- dev - passt

passt: new version 2025_04_15.2340bbf available
by Stefano Brivio 15 Apr '25

15 Apr '25

The new version with tag 2025_04_15.2340bbf includes the following changes: 2340bbf udp: Propagate errors on listening and brand new sockets cfc0ee1 udp: Minor re-organisation of udp_sock_recverr() f107a86 udp: Add udp_pktinfo() helper 0498457 udp: Deal with errors as we go in udp_sock_fwd() 3f99558 udp: Pass socket & flow information direction to error handling functions 1bb8145 udp: Be quieter about errors on UDP receive baf049f udp: Fix breakage of UDP error handling by PKTINFO support 5024908 conf: Honour --dns-forward for local resolver even with --no-map-gw bbff365 conf: Split add_dns_resolv() into separate IPv4 and IPv6 versions 59cc89f udp, udp_flow: Track our specific address on socket interfaces 695c623 inany: Improve ASSERT message for bad socket family f4b0dd8 udp: Use PKTINFO cmsgs to get destination address for received datagrams 6693fa1 tcp_splice: Don't clobber errno before checking for EAGAIN d3f33f3 tcp_splice: Don't double count bytes read on EINTR ffbef85 conf: Add missing return in conf_nat(), fix --map-guest-addr none 06ef64c udp_flow: Save 8 bytes in struct udp_flow on 64-bit architectures 9725e79 udp_flow: Don't discard packets that arrive between bind() and connect() 9eb5406 udp: Fold udp_splice_prepare and udp_splice_send into udp_sock_to_sock bd6a41e udp: Rework udp_listen_sock_data() into udp_sock_fwd() 159beef udp_flow: Take pif and port as explicit parameters to udp_flow_from_sock() fd844a9 udp: Move UDP_MAX_FRAMES to udp.c fc6ee68 udp: Merge vhost-user and "buf" listening socket paths 0304dd9 udp: Split spliced forwarding path from udp_buf_reply_sock_data() 5221e17 udp: Parameterize number of datagrams handled by udp_*_reply_sock_data() 3a0881d udp: Don't bother to batch datagrams from "listening" socket 84ab130 udp: Polish udp_vu_sock_info() and remove from vu specific code 1d7bbb1 udp: Make udp_sock_recv() take max number of frames as a parameter d74b5a7 udp: Use connect()ed sockets for initiating side a7775e9 udp: support traceroute in direction tap-socket 06784d7 passt-repair: Ensure that read buffer is NULL-terminated 684870a udp: Correct some seccomp filter annotations 76e554d udp: Simplify updates to UDP flow timestamp 8aa2d90 udp: Remove redundant udp_at_sidx() call in udp_tap_handler() 3d41e4d passt-repair: Correct off-by-one error verifying name dec3d73 migrate, tcp: bind() migrated sockets in repair mode 6bfc60b platform requirements: Add test for address conflicts with TCP_REPAIR 8e32881 platform requirements: Add attributes to die() function 2ed2d59 platform requirements: Fix clang-tidy warning 3de5af6 udp: Improve name of UDP related ICMP sending functions 025a3c2 udp: Don't attempt to forward ICMP socket errors to other sockets 42a854a pasta, passt-repair: Support multiple events per read() in inotify handlers 65cca54 