From b7d4e1a5f125358b49307256672a0db7f425a4e5 Mon Sep 17 00:00:00 2001 From: Howard Chu Date: Thu, 18 Nov 2004 16:41:04 +0000 Subject: [PATCH] First pass at macro-izing event handling method. Still only supports select(). If you think this is too ugly, feel free to revert... --- servers/slapd/daemon.c | 845 +++++++++++++++++++++++++---------------- 1 file changed, 511 insertions(+), 334 deletions(-) diff --git a/servers/slapd/daemon.c b/servers/slapd/daemon.c index 3c31b18a5f..5cc423dc98 100644 --- a/servers/slapd/daemon.c +++ b/servers/slapd/daemon.c @@ -85,19 +85,195 @@ static struct slap_daemon { ldap_pvt_thread_mutex_t sd_mutex; ber_socket_t sd_nactives; + int sd_nwriters; + +#ifdef HAVE_EPOLL + struct epoll_event *sd_epolls; + int sd_nepolls; + int *sd_index; + int sd_epfd; + int sd_nfds; +#else #ifndef HAVE_WINSOCK /* In winsock, accept() returns values higher than dtblsize so don't bother with this optimization */ int sd_nfds; #endif - int sd_nwriters; fd_set sd_actives; fd_set sd_readers; fd_set sd_writers; +#endif } slap_daemon; +#ifdef HAVE_EPOLL +#define SLAP_EVENTS_ARE_INDEXED 0 +#define SLAP_SOCK_IX(fd) (slap_daemon.sd_index[fd]) +#define SLAP_SOCK_EP(fd) (slap_daemon.sd_epolls[SLAP_SOCK_IX(fd)]) +#define SLAP_SOCK_FD(fd) (SLAP_SOCK_EP(fd).data.fd) +#define SLAP_SOCK_EV(fd) (SLAP_SOCK_EP(fd).events) +#define SLAP_SOCK_IS_ACTIVE(fd) (SLAP_SOCK_IX(fd) != -1 && SLAP_SOCK_FD(fd) == fd) +#define SLAP_SOCK_NOT_ACTIVE(fd) (SLAP_SOCK_IX(fd) == -1) +#define SLAP_SOCK_IS_SET(fd, mode) (SLAP_SOCK_EV(fd) & mode) + +#define SLAP_SOCK_IS_READ(fd) SLAP_SOCK_IS_SET(fd, EPOLLIN) +#define SLAP_SOCK_IS_WRITE(fd) SLAP_SOCK_IS_SET(fd, EPOLLOUT) + +#define SLAP_SET_SOCK(fd, events) do { \ + assert(SLAP_SOCK_IS_ACTIVE(fd)); \ + if ((SLAP_SOCK_EV(fd) & events) != events) { \ + SLAP_SOCK_EV(fd) |= events; \ + rc = epoll_ctl(slap_daemon.sd_epfd, EPOLL_CTL_MOD, fd, \ + &SLAP_SOCK_EP(fd)); \ + } \ +} while(0) + +#define SLAP_CLR_SOCK(fd, events) do { \ + assert(SLAP_SOCK_IS_ACTIVE(fd)); \ + if ((SLAP_SOCK_EV(fd) & events)) { \ + SLAP_SOCK_EV(fd) &= ~events; \ + rc = epoll_ctl(slap_daemon.sd_epfd, EPOLL_CTL_MOD, fd, \ + &SLAP_SOCK_EP(fd)); \ + } \ +} while(0) \ + +#define SLAP_SOCK_SET_READ(fd) SLAP_SET_SOCK(fd, EPOLLIN) +#define SLAP_SOCK_SET_WRITE(fd) SLAP_SET_SOCK(fd, EPOLLOUT) + +#define SLAP_SOCK_CLR_READ(fd) SLAP_CLR_SOCK(fd, EPOLLIN) +#define SLAP_SOCK_CLR_WRITE(fd) SLAP_CLR_SOCK(fd, EPOLLOUT) + +#define SLAP_CLR_EVENT(i, events) (revents[i].events &= ~events) + +#define SLAP_EVENT_CLR_READ(i) SLAP_CLR_EVENT(i, EPOLLIN) +#define SLAP_EVENT_CLR_WRITE(i) SLAP_CLR_EVENT(i, EPOLLOUT) + +#define SLAP_CHK_EVENT(i, events) (revents[i].events & events) + +#define SLAP_EVENT_IS_READ(i) SLAP_CHK_EVENT(i, EPOLLIN) +#define SLAP_EVENT_IS_WRITE(i) SLAP_CHK_EVENT(i, EPOLLOUT) + +#define SLAP_EVENT_FD(i) (revents[i].data.fd) + +#define SLAP_ADD_SOCK(fd) do { \ + SLAP_SOCK_IX(fd) = slap_daemon.sd_nfds; \ + SLAP_SOCK_FD(fd) = fd; \ + SLAP_SOCK_EV(fd) = EPOLLIN; \ + rc = epoll_ctl(slap_daemon.sd_epfd, EPOLL_CTL_ADD, fd, \ + &SLAP_SOCK_EP(fd)); \ + if ( rc == 0 ) slap_daemon.sd_nfds++; \ +} while(0) + +#define SLAP_DEL_SOCK(fd) do { \ + int index = SLAP_SOCK_IX(fd); + + rc = epoll_ctl(slap_daemon.sd_epfd, EPOLL_CTL_DEL, fd, \ + &SLAP_SOCK_EP(fd)); \ + slap_daemon.sd_epolls[index] = slap_daemon.sd_epolls[slap_daemon.sd_nfds-1]; \ + slap_daemon.sd_index[slap_daemon.sd_epools[index].data.fd] = index; \ + slap_daemon.sd_index[fd] = -1; \ + slap_daemon.sd_nfds--; \ +} while(0) + +#define