X-Git-Url: https://git.saurik.com/apple/xnu.git/blobdiff_plain/2d21ac55c334faf3a56e5634905ed6987fc787d4..7e41aa883dd258f888d0470250eead40a53ef1f5:/bsd/kern/uipc_socket2.c diff --git a/bsd/kern/uipc_socket2.c b/bsd/kern/uipc_socket2.c index 379b9afd6..40e7f1919 100644 --- a/bsd/kern/uipc_socket2.c +++ b/bsd/kern/uipc_socket2.c @@ -1,8 +1,8 @@ /* - * Copyright (c) 1998-2007 Apple Inc. All rights reserved. + * Copyright (c) 1998-2015 Apple Inc. All rights reserved. * * @APPLE_OSREFERENCE_LICENSE_HEADER_START@ - * + * * This file contains Original Code and/or Modifications of Original Code * as defined in and that are subject to the Apple Public Source License * Version 2.0 (the 'License'). You may not use this file except in @@ -11,10 +11,10 @@ * unlawful or unlicensed copies of an Apple operating system, or to * circumvent, violate, or enable the circumvention or violation of, any * terms of an Apple operating system software license agreement. - * + * * Please obtain a copy of the License at * http://www.opensource.apple.com/apsl/ and read it before using this file. - * + * * The Original Code and all software distributed under the License are * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES, @@ -22,7 +22,7 @@ * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT. * Please see the License for the specific language governing rights and * limitations under the License. - * + * * @APPLE_OSREFERENCE_LICENSE_HEADER_END@ */ /* Copyright (c) 1995 NeXT Computer, Inc. All Rights Reserved */ @@ -59,7 +59,6 @@ * SUCH DAMAGE. * * @(#)uipc_socket2.c 8.1 (Berkeley) 6/10/93 - * $FreeBSD: src/sys/kern/uipc_socket2.c,v 1.55.2.9 2001/07/26 18:53:02 peter Exp $ */ /* * NOTICE: This file was modified by SPARTA, Inc. in 2005 to introduce @@ -76,15 +75,18 @@ #include #include #include +#include #include #include #include #include #include #include +#include #include #include #include +#include #include #include #include @@ -94,11 +96,20 @@ #include #endif -/* TODO: this should be in a header file somewhere */ -extern void postevent(struct socket *, struct sockbuf *, int); +#include + +#if MPTCP +#include +#endif + +#define DBG_FNC_SBDROP NETDBG_CODE(DBG_NETSOCK, 4) +#define DBG_FNC_SBAPPEND NETDBG_CODE(DBG_NETSOCK, 5) + +SYSCTL_DECL(_kern_ipc); -#define DBG_FNC_SBDROP NETDBG_CODE(DBG_NETSOCK, 4) -#define DBG_FNC_SBAPPEND NETDBG_CODE(DBG_NETSOCK, 5) +__private_extern__ u_int32_t net_io_policy_throttle_best_effort = 0; +SYSCTL_INT(_kern_ipc, OID_AUTO, throttle_best_effort, + CTLFLAG_RW | CTLFLAG_LOCKED, &net_io_policy_throttle_best_effort, 0, ""); static inline void sbcompress(struct sockbuf *, struct mbuf *, struct mbuf *); static struct socket *sonewconn_internal(struct socket *, int); @@ -106,6 +117,7 @@ static int sbappendaddr_internal(struct sockbuf *, struct sockaddr *, struct mbuf *, struct mbuf *); static int sbappendcontrol_internal(struct sockbuf *, struct mbuf *, struct mbuf *); +static void soevent_ifdenied(struct socket *); /* * Primitive routines for operating on sockets and socket buffers @@ -113,12 +125,29 @@ static int sbappendcontrol_internal(struct sockbuf *, struct mbuf *, static int soqlimitcompat = 1; static int soqlencomp = 0; -u_long sb_max = SB_MAX; /* XXX should be static */ +/* + * Based on the number of mbuf clusters configured, high_sb_max and sb_max can + * get scaled up or down to suit that memory configuration. high_sb_max is a + * higher limit on sb_max that is checked when sb_max gets set through sysctl. + */ + +u_int32_t sb_max = SB_MAX; /* XXX should be static */ +u_int32_t high_sb_max = SB_MAX; + +static u_int32_t sb_efficiency = 8; /* parameter for sbreserve() */ +int32_t total_sbmb_cnt __attribute__((aligned(8))) = 0; +int32_t total_sbmb_cnt_peak __attribute__((aligned(8))) = 0; +int32_t total_snd_byte_count __attribute__((aligned(8))) = 0; +int64_t sbmb_limreached __attribute__((aligned(8))) = 0; -static u_long sb_efficiency = 8; /* parameter for sbreserve() */ -__private_extern__ unsigned int total_mb_cnt = 0; -__private_extern__ unsigned int total_cl_cnt = 0; -__private_extern__ int sbspace_factor = 8; +/* Control whether to throttle sockets eligible to be throttled */ +__private_extern__ u_int32_t net_io_policy_throttled = 0; +static int sysctl_io_policy_throttled SYSCTL_HANDLER_ARGS; + +u_int32_t net_io_policy_log = 0; /* log socket policy changes */ +#if CONFIG_PROC_UUID_POLICY +u_int32_t net_io_policy_uuid = 1; /* enable UUID socket policy */ +#endif /* CONFIG_PROC_UUID_POLICY */ /* * Procedures to manipulate state flags of socket @@ -167,6 +196,8 @@ soisconnected(struct socket *so) so->so_state &= ~(SS_ISCONNECTING|SS_ISDISCONNECTING|SS_ISCONFIRMING); so->so_state |= SS_ISCONNECTED; + soreserve_preconnect(so, 0); + sflt_notify(so, sock_evt_connected, NULL); if (head && (so->so_state & SS_INCOMP)) { @@ -191,14 +222,26 @@ soisconnected(struct socket *so) wakeup((caddr_t)&so->so_timeo); sorwakeup(so); sowwakeup(so); + soevent(so, SO_FILT_HINT_LOCKED | SO_FILT_HINT_CONNECTED | + SO_FILT_HINT_CONNINFO_UPDATED); } } +boolean_t +socanwrite(struct socket *so) +{ + return ((so->so_state & SS_ISCONNECTED) || + !(so->so_proto->pr_flags & PR_CONNREQUIRED) || + (so->so_flags1 & SOF1_PRECONNECT_DATA)); + +} + void soisdisconnecting(struct socket *so) { so->so_state &= ~SS_ISCONNECTING; so->so_state |= (SS_ISDISCONNECTING|SS_CANTRCVMORE|SS_CANTSENDMORE); + soevent(so, SO_FILT_HINT_LOCKED); sflt_notify(so, sock_evt_disconnecting, NULL); wakeup((caddr_t)&so->so_timeo); sowwakeup(so); @@ -210,10 +253,39 @@ soisdisconnected(struct socket *so) { so->so_state &= ~(SS_ISCONNECTING|SS_ISCONNECTED|SS_ISDISCONNECTING); so->so_state |= (SS_CANTRCVMORE|SS_CANTSENDMORE|SS_ISDISCONNECTED); + soevent(so, SO_FILT_HINT_LOCKED | SO_FILT_HINT_DISCONNECTED | + SO_FILT_HINT_CONNINFO_UPDATED); sflt_notify(so, sock_evt_disconnected, NULL); wakeup((caddr_t)&so->so_timeo); sowwakeup(so); sorwakeup(so); + +#if CONTENT_FILTER + /* Notify content filters as soon as we cannot send/receive data */ + cfil_sock_notify_shutdown(so, SHUT_RDWR); +#endif /* CONTENT_FILTER */ +} + +/* + * This function will issue a wakeup like soisdisconnected but it will not + * notify the socket filters. This will avoid unlocking the socket + * in the midst of closing it. + */ +void +sodisconnectwakeup(struct socket *so) +{ + so->so_state &= ~(SS_ISCONNECTING|SS_ISCONNECTED|SS_ISDISCONNECTING); + so->so_state |= (SS_CANTRCVMORE|SS_CANTSENDMORE|SS_ISDISCONNECTED); + soevent(so, SO_FILT_HINT_LOCKED | SO_FILT_HINT_DISCONNECTED | + SO_FILT_HINT_CONNINFO_UPDATED); + wakeup((caddr_t)&so->so_timeo); + sowwakeup(so); + sorwakeup(so); + +#if CONTENT_FILTER + /* Notify content filters as soon as we cannot send/receive data */ + cfil_sock_notify_shutdown(so, SHUT_RDWR); +#endif /* CONTENT_FILTER */ } /* @@ -260,8 +332,7 @@ sonewconn_internal(struct socket *head, int connstatus) if (so_qlen >= (soqlimitcompat ? head->so_qlimit : (3 * head->so_qlimit / 2))) return ((struct socket *)0); - so = soalloc(M_NOWAIT, head->so_proto->pr_domain->dom_family, - head->so_type); + so = soalloc(1, SOCK_DOM(head), head->so_type); if (so == NULL) return ((struct socket *)0); /* check if head was closed during the soalloc */ @@ -270,7 +341,6 @@ sonewconn_internal(struct socket *head, int connstatus) return ((struct socket *)0); } - so->so_head = head; so->so_type = head->so_type; so->so_options = head->so_options &~ SO_ACCEPTCONN; so->so_linger = head->so_linger; @@ -278,27 +348,46 @@ sonewconn_internal(struct socket *head, int connstatus) so->so_proto = head->so_proto; so->so_timeo = head->so_timeo; so->so_pgid = head->so_pgid; - so->so_uid = head->so_uid; - so->so_flags = head->so_flags & (SOF_REUSESHAREUID|SOF_NOTIFYCONFLICT); /* inherit SO_REUSESHAREUID and SO_NOTIFYCONFLICT ocket options */ + kauth_cred_ref(head->so_cred); + so->so_cred = head->so_cred; + so->last_pid = head->last_pid; + so->last_upid = head->last_upid; + memcpy(so->last_uuid, head->last_uuid, sizeof (so->last_uuid)); + if (head->so_flags & SOF_DELEGATED) { + so->e_pid = head->e_pid; + so->e_upid = head->e_upid; + memcpy(so->e_uuid, head->e_uuid, sizeof (so->e_uuid)); + } + /* inherit socket options stored in so_flags */ + so->so_flags = head->so_flags & + (SOF_NOSIGPIPE | SOF_NOADDRAVAIL | SOF_REUSESHAREUID | + SOF_NOTIFYCONFLICT | SOF_BINDRANDOMPORT | SOF_NPX_SETOPTSHUT | + SOF_NODEFUNCT | SOF_PRIVILEGED_TRAFFIC_CLASS| SOF_NOTSENT_LOWAT | + SOF_USELRO | SOF_DELEGATED); so->so_usecount = 1; so->next_lock_lr = 0; so->next_unlock_lr = 0; -#ifdef __APPLE__ so->so_rcv.sb_flags |= SB_RECV; /* XXX */ so->so_rcv.sb_so = so->so_snd.sb_so = so; TAILQ_INIT(&so->so_evlist); -#endif #if CONFIG_MACF_SOCKET mac_socket_label_associate_accept(head, so); #endif + /* inherit traffic management properties of listener */ + so->so_traffic_mgt_flags = + head->so_traffic_mgt_flags & (TRAFFIC_MGT_SO_BACKGROUND); + so->so_background_thread = head->so_background_thread; + so->so_traffic_class = head->so_traffic_class; + if (soreserve(so, head->so_snd.sb_hiwat, head->so_rcv.sb_hiwat)) { - sflt_termsock(so); sodealloc(so); return ((struct socket *)0); } + so->so_rcv.sb_flags |= (head->so_rcv.sb_flags & SB_USRSIZE); + so->so_snd.sb_flags |= (head->so_snd.sb_flags & SB_USRSIZE); /* * Must be done with head unlocked to avoid deadlock @@ -308,17 +397,37 @@ sonewconn_internal(struct socket *head, int connstatus) socket_unlock(head, 0); if (((*so->so_proto->pr_usrreqs->pru_attach)(so, 0, NULL) != 0) || error) { - sflt_termsock(so); sodealloc(so); if (head->so_proto->pr_unlock) socket_lock(head, 0); return ((struct socket *)0); } - if (head->so_proto->pr_unlock) + if (head->so_proto->pr_unlock) { socket_lock(head, 0); -#ifdef __APPLE__ - so->so_proto->pr_domain->dom_refs++; -#endif + /* + * Radar 7385998 Recheck that the head is still accepting + * to avoid race condition when head is getting closed. + */ + if ((head->so_options & SO_ACCEPTCONN) == 0) { + so->so_state &= ~SS_NOFDREF; + soclose(so); + return ((struct socket *)0); + } + } + + atomic_add_32(&so->so_proto->pr_domain->dom_refs, 1); + + /* Insert in head appropriate lists */ + so->so_head = head; + + /* + * Since this socket is going to be inserted into the incomp + * queue, it can be picked up by another thread in + * tcp_dropdropablreq to get dropped before it is setup.. + * To prevent this race, set in-progress flag which can be + * cleared later + */ + so->so_flags |= SOF_INCOMP_INPROGRESS; if (connstatus) { TAILQ_INSERT_TAIL(&head->so_comp, so, so_list); @@ -330,10 +439,8 @@ sonewconn_internal(struct socket *head, int connstatus) } head->so_qlen++; -#ifdef __APPLE__ /* Attach socket filters for this protocol */ sflt_initsock(so); -#endif if (connstatus) { so->so_state |= connstatus; @@ -347,27 +454,7 @@ sonewconn_internal(struct socket *head, int connstatus) struct socket * sonewconn(struct socket *head, int connstatus, const struct sockaddr *from) { - int error = 0; - struct socket_filter_entry *filter; - int filtered = 0; - - for (filter = head->so_filt; filter && (error == 0); - filter = filter->sfe_next_onsocket) { - if (filter->sfe_filter->sf_filter.sf_connect_in) { - if (filtered == 0) { - filtered = 1; - sflt_use(head); - socket_unlock(head, 0); - } - error = filter->sfe_filter->sf_filter. - sf_connect_in(filter->sfe_cookie, head, from); - } - } - if (filtered != 0) { - socket_lock(head, 0); - sflt_unuse(head); - } - + int error = sflt_connectin(head, from); if (error) { return (NULL); } @@ -389,6 +476,7 @@ void socantsendmore(struct socket *so) { so->so_state |= SS_CANTSENDMORE; + soevent(so, SO_FILT_HINT_LOCKED | SO_FILT_HINT_CANTSENDMORE); sflt_notify(so, sock_evt_cantsendmore, NULL); sowwakeup(so); } @@ -397,91 +485,80 @@ void socantrcvmore(struct socket *so) { so->so_state |= SS_CANTRCVMORE; + soevent(so, SO_FILT_HINT_LOCKED | SO_FILT_HINT_CANTRCVMORE); sflt_notify(so, sock_evt_cantrecvmore, NULL); sorwakeup(so); } /* * Wait for data to arrive at/drain from a socket buffer. - * - * Returns: 0 Success - * EBADF - * msleep:EINTR */ int sbwait(struct sockbuf *sb) { - int error = 0, lr_saved; + boolean_t nointr = (sb->sb_flags & SB_NOINTR); + void *lr_saved = __builtin_return_address(0); struct socket *so = sb->sb_so; lck_mtx_t *mutex_held; struct timespec ts; + int error = 0; - lr_saved = (unsigned int) __builtin_return_address(0); + if (so == NULL) { + panic("%s: null so, sb=%p sb_flags=0x%x lr=%p\n", + __func__, sb, sb->sb_flags, lr_saved); + /* NOTREACHED */ + } else if (so->so_usecount < 1) { + panic("%s: sb=%p sb_flags=0x%x sb_so=%p usecount=%d lr=%p " + "lrh= %s\n", __func__, sb, sb->sb_flags, so, + so->so_usecount, lr_saved, solockhistory_nr(so)); + /* NOTREACHED */ + } if (so->so_proto->pr_getlock != NULL) mutex_held = (*so->so_proto->pr_getlock)(so, 0); else mutex_held = so->so_proto->pr_domain->dom_mtx; - sb->sb_flags |= SB_WAIT; + lck_mtx_assert(mutex_held, LCK_MTX_ASSERT_OWNED); - if (so->so_usecount < 1) - panic("sbwait: so=%p refcount=%d\n", so, so->so_usecount); ts.tv_sec = sb->sb_timeo.tv_sec; ts.tv_nsec = sb->sb_timeo.tv_usec * 1000; + + sb->sb_waiters++; + VERIFY(sb->sb_waiters != 0); + error = msleep((caddr_t)&sb->sb_cc, mutex_held, - (sb->sb_flags & SB_NOINTR) ? PSOCK : PSOCK | PCATCH, "sbwait", &ts); + nointr ? PSOCK : PSOCK | PCATCH, + nointr ? "sbwait_nointr" : "sbwait", &ts); - lck_mtx_assert(mutex_held, LCK_MTX_ASSERT_OWNED); + VERIFY(sb->sb_waiters != 0); + sb->sb_waiters--; - if (so->so_usecount < 1) - panic("sbwait: so=%p refcount=%d\n", so, so->so_usecount); + if (so->so_usecount < 1) { + panic("%s: 2 sb=%p sb_flags=0x%x sb_so=%p usecount=%d lr=%p " + "lrh= %s\n", __func__, sb, sb->sb_flags, so, + so->so_usecount, lr_saved, solockhistory_nr(so)); + /* NOTREACHED */ + } - if ((so->so_state & SS_DRAINING)) { + if ((so->so_state & SS_DRAINING) || (so->so_flags & SOF_DEFUNCT)) { error = EBADF; + if (so->so_flags & SOF_DEFUNCT) { + SODEFUNCTLOG(("%s[%d]: defunct so 0x%llx [%d,%d] " + "(%d)\n", __func__, proc_selfpid(), + (uint64_t)VM_KERNEL_ADDRPERM(so), + SOCK_DOM(so), SOCK_TYPE(so), error)); + } } return (error); } -/* - * Lock a sockbuf already known to be locked; - * return any error returned from sleep (EINTR). - * - * Returns: 0 Success - * EINTR - */ -int -sb_lock(struct sockbuf *sb) +void +sbwakeup(struct sockbuf *sb) { - struct socket *so = sb->sb_so; - lck_mtx_t *mutex_held; - int error = 0; - - if (so == NULL) - panic("sb_lock: null so back pointer sb=%p\n", sb); - - while (sb->sb_flags & SB_LOCK) { - sb->sb_flags |= SB_WANT; - if (so->so_proto->pr_getlock != NULL) - mutex_held = (*so->so_proto->pr_getlock)(so, 0); - else - mutex_held = so->so_proto->pr_domain->dom_mtx; - if (so->so_usecount < 1) - panic("sb_lock: so=%p refcount=%d\n", so, - so->so_usecount); - - error = msleep((caddr_t)&sb->sb_flags, mutex_held, - (sb->sb_flags & SB_NOINTR) ? PSOCK : PSOCK | PCATCH, - "sb_lock", 0); - if (so->so_usecount < 1) - panic("sb_lock: 2 so=%p refcount=%d\n", so, - so->so_usecount); - if (error) - return (error); - } - sb->sb_flags |= SB_LOCK; - return (0); + if (sb->sb_waiters > 0) + wakeup((caddr_t)&sb->sb_cc); } /* @@ -492,12 +569,17 @@ sb_lock(struct sockbuf *sb) void sowakeup(struct socket *so, struct sockbuf *sb) { + if (so->so_flags & SOF_DEFUNCT) { + SODEFUNCTLOG(("%s[%d]: defunct so 0x%llx [%d,%d] si 0x%x, " + "fl 0x%x [%s]\n", __func__, proc_selfpid(), + (uint64_t)VM_KERNEL_ADDRPERM(so), SOCK_DOM(so), + SOCK_TYPE(so), (uint32_t)sb->sb_sel.si_flags, sb->sb_flags, + (sb->sb_flags & SB_RECV) ? "rcv" : "snd")); + } + sb->sb_flags &= ~SB_SEL; selwakeup(&sb->sb_sel); - if (sb->sb_flags & SB_WAIT) { - sb->sb_flags &= ~SB_WAIT; - wakeup((caddr_t)&sb->sb_cc); - } + sbwakeup(sb); if (so->so_state & SS_ASYNC) { if (so->so_pgid < 0) gsignal(-so->so_pgid, SIGIO); @@ -508,23 +590,38 @@ sowakeup(struct socket *so, struct sockbuf *sb) KNOTE(&sb->sb_sel.si_note, SO_FILT_HINT_LOCKED); } if (sb->sb_flags & SB_UPCALL) { - void (*so_upcall)(struct socket *, caddr_t, int); - caddr_t so_upcallarg; + void (*sb_upcall)(struct