From dd581730e54b934f80d5b6a820136707dc71c664 Mon Sep 17 00:00:00 2001 From: Stefano Brivio Date: Sun, 26 Sep 2021 23:38:22 +0200 Subject: tap: Completely de-serialise input message batches Until now, messages would be passed to protocol handlers in a single batch only if they happened to be dequeued in a row. Packets interleaved between different connections would result in multiple calls to the same protocol handler for a single connection. Instead, keep track of incoming packet descriptors, arrange them in sequences, and call protocol handlers only as we completely sorted input messages in batches. Signed-off-by: Stefano Brivio --- icmp.c | 20 ++- icmp.h | 2 +- passt.h | 18 +- tap.c | 618 ++++++++++++++++++++++++++++++++++++++++------------------------ tcp.c | 26 ++- udp.c | 13 +- udp.h | 4 +- 7 files changed, 435 insertions(+), 266 deletions(-) diff --git a/icmp.c b/icmp.c index 49fdf91..51848c2 100644 --- a/icmp.c +++ b/icmp.c @@ -141,23 +141,26 @@ void icmp_sock_handler(struct ctx *c, union epoll_ref ref, uint32_t events, * Return: count of consumed packets (always 1, even if malformed) */ int icmp_tap_handler(struct ctx *c, int af, void *addr, - struct tap_msg *msg, int count, struct timespec *now) + struct tap_l4_msg *msg, int count, struct timespec *now) { (void)count; if (af == AF_INET) { - struct icmphdr *ih = (struct icmphdr *)msg[0].l4h; union icmp_epoll_ref iref = { .v6 = 0 }; struct sockaddr_in sa = { .sin_family = AF_INET, .sin_addr = { .s_addr = INADDR_ANY }, - .sin_port = ih->un.echo.id, }; + struct icmphdr *ih; int id, s; + ih = (struct icmphdr *)(pkt_buf + msg[0].pkt_buf_offset); + if (msg[0].l4_len < sizeof(*ih) || ih->type != ICMP_ECHO) return 1; + sa.sin_port = ih->un.echo.id; + iref.id = id = ntohs(ih->un.echo.id); if ((s = icmp_id_map[V4][id].sock) <= 0) { @@ -171,22 +174,25 @@ int icmp_tap_handler(struct ctx *c, int af, void *addr, bitmap_set(icmp_act[V4], id); sa.sin_addr = *(struct in_addr *)addr; - sendto(s, msg[0].l4h, msg[0].l4_len, MSG_NOSIGNAL, + sendto(s, ih, msg[0].l4_len, MSG_NOSIGNAL, (struct sockaddr *)&sa, sizeof(sa)); } else if (af == AF_INET6) { - struct icmp6hdr *ih = (struct icmp6hdr *)msg[0].l4h; union icmp_epoll_ref iref = { .v6 = 1 }; struct sockaddr_in6 sa = { .sin6_family = AF_INET6, .sin6_addr = IN6ADDR_ANY_INIT, - .sin6_port = ih->icmp6_identifier, }; + struct icmp6hdr *ih; int id, s; + ih = (struct icmp6hdr *)(pkt_buf + msg[0].pkt_buf_offset); + if (msg[0].l4_len < sizeof(*ih) || (ih->icmp6_type != 128 && ih->icmp6_type != 129)) return 1; + sa.sin6_port = ih->icmp6_identifier; + iref.id = id = ntohs(ih->icmp6_identifier); if ((s = icmp_id_map[V6][id].sock) <= 0) { s = sock_l4(c, AF_INET6, IPPROTO_ICMPV6, id, 0, @@ -200,7 +206,7 @@ int icmp_tap_handler(struct ctx *c, int af, void *addr, bitmap_set(icmp_act[V6], id); sa.sin6_addr = *(struct in6_addr *)addr; - sendto(s, msg[0].l4h, msg[0].l4_len, MSG_NOSIGNAL, + sendto(s, ih, msg[0].l4_len, MSG_NOSIGNAL, (struct sockaddr *)&sa, sizeof(sa)); } diff --git a/icmp.h b/icmp.h index 27f0a5c..78f54aa 100644 --- a/icmp.h +++ b/icmp.h @@ -8,7 +8,7 @@ struct ctx; void icmp_sock_handler(struct ctx *c, union epoll_ref ref, uint32_t events, struct timespec *now); int icmp_tap_handler(struct ctx *c, int af, void *addr, - struct tap_msg *msg, int count, struct timespec *now); + struct tap_l4_msg *msg, int count, struct timespec *now); void icmp_timer(struct ctx *c, struct timespec *ts); /** diff --git a/passt.h b/passt.h index bb41e90..57b5c09 100644 --- a/passt.h +++ b/passt.h @@ -3,15 +3,21 @@ /** * struct tap_msg - Generic message descriptor for arrays of messages - * @start: Pointer to message start - * @l4_start: Pointer to L4 header - * @len: Message length, with L2 headers - * @l4_len: Message length, with L4 headers + * @pkt_buf_offset: Offset from @pkt_buf + * @len: Message length, with L2 headers */ struct tap_msg { - char *start; - char *l4h; + uint32_t pkt_buf_offset; uint16_t len; +}; + +/** + * struct tap_l4_msg - Layer-4 message descriptor for protocol handlers + * @pkt_buf_offset: Offset of message from @pkt_buf + * @l4_len: Length of Layer-4 payload, host order + */ +struct tap_l4_msg { + uint32_t pkt_buf_offset; uint16_t l4_len; }; diff --git a/tap.c b/tap.c index cac7e4f..162d02b 100644 --- a/tap.c +++ b/tap.c @@ -50,7 +50,9 @@ #include "dhcpv6.h" #include "pcap.h" -static struct tap_msg tap_msgs[TAP_MSGS]; +/* IPv4 (plus ARP) and IPv6 message batches from tap/guest to IP handlers */ +static struct tap_msg seq4[TAP_MSGS]; +static struct tap_msg seq6[TAP_MSGS]; /** * tap_send() - Send frame, with qemu socket header if needed @@ -198,257 +200,410 @@ void tap_ip_send(struct ctx *c, struct in6_addr *src, uint8_t proto, } } +/** + * struct l4_seq4_t - Message sequence for one protocol handler call, IPv4 + * @msgs: Count of messages in sequence + * @protocol: Protocol number + * @source: Source port + * @dest: Destination port + * @saddr: Source address + * @daddr: Destination address + * @msg: Array of messages that can be handled in a single call + */ +static struct tap_l4_seq4 { + uint16_t msgs; + uint8_t protocol; + + uint16_t source; + uint16_t dest; + + uint32_t saddr; + uint32_t daddr; + + struct tap_l4_msg msg[UIO_MAXIOV]; +} l4_seq4[UIO_MAXIOV /* Arbitrary: TAP_MSGS in theory, so limit in users */]; + +/** + * struct l4_seq6_t - Message sequence for one protocol handler call, IPv6 + * @msgs: Count of messages in sequence + * @protocol: Protocol number + * @source: Source port + * @dest: Destination port + * @saddr: Source address + * @daddr: Destination address + * @msg: Array of messages that can be handled in a single call + */ +static struct tap_l4_seq6 { + uint16_t msgs; + uint8_t protocol; + + uint16_t source; + uint16_t dest; + + struct in6_addr saddr; + struct in6_addr daddr; + + struct tap_l4_msg msg[UIO_MAXIOV]; +} l4_seq6[UIO_MAXIOV /* Arbitrary: TAP_MSGS in theory, so limit in users */]; + +/** + * tap_packet_debug() - Print debug message for packet(s) from guest/tap + * @iph: IPv4 header, can be NULL + * @ip6h: IPv6 header, can be NULL + * @seq4: Pointer to @struct tap_l4_seq4, can be NULL + * @proto6: IPv6 protocol, for IPv6 + * @seq6: Pointer to @struct tap_l4_seq6, can be NULL + * @count: Count of packets in this sequence + */ +static void tap_packet_debug(struct iphdr *iph, struct ipv6hdr *ip6h, + struct tap_l4_seq4 *seq4, uint8_t proto6, + struct tap_l4_seq6 *seq6, int count) +{ + char buf6s[INET6_ADDRSTRLEN], buf6d[INET6_ADDRSTRLEN]; + char buf4s[INET_ADDRSTRLEN], buf4d[INET_ADDRSTRLEN]; + uint8_t proto; + + if (iph || seq4) { + inet_ntop(AF_INET, iph ? &iph->saddr : &seq4->saddr, + buf4s, sizeof(buf4s)), + inet_ntop(AF_INET, iph ? &iph->daddr : &seq4->daddr, + buf4d, sizeof(buf4d)), + proto = iph ? iph->protocol : seq4->protocol; + } else { + inet_ntop(AF_INET6, ip6h ? &ip6h->saddr : &seq6->saddr, + buf6s, sizeof(buf6s)), + inet_ntop(AF_INET6, ip6h ? &ip6h->daddr : &seq6->daddr, + buf6d, sizeof(buf6d)), + proto = proto6; + } + + if (proto == IPPROTO_TCP || proto == IPPROTO_UDP) { + debug("protocol %i from tap: %s:%i -> %s:%i (%i packet%s)", + proto, seq4 ? buf4s : buf6s, + ntohs(seq4 ? seq4->source : seq6->source), + seq4 ? buf4d : buf6d, + ntohs(seq4 ? seq4->dest : seq6->dest), + count, count == 1 ? "" : "s"); + } else { + debug("protocol %i from tap: %s -> %s (%i packet%s)", + proto, iph ? buf4s : buf6s, iph ? buf4d : buf6d, + count, count == 1 ? "" : "s"); + } +} + /** * tap4_handler() - IPv4 and ARP packet handler for tap file descriptor * @c: Execution context - * @msg: Array of messages with the same L3 protocol - * @count: Count of messages with the same L3 protocol + * @msg: Array of messages with IPv4 or ARP protocol + * @count: Count of messages * @now: Current timestamp - * @first: First call for an IPv4 packet in this batch * * Return: count of packets consumed by handlers */ static int tap4_handler(struct ctx *c, struct tap_msg *msg, size_t count, - struct timespec *now, int first) + struct timespec *now) { - char buf_s[INET_ADDRSTRLEN] __attribute((__unused__)); - char buf_d[INET_ADDRSTRLEN] __attribute((__unused__)); - struct ethhdr *eh = (struct ethhdr *)msg[0].start; - struct iphdr *iph, *prev_iph = NULL; - struct udphdr *uh, *prev_uh = NULL; - size_t len = msg[0].len; - unsigned int i; + unsigned int i, j, seq_count; + struct tap_l4_msg *l4_msg; + struct tap_l4_seq4 *seq; + size_t len, l4_len; + struct ethhdr *eh; + struct iphdr *iph; + struct udphdr *uh; char *l4h; if (!c->v4) return count; - if (len < sizeof(*eh) + sizeof(*iph)) - return 1; + i = 0; +resume: + for (seq_count = 0, seq = NULL; i < count; i++) { + eh = (struct ethhdr *)(pkt_buf + msg[i].pkt_buf_offset); + len = msg[i].len; - if (arp(c, eh, len) || dhcp(c, eh, len)) - return 1; + if (len < sizeof(*eh)) + continue; + + if (ntohs(eh->h_proto) == ETH_P_ARP && arp(c, eh, len)) + continue; - for (i = 0; i < count; i++) { - len = msg[i].len; if (len < sizeof(*eh) + sizeof(*iph)) - return 1; + continue; - eh = (struct ethhdr *)msg[i].start; iph = (struct iphdr *)(eh + 1); - l4h = (char *)iph + iph->ihl * 4; + if ((iph->ihl * 4) + sizeof(*eh) > len) + continue; + if (iph->ihl * 4 < sizeof(*iph)) + continue; - if (first && c->addr4_seen != iph->saddr) { + if (iph->saddr && c->addr4_seen != iph->saddr) { c->addr4_seen = iph->saddr; proto_update_l2_buf(NULL, NULL, &c->addr4_seen); } - msg[i].l4h = l4h; - msg[i].l4_len = len - ((intptr_t)l4h - (intptr_t)eh); + l4h = (char *)iph + iph->ihl * 4; + l4_len = len - ((intptr_t)l4h - (intptr_t)eh); - if (iph->protocol != IPPROTO_TCP && - iph->protocol != IPPROTO_UDP) - break; + if (iph->protocol == IPPROTO_ICMP) { + struct tap_l4_msg icmp_msg = { l4h - pkt_buf, + l4_len }; - if (len < sizeof(*uh)) - break; + if (l4_len < sizeof(struct icmphdr)) + continue; + + tap_packet_debug(iph, NULL, NULL, 0, NULL, 1); + if (!c->no_icmp) { + icmp_tap_handler(c, AF_INET, &iph->daddr, + &icmp_msg, 1, now); + } + continue; + } + + if (l4_len < sizeof(*uh)) + continue; uh = (struct udphdr *)l4h; - if (!i) { - prev_iph = iph; - prev_uh = uh; + if (iph->protocol == IPPROTO_UDP && dhcp(c, eh, len)) + continue; + + if (iph->protocol != IPPROTO_TCP && + iph->protocol != IPPROTO_UDP) { + tap_packet_debug(iph, NULL, NULL, 0, NULL, 1); continue; } - if (iph->tos != prev_iph->tos || - iph->frag_off != prev_iph->frag_off || - iph->protocol != prev_iph->protocol || - iph->saddr != prev_iph->saddr || - iph->daddr != prev_iph->daddr || - uh->source != prev_uh->source || - uh->dest != prev_uh->dest) - break; +#define