From owner-dev-commits-src-branches@freebsd.org Wed Aug 11 02:34:25 2021 Return-Path: Delivered-To: dev-commits-src-branches@mailman.nyi.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) by mailman.nyi.freebsd.org (Postfix) with ESMTP id B543867335F; Wed, 11 Aug 2021 02:34:25 +0000 (UTC) (envelope-from git@FreeBSD.org) Received: from mxrelay.nyi.freebsd.org (mxrelay.nyi.freebsd.org [IPv6:2610:1c1:1:606c::19:3]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256 client-signature RSA-PSS (4096 bits) client-digest SHA256) (Client CN "mxrelay.nyi.freebsd.org", Issuer "R3" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id 4Gkv694jQGz3j4V; Wed, 11 Aug 2021 02:34:25 +0000 (UTC) (envelope-from git@FreeBSD.org) Received: from gitrepo.freebsd.org (gitrepo.freebsd.org [IPv6:2610:1c1:1:6068::e6a:5]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256) (Client did not present a certificate) by mxrelay.nyi.freebsd.org (Postfix) with ESMTPS id 85F1D27205; Wed, 11 Aug 2021 02:34:25 +0000 (UTC) (envelope-from git@FreeBSD.org) Received: from gitrepo.freebsd.org ([127.0.1.44]) by gitrepo.freebsd.org (8.16.1/8.16.1) with ESMTP id 17B2YPBu033171; Wed, 11 Aug 2021 02:34:25 GMT (envelope-from git@gitrepo.freebsd.org) Received: (from git@localhost) by gitrepo.freebsd.org (8.16.1/8.16.1/Submit) id 17B2YPHT033170; Wed, 11 Aug 2021 02:34:25 GMT (envelope-from git) Date: Wed, 11 Aug 2021 02:34:25 GMT Message-Id: <202108110234.17B2YPHT033170@gitrepo.freebsd.org> To: src-committers@FreeBSD.org, dev-commits-src-all@FreeBSD.org, dev-commits-src-branches@FreeBSD.org From: Kevin Bowling Subject: git: 664077e69e8f - stable/12 - socket: Implement SO_RERROR MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit X-Git-Committer: kbowling X-Git-Repository: src X-Git-Refname: refs/heads/stable/12 X-Git-Reftype: branch X-Git-Commit: 664077e69e8f300cd0218e3c700ee235df12fcff Auto-Submitted: auto-generated X-BeenThere: dev-commits-src-branches@freebsd.org X-Mailman-Version: 2.1.34 Precedence: list List-Id: Commits to the stable branches of the FreeBSD src repository List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 11 Aug 2021 02:34:25 -0000 The branch stable/12 has been updated by kbowling (ports committer): URL: https://cgit.FreeBSD.org/src/commit/?id=664077e69e8f300cd0218e3c700ee235df12fcff commit 664077e69e8f300cd0218e3c700ee235df12fcff Author: Roy Marples AuthorDate: 2021-07-28 15:46:59 +0000 Commit: Kevin Bowling CommitDate: 2021-08-11 01:59:36 +0000 socket: Implement SO_RERROR SO_RERROR indicates that receive buffer overflows should be handled as errors. Historically receive buffer overflows have been ignored and programs could not tell if they missed messages or messages had been truncated because of overflows. Since programs historically do not expect to get receive overflow errors, this behavior is not the default. This is really really important for programs that use route(4) to keep in sync with the system. If we loose a message then we need to reload the full system