udp: correct source address for ICMP messages 664c588 build: normalize arm targets 77883fb udp: Add helper function for creating connected UDP socket 37d78c9 udp: Always hash socket facing flowsides f67c488 udp: Better handling of failure to forward from reply socket 269cf6a udp: Share more logic between vu and non-vu reply socket paths d924b7d udp_vu: Factor things out of udp_vu_reply_sock_data() loop 5a977c2 udp: Simplify checking of epoll event bits 89b203b udp: Common invocation of udp_sock_errs() for vhost-user and "buf" paths cf4d3f0 packet: Upgrade severity of most packet errors 0857515 packet: ASSERT on signs of pool corruption 9153aca util: Add abort_with_msg() and ASSERT_WITH_MSG() helpers 38bcce9 packet: Rework packet_get() versus packet_get_try() 961aa6a packet: Move checks against PACKET_MAX_LEN to packet_check_range() 37d9f37 packet: Avoid integer overflows in packet_get_do() c48331c packet: Correct type of PACKET_MAX_LEN 9866d14 tap: Clarify calculation of TAP_MSGS a41d6d1 tap: Make size of pool_tap[46] purely a tuning parameter e43e007 packet: More cautious checks to avoid pointer arithmetic UB 4592719 vu_common: Tighten vu_packet_check_range() https://passt.top/passt/log/?qt=range&q=2025_03_20.32f6212..2025_04_15.2340… Packages: - Alpine Linux: https://pkgs.alpinelinux.org/packages?name=passt - Arch Linux: https://www.archlinux.org/packages/extra/x86_64/passt/ https://archlinuxarm.org/packages/aarch64/passt https://archlinuxarm.org/packages/armv7h/passt - Chimera: https://pkgs.chimera-linux.org/packages?name=passt - Clear Linux: https://github.com/clearlinux-pkgs/passt/ - Copr (CentOS Stream, EPEL, Fedora, Mageia): https://copr.fedorainfracloud.org/coprs/sbrivio/passt/build/8907839/ permanent mirror: https://passt.top/builds/copr/0^20250415.g2340bbf/ - Debian tracker: https://tracker.debian.org/pkg/passt - Fedora updates: https://bodhi.fedoraproject.org/updates/?packages=passt - Gentoo versions: https://packages.gentoo.org/packages/net-misc/passt - GNU Guix: https://packages.guix.gnu.org/packages/passt/ - Homebrew: https://formulae.brew.sh/formula/passt - NixOS: https://github.com/NixOS/nixpkgs/tree/nixos-unstable/pkgs/by-name/pa/passt - openSUSE: https://software.opensuse.org/package/passt - OpenMandriva: https://github.com/OpenMandrivaAssociation/passt/tree/master - PLD Linux: https://git.pld-linux.org/cgi-bin/gitweb.cgi?p=packages/passt.git - Solus: https://github.com/getsolus/packages/tree/main/packages/p/passt - Ubuntu tracker: https://launchpad.net/ubuntu/+source/passt - Void Linux: https://voidlinux.org/packages/?q=passt - Static builds: - Package for other RPM-based distributions, x86_64 only: https://passt.top/builds/latest/x86_64/passt-g2340bbf-1.x86_64.rpm - x86_64 static binaries: https://passt.top/builds/latest/x86_64/ - Debian package, from x86_64 static build: https://passt.top/builds/latest/x86_64/passt_2340bbf-1_all.deb -- Stefano