L4_MATCH(iph, uh, seq) \ + (seq->protocol == iph->protocol && \ + seq->source == uh->source && seq->dest == uh->dest && \ + seq->saddr == iph->saddr && seq->daddr == iph->daddr) + +#define L4_SET(iph, uh, seq) \ + do { \ + seq->protocol = iph->protocol; \ + seq->source = uh->source; \ + seq->dest = uh->dest; \ + seq->saddr = iph->saddr; \ + seq->daddr = iph->daddr; \ + } while (0) + + if (seq && L4_MATCH(iph, uh, seq) && seq->msgs < UIO_MAXIOV) + goto append; + + for (seq = l4_seq4 + seq_count - 1; seq >= l4_seq4; seq--) { + if (L4_MATCH(iph, uh, seq)) { + if (seq->msgs >= UIO_MAXIOV) + seq = l4_seq4 - 1; + break; + } + } - prev_iph = iph; - prev_uh = uh; - } + if (seq < l4_seq4) { + seq = l4_seq4 + seq_count++; + L4_SET(iph, uh, seq); + seq->msgs = 0; + } - eh = (struct ethhdr *)msg[0].start; - iph = (struct iphdr *)(eh + 1); +#undef L4_MATCH +#undef L4_SET - if (iph->protocol == IPPROTO_TCP || iph->protocol == IPPROTO_UDP || - iph->protocol == IPPROTO_SCTP) { - uh = (struct udphdr *)msg[0].l4h; +append: + l4_msg = &seq->msg[seq->msgs++]; - if (msg[0].len < sizeof(*uh)) - return 1; + l4_msg->pkt_buf_offset = l4h - pkt_buf; + l4_msg->l4_len = l4_len; - debug("%s (%i) from tap: %s:%i -> %s:%i (%i packet%s)", - IP_PROTO_STR(iph->protocol), iph->protocol, - inet_ntop(AF_INET, &iph->saddr, buf_s, sizeof(buf_s)), - ntohs(uh->source), - inet_ntop(AF_INET, &iph->daddr, buf_d, sizeof(buf_d)), - ntohs(uh->dest), - i, i > 1 ? "s" : ""); - } else if (iph->protocol == IPPROTO_ICMP) { - debug("icmp from tap: %s -> %s", - inet_ntop(AF_INET, &iph->saddr, buf_s, sizeof(buf_s)), - inet_ntop(AF_INET, &iph->daddr, buf_d, sizeof(buf_d))); + if (seq_count == UIO_MAXIOV) + break; /* Resume after flushing if i < count */ } - if (iph->protocol == IPPROTO_TCP) { - if (c->no_tcp) - return i; - return tcp_tap_handler(c, AF_INET, &iph->daddr, msg, i, now); - } + for (j = 0, seq = l4_seq4; j < seq_count; j++, seq++) { + int n = seq->msgs; - if (iph->protocol == IPPROTO_UDP) { - if (c->no_udp) - return i; - return udp_tap_handler(c, AF_INET, &iph->daddr, msg, i, now); - } + l4_msg = seq->msg; - if (iph->protocol == IPPROTO_ICMP) { - if (c->no_icmp) - return 1; - icmp_tap_handler(c, AF_INET, &iph->daddr, msg, 1, now); + tap_packet_debug(NULL, NULL, seq, 0, NULL, n); + + if (seq->protocol == IPPROTO_TCP) { + if (c->no_tcp) + continue; + while ((n -= tcp_tap_handler(c, AF_INET, &seq->daddr, + l4_msg, n, now))); + } else if (seq->protocol == IPPROTO_UDP) { + if (c->no_udp) + continue; + while ((n -= udp_tap_handler(c, AF_INET, &seq->daddr, + l4_msg, n, now))); + } } - return 1; + if (i < count) + goto resume; + + return count; } /** * tap6_handler() - IPv6 packet handler for tap file descriptor * @c: Execution context - * @msg: Array of messages with the same L3 protocol - * @count: Count of messages with the same L3 protocol + * @msg: Array of messages with IPv6 protocol + * @count: Count of messages * @now: Current timestamp - * @first: First call for an IPv6 packet in this batch * * Return: count of packets consumed by handlers */ static int tap6_handler(struct ctx *c, struct tap_msg *msg, size_t count, - struct timespec *now, int first) + struct timespec *now) { - char buf_s[INET6_ADDRSTRLEN], buf_d[INET6_ADDRSTRLEN]; - struct ethhdr *eh = (struct ethhdr *)msg[0].start; - struct udphdr *uh, *prev_uh = NULL; - uint8_t proto = 0, prev_proto = 0; - size_t len = msg[0].len; + unsigned int i, j, seq_count = 0; + struct tap_l4_msg *l4_msg; + struct tap_l4_seq6 *seq; struct