socket *, void *, int); + caddr_t sb_upcallarg; - so_upcall = so->so_upcall; - so_upcallarg = so->so_upcallarg; + sb_upcall = sb->sb_upcall; + sb_upcallarg = sb->sb_upcallarg; /* Let close know that we're about to do an upcall */ - so->so_flags |= SOF_UPCALLINUSE; + so->so_upcallusecount++; socket_unlock(so, 0); - (*so_upcall)(so, so_upcallarg, M_DONTWAIT); + (*sb_upcall)(so, sb_upcallarg, M_DONTWAIT); socket_lock(so, 0); - so->so_flags &= ~SOF_UPCALLINUSE; + so->so_upcallusecount--; /* Tell close that it's safe to proceed */ - if (so->so_flags & SOF_CLOSEWAIT) - wakeup((caddr_t)&so->so_upcall); + if ((so->so_flags & SOF_CLOSEWAIT) && + so->so_upcallusecount == 0) + wakeup((caddr_t)&so->so_upcallusecount); + } +#if CONTENT_FILTER + /* + * Trap disconnection events for content filters + */ + if ((so->so_flags & SOF_CONTENT_FILTER) != 0) { + if ((sb->sb_flags & SB_RECV)) { + if (so->so_state & (SS_CANTRCVMORE)) + cfil_sock_notify_shutdown(so, SHUT_RD); + } else { + if (so->so_state & (SS_CANTSENDMORE)) + cfil_sock_notify_shutdown(so, SHUT_WR); + } } +#endif /* CONTENT_FILTER */ } /* @@ -564,13 +661,19 @@ sowakeup(struct socket *so, struct sockbuf *sb) * ENOBUFS */ int -soreserve(struct socket *so, u_long sndcc, u_long rcvcc) +soreserve(struct socket *so, u_int32_t sndcc, u_int32_t rcvcc) { if (sbreserve(&so->so_snd, sndcc) == 0) goto bad; + else + so->so_snd.sb_idealsize = sndcc; + if (sbreserve(&so->so_rcv, rcvcc) == 0) goto bad2; + else + so->so_rcv.sb_idealsize = rcvcc; + if (so->so_rcv.sb_lowat == 0) so->so_rcv.sb_lowat = 1; if (so->so_snd.sb_lowat == 0) @@ -579,21 +682,28 @@ soreserve(struct socket *so, u_long sndcc, u_long rcvcc) so->so_snd.sb_lowat = so->so_snd.sb_hiwat; return (0); bad2: -#ifdef __APPLE__ + so->so_snd.sb_flags &= ~SB_SEL; selthreadclear(&so->so_snd.sb_sel); -#endif sbrelease(&so->so_snd); bad: return (ENOBUFS); } +void +soreserve_preconnect(struct socket *so, unsigned int pre_cc) +{ + /* As of now, same bytes for both preconnect read and write */ + so->so_snd.sb_preconn_hiwat = pre_cc; + so->so_rcv.sb_preconn_hiwat = pre_cc; +} + /* * Allot mbufs to a sockbuf. * Attempt to scale mbmax so that mbcnt doesn't become limiting * if buffering efficiency is near the normal case. */ int -sbreserve(struct sockbuf *sb, u_long cc) +sbreserve(struct sockbuf *sb, u_int32_t cc) { if ((u_quad_t)cc > (u_quad_t)sb_max * MCLBYTES / (MSIZE + MCLBYTES)) return (0); @@ -663,14 +773,22 @@ sbappend(struct sockbuf *sb, struct mbuf *m) if (sb->sb_lastrecord != NULL && (sb->sb_mbtail->m_flags & M_EOR)) return (sbappendrecord(sb, m)); - if (sb->sb_flags & SB_RECV) { - int error = sflt_data_in(so, NULL, &m, NULL, 0, NULL); + if (sb->sb_flags & SB_RECV && !(m && m->m_flags & M_SKIPCFIL)) { + int error = sflt_data_in(so, NULL, &m, NULL, 0); SBLASTRECORDCHK(sb, "sbappend 2"); + +#if CONTENT_FILTER + if (error == 0) + error = cfil_sock_data_in(so, NULL, m, NULL, 0); +#endif /* CONTENT_FILTER */ + if (error != 0) { if (error != EJUSTRETURN) m_freem(m); return (0); } + } else if (m) { + m->m_flags &= ~M_SKIPCFIL; } /* If this is the first record, it's also the last record */ @@ -690,26 +808,36 @@ sbappendstream(struct sockbuf *sb, struct mbuf *m) { struct socket *so = sb->sb_so; - if (m->m_nextpkt != NULL || (sb->sb_mb != sb->sb_lastrecord)) - panic("sbappendstream: nexpkt %p || mb %p != lastrecord %p\n", - m->m_nextpkt, sb->sb_mb, sb->sb_lastrecord); - - SBLASTMBUFCHK(sb, __func__); - if (m == NULL || (sb->sb_flags & SB_DROP)) { if (m != NULL) m_freem(m); return (0); } - if (sb->sb_flags & SB_RECV) { - int error = sflt_data_in(so, NULL, &m, NULL, 0, NULL); + if (m->m_nextpkt != NULL || (sb->sb_mb != sb->sb_lastrecord)) { + panic("sbappendstream: nexpkt %p || mb %p != lastrecord %p\n", + m->m_nextpkt, sb->sb_mb, sb->sb_lastrecord); + /* NOTREACHED */ + } + + SBLASTMBUFCHK(sb, __func__); + + if (sb->sb_flags & SB_RECV && !(m && m->m_flags & M_SKIPCFIL)) { + int error = sflt_data_in(so, NULL, &m, NULL, 0); SBLASTRECORDCHK(sb, "sbappendstream 1"); + +#if CONTENT_FILTER + if (error == 0) + error = cfil_sock_data_in(so, NULL, m, NULL, 0); +#endif /* CONTENT_FILTER */ + if (error != 0) { if (error != EJUSTRETURN) m_freem(m); return (0); } + } else if (m) { + m->m_flags &= ~M_SKIPCFIL; } sbcompress(sb, m, sb->sb_mbtail); @@ -724,7 +852,7 @@ sbcheck(struct sockbuf *sb) { struct mbuf *m; struct mbuf *n = 0; - u_long len = 0, mbcnt = 0; + u_int32_t len = 0, mbcnt = 0; lck_mtx_t *mutex_held; if (sb->sb_so->so_proto->pr_getlock != NULL) @@ -763,11 +891,14 @@ sblastrecordchk(struct sockbuf *sb, const char *where) m = m->m_nextpkt; if (m != sb->sb_lastrecord) { - printf("sblastrecordchk: mb %p lastrecord %p last %p\n", - sb->sb_mb, sb->sb_lastrecord, m); + printf("sblastrecordchk: mb 0x%llx lastrecord 0x%llx " + "last 0x%llx\n", + (uint64_t)VM_KERNEL_ADDRPERM(sb->sb_mb), + (uint64_t)VM_KERNEL_ADDRPERM(sb->sb_lastrecord), + (uint64_t)VM_KERNEL_ADDRPERM(m)); printf("packet chain:\n"); for (m = sb->sb_mb; m != NULL; m = m->m_nextpkt) - printf("\t%p\n", m); + printf("\t0x%llx\n", (uint64_t)VM_KERNEL_ADDRPERM(m)); panic("sblastrecordchk from %s", where); } } @@ -785,13 +916,16 @@ sblastmbufchk(struct sockbuf *sb, const char *where) m = m->m_next; if (m != sb->sb_mbtail) { - printf("sblastmbufchk: mb %p mbtail %p last %p\n", - sb->sb_mb, sb->sb_mbtail, m); + printf("sblastmbufchk: mb 0x%llx mbtail 0x%llx last 0x%llx\n", + (uint64_t)VM_KERNEL_ADDRPERM(sb->sb_mb), + (uint64_t)VM_KERNEL_ADDRPERM(sb->sb_mbtail), + (uint64_t)VM_KERNEL_ADDRPERM(m)); printf("packet tree:\n"); for (m = sb->sb_mb; m != NULL; m = m->m_nextpkt) { printf("\t"); for (n = m; n != NULL; n = n->m_next) - printf("%p ", n); + printf("0x%llx ", + (uint64_t)VM_KERNEL_ADDRPERM(n)); printf("\n"); } panic("sblastmbufchk from %s", where); @@ -821,15 +955,23 @@ sbappendrecord(struct sockbuf *sb, struct mbuf *m0) return (0); } - if (sb->sb_flags & SB_RECV) { + if (sb->sb_flags & SB_RECV && !(m0 && m0->m_flags & M_SKIPCFIL)) { int error = sflt_data_in(sb->sb_so, NULL, &m0, NULL, - sock_data_filt_flag_record, NULL); + sock_data_filt_flag_record); + +#if CONTENT_FILTER + if (error == 0) + error = cfil_sock_data_in(sb->sb_so, NULL, m0, NULL, 0); +#endif /* CONTENT_FILTER */ + if (error != 0) { SBLASTRECORDCHK(sb, "sbappendrecord 1"); if (error != EJUSTRETURN) m_freem(m0); return (0); } + } else if (m0) { + m0->m_flags &= ~M_SKIPCFIL; } /* @@ -839,10 +981,11 @@ sbappendrecord(struct sockbuf *sb, struct mbuf *m0) SBLASTRECORDCHK(sb, "sbappendrecord 2"); if (sb->sb_lastrecord != NULL) { sb->sb_lastrecord->m_nextpkt = m0; - } else { + } else { sb->sb_mb = m0; } sb->sb_lastrecord = m0; + sb->sb_mbtail = m0; m = m0->m_next; m0->m_next = 0; @@ -871,17 +1014,25 @@ sbinsertoob(struct sockbuf *sb, struct mbuf *m0) SBLASTRECORDCHK(sb, "sbinsertoob 1"); - if ((sb->sb_flags & SB_RECV) != 0) { + if ((sb->sb_flags & SB_RECV && !(m0->m_flags & M_SKIPCFIL)) != 0) { int error = sflt_data_in(sb->sb_so, NULL, &m0, NULL, - sock_data_filt_flag_oob, NULL); + sock_data_filt_flag_oob); SBLASTRECORDCHK(sb, "sbinsertoob 2"); + +#if CONTENT_FILTER + if (error == 0) + error = cfil_sock_data_in(sb->sb_so, NULL, m0, NULL, 0); +#endif /* CONTENT_FILTER */ + if (error) { if (error != EJUSTRETURN) { m_freem(m0); } return (0); } + } else if (m0) { + m0->m_flags &= ~M_SKIPCFIL; } for (mp = &sb->sb_mb; *mp; mp = &((*mp)->m_nextpkt)) { @@ -1016,10 +1167,17 @@ sbappendaddr(struct sockbuf *sb, struct sockaddr *asa, struct mbuf *m0, } /* Call socket data in filters */ - if ((sb->sb_flags & SB_RECV) != 0) { + if (sb->sb_flags & SB_RECV && !(m0 && m0->m_flags & M_SKIPCFIL)) { int error; - error = sflt_data_in(sb->sb_so, asa, &m0, &control, 0, NULL); + error = sflt_data_in(sb->sb_so, asa, &m0, &control, 0); SBLASTRECORDCHK(sb, __func__); + +#if CONTENT_FILTER + if (error == 0) + error = cfil_sock_data_in(sb->sb_so, asa, m0, control, + 0); +#endif /* CONTENT_FILTER */ + if (error) { if (error != EJUSTRETURN) { if (m0) @@ -1031,6 +1189,8 @@ sbappendaddr(struct sockbuf *sb, struct sockaddr *asa, struct mbuf *m0, } return (0); } + } else if (m0) { + m0->m_flags &= ~M_SKIPCFIL; } result = sbappendaddr_internal(sb, asa, m0, control); @@ -1067,7 +1227,6 @@ sbappendcontrol_internal(struct sockbuf *sb, struct mbuf *m0, if (space > sbspace(sb) && !