state, otherwise the behaviour from that point is undefined and can lead to chasing bogus bug reports. Reviewed by: philip (network), kbowling (transport), gbe (manpages) MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D26652 (cherry picked from commit 7045b1603bdf054145dd958a4acc17b410fb62a0) --- lib/libc/sys/getsockopt.2 | 10 +++++++- sbin/route/route.c | 13 +++++++++- sys/kern/uipc_sockbuf.c | 24 +++++++++++++++++ sys/kern/uipc_socket.c | 30 ++++++++++++++++------ sys/kern/uipc_usrreq.c | 2 +- sys/net/raw_usrreq.c | 11 ++++---- .../bluetooth/socket/ng_btsocket_hci_raw.c | 1 + sys/netgraph/ng_socket.c | 2 +- sys/netinet/ip_divert.c | 2 +- sys/netinet/ip_mroute.c | 2 +- sys/netinet/raw_ip.c | 3 +-- sys/netinet/udp_usrreq.c | 2 +- sys/netinet6/icmp6.c | 6 ++--- sys/netinet6/ip6_input.c | 1 + sys/netinet6/ip6_mroute.c | 3 ++- sys/netinet6/raw_ip6.c | 2 ++ sys/netinet6/send.c | 2 +- sys/netinet6/udp6_usrreq.c | 2 +- sys/netipsec/keysock.c | 10 ++++---- sys/sys/socket.h | 1 + sys/sys/socketvar.h | 6 ++++- 21 files changed, 100 insertions(+), 35 deletions(-) diff --git a/lib/libc/sys/getsockopt.2 b/lib/libc/sys/getsockopt.2 index dcd39e07efb8..b27f663f01bd 100644 --- a/lib/libc/sys/getsockopt.2 +++ b/lib/libc/sys/getsockopt.2 @@ -28,7 +28,7 @@ .\" @(#)getsockopt.2 8.4 (Berkeley) 5/2/95 .\" $FreeBSD$ .\" -.Dd June 03, 2020 +.Dd February 8, 2021 .Dt GETSOCKOPT 2 .Os .Sh NAME @@ -177,6 +177,7 @@ for the socket .It Dv SO_PROTOCOL Ta "get the protocol number for the socket (get only)" .It Dv SO_PROTOTYPE Ta "SunOS alias for the Linux SO_PROTOCOL (get only)" .It Dv SO_ERROR Ta "get and clear error on the socket (get only)" +.It Dv SO_RERROR Ta "enables receive error reporting" .It Dv SO_SETFIB Ta "set the associated FIB (routing table) for the socket (set only)" .El .Pp @@ -513,6 +514,13 @@ returns any pending error on the socket and clears the error status. It may be used to check for asynchronous errors on connected datagram sockets or for other asynchronous errors. +.Dv SO_RERROR +indicates that receive buffer overflows should be handled as errors. +Historically receive buffer overflows have been ignored and programs +could not tell if they missed messages or messages had been truncated +because of overflows. +Since programs historically do not expect to get receive overflow errors, +this behavior is not the default. .Pp .Dv SO_LABEL returns the MAC label of the socket. diff --git a/sbin/route/route.c b/sbin/route/route.c index 9c9e4b304848..14d1fe274ad3 100644 --- a/sbin/route/route.c +++ b/sbin/route/route.c @@ -1485,9 +1485,20 @@ monitor(int argc, char *argv[]) interfaces(); exit(0); } + +#ifdef SO_RERROR + n = 1; + if (setsockopt(s, SOL_SOCKET, SO_RERROR, &n, sizeof(n)) == -1) + warn("SO_RERROR"); +#endif + for (;;) { time_t now; - n = read(s, msg, 2048); + n = read(s, msg, sizeof(msg)); + if (n == -1) { + warn("read"); + continue; + } now = time(NULL); (void)printf("\ngot message of size %d on %s", n, ctime(&now)); print_rtmsg((struct rt_msghdr *)(void *)msg, n); diff --git a/sys/kern/uipc_sockbuf.c b/sys/kern/uipc_sockbuf.c index 598fc05f2fcf..92b6d728fcd8 100644 --- a/sys/kern/uipc_sockbuf.c +++ b/sys/kern/uipc_sockbuf.c @@ -254,6 +254,30 @@ socantrcvmore(struct socket *so) mtx_assert(SOCKBUF_MTX(&so->so_rcv), MA_NOTOWNED); } +void +soroverflow_locked(struct socket *so) +{ + + SOCKBUF_LOCK_ASSERT(&so->so_rcv); + + if (so->so_options & SO_RERROR) { + so->so_rerror = ENOBUFS; + sorwakeup_locked(so); + } else + SOCKBUF_UNLOCK(&so->so_rcv); + + mtx_assert(SOCKBUF_MTX(&so->so_rcv), MA_NOTOWNED); +} + +void +soroverflow(struct socket *so) +{ + + SOCKBUF_LOCK(&so->so_rcv); + soroverflow_locked(so); + mtx_assert(SOCKBUF_MTX(&so->so_rcv), MA_NOTOWNED); +} + /* * Wait for data to arrive at/drain from a socket buffer. */ diff --git a/sys/kern/uipc_socket.c b/sys/kern/uipc_socket.c index dd028c660d2b..620d81246062 100644 --- a/sys/kern/uipc_socket.c +++ b/sys/kern/uipc_socket.c @@ -1771,12 +1771,19 @@ restart: KASSERT(m != NULL || !sbavail(&so->so_rcv), ("receive: m == %p sbavail == %u", m, sbavail(&so->so_rcv))); - if (so->so_error) { + if (so->so_error || so->so_rerror) { if (m != NULL) goto dontblock; - error = so->so_error; - if ((flags & MSG_PEEK) == 0) - so->so_error = 0; + if (so->so_error) + error = so->so_error; + else + error = so->so_rerror; + if ((flags & MSG_PEEK) == 0) { + if (so->so_error) + so->so_error = 0; + else + so->so_rerror = 0; + } SOCKBUF_UNLOCK(&so->so_rcv); goto release; } @@ -2084,7 +2091,7 @@ dontblock: while (flags & MSG_WAITALL && m == NULL && uio->uio_resid > 0 && !sosendallatonce(so) && nextrecord == NULL) { SOCKBUF_LOCK_ASSERT(&so->so_rcv); - if (so->so_error || + if (so->so_error || so->so_rerror || so->so_rcv.sb_state & SBS_CANTRCVMORE) break; /* @@ -2805,6 +2812,7 @@ sosetopt(struct socket *so, struct sockopt *sopt) case SO_NOSIGPIPE: case SO_NO_DDP: case SO_NO_OFFLOAD: + case SO_RERROR: error = sooptcopyin(sopt, &optval, sizeof optval, sizeof optval); if (error) @@ -3026,6 +3034,7 @@ sogetopt(struct socket *so, struct sockopt *sopt) case SO_NOSIGPIPE: case SO_NO_DDP: case SO_NO_OFFLOAD: + case SO_RERROR: optval = so->so_options & sopt->sopt_name; integer: error = sooptcopyout(sopt, &optval, sizeof optval); @@ -3045,8 +3054,13 @@ integer: case SO_ERROR: SOCK_LOCK(so); - optval = so->so_error; - so->so_error = 0; + if (so->so_error) { + optval = so->so_error; + so->so_error = 0; + } else { + optval = so->so_rerror; + so->so_rerror = 0; + } SOCK_UNLOCK(so); goto integer; @@ -3595,7 +3609,7 @@ filt_soread(struct knote *kn, long hint) kn->kn_flags |= EV_EOF; kn->kn_fflags = so->so_error; return (1); - } else if (so->so_error) /* temporary udp error */ + } else if (so->so_error || so->so_rerror) return (1); if (kn->kn_sfflags & NOTE_LOWAT) { diff --git a/sys/kern/uipc_usrreq.c b/sys/kern/uipc_usrreq.c index 25f27837d6d5..329e4346c425 100644 --- a/sys/kern/uipc_usrreq.c +++ b/sys/kern/uipc_usrreq.c @@ -1047,7 +1047,7 @@ uipc_send(struct socket *so, int flags, struct mbuf *m, struct sockaddr *nam, m = NULL; control = NULL; } else { - SOCKBUF_UNLOCK(&so2->so_rcv); + soroverflow_locked(so2); error = ENOBUFS; } if (nam != NULL) diff --git a/sys/net/raw_usrreq.c b/sys/net/raw_usrreq.c index f43de7dae9ca..5d4e223e5a0a 100644 --- a/sys/net/raw_usrreq.c +++ b/sys/net/raw_usrreq.c @@ -100,10 +100,10 @@ raw_input_ext(struct mbuf *m0, struct sockproto *proto, struct sockaddr *src, n = m_copym(m, 0, M_COPYALL, M_NOWAIT); if (n) { if (sbappendaddr(&last->so_rcv, src, - n, (struct mbuf *)0) == 0) - /* should notify about lost packet */ + n, (struct mbuf *)0) == 0) { + soroverflow(last); m_freem(n); - else + } else sorwakeup(last); } } @@ -111,9 +111,10 @@ raw_input_ext(struct mbuf *m0, struct sockproto *proto, struct sockaddr *src, } if (last) { if (sbappendaddr(&last->so_rcv, src, - m, (struct mbuf *)0) == 0) + m, (struct mbuf *)0) == 0) { + soroverflow(last); m_freem(m); - else + } else sorwakeup(last); } else m_freem(m); diff --git a/sys/netgraph/bluetooth/socket/ng_btsocket_hci_raw.c b/sys/netgraph/bluetooth/socket/ng_btsocket_hci_raw.c index 9523d88af633..a78905705fd0 100644 --- a/sys/netgraph/bluetooth/socket/ng_btsocket_hci_raw.c +++ b/sys/netgraph/bluetooth/socket/ng_btsocket_hci_raw.c @@ -539,6 +539,7 @@ ng_btsocket_hci_raw_data_input(struct mbuf *nam) NG_FREE_M(m); NG_FREE_M(ctl); + soroverflow(pcb->so); } } next: diff --git a/sys/netgraph/ng_socket.c b/sys/netgraph/ng_socket.c index 08ee76300ae8..a10993a8fab6 100644 --- a/sys/netgraph/ng_socket.c +++ b/sys/netgraph/ng_socket.c @@ -974,7 +974,7 @@ ngs_rcvmsg(node_p node, item_p item, hook_p lasthook) /* Send it up to the socket. */ if (sbappendaddr_locked(&so->so_rcv, (struct sockaddr *)&addr, m, NULL) == 0) { - SOCKBUF_UNLOCK(&so->so_rcv); + soroverflow_locked(so); TRAP_ERROR; m_freem(m); return (ENOBUFS); diff --git a/sys/netinet/ip_divert.c b/sys/netinet/ip_divert.c index 81e70177e641..54cd0f509b51 100644 --- a/sys/netinet/ip_divert.c +++ b/sys/netinet/ip_divert.c @@ -287,7 +287,7 @@ divert_packet(struct mbuf *m, int incoming) if (sbappendaddr_locked(&sa->so_rcv, (struct sockaddr *)&divsrc, m, (struct mbuf *)0) == 0) { - SOCKBUF_UNLOCK(&sa->so_rcv); + soroverflow_locked(sa); sa = NULL; /* force mbuf reclaim below */ } else sorwakeup_locked(sa); diff --git a/sys/netinet/ip_mroute.c b/sys/netinet/ip_mroute.c index d024cdb16772..60b18cdbd35c 100644 --- a/sys/netinet/ip_mroute.c +++ b/sys/netinet/ip_mroute.c @@ -1195,7 +1195,7 @@ socket_send(struct socket *s, struct mbuf *mm, struct sockaddr_in *src) sorwakeup_locked(s); return 0; } - SOCKBUF_UNLOCK(&s->so_rcv); + soroverflow_locked(s); } m_freem(mm); return -1; diff --git a/sys/netinet/raw_ip.c b/sys/netinet/raw_ip.c index f15a3bca3a90..84dd4d7a764c 100644 --- a/sys/netinet/raw_ip.c +++ b/sys/netinet/raw_ip.c @@ -261,11 +261,10 @@ rip_append(struct inpcb *last, struct ip *ip, struct mbuf *n, SOCKBUF_LOCK(&so->so_rcv); if (sbappendaddr_locked(&so->so_rcv, (struct sockaddr *)ripsrc, n, opts) == 0) { - /* should notify about lost packet */ + soroverflow_locked(so); m_freem(n); if (opts) m_freem(opts); - SOCKBUF_UNLOCK(&so->so_rcv); } else sorwakeup_locked(so); } else diff --git a/sys/netinet/udp_usrreq.c b/sys/netinet/udp_usrreq.c index c271392c225d..2a5e8fdd25c2 100644 --- a/sys/netinet/udp_usrreq.c +++ b/sys/netinet/udp_usrreq.c @@ -376,7 +376,7 @@ udp_append(struct inpcb *inp, struct ip *ip, struct mbuf *n, int off, so = inp->inp_socket; SOCKBUF_LOCK(&so->so_rcv); if (sbappendaddr_locked(&so->so_rcv, append_sa, n, opts) == 0) { - SOCKBUF_UNLOCK(&so->so_rcv); + soroverflow(so); m_freem(n); if (opts) m_freem(opts); diff --git a/sys/netinet6/icmp6.c b/sys/netinet6/icmp6.c index 75e7ef7c2039..923f5dd63ddc 100644 --- a/sys/netinet6/icmp6.c +++ b/sys/netinet6/icmp6.c @@ -1976,13 +1976,11 @@ icmp6_rip6_input(struct mbuf **mp, int off) &last->inp_socket->so_rcv, (struct sockaddr *)&fromsa, n, opts) == 0) { - /* should notify about lost packet */ + soroverflow_locked(last->inp_socket); m_freem(n); if (opts) { m_freem(opts); } - SOCKBUF_UNLOCK( - &last->inp_socket->so_rcv); } else sorwakeup_locked(last->inp_socket); opts = NULL; @@ -2023,7 +2021,7 @@ icmp6_rip6_input(struct mbuf **mp, int off) m_freem(m); if (opts) m_freem(opts); - SOCKBUF_UNLOCK(&last->inp_socket->so_rcv); + soroverflow_locked(last->inp_socket); } else sorwakeup_locked(last->inp_socket); INP_RUNLOCK(last); diff --git a/sys/netinet6/ip6_input.c b/sys/netinet6/ip6_input.c index 74d275f7e66e..9ef4620565a1 100644 --- a/sys/netinet6/ip6_input.c +++ b/sys/netinet6/ip6_input.c @@ -1579,6 +1579,7 @@ ip6_notify_pmtu(struct inpcb *inp, struct sockaddr_in6 *dst, u_int32_t mtu) so = inp->inp_socket; if (sbappendaddr(&so->so_rcv, (struct sockaddr *)dst, NULL, m_mtu) == 0) { + soroverflow(so); m_freem(m_mtu); /* XXX: should count statistics */ } else diff --git a/sys/netinet6/ip6_mroute.c b/sys/netinet6/ip6_mroute.c index 82ca908d4553..bce491d7fa5b 100644 --- a/sys/netinet6/ip6_mroute.c +++ b/sys/netinet6/ip6_mroute.c @@ -1039,7 +1039,8 @@ socket_send(struct socket *s, struct mbuf *mm, struct sockaddr_in6 *src) mm, (struct mbuf *)0) != 0) { sorwakeup(s); return (0); - } + } else + soroverflow(s); } m_freem(mm); return (-1); diff --git a/sys/netinet6/raw_ip6.c b/sys/netinet6/raw_ip6.c index 304ee4a8899b..47e106f02651 100644 --- a/sys/netinet6/raw_ip6.c +++ b/sys/netinet6/raw_ip6.c @@ -212,6 +212,7 @@ rip6_input(struct mbuf **mp, int *offp, int proto) if (sbappendaddr(&last->inp_socket->so_rcv, (struct sockaddr *)&fromsa, n, opts) == 0) { + soroverflow(last->inp_socket); m_freem(n); if (opts) m_freem(opts); @@ -324,6 +325,7 @@ skip_2: m_adj(m, *offp); if (sbappendaddr(&last->inp_socket->so_rcv, (struct sockaddr *)&fromsa, m, opts) == 0) { + soroverflow(last->inp_socket); m_freem(m); if (opts) m_freem(opts); diff --git a/sys/netinet6/send.c b/sys/netinet6/send.c index bc9880c82267..8458ef367cba 100644 --- a/sys/netinet6/send.c +++ b/sys/netinet6/send.c @@ -291,7 +291,7 @@ send_input(struct mbuf *m, struct ifnet *ifp, int direction, int msglen __unused