ipv6hdr *ip6h; - unsigned int i; + size_t len, l4_len; + struct ethhdr *eh; + struct udphdr *uh; + uint8_t proto; char *l4h; if (!c->v6) return count; - if (len < sizeof(*eh) + sizeof(*ip6h)) - return 1; - - if (ndp(c, eh, len) || dhcpv6(c, eh, len)) - return 1; + i = 0; +resume: + for (seq_count = 0, seq = NULL; i < count; i++) { + eh = (struct ethhdr *)(pkt_buf + msg[i].pkt_buf_offset); + len = msg[i].len; - for (i = 0; i < count; i++) { - struct ipv6hdr *p_ip6h; + if (len < sizeof(*eh)) + continue; - len = msg[i].len; if (len < sizeof(*eh) + sizeof(*ip6h)) return 1; - eh = (struct ethhdr *)msg[i].start; ip6h = (struct ipv6hdr *)(eh + 1); - l4h = ipv6_l4hdr(ip6h, &proto); - - msg[i].l4h = l4h; - msg[i].l4_len = len - ((intptr_t)l4h - (intptr_t)eh); - if (first) { - if (IN6_IS_ADDR_LINKLOCAL(&ip6h->saddr)) { - c->addr6_ll_seen = ip6h->saddr; + if (IN6_IS_ADDR_LINKLOCAL(&ip6h->saddr)) { + c->addr6_ll_seen = ip6h->saddr; - if (IN6_IS_ADDR_UNSPECIFIED(&c->addr6_seen)) { - c->addr6_seen = ip6h->saddr; - } - } else { + if (IN6_IS_ADDR_UNSPECIFIED(&c->addr6_seen)) { c->addr6_seen = ip6h->saddr; } + } else { + c->addr6_seen = ip6h->saddr; } - ip6h->saddr = c->addr6; + if (ntohs(ip6h->payload_len) > + len - sizeof(*eh) - sizeof(*ip6h)) + continue; - if (proto != IPPROTO_TCP && proto != IPPROTO_UDP) - break; + if (!(l4h = ipv6_l4hdr(ip6h, &proto))) + continue; - if (len < sizeof(*uh)) - break; + l4_len = len - ((intptr_t)l4h - (intptr_t)eh); - uh = (struct udphdr *)l4h; + if (proto == IPPROTO_ICMPV6) { + struct tap_l4_msg icmpv6_msg = { l4h - pkt_buf, + l4_len }; - if (!i) { - p_ip6h = ip6h; - prev_proto = proto; - prev_uh = uh; + if (l4_len < sizeof(struct icmp6hdr)) + continue; + + if (ndp(c, eh, len)) + continue; + + tap_packet_debug(NULL, ip6h, NULL, proto, NULL, 1); + if (!c->no_icmp) { + icmp_tap_handler(c, AF_INET6, &ip6h->daddr, + &icmpv6_msg, 1, now); + } continue; } - if (proto != prev_proto || - memcmp(&ip6h->saddr, &p_ip6h->saddr, sizeof(ip6h->saddr)) || - memcmp(&ip6h->daddr, &p_ip6h->daddr, sizeof(ip6h->daddr)) || - uh->source != prev_uh->source || - uh->dest != prev_uh->dest) - break; + if (l4_len < sizeof(*uh)) + continue; - p_ip6h = ip6h; - prev_proto = proto; - prev_uh = uh; - } + uh = (struct udphdr *)l4h; - if (prev_proto) - proto = prev_proto; + if (proto == IPPROTO_UDP && dhcpv6(c, eh, len)) + continue; - eh = (struct ethhdr *)msg[0].start; - ip6h = (struct ipv6hdr *)(eh + 1); + ip6h->saddr = c->addr6; - if (proto == IPPROTO_ICMPV6) { - debug("icmpv6 from tap: %s ->\n\t%s", - inet_ntop(AF_INET6, &ip6h->saddr, buf_s, sizeof(buf_s)), - inet_ntop(AF_INET6, &ip6h->daddr, buf_d, sizeof(buf_d))); - } else if (proto == IPPROTO_TCP || proto == IPPROTO_UDP || - proto == IPPROTO_SCTP) { - uh = (struct udphdr *)msg[0].l4h; + if (proto != IPPROTO_TCP && proto != IPPROTO_UDP) { + tap_packet_debug(NULL, ip6h, NULL, proto, NULL, 1); + continue; + } - if (msg[0].len < sizeof(*uh)) - return 1; +#define L4_MATCH(ip6h, proto, uh, seq) \ + (seq->protocol == proto && \ + seq->source == uh->source && seq->dest == uh->dest && \ + !memcmp(&seq->saddr, &ip6h->saddr, sizeof(seq->saddr)) && \ + !memcmp(&seq->daddr, &ip6h->daddr, sizeof(seq->daddr))) + +#define L4_SET(ip6h, proto, uh, seq) \ + do { \ + seq->protocol = proto; \ + seq->source = uh->source; \ + seq->dest = uh->dest; \ + seq->saddr = ip6h->saddr; \ + seq->daddr = ip6h->daddr; \ + } while (0) + + if (seq && L4_MATCH(ip6h, proto, uh, seq) && + seq->msgs < UIO_MAXIOV) + goto append; + + for (seq = l4_seq6 + seq_count - 1; seq >= l4_seq6; seq--) { + if (L4_MATCH(ip6h, proto, uh, seq)) { + if (seq->msgs >= UIO_MAXIOV) + seq = l4_seq6 - 1; + break; + } + } - debug("%s (%i) from tap: [%s]:%i\n\t-> [%s]:%i (%i packet%s)", - IP_PROTO_STR(proto), proto, - inet_ntop(AF_INET6, &ip6h->saddr, buf_s, sizeof(buf_s)), - ntohs(uh->source), - inet_ntop(AF_INET6, &ip6h->daddr, buf_d, sizeof(buf_d)), - ntohs(uh->dest), - i, i > 1 ? "s" : ""); - } + if (seq < l4_seq6) { + seq = l4_seq6 + seq_count++; + L4_SET(ip6h, proto, uh, seq); + seq->msgs = 0; + } - if (proto == IPPROTO_TCP) { - if (c->no_tcp) - return i; - return tcp_tap_handler(c, AF_INET6, &ip6h->daddr, msg, i, now); - } +#undef L4_MATCH +#undef L4_SET - if (proto == IPPROTO_UDP) { - if (c->no_udp) - return i; - return udp_tap_handler(c, AF_INET6, &ip6h->daddr, msg, i, now); +append: + l4_msg = &seq->msg[seq->msgs++]; + + l4_msg->pkt_buf_offset = l4h - pkt_buf; + l4_msg->l4_len = l4_len; + + if (seq_count == UIO_MAXIOV) + break; /* Resume after flushing if i < count */ } - if (proto == IPPROTO_ICMPV6) { - if (c->no_icmp) - return 1; - icmp_tap_handler(c, AF_INET6, &ip6h->daddr, msg, 1, now); + for (j = 0, seq = l4_seq6; j < seq_count; j++, seq++) { + int n = seq->msgs; + + l4_msg = seq->msg; + + tap_packet_debug(NULL, NULL, NULL, seq->protocol, seq, n); + + if (seq->protocol == IPPROTO_TCP) { + if (c->no_tcp) + continue; + while ((n -= tcp_tap_handler(c, AF_INET6, &seq->daddr, + l4_msg, n, now))); + } else if (seq->protocol == IPPROTO_UDP) { + if (c->no_udp) + continue; + while ((n -= udp_tap_handler(c, AF_INET6, &seq->daddr, + l4_msg, n, now))); + } } - return 1; + if (i < count) + goto resume; + + return count; } /** @@ -460,10 +615,14 @@ static int tap6_handler(struct ctx *c, struct tap_msg *msg, size_t count, */ static int tap_handler_passt(struct ctx *c, struct timespec *now) { - int msg_count = 0, same, i = 0, first_v4 = 1, first_v6 = 1; + int seq4_i, seq6_i; struct ethhdr *eh; - char *p = pkt_buf; ssize_t n, rem; + char *p; + +redo: + p = pkt_buf; + seq4_i = seq6_i = rem = 0; n = recv(c->fd_tap, p, TAP_BUF_FILL, MSG_DONTWAIT); if (n < 0) { @@ -479,30 +638,27 @@ static int tap_handler_passt(struct ctx *c, struct timespec *now) while (n > (ssize_t)sizeof(uint32_t)) { ssize_t len = ntohl(*(uint32_t *)p); - if (len < (ssize_t)sizeof(*eh) || len > ETH_MAX_MTU) - return 0; - p += sizeof(uint32_t); n -= sizeof(uint32_t); - /* At most one packet might not fit in a single read */ + /* At most one packet might not fit in a single read, and this + * needs to be blocking. + */ if (len > n) { - rem = recv(c->fd_tap, p + n, len - n, MSG_DONTWAIT); + rem = recv(c->fd_tap, p + n, len - n, 0); if ((n += rem) != len) return 0; } - pcap(p, len); - - tap_msgs[msg_count].start = p; - tap_msgs[msg_count++].len = len; + /* Complete the partial read above before discarding a malformed + * frame, otherwise the stream will be inconsistent. + */ + if (len < (ssize_t)sizeof(*eh) || len > ETH_MAX_MTU) + goto next; - n -= len; - p += len; - } + pcap(p, len); - while (i < msg_count) { - eh = (struct ethhdr *)tap_msgs[i].start; + eh = (struct ethhdr *)p; if (memcmp(c->mac_guest, eh->h_source, ETH_ALEN)) { memcpy(c->mac_guest, eh->h_source, ETH_ALEN); @@ -511,52 +667,33 @@ static int tap_handler_passt(struct ctx *c, struct timespec *now) switch (ntohs(eh->h_proto)) { case ETH_P_ARP: - if (c->v4) - tap4_handler(c, tap_msgs + i, 1, now, 1); - i++; - break; case ETH_P_IP: - for (same = 1; i + same < msg_count && - same < UIO_MAXIOV; same++) { - struct tap_msg *next = &tap_msgs[i + same]; - - eh = (struct ethhdr *)next->start; - if (ntohs(eh->h_proto) != ETH_P_IP) - break; - } - - if (!c->v4) { - i += same; - break; - } - - i += tap4_handler(c, tap_msgs + i, same, now, first_v4); - first_v4 = 0; + seq4[seq4_i].pkt_buf_offset = p - pkt_buf; + seq4[seq4_i++].len = len; break; case ETH_P_IPV6: - for (same = 1; i + same < msg_count && - same < UIO_MAXIOV; same++) { - struct tap_msg *next = &tap_msgs[i + same]; - - eh = (struct ethhdr *)next->start; - if (ntohs(eh->h_proto) != ETH_P_IPV6) - break; - } - - if (!c->v6) { - i += same; - break; - } - - i += tap6_handler(c, tap_msgs + i, same, now, first_v6); - first_v6 = 0; + seq6[seq6_i].pkt_buf_offset = p - pkt_buf; + seq6[seq6_i++].len = len; break; default: - i++; break; } + +next: + p += len; + n -= len; } + if (seq4_i) + tap4_handler(c, seq4, seq4_i, now); + + if (seq6_i) + tap6_handler(c, seq6, seq6_i, now); + + /* We can't use EPOLLET otherwise. */ + if (rem) + goto redo; + return 0; } @@ -569,14 +706,19 @@ static int tap_handler_passt(struct ctx *c, struct timespec *now) */ static int tap_handler_pasta(struct ctx *c, struct timespec *now) { - struct tap_msg msg = { .start = pkt_buf }; - ssize_t n; + ssize_t n = 0, len; + int err, seq4_i = 0, seq6_i = 0; + +restart: + while ((len = read(c->fd_tap, pkt_buf + n, TAP_BUF_BYTES - n)) > 0) { + struct ethhdr *eh = (struct ethhdr *)(pkt_buf + n); - while ((n = read(c->fd_tap, pkt_buf, TAP_BUF_BYTES)) > 0) { - struct ethhdr *eh = (struct ethhdr *)pkt_buf; - msg.len = n; + if (len < (ssize_t)sizeof(*eh) || len > ETH_MAX_MTU) { + n += len; + continue; + } - pcap(msg.start, msg.len); + pcap(pkt_buf + n, len); if (memcmp(c->mac_guest, eh->h_source, ETH_ALEN)) { memcpy(c->mac_guest, eh->h_source, ETH_ALEN); @@ -585,21 +727,33 @@ static int tap_handler_pasta(struct ctx *c, struct timespec *now) switch (ntohs(eh->h_proto)) { case ETH_P_ARP: - if (c->v4) - tap4_handler(c, &msg, 1, now, 1); - break; case ETH_P_IP: - if (c->v4) - tap4_handler(c, &msg, 1, now, 1); + seq4[seq4_i].pkt_buf_offset = n; + seq4[seq4_i++].len = len; break; case ETH_P_IPV6: - if (c->v6) - tap6_handler(c, &msg, 1, now, 1); + seq6[seq6_i].pkt_buf_offset = n; + seq6[seq6_i++].len = len; + break; + default: break; } + + n += len; } - if (!n || errno == EINTR || errno == EAGAIN || errno == EWOULDBLOCK) + if (len < 0 && errno == EINTR) + goto restart; + + err = errno; + + if (seq4_i) + tap4_handler(c, seq4, seq4_i, now); + + if (seq6_i) + tap6_handler(c, seq6, seq6_i, now); + + if (len > 0 || err == EAGAIN) return 0; epoll_ctl(c->epollfd, EPOLL_CTL_DEL, c->fd_tap, NULL); @@ -753,12 +907,14 @@ void tap_sock_init(struct ctx *c) close(c->fd_tap); } - if (c->mode == MODE_PASST) + if (c->mode == MODE_PASST) { tap_sock_init_unix(c); - else + ev.events = EPOLLIN | EPOLLET | EPOLLRDHUP; + } else { tap_sock_init_tun(c); + ev.events = EPOLLIN | EPOLLRDHUP; + } - ev.events = EPOLLIN | EPOLLRDHUP; ev.data.fd = c->fd_tap; epoll_ctl(c->epollfd, EPOLL_CTL_ADD, c->fd_tap, &ev); } diff --git a/tcp.c b/tcp.c index 645d525..b269e0e 100644 --- a/tcp.c +++ b/tcp.c @@ -333,6 +333,7 @@ #include #include #include +#include #include #include #include @@ -645,7 +646,7 @@ static struct mmsghdr tcp_l2_mh_tap [TCP_TAP_FRAMES] = { }; /* sendmsg() to socket */ -static struct iovec tcp_tap_iov [TAP_MSGS]; +static struct iovec tcp_tap_iov [UIO_MAXIOV]; /* Bitmap, activity monitoring needed for connection via tap */ static uint8_t tcp_act[MAX_TAP_CONNS / 8] = { 0 }; @@ -1968,7 +1969,7 @@ out_restore_iov: * @now: Current timestamp */ static void tcp_data_from_tap(struct ctx *c, struct tcp_tap_conn *conn, - struct tap_msg *msg, int count, + struct tap_l4_msg *msg, int count, struct timespec *now) { int i, iov_i, ack = 0, fin = 0, retr = 0, keep = -1; @@ -1979,10 +1980,13 @@ static void tcp_data_from_tap(struct ctx *c, struct tcp_tap_conn *conn, ssize_t len; for (i = 0, iov_i = 0; i < count; i++) { - struct tcphdr *th = (struct tcphdr *)msg[i].l4h; uint32_t seq, seq_offset, ack_seq; - size_t len = msg[i].l4_len, off; + struct tcphdr *th; char *data; + size_t off; + + th = (struct tcphdr *)(pkt_buf + msg[i].pkt_buf_offset); + len = msg[i].l4_len; if (len < sizeof(*th)) { tcp_rst(c, conn); @@ -2152,19 +2156,11 @@ out: * Return: count of consumed packets */ int tcp_tap_handler(struct ctx *c, int af, void *addr, - struct tap_msg *msg, int count, struct timespec *now) + struct tap_l4_msg *msg, int count, struct timespec *now) { - struct tcphdr *th = (struct tcphdr *)msg[0].l4h; - size_t len = msg[0].l4_len, off; + struct tcphdr *th = (struct tcphdr *)(pkt_buf + msg[0].pkt_buf_offset); + uint16_t len = msg[0].l4_len; struct tcp_tap_conn *conn; - int ws; - - if (len < sizeof(*th)) - return 1; - - off = th->doff * 4; - if (off < sizeof(*th) || off > len) - return 1; conn = tcp_hash_lookup(c, af, addr, htons(th->source), htons(th->dest)); if (!conn) { diff --git a/udp.c b/udp.c index e640f16..6b2ee36 100644 --- a/udp.c +++ b/udp.c @@ -879,12 +879,12 @@ void udp_sock_handler(struct ctx *c, union epoll_ref ref, uint32_t events, * Return: count of consumed packets */ int udp_tap_handler(struct ctx *c, int af, void *addr, - struct tap_msg *msg, int count, struct timespec *now) + struct tap_l4_msg *msg, int count, struct timespec *now) { /* The caller already checks that all the messages have the same source * and destination, so we can just take those from the first message. */ - struct udphdr *uh = (struct udphdr *)msg[0].l4h; + struct udphdr *uh = (struct udphdr *)(pkt_buf + msg[0].pkt_buf_offset); struct mmsghdr mm[UIO_MAXIOV] = { 0 }; struct iovec m[UIO_MAXIOV]; struct sockaddr_in6 s_in6; @@ -972,7 +972,10 @@ int udp_tap_handler(struct ctx *c, int af, void *addr, } for (i = 0; i < count; i++) { - m[i].iov_base = (char *)((struct udphdr *)msg[i].l4h + 1); + struct udphdr *uh; + + uh = (struct udphdr *)(msg[i].pkt_buf_offset + pkt_buf); + m[i].iov_base = (char *)(uh + 1); m[i].iov_len = msg[i].l4_len - sizeof(*uh); mm[i].msg_hdr.msg_name = sa; @@ -1084,12 +1087,14 @@ static void udp_splice_iov_init(void) * * Return: 0 on success, -1 on failure */ -int udp_sock_init(struct ctx *c) +int udp_sock_init(struct ctx *c, struct timespec *now) { union udp_epoll_ref uref = { .bound = 1 }; in_port_t dst; int s; + (void)now; + for (dst = 0; dst < USHRT_MAX; dst++) { if (!bitmap_isset(c->udp.port_to_tap, dst)) continue; diff --git a/udp.h b/udp.h index bdafcaf..c20f936 100644 --- a/udp.h +++ b/udp.h @@ -6,8 +6,8 @@ void udp_sock_handler(struct ctx *c, union epoll_ref ref, uint32_t events, struct timespec *now); int udp_tap_handler(struct ctx *c, int af, void *addr, - struct tap_msg *msg, int count, struct timespec *now); -int udp_sock_init(struct ctx *c); + struct tap_l4_msg *msg, int count, struct timespec *now); +int udp_sock_init(struct ctx *c, struct timespec *now); void udp_timer(struct ctx *c, struct timespec *ts); void udp_update_l2_buf(unsigned char *eth_d, unsigned char *eth_s, uint32_t *ip_da); -- cgit v1.2.3