(sb->sb_flags & SB_UNIX)) return (0); n->m_next = m0; /* concatenate data to control */ - SBLASTRECORDCHK(sb, "sbappendcontrol 1"); for (m = control; m->m_next != NULL; m = m->m_next) @@ -1110,11 +1269,18 @@ sbappendcontrol(struct sockbuf *sb, struct mbuf *m0, struct mbuf *control, return (0); } - if (sb->sb_flags & SB_RECV) { + if (sb->sb_flags & SB_RECV && !(m0 && m0->m_flags & M_SKIPCFIL)) { int error; - error = sflt_data_in(sb->sb_so, NULL, &m0, &control, 0, NULL); + error = sflt_data_in(sb->sb_so, NULL, &m0, &control, 0); SBLASTRECORDCHK(sb, __func__); + +#if CONTENT_FILTER + if (error == 0) + error = cfil_sock_data_in(sb->sb_so, NULL, m0, control, + 0); +#endif /* CONTENT_FILTER */ + if (error) { if (error != EJUSTRETURN) { if (m0) @@ -1126,6 +1292,8 @@ sbappendcontrol(struct sockbuf *sb, struct mbuf *m0, struct mbuf *control, } return (0); } + } else if (m0) { + m0->m_flags &= ~M_SKIPCFIL; } result = sbappendcontrol_internal(sb, m0, control); @@ -1142,145 +1310,555 @@ sbappendcontrol(struct sockbuf *sb, struct mbuf *m0, struct mbuf *control, } /* - * Compress mbuf chain m into the socket - * buffer sb following mbuf n. If n - * is null, the buffer is presumed empty. + * Append a contiguous TCP data blob with TCP sequence number as control data + * as a new msg to the receive socket buffer. */ -static inline void -sbcompress(struct sockbuf *sb, struct mbuf *m, struct mbuf *n) +int +sbappendmsgstream_rcv(struct sockbuf *sb, struct mbuf *m, uint32_t seqnum, + int unordered) { - int eor = 0; - struct mbuf *o; + struct mbuf *m_eor = NULL; + u_int32_t data_len = 0; + int ret = 0; + struct socket *so = sb->sb_so; - if (m == NULL) { - /* There is nothing to compress; just update the tail */ - for (; n->m_next != NULL; n = n->m_next) - ; - sb->sb_mbtail = n; - goto done; - } + VERIFY((m->m_flags & M_PKTHDR) && m_pktlen(m) > 0); + VERIFY(so->so_msg_state != NULL); + VERIFY(sb->sb_flags & SB_RECV); - while (m) { - eor |= m->m_flags & M_EOR; - if (m->m_len == 0 && (eor == 0 || - (((o = m->m_next) || (o = n)) && o->m_type == m->m_type))) { - if (sb->sb_lastrecord == m) - sb->sb_lastrecord = m->m_next; - m = m_free(m); - continue; - } - if (n && (n->m_flags & M_EOR) == 0 && -#ifndef __APPLE__ - M_WRITABLE(n) && -#endif - m->m_len <= MCLBYTES / 4 && /* XXX: Don't copy too much */ - m->m_len <= M_TRAILINGSPACE(n) && - n->m_type == m->m_type) { - bcopy(mtod(m, caddr_t), mtod(n, caddr_t) + n->m_len, - (unsigned)m->m_len); - n->m_len += m->m_len; - sb->sb_cc += m->m_len; - if (m->m_type != MT_DATA && m->m_type != MT_HEADER && - m->m_type != MT_OOBDATA) - /* XXX: Probably don't need.*/ - sb->sb_ctl += m->m_len; - m = m_free(m); - continue; + /* Keep the TCP sequence number in the mbuf pkthdr */ + m->m_pkthdr.msg_seq = seqnum; + + /* find last mbuf and set M_EOR */ + for (m_eor = m; ; m_eor = m_eor->m_next) { + /* + * If the msg is unordered, we need to account for + * these bytes in receive socket buffer size. Otherwise, + * the receive window advertised will shrink because + * of the additional unordered bytes added to the + * receive buffer. + */ + if (unordered) { + m_eor->m_flags |= M_UNORDERED_DATA; + data_len += m_eor->m_len; + so->so_msg_state->msg_uno_bytes += m_eor->m_len; + } else { + m_eor->m_flags &= ~M_UNORDERED_DATA; } - if (n) - n->m_next = m; - else - sb->sb_mb = m; - sb->sb_mbtail = m; - sballoc(sb, m); - n = m; - m->m_flags &= ~M_EOR; - m = m->m_next; - n->m_next = 0; + if (m_eor->m_next == NULL) + break; } - if (eor) { - if (n) - n->m_flags |= eor; - else - printf("semi-panic: sbcompress\n"); + + /* set EOR flag at end of byte blob */ + m_eor->m_flags |= M_EOR; + + /* expand the receive socket buffer to allow unordered data */ + if (unordered && !sbreserve(sb, sb->sb_hiwat + data_len)) { + /* + * Could not allocate memory for unordered data, it + * means this packet will have to be delivered in order + */ + printf("%s: could not reserve space for unordered data\n", + __func__); } -done: - SBLASTMBUFCHK(sb, __func__); - postevent(0, sb, EV_RWBYTES); -} -void -sb_empty_assert(struct sockbuf *sb, const char *where) -{ - if (!(sb->sb_cc == 0 && sb->sb_mb == NULL && sb->sb_mbcnt == 0 && - sb->sb_mbtail == NULL && sb->sb_lastrecord == NULL)) { - panic("%s: sb %p so %p cc %ld mbcnt %ld mb %p mbtail %p " - "lastrecord %p\n", where, sb, sb->sb_so, sb->sb_cc, - sb->sb_mbcnt, sb->sb_mb, sb->sb_mbtail, sb->sb_lastrecord); - /* NOTREACHED */ + if (!unordered && (sb->sb_mbtail != NULL) && + !(sb->sb_mbtail->m_flags & M_UNORDERED_DATA)) { + sb->sb_mbtail->m_flags &= ~M_EOR; + sbcompress(sb, m, sb->sb_mbtail); + ret = 1; + } else { + ret = sbappendrecord(sb, m); } + VERIFY(sb->sb_mbtail->m_flags & M_EOR); + return (ret); } /* - * Free all mbufs in a sockbuf. - * Check that all resources are reclaimed. + * TCP streams have message based out of order delivery support, or have + * Multipath TCP support, or are regular TCP sockets */ -void -sbflush(struct sockbuf *sb) +int +sbappendstream_rcvdemux(struct socket *so, struct mbuf *m, uint32_t seqnum, + int unordered) { - if (sb->sb_so == NULL) - panic("sbflush sb->sb_so already null sb=%p\n", sb); - (void) sblock(sb, M_WAIT); - while (sb->sb_mbcnt) { - /* - * Don't call sbdrop(sb, 0) if the leading mbuf is non-empty: - * we would loop forever. Panic instead. - */ - if (!sb->sb_cc && (sb->sb_mb == NULL || sb->sb_mb->m_len)) - break; - sbdrop(sb, (int)sb->sb_cc); + int ret = 0; + + if ((m != NULL) && (m_pktlen(m) <= 0)) { + m_freem(m); + return (ret); } - sb_empty_assert(sb, __func__); - postevent(0, sb, EV_RWBYTES); - sbunlock(sb, 1); /* keep socket locked */ + if (so->so_flags & SOF_ENABLE_MSGS) { + ret = sbappendmsgstream_rcv(&so->so_rcv, m, seqnum, unordered); + } +#if MPTCP + else if (so->so_flags & SOF_MPTCP_TRUE) { + ret = sbappendmptcpstream_rcv(&so->so_rcv, m); + } +#endif /* MPTCP */ + else { + ret = sbappendstream(&so->so_rcv, m); + } + return (ret); } -/* - * Drop data from (the front of) a sockbuf. - * use m_freem_list to free the mbuf structures - * under a single lock... this is done by pruning - * the top of the tree from the body by keeping track - * of where we get to in the tree and then zeroing the - * two pertinent pointers m_nextpkt and m_next - * the socket buffer is then updated to point at the new - * top of the tree and the pruned area is released via - * m_freem_list. - */ -void -sbdrop(struct sockbuf *sb, int len) +#if MPTCP +int +sbappendmptcpstream_rcv(struct sockbuf *sb, struct mbuf *m) { - struct mbuf *m, *free_list, *ml; - struct mbuf *next, *last; + struct socket *so = sb->sb_so; - KERNEL_DEBUG((DBG_FNC_SBDROP | DBG_FUNC_START), sb, len, 0, 0, 0); + VERIFY(m == NULL || (m->m_flags & M_PKTHDR)); + /* SB_NOCOMPRESS must be set prevent loss of M_PKTHDR data */ + VERIFY((sb->sb_flags & (SB_RECV|SB_NOCOMPRESS)) == + (SB_RECV|SB_NOCOMPRESS)); - next = (m = sb->sb_mb) ? m->m_nextpkt : 0; - free_list = last = m; - ml = (struct mbuf *)0; + if (m == NULL || m_pktlen(m) == 0 || (sb->sb_flags & SB_DROP) || + (so->so_state & SS_CANTRCVMORE)) { + if (m != NULL) + m_freem(m); + return (0); + } + /* the socket is not closed, so SOF_MP_SUBFLOW must be set */ + VERIFY(so->so_flags & SOF_MP_SUBFLOW); - while (len > 0) { - if (m == 0) { - if (next == 0) { - /* - * temporarily replacing this panic with printf - * because it occurs occasionally when closing - * a socket when there is no harm in ignoring - * it. This problem will be investigated - * further. - */ - /* panic("sbdrop"); */ - printf("sbdrop - count not zero\n"); + if (m->m_nextpkt != NULL || (sb->sb_mb != sb->sb_lastrecord)) { + panic("%s: nexpkt %p || mb %p != lastrecord %p\n", __func__, + m->m_nextpkt, sb->sb_mb, sb->sb_lastrecord); + /* NOTREACHED */ + } + + SBLASTMBUFCHK(sb, __func__); + + if (mptcp_adj_rmap(so, m) != 0) + return (0); + + /* No filter support (SB_RECV) on mptcp subflow sockets */ + + sbcompress(sb, m, sb->sb_mbtail); + sb->sb_lastrecord = sb->sb_mb; + SBLASTRECORDCHK(sb, __func__); + return (1); +} +#endif /* MPTCP */ + +/* + * Append message to send socket buffer based on priority. + */ +int +sbappendmsg_snd(struct sockbuf *sb, struct mbuf *m) +{ + struct socket *so = sb->sb_so; + struct msg_priq *priq; + int set_eor = 0; + + VERIFY(so->so_msg_state != NULL); + + if (m->m_nextpkt != NULL || (sb->sb_mb != sb->sb_lastrecord)) + panic("sbappendstream: nexpkt %p || mb %p != lastrecord %p\n", + m->m_nextpkt, sb->sb_mb, sb->sb_lastrecord); + + SBLASTMBUFCHK(sb, __func__); + + if (m == NULL || (sb->sb_flags & SB_DROP) || so->so_msg_state == NULL) { + if (m != NULL) + m_freem(m); + return (0); + } + + priq = &so->so_msg_state->msg_priq[m->m_pkthdr.msg_pri]; + + /* note if we need to propogate M_EOR to the last mbuf */ + if (m->m_flags & M_EOR) { + set_eor = 1; + + /* Reset M_EOR from the first mbuf */ + m->m_flags &= ~(M_EOR); + } + + if (priq->msgq_head == NULL) { + VERIFY(priq->msgq_tail == NULL && priq->msgq_lastmsg == NULL); + priq->msgq_head = priq->msgq_lastmsg = m; + } else { + VERIFY(priq->msgq_tail->m_next == NULL); + + /* Check if the last message has M_EOR flag set */ + if (priq->msgq_tail->m_flags & M_EOR) { + /* Insert as a new message */ + priq->msgq_lastmsg->m_nextpkt = m; + + /* move the lastmsg pointer */ + priq->msgq_lastmsg = m; + } else { + /* Append to the existing message */ + priq->msgq_tail->m_next = m; + } + } + + /* Update accounting and the queue tail pointer */ + + while (m->m_next != NULL) { + sballoc(sb, m); + priq->msgq_bytes += m->m_len; + m = m->m_next; + } + sballoc(sb, m); + priq->msgq_bytes += m->m_len; + + if (set_eor) { + m->m_flags |= M_EOR; + + /* + * Since the user space can not write a new msg + * without completing the previous one, we can + * reset this flag to start sending again. + */ + priq->msgq_flags &= ~(MSGQ_MSG_NOTDONE); + } + + priq->msgq_tail = m; + + SBLASTRECORDCHK(sb, "sbappendstream 2"); + postevent(0, sb, EV_RWBYTES); + return (1); +} + +/* + * Pull data from priority queues to the serial snd queue + * right before sending. + */ +void +sbpull_unordered_data(struct socket *so, int32_t off, int32_t len) +{ + int32_t topull, i; + struct msg_priq *priq = NULL; + + VERIFY(so->so_msg_state != NULL); + + topull = (off + len) - so->so_msg_state->msg_serial_bytes; + + i = MSG_PRI_MAX; + while (i >= MSG_PRI_MIN && topull > 0) { + struct mbuf *m = NULL, *mqhead = NULL, *mend = NULL; + priq = &so->so_msg_state->msg_priq[i]; + if ((priq->msgq_flags & MSGQ_MSG_NOTDONE) && + priq->msgq_head == NULL) { + /* + * We were in the middle of sending + * a message and we have not seen the + * end of it. + */ + VERIFY(priq->msgq_lastmsg == NULL && + priq->msgq_tail == NULL); + return; + } + if (priq->msgq_head != NULL) { + int32_t bytes = 0, topull_tmp = topull; + /* + * We found a msg while scanning the priority + * queue from high to low priority. + */ + m = priq->msgq_head; + mqhead = m; + mend = m; + + /* + * Move bytes from the priority queue to the + * serial queue. Compute the number of bytes + * being added. + */ + while (mqhead->m_next != NULL && topull_tmp > 0) { + bytes += mqhead->m_len; + topull_tmp -= mqhead->m_len; + mend = mqhead; + mqhead = mqhead->m_next; + } + + if (mqhead->m_next == NULL) { + /* + * If we have only one more mbuf left, + * move the last mbuf of this message to + * serial queue and set the head of the + * queue to be the next message. + */ + bytes += mqhead->m_len; + mend = mqhead; + mqhead = m->m_nextpkt; + if (!(mend->m_flags & M_EOR)) { + /* + * We have not seen the end of + * this message, so we can not + * pull anymore. + */ + priq->msgq_flags |= MSGQ_MSG_NOTDONE; + } else { + /* Reset M_EOR */ + mend->m_flags &= ~(M_EOR); + } + } else { + /* propogate the next msg pointer */ + mqhead->m_nextpkt = m->m_nextpkt; + } + priq->msgq_head = mqhead; + + /* + * if the lastmsg pointer points to + * the mbuf that is being dequeued, update + * it to point to the new head. + */ + if (priq->msgq_lastmsg == m) + priq->msgq_lastmsg = priq->msgq_head; + + m->m_nextpkt = NULL; + mend->m_next = NULL; + + if (priq->msgq_head == NULL) { + /* Moved all messages, update tail */ + priq->msgq_tail = NULL; + VERIFY(priq->msgq_lastmsg == NULL); + } + + /* Move it to serial sb_mb queue */ + if (so->so_snd.sb_mb == NULL) { + so->so_snd.sb_mb = m; + } else { + so->so_snd.sb_mbtail->m_next = m; + } + + priq->msgq_bytes -= bytes; + VERIFY(priq->msgq_bytes >= 0); + sbwakeup(&so->so_snd); + + so->so_msg_state->msg_serial_bytes += bytes; + so->so_snd.sb_mbtail = mend; + so->so_snd.sb_lastrecord = so->so_snd.sb_mb; + + topull = + (off + len) - so->so_msg_state->msg_serial_bytes; + + if (priq->msgq_flags & MSGQ_MSG_NOTDONE) + break; + } else { + --i; + } + } + sblastrecordchk(&so->so_snd, "sbpull_unordered_data"); + sblastmbufchk(&so->so_snd, "sbpull_unordered_data"); +} + +/* + * Compress mbuf chain m into the socket + * buffer sb following mbuf n. If n + * is null, the buffer is presumed empty. + */ +static inline void +sbcompress(struct sockbuf *sb, struct mbuf *m, struct mbuf *n) +{ + int eor = 0, compress = (!(sb->sb_flags & SB_NOCOMPRESS)); + struct mbuf *o; + + if (m == NULL) { + /* There is nothing to compress; just update the tail */ + for (; n->m_next != NULL; n = n->m_next) + ; + sb->sb_mbtail = n; + goto done; + } + + while (m != NULL) { + eor |= m->m_flags & M_EOR; + if (compress && m->m_len == 0 && (eor == 0 || + (((o = m->m_next) || (o = n)) && o->m_type == m->m_type))) { + if (sb->sb_lastrecord == m) + sb->sb_lastrecord = m->m_next; + m = m_free(m); + continue; + } + if (compress && n != NULL && (n->m_flags & M_EOR) == 0 && +#ifndef __APPLE__ + M_WRITABLE(n) && +#endif + m->m_len <= MCLBYTES / 4 && /* XXX: Don't copy too much */ + m->m_len <= M_TRAILINGSPACE(n) && + n->m_type == m->m_type) { + bcopy(mtod(m, caddr_t), mtod(n, caddr_t) + n->m_len, + (unsigned)m->m_len); + n->m_len += m->m_len; + sb->sb_cc += m->m_len; + if (m->m_type != MT_DATA && m->m_type != MT_HEADER && + m->m_type != MT_OOBDATA) { + /* XXX: Probably don't need */ + sb->sb_ctl += m->m_len; + } + m = m_free(m); + continue; + } + if (n != NULL) + n->m_next = m; + else + sb->sb_mb = m; + sb->sb_mbtail = m; + sballoc(sb, m); + n = m; + m->m_flags &= ~M_EOR; + m = m->m_next; + n->m_next = NULL; + } + if (eor != 0) { + if (n != NULL) + n->m_flags |= eor; + else + printf("semi-panic: sbcompress\n"); + } +done: + SBLASTMBUFCHK(sb, __func__); + postevent(0, sb, EV_RWBYTES); +} + +void +sb_empty_assert(struct sockbuf *sb, const char *where) +{ + if (!(sb->sb_cc == 0 && sb->sb_mb == NULL && sb->sb_mbcnt == 0 && + sb->sb_mbtail == NULL && sb->sb_lastrecord == NULL)) { + panic("%s: sb %p so %p cc %d mbcnt %d mb %p mbtail %p " + "lastrecord %p\n", where, sb, sb->sb_so, sb->sb_cc, + sb->sb_mbcnt, sb->sb_mb, sb->sb_mbtail, + sb->sb_lastrecord); + /* NOTREACHED */ + } +} + +static void +sbflush_priq(struct msg_priq *priq) +{ + struct mbuf *m; + m = priq->msgq_head; + if (m != NULL) + m_freem_list(m); + priq->msgq_head = priq->msgq_tail = priq->msgq_lastmsg = NULL; + priq->msgq_bytes = priq->msgq_flags = 0; +} + +/* + * Free all mbufs in a sockbuf. + * Check that all resources are reclaimed. + */ +void +sbflush(struct sockbuf *sb) +{ + void *lr_saved = __builtin_return_address(0); + struct socket *so = sb->sb_so; +#ifdef notyet + lck_mtx_t *mutex_held; +#endif + u_int32_t i; + + /* so_usecount may be 0 if we get here from sofreelastref() */ + if (so == NULL) { + panic("%s: null so, sb=%p sb_flags=0x%x lr=%p\n", + __func__, sb, sb->sb_flags, lr_saved); + /* NOTREACHED */ + } else if (so->so_usecount < 0) { + panic("%s: sb=%p sb_flags=0x%x sb_so=%p usecount=%d lr=%p " + "lrh= %s\n", __func__, sb, sb->sb_flags, so, + so->so_usecount, lr_saved, solockhistory_nr(so)); + /* NOTREACHED */ + } +#ifdef notyet + /* + * XXX: This code is currently commented out, because we may get here + * as part of sofreelastref(), and at that time, pr_getlock() may no + * longer be able to return us the lock; this will be fixed in future. + */ + if (so->so_proto->pr_getlock != NULL) + mutex_held = (*so->so_proto->pr_getlock)(so, 0); + else + mutex_held = so->so_proto->pr_domain->dom_mtx; + + lck_mtx_assert(mutex_held, LCK_MTX_ASSERT_OWNED); +#endif + + /* + * Obtain lock on the socket buffer (SB_LOCK). This is required + * to prevent the socket buffer from being unexpectedly altered + * while it is used by another thread in socket send/receive. + * + * sblock() must not fail here, hence the assertion. + */ + (void) sblock(sb, SBL_WAIT | SBL_NOINTR | SBL_IGNDEFUNCT); + VERIFY(sb->sb_flags & SB_LOCK); + + while (sb->sb_mbcnt > 0) { + /* + * Don't call sbdrop(sb, 0) if the leading mbuf is non-empty: + * we would loop forever. Panic instead. + */ + if (!sb->sb_cc && (sb->sb_mb == NULL || sb->sb_mb->m_len)) + break; + sbdrop(sb, (int)sb->sb_cc); + } + + if (!