1 0

[PATCH] udp: Fix breakage of UDP error handling by PKTINFO support
by David Gibson 15 Apr '25

15 Apr '25

We recently enabled the IP_PKTINFO / IPV6_RECVPKTINFO socket options on our UDP sockets. This lets us obtain and properly handle the specific local address used when we're "listening" with a socket on 0.0.0.0 or ::. However, the PKTINFO cmsgs this option generates appear on error queue messages as well as regular datagrams. udp_sock_recverr() doesn't expect this and so flags an unrecoverable error when it can't parse the control message. Correct this by adding space in udp_sock_recverr()s control buffer for the additional PKTINFO data, and scan through all cmsgs for the RECVERR, rather than only looking at the first one. Link: https://bugs.passt.top/show_bug.cgi?id=99 Fixes: f4b0dd8b06 ("udp: Use PKTINFO cmsgs to get destination address..") Reported-by: Stefano Brivio <sbrivio(a)redhat.com> Signed-off-by: David Gibson <david(a)gibson.dropbear.id.au> --- udp.c | 30 +++++++++++++++++------------- 1 file changed, 17 insertions(+), 13 deletions(-) diff --git a/udp.c b/udp.c index 40af7dfc..f5fb98c2 100644 --- a/udp.c +++ b/udp.c @@ -155,6 +155,10 @@ __attribute__ ((aligned(32))) #endif udp_meta[UDP_MAX_FRAMES]; +#define PKTINFO_SPACE \ + MAX(CMSG_SPACE(sizeof(struct in_pktinfo)), \ + CMSG_SPACE(sizeof(struct in6_pktinfo))) + /** * enum udp_iov_idx - Indices for the buffers making up a single UDP frame * @UDP_IOV_TAP tap specific header @@ -476,10 +480,10 @@ static int udp_sock_recverr(const struct ctx *c, union epoll_ref ref) struct sock_extended_err ee; union sockaddr_inany saddr; }; - const struct errhdr *eh; - const struct cmsghdr *hdr; - char buf[CMSG_SPACE(sizeof(struct errhdr))]; + char buf[PKTINFO_SPACE + CMSG_SPACE(sizeof(struct errhdr))]; char data[ICMP6_MAX_DLEN]; + const struct errhdr *eh; + struct cmsghdr *hdr; int s = ref.fd; struct iovec iov = { .iov_base = data, @@ -507,12 +511,16 @@ static int udp_sock_recverr(const struct ctx *c, union epoll_ref ref) return -1; } - hdr = CMSG_FIRSTHDR(&mh); - if (!((hdr->cmsg_level == IPPROTO_IP && - hdr->cmsg_type == IP_RECVERR) || - (hdr->cmsg_level == IPPROTO_IPV6 && - hdr->cmsg_type == IPV6_RECVERR))) { - err("Unexpected cmsg reading error queue"); + for (hdr = CMSG_FIRSTHDR(&mh); hdr; hdr = CMSG_NXTHDR(&mh, hdr)) { + if ((hdr->cmsg_level == IPPROTO_IP && + hdr->cmsg_type == IP_RECVERR) || + (hdr->cmsg_level == IPPROTO_IPV6 && + hdr->cmsg_type == IPV6_RECVERR)) + break; + } + + if (!hdr) { + err("Missing RECVERR cmsg in error queue"); return -1; } @@ -587,10 +595,6 @@ static int udp_sock_errs(const struct ctx *c, union epoll_ref ref) return n_err; } -#define PKTINFO_SPACE \ - MAX(CMSG_SPACE(sizeof(struct in_pktinfo)), \ - CMSG_SPACE(sizeof(struct in6_pktinfo))) - /** * udp_peek_addr() - Get source address for next packet * @s: Socket to get information from -- 2.49.0

2 3

[PATCH 0/2] Fix --dns-forward with --no-map-gw
by Stefano Brivio 14 Apr '25

14 Apr '25

Patch 2/2 is the actual fix, patch 1/2 makes it less awkward. Stefano Brivio (2): conf: Split add_dns_resolv() into separate IPv4 and IPv6 versions conf: Honour --dns-forward for local resolver even with --no-map-gw conf.c | 115 +++++++++++++++++++++++++++++++++++++-------------------- 1 file changed, 74 insertions(+), 41 deletions(-) -- 2.43.0

3 6

[PATCH v2 00/20] Introduce discontiguous frames management
by Laurent Vivier 14 Apr '25

14 Apr '25

This series introduces iov_tail to convey frame information between functions. This is only an API change, for the moment the memory pool is only able to store contiguous buffer, so, except for vhost-user in a special case, we only play with iovec array with only one entry. Laurent Vivier (20): arp: Don't mix incoming and outgoing buffers iov: Introduce iov_tail_drop() and iov_slice() iov: Update IOV_REMOVE_HEADER() and IOV_PEEK_HEADER() tap: Use iov_tail with tap_add_packet() packet: Use iov_tail with packet_add() packet: Add packet_base() arp: Convert to iov_tail ndp: Convert to iov_tail icmp: Convert to iov_tail udp: Convert to iov_tail tcp: Convert tcp_tap_handler() to use iov_tail tcp: Convert tcp_data_from_tap() to use iov_tail dhcpv6: move offset initialization out of dhcpv6_opt() dhcpv6: Extract sending of NotOnLink status dhcpv6: Convert to iov_tail dhcpv6: Use iov_tail in dhcpv6_opt() dhcp: Convert to iov_tail ip: Use iov_tail in ipv6_l4hdr() tap: Convert tap4_handler() to iov_tail tap: Convert tap6_handler() to iov_tail arp.c | 90 ++++++++++++++------- dhcp.c | 45 ++++++----- dhcpv6.c | 222 ++++++++++++++++++++++++++++++++-------------------- icmp.c | 32 ++++---- iov.c | 139 +++++++++++++++++++++++++++++--- iov.h | 59 ++++++++++---- ip.c | 27 +++---- ip.h | 3 +- ndp.c | 7 +- packet.c | 84 ++++++++------------ packet.h | 21 ++--- pcap.c | 1 + tap.c | 88 ++++++++++++++------- tap.h | 3 +- tcp.c | 61 ++++++++++----- tcp_buf.c | 2 +- udp.c | 34 +++++--- vu_common.c | 26 ++---- 18 files changed, 609 insertions(+), 335 deletions(-) -- 2.49.0

