where some error means
we have to discard a connection. We destroy the connection with tcp_rst()
in that case, but because the guest is stopped, we may not be able to
deliver the RST packet on the tap interface immediately. This change
ensures an RST will be sent if the guest tries to use the connection again.
A similar situation can arise if a passt/pasta instance is killed or
crashes, but is then replaced with another attached to the same guest.
This can leave the guest with stale connections that the new passt instance
isn't aware of. It's better to send an RST so the guest knows quickly
these are broken, rather than letting them linger until they time out.
Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au>
---
tap.c | 17 +++++++-------
tap.h | 6 +++++
tcp.c | 74 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-
tcp.h | 2 +-
4 files changed, 88 insertions(+), 11 deletions(-)
diff --git a/tap.c b/tap.c
index 202abae9..86d051e3 100644
--- a/tap.c
+++ b/tap.c
@@ -122,7 +122,7 @@ const struct in6_addr *tap_ip6_daddr(const struct ctx *c,
*
* Return: pointer at which to write the packet's payload
*/
-static void *tap_push_l2h(const struct ctx *c, void *buf, uint16_t proto)
+void *tap_push_l2h(const struct ctx *c, void *buf, uint16_t proto)
{
struct ethhdr *eh = (struct ethhdr *)buf;
@@ -143,8 +143,8 @@ static void *tap_push_l2h(const struct ctx *c, void *buf, uint16_t
proto)
*
* Return: pointer at which to write the packet's payload
*/
-static void *tap_push_ip4h(struct iphdr *ip4h, struct in_addr src,
- struct in_addr dst, size_t l4len, uint8_t proto)
+void *tap_push_ip4h(struct iphdr *ip4h, struct in_addr src,
+ struct in_addr dst, size_t l4len, uint8_t proto)
{
uint16_t l3len = l4len + sizeof(*ip4h);
@@ -229,10 +229,9 @@ void tap_icmp4_send(const struct ctx *c, struct in_addr src, struct
in_addr dst,
*
* Return: pointer at which to write the packet's payload
*/
-static void *tap_push_ip6h(struct ipv6hdr *ip6h,
- const struct in6_addr *src,
- const struct in6_addr *dst,
- size_t l4len, uint8_t proto, uint32_t flow)
+void *tap_push_ip6h(struct ipv6hdr *ip6h,
+ const struct in6_addr *src, const struct in6_addr *dst,
+ size_t l4len, uint8_t proto, uint32_t flow)
{
ip6h->payload_len = htons(l4len);
ip6h->priority = 0;
@@ -744,7 +743,7 @@ append:
for (k = 0; k < p->count; )
k += tcp_tap_handler(c, PIF_TAP, AF_INET,
&seq->saddr, &seq->daddr,
- p, k, now);
+ 0, p, k, now);
} else if (seq->protocol == IPPROTO_UDP) {
if (c->no_udp)
continue;
@@ -927,7 +926,7 @@ append:
for (k = 0; k < p->count; )
k += tcp_tap_handler(c, PIF_TAP, AF_INET6,
&seq->saddr, &seq->daddr,
- p, k, now);
+ seq->flow_lbl, p, k, now);
} else if (seq->protocol == IPPROTO_UDP) {
if (c->no_udp)
continue;
diff --git a/tap.h b/tap.h
index a476a125..390ac127 100644
--- a/tap.h
+++ b/tap.h
@@ -42,6 +42,9 @@ static inline void tap_hdr_update(struct tap_hdr *thdr, size_t l2len)
thdr->vnet_len = htonl(l2len);
}
+void *tap_push_l2h(const struct ctx *c, void *buf, uint16_t proto);
+void *tap_push_ip4h(struct iphdr *ip4h, struct in_addr src,
+ struct in_addr dst, size_t l4len, uint8_t proto);
void tap_udp4_send(const struct ctx *c, struct in_addr src, in_port_t sport,
struct in_addr dst, in_port_t dport,
const void *in, size_t dlen);
@@ -49,6 +52,9 @@ void tap_icmp4_send(const struct ctx *c, struct in_addr src, struct
in_addr dst,
const void *in, size_t l4len);
const struct in6_addr *tap_ip6_daddr(const struct ctx *c,
const struct in6_addr *src);
+void *tap_push_ip6h(struct ipv6hdr *ip6h,
+ const struct in6_addr *src, const struct in6_addr *dst,
+ size_t l4len, uint8_t proto, uint32_t flow);
void tap_udp6_send(const struct ctx *c,
const struct in6_addr *src, in_port_t sport,
const struct in6_addr *dst, in_port_t dport,
diff --git a/tcp.c b/tcp.c
index 74598039..fb04e2e7 100644