(sb->sb_flags & SB_RECV) && (so->so_flags & SOF_ENABLE_MSGS)) { + VERIFY(so->so_msg_state != NULL); + for (i = MSG_PRI_MIN; i <= MSG_PRI_MAX; ++i) { + sbflush_priq(&so->so_msg_state->msg_priq[i]); + } + so->so_msg_state->msg_serial_bytes = 0; + so->so_msg_state->msg_uno_bytes = 0; + } + + sb_empty_assert(sb, __func__); + postevent(0, sb, EV_RWBYTES); + + sbunlock(sb, TRUE); /* keep socket locked */ +} + +/* + * Drop data from (the front of) a sockbuf. + * use m_freem_list to free the mbuf structures + * under a single lock... this is done by pruning + * the top of the tree from the body by keeping track + * of where we get to in the tree and then zeroing the + * two pertinent pointers m_nextpkt and m_next + * the socket buffer is then updated to point at the new + * top of the tree and the pruned area is released via + * m_freem_list. + */ +void +sbdrop(struct sockbuf *sb, int len) +{ + struct mbuf *m, *free_list, *ml; + struct mbuf *next, *last; + + next = (m = sb->sb_mb) ? m->m_nextpkt : 0; +#if MPTCP + if ((m != NULL) && (len > 0) && + (!(sb->sb_flags & SB_RECV)) && + ((sb->sb_so->so_flags & SOF_MP_SUBFLOW) || + ((SOCK_CHECK_DOM(sb->sb_so, PF_MULTIPATH)) && + (SOCK_CHECK_PROTO(sb->sb_so, IPPROTO_TCP)))) && + (!(sb->sb_so->so_flags1 & SOF1_POST_FALLBACK_SYNC))) { + mptcp_preproc_sbdrop(sb->sb_so, m, (unsigned int)len); + } +#endif /* MPTCP */ + KERNEL_DEBUG((DBG_FNC_SBDROP | DBG_FUNC_START), sb, len, 0, 0, 0); + + free_list = last = m; + ml = (struct mbuf *)0; + + while (len > 0) { + if (m == 0) { + if (next == 0) { + /* + * temporarily replacing this panic with printf + * because it occurs occasionally when closing + * a socket when there is no harm in ignoring + * it. This problem will be investigated + * further. + */ + /* panic("sbdrop"); */ + printf("sbdrop - count not zero\n"); len = 0; /* * zero the counts. if we have no mbufs, @@ -1288,6 +1866,11 @@ sbdrop(struct sockbuf *sb, int len) */ sb->sb_cc = 0; sb->sb_mbcnt = 0; + if (!(sb->sb_flags & SB_RECV) && + (sb->sb_so->so_flags & SOF_ENABLE_MSGS)) { + sb->sb_so->so_msg_state-> + msg_serial_bytes = 0; + } break; } m = last = next; @@ -1299,7 +1882,7 @@ sbdrop(struct sockbuf *sb, int len) m->m_data += len; sb->sb_cc -= len; if (m->m_type != MT_DATA && m->m_type != MT_HEADER && - m->m_type != MT_OOBDATA) + m->m_type != MT_OOBDATA) sb->sb_ctl -= len; break; } @@ -1340,6 +1923,10 @@ sbdrop(struct sockbuf *sb, int len) sb->sb_lastrecord = m; } +#if CONTENT_FILTER + cfil_sock_buf_update(sb); +#endif /* CONTENT_FILTER */ + postevent(0, sb, EV_RWBYTES); KERNEL_DEBUG((DBG_FNC_SBDROP | DBG_FUNC_END), sb, 0, 0, 0, 0); @@ -1382,6 +1969,7 @@ sbcreatecontrol(caddr_t p, int size, int type, int level) if ((m = m_get(M_DONTWAIT, MT_CONTROL)) == NULL) return ((struct mbuf *)NULL); cp = mtod(m, struct cmsghdr *); + VERIFY(IS_P2ALIGNED(cp, sizeof (u_int32_t))); /* XXX check size? */ (void) memcpy(CMSG_DATA(cp), p, size); m->m_len = CMSG_SPACE(size); @@ -1391,102 +1979,175 @@ sbcreatecontrol(caddr_t p, int size, int type, int level) return (m); } +struct mbuf ** +sbcreatecontrol_mbuf(caddr_t p, int size, int type, int level, struct mbuf **mp) +{ + struct mbuf *m; + struct cmsghdr *cp; + + if (*mp == NULL) { + *mp = sbcreatecontrol(p, size, type, level); + return (mp); + } + + if (CMSG_SPACE((u_int)size) + (*mp)->m_len > MLEN) { + mp = &(*mp)->m_next; + *mp = sbcreatecontrol(p, size, type, level); + return (mp); + } + + m = *mp; + + cp = (struct cmsghdr *)(void *)(mtod(m, char *) + m->m_len); + /* CMSG_SPACE ensures 32-bit alignment */ + VERIFY(IS_P2ALIGNED(cp, sizeof (u_int32_t))); + m->m_len += CMSG_SPACE(size); + + /* XXX check size? */ + (void) memcpy(CMSG_DATA(cp), p, size); + cp->cmsg_len = CMSG_LEN(size); + cp->cmsg_level = level; + cp->cmsg_type = type; + + return (mp); +} + + /* * Some routines that return EOPNOTSUPP for entry points that are not * supported by a protocol. Fill in as needed. */ int -pru_abort_notsupp(__unused struct socket *so) +pru_abort_notsupp(struct socket *so) { +#pragma unused(so) return (EOPNOTSUPP); } int -pru_accept_notsupp(__unused struct socket *so, __unused struct sockaddr **nam) +pru_accept_notsupp(struct socket *so, struct sockaddr **nam) { +#pragma unused(so, nam) return (EOPNOTSUPP); } int -pru_attach_notsupp(__unused struct socket *so, __unused int proto, - __unused struct proc *p) +pru_attach_notsupp(struct socket *so, int proto, struct proc *p) { +#pragma unused(so, proto, p) return (EOPNOTSUPP); } int -pru_bind_notsupp(__unused struct socket *so, __unused struct sockaddr *nam, - __unused struct proc *p) +pru_bind_notsupp(struct socket *so, struct sockaddr *nam, struct proc *p) { +#pragma unused(so, nam, p) return (EOPNOTSUPP); } int -pru_connect_notsupp(__unused struct socket *so, __unused struct sockaddr *nam, - __unused struct proc *p) +pru_connect_notsupp(struct socket *so, struct sockaddr *nam, struct proc *p) { +#pragma unused(so, nam, p) return (EOPNOTSUPP); } int -pru_connect2_notsupp(__unused struct socket *so1, __unused struct socket *so2) +pru_connect2_notsupp(struct socket *so1, struct socket *so2) { +#pragma unused(so1, so2) return (EOPNOTSUPP); } int -pru_control_notsupp(__unused struct socket *so, __unused u_long cmd, - __unused caddr_t data, __unused struct ifnet *ifp, __unused struct proc *p) +pru_connectx_notsupp(struct socket *so, struct sockaddr_list **src_sl, + struct sockaddr_list **dst_sl, struct proc *p, uint32_t ifscope, + sae_associd_t aid, sae_connid_t *pcid, uint32_t flags, void *arg, + uint32_t arglen, struct uio *uio, user_ssize_t *bytes_written) { +#pragma unused(so, src_sl, dst_sl, p, ifscope, aid, pcid, flags, arg, arglen, uio, bytes_written) return (EOPNOTSUPP); } int -pru_detach_notsupp(__unused struct socket *so) +pru_control_notsupp(struct socket *so, u_long cmd, caddr_t data, + struct ifnet *ifp, struct proc *p) { +#pragma unused(so, cmd, data, ifp, p) return (EOPNOTSUPP); } int -pru_disconnect_notsupp(__unused struct socket *so) +pru_detach_notsupp(struct socket *so) { +#pragma unused(so) return (EOPNOTSUPP); } int -pru_listen_notsupp(__unused struct socket *so, __unused struct proc *p) +pru_disconnect_notsupp(struct socket *so) { +#pragma unused(so) return (EOPNOTSUPP); } int -pru_peeraddr_notsupp(__unused struct socket *so, __unused struct sockaddr **nam) +pru_disconnectx_notsupp(struct socket *so, sae_associd_t aid, sae_connid_t cid) { +#pragma unused(so, aid, cid) return (EOPNOTSUPP); } int -pru_rcvd_notsupp(__unused struct socket *so, __unused int flags) +pru_listen_notsupp(struct socket *so, struct proc *p) { +#pragma unused(so, p) return (EOPNOTSUPP); } int -pru_rcvoob_notsupp(__unused struct socket *so, __unused struct mbuf *m, - __unused int flags) +pru_peeloff_notsupp(struct socket *so, sae_associd_t aid, struct socket **psop) { +#pragma unused(so, aid, psop) return (EOPNOTSUPP); } int -pru_send_notsupp(__unused struct socket *so, __unused int flags, - __unused struct mbuf *m, __unused