2 31

tcp_splice SO_RCVLOWAT code; never invoked?
by David Gibson 11 Apr '25

11 Apr '25

Hi Stefano, When debugging the splice EINTR bug I fixed the other day, I found the whole tcp_splice_sock_handler() pretty confusing to follow. So, I was working on some cleanups. But then I noticed something more specifically odd here. We've discussed the use of SO_RCVLOWAT previously. AIUI, you found it essential to achieve reasonable throughput and load for spliced connections. I think we've agreed before that it's not entirely the right tool for the job; just the only one available. Except... as far as I can tell, it's never invoked. AFAICT the only place we enable the RCVLOWAT stuff is in a block under this if: if (!(conn->flags & lowat_set_flag) && readlen > (long)c->tcp.pipe_size / 10) { But... this occurs immediately after: if (readlen >= (long)c->tcp.pipe_size * 10 / 100) continue; .. which is a strictly more inclusive condition, so we'll never reach the RCVLOWAT block. To confirm, I tried putting an ASSERT(0) in that block, and didn't hit it with spliced iperf3 runs. Am I missing something? -- David Gibson (he or they) | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you, not the other way | around. http://www.ozlabs.org/~dgibson

2 2

[PATCH 0/3] Properly preseve local addresses for UDP flows
by David Gibson 11 Apr '25

11 Apr '25

Add the pieces we need so that we'll reliably send replies from the same local address that the originating datagram was delivered to. Fixes bug 99, and will (I think) be necessary for a bunch more of the flexible forwarding stuff we have in mind. David Gibson (3): udp: Use PKTINFO cmsgs to get destination address for received datagrams inany: Improve ASSERT message for bad socket family udp, udp_flow: Track our specific address on socket interfaces flow.c | 14 +++++++++++--- flow.h | 3 ++- flow_table.h | 1 + inany.h | 3 ++- tcp.c | 2 +- udp.c | 41 +++++++++++++++++++++++++++++++++++++---- udp_flow.c | 36 ++++++++++++++++++++++++++++++++---- udp_flow.h | 3 ++- util.c | 8 ++++++-- util.h | 10 ++++++++++ 10 files changed, 104 insertions(+), 17 deletions(-) -- 2.49.0

2 4

[PATCH 0/2] Fix bugs in tcp_splice_sock_handler()
by David Gibson 10 Apr '25

10 Apr '25

Patch 1/2 here fixes what I think is the bug causing the latest issue in: https://github.com/containers/podman/issues/23686 While fixing that I spotted the bug fixed in 2/2. Definitely a bug, although I suspect we'll mostly get away with it if the --trace option is not enabled. David Gibson (2): tcp_splice: Don't double could bytes read on EINTR tcp_splice: Don't clobber errno before checking for EAGAIN tcp_splice.c | 43 ++++++++++++++++++++++--------------------- 1 file changed, 22 insertions(+), 21 deletions(-) -- 2.49.0

2 3

[PATCH] conf: Add missing return in conf_nat(), fix --map-guest-addr none
by Stefano Brivio 08 Apr '25

08 Apr '25

As reported by somebody on IRC: $ pasta --map-guest-addr none Invalid address to remap to host: none that's because once we parsed "none", we try to parse it as an address as well. But we already handled it, so stop once we're done. Fixes: e813a4df7da2 ("conf: Allow address remapped to host to be configured") Signed-off-by: Stefano Brivio <sbrivio(a)redhat.com> --- conf.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/conf.c b/conf.c index b54c55d..168646f 100644 --- a/conf.c +++ b/conf.c @@ -1272,6 +1272,8 @@ static void conf_nat(const char *arg, struct in_addr *addr4, *addr6 = in6addr_any; if (no_map_gw) *no_map_gw = 1; + + return; } if (inet_pton(AF_INET6, arg, addr6) && -- 2.43.0

2 1

[PATCH] udp_flow: Save 8 bytes in struct udp_flow on 64-bit architectures
by Stefano Brivio 08 Apr '25