SLAP_SOCK_SET_MUTE(fd) SLAP_SOCK_CLR_READ(fd) +#define SLAP_SOCK_CLR_MUTE(fd) SLAP_SOCK_SET_READ(fd) +#define SLAP_SOCK_IS_MUTE(fd) !SLAP_SOCK_IS_READ(fd) + +#define SLAP_SOCK_SET_INIT \ + slap_daemon.sd_epolls = ch_malloc(sizeof(struct epoll_event) * dtblsize * 2); \ + slap_daemon.sd_index = ch_malloc(sizeof(int) * dtblsize); \ + slap_daemon.sd_epfd = epoll_create( dtblsize ) + +#define SLAP_EVENT_DECL \ + struct epoll_event *revents + +#define SLAP_EVENT_INIT \ + revents = slap_daemon.sd_epolls + dtblsize; \ + +#define SLAP_EVENT_WAIT(tvp) \ + epoll_wait( slap_daemon.sd_epfd, revents, dtblsize, tvp ? tvp->tv_sec * 1000 : -1 ) + +#else + +/* select */ +#define SLAP_EVENTS_ARE_INDEXED 1 +#define SLAP_EVENT_DECL \ + fd_set readfds, writefds + +#define SLAP_EVENT_INIT \ + AC_MEMCPY( &readfds, &slap_daemon.sd_readers, sizeof(fd_set) ); \ + if ( nwriters ) \ + AC_MEMCPY( &writefds, &slap_daemon.sd_writers, sizeof(fd_set) ); + +#ifdef FD_SETSIZE +#define CHK_SETSIZE \ + if (dtblsize > FD_SETSIZE) dtblsize = FD_SETSIZE +#else +#define CHK_SETSIZE +#endif + +#define SLAP_SOCK_SET_INIT \ + CHK_SETSIZE; \ + FD_ZERO(&slap_daemon.sd_readers); \ + FD_ZERO(&slap_daemon.sd_writers) + +#define SLAP_SOCK_IS_ACTIVE(fd) FD_ISSET(fd, &slap_daemon.sd_actives) + +#define SLAP_SOCK_IS_READ(fd) FD_ISSET(fd, &slap_daemon.sd_readers) +#define SLAP_SOCK_IS_WRITE(fd) FD_ISSET(fd, &slap_daemon.sd_writers) + +#define SLAP_SOCK_NOT_ACTIVE(fd) (!SLAP_SOCK_IS_ACTIVE(fd) && \ + !SLAP_SOCK_IS_READ(fd) && !SLAP_SOCK_IS_WRITE(fd)) + +#ifdef HAVE_WINSOCK +#define SLAP_SOCK_SET_READ(fd) do { \ + if (!SLAP_SOCK_IS_READ(fd)) {FD_SET(fd, &slap_daemon.sd_readers);} \ +} while(0) +#define SLAP_SOCK_SET_WRITE(fd) do { \ + if (!SLAP_SOCK_IS_WRITE(fd)) {FD_SET(fd, &slap_daemon.sd_writers);} \ +} while(0) + +#define SLAP_ADDTEST(s) +#define SLAP_MAXWAIT dtblsize +#else +#define SLAP_SOCK_SET_READ(fd) FD_SET(fd, &slap_daemon.sd_readers) +#define SLAP_SOCK_SET_WRITE(fd) FD_SET(fd, &slap_daemon.sd_writers) + +#define SLAP_MAXWAIT slap_daemon.sd_nfds +#define SLAP_ADDTEST(s) if (s >= slap_daemon.sd_nfds) slap_daemon.sd_nfds = s+1 +#endif + +#define SLAP_SOCK_CLR_READ(fd) FD_CLR(fd, &slap_daemon.sd_readers) +#define SLAP_SOCK_CLR_WRITE(fd) FD_CLR(fd, &slap_daemon.sd_writers) + +#define SLAP_ADD_SOCK(s) do { \ + SLAP_ADDTEST(s); \ + FD_SET(s, &slap_daemon.sd_actives); \ + FD_SET(s, &slap_daemon.sd_readers); \ +} while(0) + +#define SLAP_DEL_SOCK(s) do { \ + FD_CLR(s, &slap_daemon.sd_actives); \ + FD_CLR(s, &slap_daemon.sd_readers); \ + FD_CLR(s, &slap_daemon.sd_writers); \ +} while(0) + +#define SLAP_EVENT_IS_READ(fd) FD_ISSET(fd, &readfds) +#define SLAP_EVENT_IS_WRITE(fd) FD_ISSET(fd, &writefds) + +#define SLAP_EVENT_CLR_READ(fd) FD_CLR(fd, &readfds) +#define SLAP_EVENT_CLR_WRITE(fd) FD_CLR(fd, &writefds) + +#define SLAP_EVENT_WAIT(tvp) \ + select( SLAP_MAXWAIT, &readfds, \ + nwriters > 0 ? &writefds : NULL, NULL, tvp ) + +#define SLAP_SOCK_SET_MUTE(s) FD_CLR(s, &readfds) +#define SLAP_SOCK_CLR_MUTE(s) FD_SET(s, &readfds) +#define SLAP_SOCK_IS_MUTE(s) FD_ISSET(s, &readfds) + +#endif #ifdef HAVE_SLP @@ -217,27 +393,16 @@ void slapd_slp_dereg() { static void slapd_add(ber_socket_t s, int isactive) { ldap_pvt_thread_mutex_lock( &slap_daemon.sd_mutex ); - assert( !FD_ISSET( s, &slap_daemon.sd_actives )); - assert( !FD_ISSET( s, &slap_daemon.sd_readers )); - assert( !FD_ISSET( s, &slap_daemon.sd_writers )); - -#ifndef HAVE_WINSOCK - if (s >= slap_daemon.sd_nfds) { - slap_daemon.sd_nfds = s + 1; - } -#endif + assert( SLAP_SOCK_NOT_ACTIVE(s) ); if ( isactive ) { slap_daemon.sd_nactives++; } - FD_SET( s, &slap_daemon.sd_actives ); - FD_SET( s, &slap_daemon.sd_readers ); + SLAP_ADD_SOCK(s); - Debug( LDAP_DEBUG_CONNS, "daemon: added %ld%s%s\n", - (long) s, - FD_ISSET(s, &slap_daemon.sd_readers) ? "r" : "", - FD_ISSET(s, &slap_daemon.sd_writers) ? "w" : "" ); + Debug( LDAP_DEBUG_CONNS, "daemon: added %ldr\n", + (long) s, 0, 0 ); ldap_pvt_thread_mutex_unlock( &slap_daemon.sd_mutex ); } @@ -251,16 +416,15 @@ void slapd_remove(ber_socket_t s, int wasactive, int wake) { if ( wasactive ) { slap_daemon.sd_nactives--; } - waswriter = FD_ISSET(s, &slap_daemon.sd_writers); + waswriter = SLAP_SOCK_IS_WRITE(s); Debug( LDAP_DEBUG_CONNS, "daemon: removing %ld%s%s\n", (long) s, - FD_ISSET(s, &slap_daemon.sd_readers) ? "r" : "", + SLAP_SOCK_IS_READ(s) ? "r" : "", waswriter ? "w" : "" ); if ( waswriter ) slap_daemon.sd_nwriters--; - FD_CLR( s, &slap_daemon.sd_actives ); - FD_CLR( s, &slap_daemon.sd_readers ); - FD_CLR( s, &slap_daemon.sd_writers ); + + SLAP_DEL_SOCK(s); /* If we ran out of file descriptors, we dropped a listener from * the select() loop. Now that we're removing a session from our @@ -291,9 +455,9 @@ void slapd_remove(ber_socket_t s, int wasactive, int wake) { void slapd_clr_write(ber_socket_t s, int wake) { ldap_pvt_thread_mutex_lock( &slap_daemon.sd_mutex ); - assert( FD_ISSET( s, &slap_daemon.sd_actives) ); - if ( FD_ISSET( s, &slap_daemon.sd_writers )) { - FD_CLR( s, &slap_daemon.sd_writers ); + assert( SLAP_SOCK_IS_ACTIVE( s )); + if ( SLAP_SOCK_IS_WRITE( s )) { + SLAP_SOCK_CLR_WRITE( s ); slap_daemon.sd_nwriters--; } @@ -304,9 +468,9 @@ void slapd_clr_write(ber_socket_t s, int wake) { void slapd_set_write(ber_socket_t s, int wake) { ldap_pvt_thread_mutex_lock( &slap_daemon.sd_mutex ); - assert( FD_ISSET( s, &slap_daemon.sd_actives) ); - if (!FD_ISSET(s, &slap_daemon.sd_writers)) { - FD_SET( (unsigned) s, &slap_daemon.sd_writers ); + assert( SLAP_SOCK_IS_ACTIVE( s )); + if ( !SLAP_SOCK_IS_WRITE( s )) { + SLAP_SOCK_SET_WRITE( s ); slap_daemon.sd_nwriters++; } @@ -317,8 +481,8 @@ void slapd_set_write(ber_socket_t s, int wake) { void slapd_clr_read(ber_socket_t s, int wake) { ldap_pvt_thread_mutex_lock( &slap_daemon.sd_mutex ); - assert( FD_ISSET( s, &slap_daemon.sd_actives) ); - FD_CLR( s, &slap_daemon.sd_readers ); + assert( SLAP_SOCK_IS_ACTIVE( s )); + SLAP_SOCK_CLR_READ( s ); ldap_pvt_thread_mutex_unlock( &slap_daemon.sd_mutex ); WAKE_LISTENER(wake); @@ -327,9 +491,9 @@ void slapd_clr_read(ber_socket_t s, int wake) { void slapd_set_read(ber_socket_t s, int wake) { ldap_pvt_thread_mutex_lock( &slap_daemon.sd_mutex ); - assert( FD_ISSET( s, &slap_daemon.sd_actives) ); - if (!FD_ISSET(s, &slap_daemon.sd_readers)) - FD_SET( s, &slap_daemon.sd_readers ); + assert( SLAP_SOCK_IS_ACTIVE( s )); + if (!SLAP_SOCK_IS_READ( s )) + SLAP_SOCK_SET_READ( s ); ldap_pvt_thread_mutex_unlock( &slap_daemon.sd_mutex ); WAKE_LISTENER(wake); @@ -906,12 +1070,6 @@ int slapd_daemon_init( const char *urls ) dtblsize = FD_SETSIZE; #endif -#ifdef FD_SETSIZE - if(dtblsize > FD_SETSIZE) { - dtblsize = FD_SETSIZE; - } -#endif /* !FD_SETSIZE */ - /* open a pipe (or something equivalent connected to itself). * we write a byte on this fd whenever we catch a signal. The main * loop will be select'ing on this socket, and will wake up when @@ -923,8 +1081,7 @@ int slapd_daemon_init( const char *urls ) return rc; } - FD_ZERO( &slap_daemon.sd_readers ); - FD_ZERO( &slap_daemon.sd_writers ); + SLAP_SOCK_SET_INIT; if( urls == NULL ) { urls = "ldap:///"; @@ -1024,6 +1181,272 @@ close_listeners( } } +static int +slapd_handle_listener( + Listener *sl +) +{ + Sockaddr from; + + ber_socket_t s; + socklen_t len = sizeof(from); + long id; + slap_ssf_t ssf = 0; + struct berval authid = BER_BVNULL; +#ifdef SLAPD_RLOOKUPS + char hbuf[NI_MAXHOST]; +#endif + + char *dnsname = NULL; + char *peeraddr = NULL; +#ifdef LDAP_PF_LOCAL + char peername[MAXPATHLEN + sizeof("PATH=")]; +#elif defined(LDAP_PF_INET6) + char peername[sizeof("IP=ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff 65535")]; +#else + char peername[sizeof("IP=255.255.255.255:65336")]; +#endif /* LDAP_PF_LOCAL */ + + peername[0] = '\0'; + +#ifdef LDAP_CONNECTIONLESS + if ( sl->sl_is_udp ) { + /* The first time we receive a query, we set this + * up as a "connection". It remains open for the life + * of the slapd. + */ + if ( sl->sl_is_udp < 2 ) { + id = connection_init( sl->sl_sd, sl, "", "", + CONN_IS_UDP, ssf, NULL ); + sl->sl_is_udp++; + } + return 1; + } +#endif + +# ifdef LDAP_PF_LOCAL + /* FIXME: apparently accept doesn't fill + * the sun_path sun_path member */ + from.sa_un_addr.sun_path[0] = '\0'; +# endif /* LDAP_PF_LOCAL */ + + s = accept( sl->sl_sd, (struct sockaddr *) &from, &len ); + if ( s == AC_SOCKET_INVALID ) { + int err = sock_errno(); + + if( +#ifdef EMFILE + err == EMFILE || +#endif +#ifdef ENFILE + err == ENFILE || +#endif + 0 ) + { + ldap_pvt_thread_mutex_lock( &slap_daemon.sd_mutex ); + emfile++; + /* Stop listening until an existing session closes */ + sl->sl_is_mute = 1; + ldap_pvt_thread_mutex_unlock( &slap_daemon.sd_mutex ); + } + + Debug( LDAP_DEBUG_ANY, + "daemon: accept(%ld) failed errno=%d (%s)\n", + (long) sl->sl_sd, err, + sock_errstr(err) ); + ldap_pvt_thread_yield(); + return 0; + } + +#ifndef HAVE_WINSOCK + /* make sure descriptor number isn't too great */ + if ( s >= dtblsize ) { + Debug( LDAP_DEBUG_ANY, + "daemon: %ld beyond descriptor table size %ld\n", + (long) s, (long) dtblsize, 0 ); + + slapd_close(s); + ldap_pvt_thread_yield(); + return 0; + } +#endif + +#ifdef LDAP_DEBUG + ldap_pvt_thread_mutex_lock( &slap_daemon.sd_mutex ); + + /* newly accepted stream should not be in any of the FD SETS */ + assert( SLAP_SOCK_NOT_ACTIVE( s )); + + ldap_pvt_thread_mutex_unlock( &slap_daemon.sd_mutex ); +#endif + +#if defined( SO_KEEPALIVE ) || defined( TCP_NODELAY ) +#ifdef LDAP_PF_LOCAL + /* for IPv4 and IPv6 sockets only */ + if ( from.sa_addr.sa_family != AF_LOCAL ) +#endif /* LDAP_PF_LOCAL */ + { + int rc; + int tmp; +#ifdef SO_KEEPALIVE + /* enable keep alives */ + tmp = 1; + rc = setsockopt( s, SOL_SOCKET, SO_KEEPALIVE, + (char *) &tmp, sizeof(tmp) ); + if ( rc == AC_SOCKET_ERROR ) { + int err = sock_errno(); + Debug( LDAP_DEBUG_ANY, + "slapd(%ld): setsockopt(SO_KEEPALIVE) failed " + "errno=%d (%s)\n", (long) s, err, sock_errstr(err) ); + } +#endif +#ifdef TCP_NODELAY + /* enable no delay */ + tmp = 1; + rc = setsockopt( s, IPPROTO_TCP, TCP_NODELAY, + (char *)&tmp, sizeof(tmp) ); + if ( rc == AC_SOCKET_ERROR ) { + int err = sock_errno(); + Debug( LDAP_DEBUG_ANY, + "slapd(%ld): setsockopt(TCP_NODELAY) failed " + "errno=%d (%s)\n", (long) s, err, sock_errstr(err) ); + } +#endif + } +#endif + + Debug( LDAP_DEBUG_CONNS, "daemon: new connection on %ld\n", + (long) s, 0, 0 ); + switch ( from.sa_addr.sa_family ) { +# ifdef LDAP_PF_LOCAL + case AF_LOCAL: + /* FIXME: apparently accept doesn't fill + * the sun_path sun_path member */ + if ( from.sa_un_addr.sun_path[0] == '\0' ) { + AC_MEMCPY( from.sa_un_addr.sun_path, + sl->sl_sa.sa_un_addr.sun_path, + sizeof( from.sa_un_addr.sun_path ) ); + } + + sprintf( peername, "PATH=%s", from.sa_un_addr.sun_path ); + ssf = local_ssf; + { + uid_t uid; + gid_t gid; + + if( getpeereid( s, &uid, &gid ) == 0 ) { + authid.bv_val = ch_malloc( + sizeof("uidnumber=4294967295+gidnumber=4294967295," + "cn=peercred,cn=external,cn=auth")); + authid.bv_len = sprintf( authid.bv_val, + "uidnumber=%d+gidnumber=%d," + "cn=peercred,cn=external,cn=auth", + (int) uid, (int) gid); + } + } + dnsname = "local"; + break; +#endif /* LDAP_PF_LOCAL */ + +# ifdef LDAP_PF_INET6 + case AF_INET6: + if ( IN6_IS_ADDR_V4MAPPED(&from.sa_in6_addr.sin6_addr) ) { + peeraddr = inet_ntoa( *((struct in_addr *) + &from.sa_in6_addr.sin6_addr.s6_addr[12]) ); + sprintf( peername, "IP=%s:%d", + peeraddr != NULL ? peeraddr : SLAP_STRING_UNKNOWN, + (unsigned) ntohs( from.sa_in6_addr.sin6_port ) ); + } else { + char addr[INET6_ADDRSTRLEN]; + + peeraddr = (char *) inet_ntop( AF_INET6, + &from.sa_in6_addr.sin6_addr, + addr, sizeof addr ); + sprintf( peername, "IP=%s %d", + peeraddr != NULL ? peeraddr : SLAP_STRING_UNKNOWN, + (unsigned) ntohs( from.sa_in6_addr.sin6_port ) ); + } + break; +# endif /* LDAP_PF_INET6 */ + + case AF_INET: + peeraddr = inet_ntoa( from.sa_in_addr.sin_addr ); + sprintf( peername, "IP=%s:%d", + peeraddr != NULL ? peeraddr : SLAP_STRING_UNKNOWN, + (unsigned) ntohs( from.sa_in_addr.sin_port ) ); + break; + + default: + slapd_close(s); + return 0; + } + + if ( ( from.sa_addr.sa_family == AF_INET ) +#ifdef LDAP_PF_INET6 + || ( from.sa_addr.sa_family == AF_INET6 ) +#endif + ) { +#ifdef SLAPD_RLOOKUPS + if ( use_reverse_lookup ) { + char *herr; + if (ldap_pvt_get_hname( (const struct sockaddr *)&from, len, hbuf, + sizeof(hbuf), &herr ) == 0) { + ldap_pvt_str2lower( hbuf ); + dnsname = hbuf; + } + } +#else + dnsname = NULL; +#endif /* SLAPD_RLOOKUPS */ + +#ifdef HAVE_TCPD + if ( !hosts_ctl("slapd", + dnsname != NULL ? dnsname : SLAP_STRING_UNKNOWN, + peeraddr != NULL ? peeraddr : SLAP_STRING_UNKNOWN, + SLAP_STRING_UNKNOWN )) + { + /* DENY ACCESS */ + Statslog( LDAP_DEBUG_STATS, + "fd=%ld DENIED from %s (%s)\n", + (long) s, + dnsname != NULL ? dnsname : SLAP_STRING_UNKNOWN, + peeraddr != NULL ? peeraddr : SLAP_STRING_UNKNOWN, + 0, 0 ); + slapd_close(s); + return 0; + } +#endif /* HAVE_TCPD */ + } + + id = connection_init(s, sl, + dnsname != NULL ? dnsname : SLAP_STRING_UNKNOWN, + peername, +#ifdef HAVE_TLS + sl->sl_is_tls ? CONN_IS_TLS : 0, +#else + 0, +#endif + ssf, + authid.bv_val ? &authid : NULL ); + + if( authid.bv_val ) ch_free(authid.bv_val); + + if( id < 0 ) { + Debug( LDAP_DEBUG_ANY, + "daemon: connection_init(%ld, %s, %s) failed.\n", + (long) s, peername, sl->sl_name.bv_val ); + slapd_close(s); + return 0; + } + + Statslog( LDAP_DEBUG_STATS, + "conn=%ld fd=%ld ACCEPT from %s (%s)\n", + id, (long) s, peername, sl->sl_name.bv_val, + 0 ); + + slapd_add( s, 1 ); + return 0; +} static void * slapd_daemon_task( @@ -1050,6 +1473,8 @@ slapd_daemon_task( idle.tv_usec = 0; } + slapd_add( wake_sds[0], 0 ); + for ( l = 0; slap_listeners[l] != NULL; l++ ) { if ( slap_listeners[l]->sl_sd == AC_SOCKET_INVALID ) continue; @@ -1125,9 +1550,7 @@ slapd_daemon_task( time_t now; - fd_set readfds; - fd_set writefds; - Sockaddr from; + SLAP_EVENT_DECL; struct timeval tv; struct timeval *tvp; @@ -1171,27 +1594,20 @@ slapd_daemon_task( ldap_pvt_thread_mutex_lock( &slap_daemon.sd_mutex ); nwriters = slap_daemon.sd_nwriters; - AC_MEMCPY( &readfds, &slap_daemon.sd_readers, sizeof(fd_set) ); - if ( nwriters ) - AC_MEMCPY( &writefds, &slap_daemon.sd_writers, sizeof(fd_set) ); - assert(!FD_ISSET(wake_sds[0], &readfds)); - FD_SET( wake_sds[0], &readfds ); + SLAP_EVENT_INIT; for ( l = 0; slap_listeners[l] != NULL; l++ ) { if ( slap_listeners[l]->sl_sd == AC_SOCKET_INVALID ) continue; if ( slap_listeners[l]->sl_is_mute ) - FD_CLR( slap_listeners[l]->sl_sd, &readfds ); + SLAP_SOCK_SET_MUTE( slap_listeners[l]->sl_sd ); else - if (!FD_ISSET(slap_listeners[l]->sl_sd, &readfds)) - FD_SET( slap_listeners[l]->sl_sd, &readfds ); + if ( SLAP_SOCK_IS_MUTE( slap_listeners[l]->sl_sd )) + SLAP_SOCK_CLR_MUTE( slap_listeners[l]->sl_sd ); } -#ifndef HAVE_WINSOCK - nfds = slap_daemon.sd_nfds; -#else - nfds = dtblsize; -#endif + nfds = SLAP_MAXWAIT; + if ( global_idletimeout && slap_daemon.sd_nactives ) at = 1; @@ -1245,8 +1661,7 @@ slapd_daemon_task( tvp == NULL ? "NULL" : "zero" ); } - switch(ns = select( nfds, &readfds, nwriters > 0 ? &writefds : NULL, - NULL, tvp )) + switch(ns = SLAP_EVENT_WAIT(tvp)) { case -1: { /* failure - try again */ int err = sock_errno(); @@ -1287,7 +1702,11 @@ slapd_daemon_task( /* FALL THRU */ } - if( FD_ISSET( wake_sds[0], &readfds ) ) { + /* We don't need to examine the event status for wake_sds; + * if waking is set and we woke up, we'll read whatever + * we can. + */ + if ( waking ) { char c[BUFSIZ]; tcp_read( wake_sds[0], c, sizeof(c) ); waking = 0; @@ -1295,290 +1714,33 @@ slapd_daemon_task( continue; } +#if SLAP_EVENTS_ARE_INDEXED + /* The event slot equals the descriptor number - this is + * true for Unix select and poll. We treat Windows select + * like this too, even though it's a kludge. + */ for ( l = 0; slap_listeners[l] != NULL; l++ ) { - ber_socket_t s; - socklen_t len = sizeof(from); - long id; - slap_ssf_t ssf = 0; - struct berval authid = BER_BVNULL; -#ifdef SLAPD_RLOOKUPS - char hbuf[NI_MAXHOST]; -#endif - - char *dnsname = NULL; - char *peeraddr = NULL; -#ifdef LDAP_PF_LOCAL - char peername[MAXPATHLEN + sizeof("PATH=")]; -#elif defined(LDAP_PF_INET6) - char peername[sizeof( - "IP=ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff 65535")]; -#else - char peername[sizeof("IP=255.255.255.255:65336")]; -#endif /* LDAP_PF_LOCAL */ - - peername[0] = '\0'; + int rc; if ( ns <= 0 ) break; if ( slap_listeners[l]->sl_sd == AC_SOCKET_INVALID ) continue; - if ( !FD_ISSET( slap_listeners[l]->sl_sd, &readfds ) ) + if ( !SLAP_EVENT_IS_READ( slap_listeners[l]->sl_sd )) continue; ns--; + rc = slapd_handle_listener(slap_listeners[l]); + #ifdef LDAP_CONNECTIONLESS - if ( slap_listeners[l]->sl_is_udp ) { - /* The first time we receive a query, we set this - * up as a "connection". It remains open for the life - * of the slapd. - */ - if ( slap_listeners[l]->sl_is_udp < 2 ) { - id = connection_init( - slap_listeners[l]->sl_sd, - slap_listeners[l], "", "", - CONN_IS_UDP, ssf, NULL ); - slap_listeners[l]->sl_is_udp++; - } - continue; - } + if ( rc ) continue; #endif /* Don't need to look at this in the data loops */ - FD_CLR( slap_listeners[l]->sl_sd, &readfds ); - FD_CLR( slap_listeners[l]->sl_sd, &writefds ); - -# ifdef LDAP_PF_LOCAL - /* FIXME: apparently accept doesn't fill - * the sun_path sun_path member */ - from.sa_un_addr.sun_path[0] = '\0'; -# endif /* LDAP_PF_LOCAL */ - s = accept( slap_listeners[l]->sl_sd, - (struct sockaddr *) &from, &len ); - if ( s == AC_SOCKET_INVALID ) { - int err = sock_errno(); - - if( -#ifdef EMFILE - err == EMFILE || -#endif -#ifdef ENFILE - err == ENFILE || -#endif - 0 ) - { - ldap_pvt_thread_mutex_lock( &slap_daemon.sd_mutex ); - emfile++; - /* Stop listening until an existing