struct sockaddr *addr, - __unused struct mbuf *control, __unused struct proc *p) +pru_peeraddr_notsupp(struct socket *so, struct sockaddr **nam) +{ +#pragma unused(so, nam) + return (EOPNOTSUPP); +} +int +pru_rcvd_notsupp(struct socket *so, int flags) { +#pragma unused(so, flags) return (EOPNOTSUPP); } +int +pru_rcvoob_notsupp(struct socket *so, struct mbuf *m, int flags) +{ +#pragma unused(so, m, flags) + return (EOPNOTSUPP); +} + +int +pru_send_notsupp(struct socket *so, int flags, struct mbuf *m, + struct sockaddr *addr, struct mbuf *control, struct proc *p) +{ +#pragma unused(so, flags, m, addr, control, p) + return (EOPNOTSUPP); +} + +int +pru_send_list_notsupp(struct socket *so, int flags, struct mbuf *m, + struct sockaddr *addr, struct mbuf *control, struct proc *p) +{ +#pragma unused(so, flags, m, addr, control, p) + return (EOPNOTSUPP); +} /* * This isn't really a ``null'' operation, but it's the default one @@ -1512,44 +2173,110 @@ pru_sense_null(struct socket *so, void *ub, int isstat64) int -pru_sosend_notsupp(__unused struct socket *so, __unused struct sockaddr *addr, - __unused struct uio *uio, __unused struct mbuf *top, - __unused struct mbuf *control, __unused int flags) +pru_sosend_notsupp(struct socket *so, struct sockaddr *addr, struct uio *uio, + struct mbuf *top, struct mbuf *control, int flags) +{ +#pragma unused(so, addr, uio, top, control, flags) + return (EOPNOTSUPP); +} + +int +pru_sosend_list_notsupp(struct socket *so, struct uio **uio, + u_int uiocnt, int flags) +{ +#pragma unused(so, uio, uiocnt, flags) + return (EOPNOTSUPP); +} +int +pru_soreceive_notsupp(struct socket *so, struct sockaddr **paddr, + struct uio *uio, struct mbuf **mp0, struct mbuf **controlp, int *flagsp) { +#pragma unused(so, paddr, uio, mp0, controlp, flagsp) return (EOPNOTSUPP); } int -pru_soreceive_notsupp(__unused struct socket *so, - __unused struct sockaddr **paddr, - __unused struct uio *uio, __unused struct mbuf **mp0, - __unused struct mbuf **controlp, __unused int *flagsp) +pru_soreceive_list_notsupp(struct socket *so, + struct recv_msg_elem *recv_msg_array, u_int uiocnt, int *flagsp) { +#pragma unused(so, recv_msg_array, uiocnt, flagsp) return (EOPNOTSUPP); } int -pru_shutdown_notsupp(__unused struct socket *so) +pru_shutdown_notsupp(struct socket *so) { +#pragma unused(so) return (EOPNOTSUPP); } int -pru_sockaddr_notsupp(__unused struct socket *so, __unused struct sockaddr **nam) +pru_sockaddr_notsupp(struct socket *so, struct sockaddr **nam) { +#pragma unused(so, nam) return (EOPNOTSUPP); } int -pru_sopoll_notsupp(__unused struct socket *so, __unused int events, - __unused kauth_cred_t cred, __unused void *wql) +pru_sopoll_notsupp(struct socket *so, int events, kauth_cred_t cred, void *wql) { +#pragma unused(so, events, cred, wql) return (EOPNOTSUPP); } +int +pru_socheckopt_null(struct socket *so, struct sockopt *sopt) +{ +#pragma unused(so, sopt) + /* + * Allow all options for set/get by default. + */ + return (0); +} + +static int +pru_preconnect_null(struct socket *so) +{ +#pragma unused(so) + return (0); +} + +void +pru_sanitize(struct pr_usrreqs *pru) +{ +#define DEFAULT(foo, bar) if ((foo) == NULL) (foo) = (bar) + DEFAULT(pru->pru_abort, pru_abort_notsupp); + DEFAULT(pru->pru_accept, pru_accept_notsupp); + DEFAULT(pru->pru_attach, pru_attach_notsupp); + DEFAULT(pru->pru_bind, pru_bind_notsupp); + DEFAULT(pru->pru_connect, pru_connect_notsupp); + DEFAULT(pru->pru_connect2, pru_connect2_notsupp); + DEFAULT(pru->pru_connectx, pru_connectx_notsupp); + DEFAULT(pru->pru_control, pru_control_notsupp); + DEFAULT(pru->pru_detach, pru_detach_notsupp); + DEFAULT(pru->pru_disconnect, pru_disconnect_notsupp); + DEFAULT(pru->pru_disconnectx, pru_disconnectx_notsupp); + DEFAULT(pru->pru_listen, pru_listen_notsupp); + DEFAULT(pru->pru_peeloff, pru_peeloff_notsupp); + DEFAULT(pru->pru_peeraddr, pru_peeraddr_notsupp); + DEFAULT(pru->pru_rcvd, pru_rcvd_notsupp); + DEFAULT(pru->pru_rcvoob, pru_rcvoob_notsupp); + DEFAULT(pru->pru_send, pru_send_notsupp); + DEFAULT(pru->pru_send_list, pru_send_list_notsupp); + DEFAULT(pru->pru_sense, pru_sense_null); + DEFAULT(pru->pru_shutdown, pru_shutdown_notsupp); + DEFAULT(pru->pru_sockaddr, pru_sockaddr_notsupp); + DEFAULT(pru->pru_sopoll, pru_sopoll_notsupp); + DEFAULT(pru->pru_soreceive, pru_soreceive_notsupp); + DEFAULT(pru->pru_soreceive_list, pru_soreceive_list_notsupp); + DEFAULT(pru->pru_sosend, pru_sosend_notsupp); + DEFAULT(pru->pru_sosend_list, pru_sosend_list_notsupp); + DEFAULT(pru->pru_socheckopt, pru_socheckopt_null); + DEFAULT(pru->pru_preconnect, pru_preconnect_null); +#undef DEFAULT +} -#ifdef __APPLE__ /* * The following are macros on BSD and functions on Darwin */ @@ -1561,21 +2288,64 @@ pru_sopoll_notsupp(__unused struct socket *so, __unused int events, int sb_notify(struct sockbuf *sb) { - return ((sb->sb_flags & - (SB_WAIT|SB_SEL|SB_ASYNC|SB_UPCALL|SB_KNOTE)) != 0); + return (sb->sb_waiters > 0 || + (sb->sb_flags & (SB_SEL|SB_ASYNC|SB_UPCALL|SB_KNOTE))); } /* * How much space is there in a socket buffer (so->so_snd or so->so_rcv)? * This is problematical if the fields are unsigned, as the space might * still be negative (cc > hiwat or mbcnt > mbmax). Should detect - * overflow and return 0. Should use "lmin" but it doesn't exist now. + * overflow and return 0. */ -long +int sbspace(struct sockbuf *sb) { - return ((long)imin((int)(sb->sb_hiwat - sb->sb_cc), - (int)(sb->sb_mbmax - sb->sb_mbcnt))); + int pending = 0; + int space = imin((int)(sb->sb_hiwat - sb->sb_cc), + (int)(sb->sb_mbmax - sb->sb_mbcnt)); + + if (sb->sb_preconn_hiwat != 0) + space = imin((int)(sb->sb_preconn_hiwat - sb->sb_cc), space); + + if (space < 0) + space = 0; + + /* Compensate for data being processed by content filters */ +#if CONTENT_FILTER + pending = cfil_sock_data_space(sb); +#endif /* CONTENT_FILTER */ + if (pending > space) + space = 0; + else + space -= pending; + + return (space); +} + +/* + * If this socket has priority queues, check if there is enough + * space in the priority queue for this msg. + */ +int +msgq_sbspace(struct socket *so, struct mbuf *control) +{ + int space = 0, error; + u_int32_t msgpri; + VERIFY(so->so_type == SOCK_STREAM && + SOCK_PROTO(so) == IPPROTO_TCP); + if (control != NULL) { + error = tcp_get_msg_priority(control, &msgpri); + if (error) + return (0); + } else { + msgpri = MSG_PRI_0; + } + space = (so->so_snd.sb_idealsize / MSG_PRI_COUNT) - + so->so_msg_state->msg_priq[msgpri].msgq_bytes; + if (space < 0) + space = 0; + return (space); } /* do we have to send all at once on a socket? */ @@ -1590,7 +2360,11 @@ int soreadable(struct socket *so) { return (so->so_rcv.sb_cc >= so->so_rcv.sb_lowat || - (so->so_state & SS_CANTRCVMORE) || + ((so->so_state & SS_CANTRCVMORE) +#if CONTENT_FILTER + && cfil_sock_data_pending(&so->so_rcv) == 0 +#endif /* CONTENT_FILTER */ + ) || so->so_comp.tqh_first || so->so_error); } @@ -1599,11 +2373,35 @@ soreadable(struct socket *so) int sowriteable(struct socket *so) { - return ((sbspace(&(so)->so_snd) >= (long)(so)->so_snd.sb_lowat && - ((so->so_state&SS_ISCONNECTED) || - (so->so_proto->pr_flags&PR_CONNREQUIRED) == 0)) || - (so->so_state & SS_CANTSENDMORE) || - so->so_error); + if ((so->so_state & SS_CANTSENDMORE) || + so->so_error > 0) + return (1); + if (so_wait_for_if_feedback(so) || !socanwrite(so)) + return (0); + if (so->so_flags1 & SOF1_PRECONNECT_DATA) + return(1); + + if (sbspace(&(so)->so_snd) >= (so)->so_snd.sb_lowat) { + if (so->so_flags & SOF_NOTSENT_LOWAT) { + if ((SOCK_DOM(so) == PF_INET6 || + SOCK_DOM(so) == PF_INET) && + so->so_type == SOCK_STREAM) { + return (tcp_notsent_lowat_check(so)); + } +#if MPTCP + else if ((SOCK_DOM(so) == PF_MULTIPATH) && + (SOCK_PROTO(so) == IPPROTO_TCP)) { + return (mptcp_notsent_lowat_check(so)); + } +#endif + else { + return (1); + } + } else { + return (1); + } + } + return (0); } /* adjust counters in sb reflecting allocation of m */ @@ -1611,18 +2409,28 @@ sowriteable(struct socket *so) void sballoc(struct sockbuf *sb, struct mbuf *m) { - int cnt = 1; - sb->sb_cc += m->m_len; - if (m->m_type != MT_DATA && m->m_type != MT_HEADER && - m->m_type != MT_OOBDATA) + u_int32_t cnt = 1; + sb->sb_cc += m->m_len; + if (m->m_type != MT_DATA && m->m_type != MT_HEADER && + m->m_type != MT_OOBDATA) sb->sb_ctl += m->m_len; - sb->sb_mbcnt += MSIZE; - + sb->sb_mbcnt += MSIZE; + if (m->m_flags & M_EXT) { - sb->sb_mbcnt += m->m_ext.ext_size; - cnt += m->m_ext.ext_size / MSIZE ; + sb->sb_mbcnt += m->m_ext.ext_size; + cnt += (m->m_ext.ext_size >> MSIZESHIFT); } - OSAddAtomic(cnt, (SInt32*)&total_mb_cnt); + OSAddAtomic(cnt, &total_sbmb_cnt); + VERIFY(total_sbmb_cnt > 0); + if (total_sbmb_cnt > total_sbmb_cnt_peak) + total_sbmb_cnt_peak = total_sbmb_cnt; + + /* + * If data is being appended to the send socket buffer, + * update the send byte count + */ + if (!