08 Apr '25

Shuffle the fields just added by commits a7775e9550fa ("udp: support traceroute in direction tap-socket") and 9725e7988837 ("udp_flow: Don't discard packets that arrive between bind() and connect()"). On x86_64, as reported by pahole(1), before: struct udp_flow { struct flow_common f; /* 0 76 */ /* --- cacheline 1 boundary (64 bytes) was 12 bytes ago --- */ _Bool closed:1; /* 76: 0 1 */ /* XXX 7 bits hole, try to pack */ _Bool flush0; /* 77 1 */ _Bool flush1:1; /* 78: 0 1 */ /* XXX 7 bits hole, try to pack */ /* XXX 1 byte hole, try to pack */ time_t ts; /* 80 8 */ int s[2]; /* 88 8 */ uint8_t ttl[2]; /* 96 2 */ /* size: 104, cachelines: 2, members: 7 */ /* sum members: 95, holes: 1, sum holes: 1 */ /* sum bitfield members: 2 bits, bit holes: 2, sum bit holes: 14 bits */ /* padding: 6 */ /* last cacheline: 40 bytes */ }; and after: struct udp_flow { struct flow_common f; /* 0 76 */ /* --- cacheline 1 boundary (64 bytes) was 12 bytes ago --- */ uint8_t ttl[2]; /* 76 2 */ _Bool closed:1; /* 78: 0 1 */ _Bool flush0:1; /* 78: 1 1 */ _Bool flush1:1; /* 78: 2 1 */ /* XXX 5 bits hole, try to pack */ /* XXX 1 byte hole, try to pack */ time_t ts; /* 80 8 */ int s[2]; /* 88 8 */ /* size: 96, cachelines: 2, members: 7 */ /* sum members: 94, holes: 1, sum holes: 1 */ /* sum bitfield members: 3 bits, bit holes: 1, sum bit holes: 5 bits */ /* last cacheline: 32 bytes */ }; It doesn't matter much because anyway the typical storage for struct udp_flow is given by union flow: union flow { struct flow_common f; /* 0 76 */ struct flow_free_cluster free; /* 0 84 */ struct tcp_tap_conn tcp; /* 0 120 */ struct tcp_splice_conn tcp_splice; /* 0 120 */ struct icmp_ping_flow ping; /* 0 96 */ struct udp_flow udp; /* 0 96 */ }; but it still improves data locality somewhat, so let me fix this up now that commits are fresh. Signed-off-by: Stefano Brivio <sbrivio(a)redhat.com> --- udp_flow.h | 11 +++++++---- 1 file changed, 7 insertions(+), 4 deletions(-) diff --git a/udp_flow.h b/udp_flow.h index 90d3b29..e289122 100644 --- a/udp_flow.h +++ b/udp_flow.h @@ -10,22 +10,25 @@ /** * struct udp_flow - Descriptor for a flow of UDP packets * @f: Generic flow information + * @ttl: TTL or hop_limit for both sides * @closed: Flow is already closed * @flush0: @s[0] may have datagrams queued for other flows * @flush1: @s[1] may have datagrams queued for other flows * @ts: Activity timestamp * @s: Socket fd (or -1) for each side of the flow - * @ttl: TTL or hop_limit for both sides */ struct udp_flow { /* Must be first element */ struct flow_common f; - bool closed :1; - bool flush0, flush1 :1; + uint8_t ttl[SIDES]; + + bool closed :1, + flush0 :1, + flush1 :1; + time_t ts; int s[SIDES]; - uint8_t ttl[SIDES]; }; struct udp_flow *udp_at_sidx(flow_sidx_t sidx); -- 2.43.0

2 1

[PATCH v7] udp: support traceroute in direction tap-socket
by Jon Maloy 07 Apr '25