SOCKBUF_LOCK(&V_send_so->so_rcv); if (sbappendaddr_locked(&V_send_so->so_rcv, (struct sockaddr *)&sendsrc, m, NULL) == 0) { - SOCKBUF_UNLOCK(&V_send_so->so_rcv); + soroverflow_locked(V_send_so); /* XXX stats. */ m_freem(m); } else { diff --git a/sys/netinet6/udp6_usrreq.c b/sys/netinet6/udp6_usrreq.c index 98588e48c511..02d4359dc92f 100644 --- a/sys/netinet6/udp6_usrreq.c +++ b/sys/netinet6/udp6_usrreq.c @@ -191,7 +191,7 @@ udp6_append(struct inpcb *inp, struct mbuf *n, int off, SOCKBUF_LOCK(&so->so_rcv); if (sbappendaddr_locked(&so->so_rcv, (struct sockaddr *)&fromsa[0], n, opts) == 0) { - SOCKBUF_UNLOCK(&so->so_rcv); + soroverflow_locked(so); m_freem(n); if (opts) m_freem(opts); diff --git a/sys/netipsec/keysock.c b/sys/netipsec/keysock.c index a216f085c1bc..84a65967eea2 100644 --- a/sys/netipsec/keysock.c +++ b/sys/netipsec/keysock.c @@ -141,7 +141,6 @@ end: static int key_sendup0(struct rawcb *rp, struct mbuf *m, int promisc) { - int error; if (promisc) { struct sadb_msg *pmsg; @@ -165,11 +164,12 @@ key_sendup0(struct rawcb *rp, struct mbuf *m, int promisc) m, NULL)) { PFKEYSTAT_INC(in_nomem); m_freem(m); - error = ENOBUFS; - } else - error = 0; + soroverflow(rp->rcb_socket); + return ENOBUFS; + } + sorwakeup(rp->rcb_socket); - return error; + return 0; } /* so can be NULL if target != KEY_SENDUP_ONE */ diff --git a/sys/sys/socket.h b/sys/sys/socket.h index eaad9b1bacdb..17a5099424f4 100644 --- a/sys/sys/socket.h +++ b/sys/sys/socket.h @@ -147,6 +147,7 @@ typedef __uintptr_t uintptr_t; #define SO_NO_OFFLOAD 0x00004000 /* socket cannot be offloaded */ #define SO_NO_DDP 0x00008000 /* disable direct data placement */ #define SO_REUSEPORT_LB 0x00010000 /* reuse with load balancing */ +#define SO_RERROR 0x00020000 /* keep track of receive errors */ /* * Additional options, not kept in so_options. diff --git a/sys/sys/socketvar.h b/sys/sys/socketvar.h index 99d8839f8beb..99d2d554143e 100644 --- a/sys/sys/socketvar.h +++ b/sys/sys/socketvar.h @@ -99,6 +99,7 @@ struct socket { struct protosw *so_proto; /* (a) protocol handle */ short so_timeo; /* (g) connection timeout */ u_short so_error; /* (f) error affecting connection */ + u_short so_rerror; /* (f) error affecting connection */ struct sigio *so_sigio; /* [sg] information for async I/O or out of band data (SIGURG) */ struct ucred *so_cred; /* (a) user credentials */ @@ -258,7 +259,8 @@ struct socket { /* can we read something from so? */ #define soreadabledata(so) \ - (sbavail(&(so)->so_rcv) >= (so)->so_rcv.sb_lowat || (so)->so_error) + (sbavail(&(so)->so_rcv) >= (so)->so_rcv.sb_lowat || \ + (so)->so_error || (so)->so_rerror) #define soreadable(so) \ (soreadabledata(so) || ((so)->so_rcv.sb_state & SBS_CANTRCVMORE)) @@ -471,6 +473,8 @@ void socantrcvmore(struct socket *so); void socantrcvmore_locked(struct socket *so); void socantsendmore(struct socket *so); void socantsendmore_locked(struct socket *so); +void soroverflow(struct socket *so); +void soroverflow_locked(struct socket *so); /* * Accept filter functions (duh).