--- a/tcp.c
+++ b/tcp.c
@@ -1866,6 +1866,75 @@ static void tcp_conn_from_sock_finish(const struct ctx *c,
tcp_data_from_sock(c, conn);
}
+/**
+ * tcp_rst_no_conn() - Send RST in response to a packet with no connection
+ * @c: Execution context
+ * @af: Address family, AF_INET or AF_INET6
+ * @saddr: Source address of the packet we're responding to
+ * @daddr: Destination address of the packet we're responding to
+ * @flow_lbl: IPv6 flow label (ignored for IPv4)
+ * @th: TCP header of the packet we're responding to
+ * @l4len: Packet length, including TCP header
+ */
+static void tcp_rst_no_conn(const struct ctx *c, int af,
+ const void *saddr, const void *daddr,
+ uint32_t flow_lbl,
+ const struct tcphdr *th, size_t l4len)
+{
+ struct iov_tail payload = IOV_TAIL(NULL, 0, 0);
+ struct tcphdr *rsth;
+ char buf[USHRT_MAX];
+ uint32_t psum = 0;
+ size_t rst_l2len;
+
+ /* Don't respond to RSTs without a connection */
+ if (th->rst)
+ return;
+
+ if (af == AF_INET) {
+ struct iphdr *ip4h = tap_push_l2h(c, buf, ETH_P_IP);
+ const struct in_addr *rst_src = daddr;
+ const struct in_addr *rst_dst = saddr;
+
+ rsth = tap_push_ip4h(ip4h, *rst_src, *rst_dst,
+ sizeof(*rsth), IPPROTO_TCP);
+ psum = proto_ipv4_header_psum(sizeof(*rsth), IPPROTO_TCP,
+ *rst_src, *rst_dst);
+
+ } else {
+ struct ipv6hdr *ip6h = tap_push_l2h(c, buf, ETH_P_IPV6);
+ const struct in6_addr *rst_src = daddr;
+ const struct in6_addr *rst_dst = saddr;
+
+ rsth = tap_push_ip6h(ip6h, rst_src, rst_dst,
+ sizeof(*rsth), IPPROTO_TCP, flow_lbl);
+ psum = proto_ipv6_header_psum(sizeof(*rsth), IPPROTO_TCP,
+ rst_src, rst_dst);
+ }
+
+ memset(rsth, 0, sizeof(*rsth));
+
+ rsth->source = th->dest;
+ rsth->dest = th->source;
+ rsth->rst = 1;
+ rsth->doff = sizeof(*rsth) / 4UL;
+
+ /* Sequence matching logic from RFC 9293 section 3.10.7.1 */
+ if (th->ack) {
+ rsth->seq = th->ack_seq;
+ } else {
+ size_t dlen = l4len - th->doff * 4UL;
+ uint32_t ack = ntohl(th->seq) + dlen;
+
+ rsth->ack_seq = htonl(ack);
+ rsth->ack = 1;
+ }
+
+ tcp_update_csum(psum, rsth, &payload);
+ rst_l2len = ((char *)rsth - buf) + sizeof(*rsth);
+ tap_send_single(c, buf, rst_l2len);
+}
+
/**
* tcp_tap_handler() - Handle packets from tap and state transitions
* @c: Execution context
@@ -1873,6 +1942,7 @@ static void tcp_conn_from_sock_finish(const struct ctx *c,
* @af: Address family, AF_INET or AF_INET6
* @saddr: Source address
* @daddr: Destination address
+ * @flow_lbl: IPv6 flow label (ignored for IPv4)
* @p: Pool of TCP packets, with TCP headers
* @idx: Index of first packet in pool to process
* @now: Current timestamp
@@ -1880,7 +1950,7 @@ static void tcp_conn_from_sock_finish(const struct ctx *c,
* Return: count of consumed packets
*/
int tcp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af,
- const void *saddr, const void *daddr,
+ const void *saddr, const void *daddr, uint32_t flow_lbl,
const struct pool *p, int idx, const struct timespec *now)
{
struct tcp_tap_conn *conn;
@@ -1913,6 +1983,8 @@ int tcp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t
af,
if (opts && th->syn && !th->ack)
tcp_conn_from_tap(c, af, saddr, daddr, th,
opts, optlen, now);
+ else
+ tcp_rst_no_conn(c, af, saddr, daddr, flow_lbl, th, len);
return 1;
}
diff --git a/tcp.h b/tcp.h
index cf30744d..9142ecac 100644
--- a/tcp.h
+++ b/tcp.h
@@ -16,7 +16,7 @@ void tcp_listen_handler(const struct ctx *c, union epoll_ref ref,
void tcp_sock_handler(const struct ctx *c, union epoll_ref ref,
uint32_t events);
int tcp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af,
- const void *saddr, const void *daddr,
+ const void *saddr, const void *daddr, uint32_t flow_lbl,
const struct pool *p, int idx, const struct timespec *now);
int tcp_sock_init(const struct ctx *c, const union inany_addr *addr,
const char *ifname, in_port_t port);
--
2.48.1