07 Apr '25

Now that ICMP pass-through from socket-to-tap is in place, it is easy to support UDP based traceroute functionality in direction tap-to-socket. We fix that in this commit. Link: https://bugs.passt.top/show_bug.cgi?id=64 Signed-off-by: Jon Maloy <jmaloy(a)redhat.com> --- v2: - Using ancillary data instead of setsockopt to transfer outgoing TTL. - Support IPv6 v3: - Storing ttl per packet instead of per flow. This may not be elegant, but much less intrusive than changing the flow criteria. This eliminates the need for the extra, flow-changing patch we introduced in v2. v4: - Going back to something similar to the original solution, but storing current ttl in struct udp_flow, plus ensuring that all packets in a struct tap4_l4_t/tap6_l4_t instance have the same ttl. After input from David Gibson. v5: - Some minor fixes after feedback from Stefano Brivio. v6: - Ensured that socket TTL is initialized at flow creation. - Ensured that setting of hop limit on IPv6 socket works. v7: - Setting of IPv6 hop_limit fails sometimes unless it is 4 bytes. --- tap.c | 17 +++++++++++++---- udp.c | 22 +++++++++++++++++++++- udp.h | 3 ++- udp_flow.c | 1 + udp_flow.h | 4 +++- 5 files changed, 40 insertions(+), 7 deletions(-) diff --git a/tap.c b/tap.c index 3a6fcbe..d630f6d 100644 --- a/tap.c +++ b/tap.c @@ -559,6 +559,7 @@ PACKET_POOL_DECL(pool_l4, UIO_MAXIOV, pkt_buf); * struct l4_seq4_t - Message sequence for one protocol handler call, IPv4 * @msgs: Count of messages in sequence * @protocol: Protocol number + * @ttl: Time to live * @source: Source port * @dest: Destination port * @saddr: Source address @@ -567,6 +568,7 @@ PACKET_POOL_DECL(pool_l4, UIO_MAXIOV, pkt_buf); */ static struct tap4_l4_t { uint8_t protocol; + uint8_t ttl; uint16_t source; uint16_t dest; @@ -586,6 +588,7 @@ static struct tap4_l4_t { * @dest: Destination port * @saddr: Source address * @daddr: Destination address + * @hop_limit: Hop limit * @msg: Array of messages that can be handled in a single call */ static struct tap6_l4_t { @@ -598,6 +601,8 @@ static struct tap6_l4_t { struct in6_addr saddr; struct in6_addr daddr; + uint8_t hop_limit; + struct pool_l4_t p; } tap6_l4[TAP_SEQS /* Arbitrary: TAP_MSGS in theory, so limit in users */]; @@ -786,7 +791,8 @@ resume: #define L4_MATCH(iph, uh, seq) \ ((seq)->protocol == (iph)->protocol && \ (seq)->source == (uh)->source && (seq)->dest == (uh)->dest && \ - (seq)->saddr.s_addr == (iph)->saddr && (seq)->daddr.s_addr == (iph)->daddr) + (seq)->saddr.s_addr == (iph)->saddr && \ + (seq)->daddr.s_addr == (iph)->daddr && (seq)->ttl == (iph)->ttl) #define L4_SET(iph, uh, seq) \ do { \ @@ -795,6 +801,7 @@ resume: (seq)->dest = (uh)->dest; \ (seq)->saddr.s_addr = (iph)->saddr; \ (seq)->daddr.s_addr = (iph)->daddr; \ + (seq)->ttl = (iph)->ttl; \ } while (0) if (seq && L4_MATCH(iph, uh, seq) && seq->p.count < UIO_MAXIOV) @@ -843,7 +850,7 @@ append: for (k = 0; k < p->count; ) k += udp_tap_handler(c, PIF_TAP, AF_INET, &seq->saddr, &seq->daddr, - p, k, now); + seq->ttl, p, k, now); } } @@ -966,7 +973,8 @@ resume: (seq)->dest == (uh)->dest && \ (seq)->flow_lbl == ip6_get_flow_lbl(ip6h) && \ IN6_ARE_ADDR_EQUAL(&(seq)->saddr, saddr) && \ - IN6_ARE_ADDR_EQUAL(&(seq)->daddr, daddr)) + IN6_ARE_ADDR_EQUAL(&(seq)->daddr, daddr) && \ + (seq)->hop_limit == (ip6h)->hop_limit) #define L4_SET(ip6h, proto, uh, seq) \ do { \ @@ -976,6 +984,7 @@ resume: (seq)->flow_lbl = ip6_get_flow_lbl(ip6h); \ (seq)->saddr = *saddr; \ (seq)->daddr = *daddr; \ + (seq)->hop_limit = (ip6h)->hop_limit; \ } while (0) if (seq && L4_MATCH(ip6h, proto, uh, seq) && @@ -1026,7 +1035,7 @@ append: for (k = 0; k < p->count; ) k += udp_tap_handler(c, PIF_TAP, AF_INET6, &seq->saddr, &seq->daddr, - p, k, now); + seq->hop_limit, p, k, now); } } diff --git a/udp.c b/udp.c index b275db3..3600336 100644 --- a/udp.c +++ b/udp.c @@ -818,6 +818,7 @@ fail: * @af: Address family, AF_INET or AF_INET6 * @saddr: Source address * @daddr: Destination address + * @ttl: TTL or hop limit for packets to be sent in this call * @p: Pool of UDP packets, with UDP headers * @idx: Index of first packet to process * @now: Current timestamp @@ -828,7 +829,8 @@ fail: */ int udp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af, const void *saddr, const void *daddr, - const struct pool *p, int idx, const struct timespec *now) + uint8_t ttl, const struct pool *p, int idx, + const struct timespec *now) { const struct flowside *toside; struct mmsghdr mm[UIO_MAXIOV]; @@ -907,6 +909,24 @@ int udp_tap_handler(const struct ctx *c, uint8_t pif, mm[i].msg_hdr.msg_controllen = 0; mm[i].msg_hdr.msg_flags = 0; + if (ttl != uflow->ttl[tosidx.sidei]) { + uflow->ttl[tosidx.sidei] = ttl; + if (af == AF_INET) { + if (setsockopt(s, IPPROTO_IP, IP_TTL, + &ttl, sizeof(ttl)) < 0) + flow_perror(uflow, + "setsockopt IP_TTL"); + } else { + /* IPv6 hop_limit cannot be only 1 byte */ + int hop_limit = ttl; + + if (setsockopt(s, SOL_IPV6, IPV6_UNICAST_HOPS, + &hop_limit, sizeof(hop_limit)) < 0) + flow_perror(uflow, + "setsockopt IPV6_UNICAST_HOPS"); + } + } + count++; } diff --git a/udp.h b/udp.h index 8fc4283..8f8531a 100644 --- a/udp.h +++ b/udp.h @@ -15,7 +15,8 @@ void udp_sock_handler(const struct ctx *c, union epoll_ref ref, uint32_t events, const struct timespec *now); int udp_tap_handler(const struct ctx *c, uint8_t pif, sa_family_t af, const void *saddr, const void *daddr, - const struct pool *p, int idx, const struct timespec *now); + uint8_t ttl, const struct pool *p, int idx, + const struct timespec *now); int udp_sock_init(const struct ctx *c, int ns, const union inany_addr *addr, const char *ifname, in_port_t port); int udp_init(struct ctx *c); diff --git a/udp_flow.c b/udp_flow.c index af15d7f..d84d170 100644 --- a/udp_flow.c +++ b/udp_flow.c @@ -136,6 +136,7 @@ static flow_sidx_t udp_flow_new(const struct ctx *c, union flow *flow, uflow = FLOW_SET_TYPE(flow, FLOW_UDP, udp); uflow->ts = now->tv_sec; uflow->s[INISIDE] = uflow->s[TGTSIDE] = -1; + uflow->ttl[INISIDE] = uflow->ttl[TGTSIDE] = 0; flow_foreach_sidei(sidei) { if (pif_is_socket(uflow->f.pif[sidei])) diff --git a/udp_flow.h b/udp_flow.h index d518737..90d3b29 100644 --- a/udp_flow.h +++ b/udp_flow.h @@ -8,13 +8,14 @@ #define UDP_FLOW_H /** - * struct udp - Descriptor for a flow of UDP packets + * struct udp_flow - Descriptor for a flow of UDP packets * @f: Generic flow information * @closed: Flow is already closed * @flush0: @s[0] may have datagrams queued for other flows * @flush1: @s[1] may have datagrams queued for other flows * @ts: Activity timestamp * @s: Socket fd (or -1) for each side of the flow + * @ttl: TTL or hop_limit for both sides */ struct udp_flow { /* Must be first element */ @@ -24,6 +25,7 @@ struct udp_flow { bool flush0, flush1 :1; time_t ts; int s[SIDES]; + uint8_t ttl[SIDES]; }; struct udp_flow *udp_at_sidx(flow_sidx_t sidx); -- 2.48.1

3 2