session closes */ - slap_listeners[l]->sl_is_mute = 1; - ldap_pvt_thread_mutex_unlock( &slap_daemon.sd_mutex ); - } - - Debug( LDAP_DEBUG_ANY, - "daemon: accept(%ld) failed errno=%d (%s)\n", - (long) slap_listeners[l]->sl_sd, err, - sock_errstr(err) ); - ldap_pvt_thread_yield(); - continue; - } - -#ifndef HAVE_WINSOCK - /* make sure descriptor number isn't too great */ - if ( s >= dtblsize ) { - Debug( LDAP_DEBUG_ANY, - "daemon: %ld beyond descriptor table size %ld\n", - (long) s, (long) dtblsize, 0 ); - - slapd_close(s); - ldap_pvt_thread_yield(); - continue; - } -#endif - -#ifdef LDAP_DEBUG - ldap_pvt_thread_mutex_lock( &slap_daemon.sd_mutex ); - - /* newly accepted stream should not be in any of the FD SETS */ - assert( !FD_ISSET( s, &slap_daemon.sd_actives) ); - assert( !FD_ISSET( s, &slap_daemon.sd_readers) ); - assert( !FD_ISSET( s, &slap_daemon.sd_writers) ); - - ldap_pvt_thread_mutex_unlock( &slap_daemon.sd_mutex ); -#endif - -#if defined( SO_KEEPALIVE ) || defined( TCP_NODELAY ) -#ifdef LDAP_PF_LOCAL - /* for IPv4 and IPv6 sockets only */ - if ( from.sa_addr.sa_family != AF_LOCAL ) -#endif /* LDAP_PF_LOCAL */ - { - int rc; - int tmp; -#ifdef SO_KEEPALIVE - /* enable keep alives */ - tmp = 1; - rc = setsockopt( s, SOL_SOCKET, SO_KEEPALIVE, - (char *) &tmp, sizeof(tmp) ); - if ( rc == AC_SOCKET_ERROR ) { - int err = sock_errno(); - Debug( LDAP_DEBUG_ANY, - "slapd(%ld): setsockopt(SO_KEEPALIVE) failed " - "errno=%d (%s)\n", (long) s, err, sock_errstr(err) ); - } -#endif -#ifdef TCP_NODELAY - /* enable no delay */ - tmp = 1; - rc = setsockopt( s, IPPROTO_TCP, TCP_NODELAY, - (char *)&tmp, sizeof(tmp) ); - if ( rc == AC_SOCKET_ERROR ) { - int err = sock_errno(); - Debug( LDAP_DEBUG_ANY, - "slapd(%ld): setsockopt(TCP_NODELAY) failed " - "errno=%d (%s)\n", (long) s, err, sock_errstr(err) ); - } -#endif - } -#endif - - Debug( LDAP_DEBUG_CONNS, "daemon: new connection on %ld\n", - (long) s, 0, 0 ); - switch ( from.sa_addr.sa_family ) { -# ifdef LDAP_PF_LOCAL - case AF_LOCAL: - /* FIXME: apparently accept doesn't fill - * the sun_path sun_path member */ - if ( from.sa_un_addr.sun_path[0] == '\0' ) { - AC_MEMCPY( from.sa_un_addr.sun_path, - slap_listeners[l]->sl_sa.sa_un_addr.sun_path, - sizeof( from.sa_un_addr.sun_path ) ); - } - - sprintf( peername, "PATH=%s", from.sa_un_addr.sun_path ); - ssf = local_ssf; - { - uid_t uid; - gid_t gid; - - if( getpeereid( s, &uid, &gid ) == 0 ) { - authid.bv_val = ch_malloc( - sizeof("uidnumber=4294967295+gidnumber=4294967295," - "cn=peercred,cn=external,cn=auth")); - authid.bv_len = sprintf( authid.bv_val, - "uidnumber=%d+gidnumber=%d," - "cn=peercred,cn=external,cn=auth", - (int) uid, (int) gid); - } - } - dnsname = "local"; - break; -#endif /* LDAP_PF_LOCAL */ - -# ifdef LDAP_PF_INET6 - case AF_INET6: - if ( IN6_IS_ADDR_V4MAPPED(&from.sa_in6_addr.sin6_addr) ) { - peeraddr = inet_ntoa( *((struct in_addr *) - &from.sa_in6_addr.sin6_addr.s6_addr[12]) ); - sprintf( peername, "IP=%s:%d", - peeraddr != NULL ? peeraddr : SLAP_STRING_UNKNOWN, - (unsigned) ntohs( from.sa_in6_addr.sin6_port ) ); - } else { - char addr[INET6_ADDRSTRLEN]; - - peeraddr = (char *) inet_ntop( AF_INET6, - &from.sa_in6_addr.sin6_addr, - addr, sizeof addr ); - sprintf( peername, "IP=%s %d", - peeraddr != NULL ? peeraddr : SLAP_STRING_UNKNOWN, - (unsigned) ntohs( from.sa_in6_addr.sin6_port ) ); - } - break; -# endif /* LDAP_PF_INET6 */ - - case AF_INET: - peeraddr = inet_ntoa( from.sa_in_addr.sin_addr ); - sprintf( peername, "IP=%s:%d", - peeraddr != NULL ? peeraddr : SLAP_STRING_UNKNOWN, - (unsigned) ntohs( from.sa_in_addr.sin_port ) ); - break; - - default: - slapd_close(s); - continue; - } - - if ( ( from.sa_addr.sa_family == AF_INET ) -#ifdef LDAP_PF_INET6 - || ( from.sa_addr.sa_family == AF_INET6 ) -#endif - ) { -#ifdef SLAPD_RLOOKUPS - if ( use_reverse_lookup ) { - char *herr; - if (ldap_pvt_get_hname( (const struct sockaddr *)&from, len, hbuf, - sizeof(hbuf), &herr ) == 0) { - ldap_pvt_str2lower( hbuf ); - dnsname = hbuf; - } - } -#else - dnsname = NULL; -#endif /* SLAPD_RLOOKUPS */ - -#ifdef HAVE_TCPD - if ( !hosts_ctl("slapd", - dnsname != NULL ? dnsname : SLAP_STRING_UNKNOWN, - peeraddr != NULL ? peeraddr : SLAP_STRING_UNKNOWN, - SLAP_STRING_UNKNOWN )) - { - /* DENY ACCESS */ - Statslog( LDAP_DEBUG_STATS, - "fd=%ld DENIED from %s (%s)\n", - (long) s, - dnsname != NULL ? dnsname : SLAP_STRING_UNKNOWN, - peeraddr != NULL ? peeraddr : SLAP_STRING_UNKNOWN, - 0, 0 ); - slapd_close(s); - continue; - } -#endif /* HAVE_TCPD */ - } - - id = connection_init(s, - slap_listeners[l], - dnsname != NULL ? dnsname : SLAP_STRING_UNKNOWN, - peername, -#ifdef HAVE_TLS - slap_listeners[l]->sl_is_tls ? CONN_IS_TLS : 0, -#else - 0, -#endif - ssf, - authid.bv_val ? &authid : NULL ); - - if( authid.bv_val ) ch_free(authid.bv_val); - - if( id < 0 ) { - Debug( LDAP_DEBUG_ANY, - "daemon: connection_init(%ld, %s, %s) " - "failed.\n", - (long) s, - peername, - slap_listeners[l]->sl_name.bv_val ); - slapd_close(s); - continue; - } - - Statslog( LDAP_DEBUG_STATS, - "conn=%ld fd=%ld ACCEPT from %s (%s)\n", - id, (long) s, - peername, - slap_listeners[l]->sl_name.bv_val, - 0 ); - - slapd_add( s, 1 ); - continue; + SLAP_EVENT_CLR_READ( slap_listeners[l]->sl_sd ); + SLAP_EVENT_CLR_WRITE( slap_listeners[l]->sl_sd ); } /* bypass the following tests if no descriptors left */ @@ -1606,8 +1768,8 @@ slapd_daemon_task( for ( i = 0; i < nfds; i++ ) { int r, w; - r = FD_ISSET( i, &readfds ); - w = FD_ISSET( i, &writefds ); + r = SLAP_EVENT_IS_READ( i ); + w = SLAP_EVENT_IS_WRITE( i ); if ( r || w ) { Debug( LDAP_DEBUG_CONNS, " %d%s%s", i, r ? "r" : "", w ? "w" : "" ); @@ -1633,7 +1795,7 @@ slapd_daemon_task( #ifdef HAVE_WINSOCK wd = writefds.fd_array[i]; #else - if( ! FD_ISSET( i, &writefds ) ) { + if( ! SLAP_EVENT_IS_WRITE( i ) ) { continue; } wd = i; @@ -1651,8 +1813,8 @@ slapd_daemon_task( */ if ( connection_write( wd ) < 0 ) { - if ( FD_ISSET( wd, &readfds )) { - FD_CLR( (unsigned) wd, &readfds ); + if ( SLAP_EVENT_IS_READ( wd )) { + SLAP_EVENT_CLR_READ( (unsigned) wd ); nrfds--; } slapd_close( wd ); @@ -1665,7 +1827,7 @@ slapd_daemon_task( #ifdef HAVE_WINSOCK rd = readfds.fd_array[i]; #else - if( ! FD_ISSET( i, &readfds ) ) { + if( ! SLAP_EVENT_IS_READ( i ) ) { continue; } rd = i; @@ -1685,6 +1847,21 @@ slapd_daemon_task( slapd_close( rd ); } } +#else /* !SLAP_EVENTS_ARE_INDEXED */ + /* FIXME */ + /* The events are returned in an arbitrary list. This is true + * for /dev/poll, epoll and kqueue. In order to prioritize things + * so that we can handle wake_sds first, listeners second, and then + * all other connections last (as we do for select), we would need + * to use multiple event handles and cascade them. + * + * That seems like a bit of hassle. So the wake_sds check has moved + * above. For epoll and kqueue we can associate arbitrary data with + * an event, so we could use pointers to the listener structure + * instead of just the file descriptor. For /dev/poll we have to + * search the listeners array for a matching descriptor. + */ +#endif ldap_pvt_thread_yield(); } -- 2.39.5