(sb->sb_flags & SB_RECV)) + OSAddAtomic(cnt, &total_snd_byte_count); } /* adjust counters in sb reflecting freeing of m */ @@ -1630,76 +2438,201 @@ void sbfree(struct sockbuf *sb, struct mbuf *m) { int cnt = -1; + sb->sb_cc -= m->m_len; - if (m->m_type != MT_DATA && m->m_type != MT_HEADER && - m->m_type != MT_OOBDATA) + if (m->m_type != MT_DATA && m->m_type != MT_HEADER && + m->m_type != MT_OOBDATA) sb->sb_ctl -= m->m_len; - sb->sb_mbcnt -= MSIZE; + sb->sb_mbcnt -= MSIZE; if (m->m_flags & M_EXT) { - sb->sb_mbcnt -= m->m_ext.ext_size; - cnt -= m->m_ext.ext_size / MSIZE ; + sb->sb_mbcnt -= m->m_ext.ext_size; + cnt -= (m->m_ext.ext_size >> MSIZESHIFT); + } + OSAddAtomic(cnt, &total_sbmb_cnt); + VERIFY(total_sbmb_cnt >= 0); + + /* + * If data is being removed from the send socket buffer, + * update the send byte count + */ + if (!(sb->sb_flags & SB_RECV)) { + OSAddAtomic(cnt, &total_snd_byte_count); } - OSAddAtomic(cnt, (SInt32*)&total_mb_cnt); } /* * Set lock on sockbuf sb; sleep if lock is already held. * Unless SB_NOINTR is set on sockbuf, sleep is interruptible. * Returns error without lock if sleep is interrupted. - * - * Returns: 0 Success - * EWOULDBLOCK - * sb_lock:EINTR */ int -sblock(struct sockbuf *sb, int wf) +sblock(struct sockbuf *sb, uint32_t flags) { + boolean_t nointr = ((sb->sb_flags & SB_NOINTR) || (flags & SBL_NOINTR)); + void *lr_saved = __builtin_return_address(0); + struct socket *so = sb->sb_so; + void * wchan; int error = 0; + thread_t tp = current_thread(); - if (sb->sb_flags & SB_LOCK) - error = (wf == M_WAIT) ? sb_lock(sb) : EWOULDBLOCK; - else - sb->sb_flags |= SB_LOCK; + VERIFY((flags & SBL_VALID) == flags); - return (error); + /* so_usecount may be 0 if we get here from sofreelastref() */ + if (so == NULL) { + panic("%s: null so, sb=%p sb_flags=0x%x lr=%p\n", + __func__, sb, sb->sb_flags, lr_saved); + /* NOTREACHED */ + } else if (so->so_usecount < 0) { + panic("%s: sb=%p sb_flags=0x%x sb_so=%p usecount=%d lr=%p " + "lrh= %s\n", __func__, sb, sb->sb_flags, so, + so->so_usecount, lr_saved, solockhistory_nr(so)); + /* NOTREACHED */ + } + + /* + * The content filter thread must hold the sockbuf lock + */ + if ((so->so_flags & SOF_CONTENT_FILTER) && sb->sb_cfil_thread == tp) { + /* + * Don't panic if we are defunct because SB_LOCK has + * been cleared by sodefunct() + */ + if (!(so->so_flags & SOF_DEFUNCT) && !(sb->sb_flags & SB_LOCK)) + panic("%s: SB_LOCK not held for %p\n", + __func__, sb); + + /* Keep the sockbuf locked */ + return (0); + } + + if ((sb->sb_flags & SB_LOCK) && !(flags & SBL_WAIT)) + return (EWOULDBLOCK); + /* + * We may get here from sorflush(), in which case "sb" may not + * point to the real socket buffer. Use the actual socket buffer + * address from the socket instead. + */ + wchan = (sb->sb_flags & SB_RECV) ? + &so->so_rcv.sb_flags : &so->so_snd.sb_flags; + + /* + * A content filter thread has exclusive access to the sockbuf + * until it clears the + */ + while ((sb->sb_flags & SB_LOCK) || + ((so->so_flags & SOF_CONTENT_FILTER) && + sb->sb_cfil_thread != NULL)) { + lck_mtx_t *mutex_held; + + /* + * XXX: This code should be moved up above outside of this loop; + * however, we may get here as part of sofreelastref(), and + * at that time pr_getlock() may no longer be able to return + * us the lock. This will be fixed in future. + */ + if (so->so_proto->pr_getlock != NULL) + mutex_held = (*so->so_proto->pr_getlock)(so, 0); + else + mutex_held = so->so_proto->pr_domain->dom_mtx; + + lck_mtx_assert(mutex_held, LCK_MTX_ASSERT_OWNED); + + sb->sb_wantlock++; + VERIFY(sb->sb_wantlock != 0); + + error = msleep(wchan, mutex_held, + nointr ? PSOCK : PSOCK | PCATCH, + nointr ? "sb_lock_nointr" : "sb_lock", NULL); + + VERIFY(sb->sb_wantlock != 0); + sb->sb_wantlock--; + + if (error == 0 && (so->so_flags & SOF_DEFUNCT) && + !(flags & SBL_IGNDEFUNCT)) { + error = EBADF; + SODEFUNCTLOG(("%s[%d]: defunct so 0x%llx [%d,%d] " + "(%d)\n", __func__, proc_selfpid(), + (uint64_t)VM_KERNEL_ADDRPERM(so), + SOCK_DOM(so), SOCK_TYPE(so), error)); + } + + if (error != 0) + return (error); + } + sb->sb_flags |= SB_LOCK; + return (0); } -/* release lock on sockbuf sb */ +/* + * Release lock on sockbuf sb + */ void -sbunlock(struct sockbuf *sb, int keeplocked) +sbunlock(struct sockbuf *sb, boolean_t keeplocked) { + void *lr_saved = __builtin_return_address(0); struct socket *so = sb->sb_so; - int lr_saved; - lck_mtx_t *mutex_held; + thread_t tp = current_thread(); - lr_saved = (unsigned int) __builtin_return_address(0); + /* so_usecount may be 0 if we get here from sofreelastref() */ + if (so == NULL) { + panic("%s: null so, sb=%p sb_flags=0x%x lr=%p\n", + __func__, sb, sb->sb_flags, lr_saved); + /* NOTREACHED */ + } else if (so->so_usecount < 0) { + panic("%s: sb=%p sb_flags=0x%x sb_so=%p usecount=%d lr=%p " + "lrh= %s\n", __func__, sb, sb->sb_flags, so, + so->so_usecount, lr_saved, solockhistory_nr(so)); + /* NOTREACHED */ + } - sb->sb_flags &= ~SB_LOCK; + /* + * The content filter thread must hold the sockbuf lock + */ + if ((so->so_flags & SOF_CONTENT_FILTER) && sb->sb_cfil_thread == tp) { + /* + * Don't panic if we are defunct because SB_LOCK has + * been cleared by sodefunct() + */ + if (!(so->so_flags & SOF_DEFUNCT) && + !(sb->sb_flags & SB_LOCK) && + !(so->so_state & SS_DEFUNCT) && + !(so->so_flags1 & SOF1_DEFUNCTINPROG)) { + panic("%s: SB_LOCK not held for %p\n", + __func__, sb); + } + /* Keep the sockbuf locked and proceed */ + } else { + VERIFY((sb->sb_flags & SB_LOCK) || + (so->so_state & SS_DEFUNCT) || + (so->so_flags1 & SOF1_DEFUNCTINPROG)); + + sb->sb_flags &= ~SB_LOCK; + + if (sb->sb_wantlock > 0) { + /* + * We may get here from sorflush(), in which case "sb" + * may not point to the real socket buffer. Use the + * actual socket buffer address from the socket instead. + */ + wakeup((sb->sb_flags & SB_RECV) ? &so->so_rcv.sb_flags : + &so->so_snd.sb_flags); + } + } - if (sb->sb_flags & SB_WANT) { - sb->sb_flags &= ~SB_WANT; - if (so->so_usecount < 0) - panic("sbunlock: b4 wakeup so=%p ref=%d lr=%x " - "sb_flags=%x\n", sb->sb_so, so->so_usecount, - lr_saved, sb->sb_flags); + if (!keeplocked) { /* unlock on exit */ + lck_mtx_t *mutex_held; - wakeup((caddr_t)&(sb)->sb_flags); - } - if (keeplocked == 0) { /* unlock on exit */ - if (so->so_proto->pr_getlock != NULL) + if (so->so_proto->pr_getlock != NULL) mutex_held = (*so->so_proto->pr_getlock)(so, 0); - else + else mutex_held = so->so_proto->pr_domain->dom_mtx; - + lck_mtx_assert(mutex_held, LCK_MTX_ASSERT_OWNED); + VERIFY(so->so_usecount != 0); so->so_usecount--; - if (so->so_usecount < 0) - panic("sbunlock: unlock on exit so=%p ref=%d lr=%x " - "sb_flags=%x\n", so, so->so_usecount, lr_saved, - sb->sb_flags); - so->unlock_lr[so->next_unlock_lr] = (u_int32_t)lr_saved; - so->next_unlock_lr = (so->next_unlock_lr+1) % SO_LCKDBG_MAX; + so->unlock_lr[so->next_unlock_lr] = lr_saved; + so->next_unlock_lr = (so->next_unlock_lr + 1) % SO_LCKDBG_MAX; lck_mtx_unlock(mutex_held); } } @@ -1717,7 +2650,103 @@ sowwakeup(struct socket *so) if (sb_notify(&so->so_snd)) sowakeup(so, &so->so_snd); } -#endif /* __APPLE__ */ + +void +soevent(struct socket *so, long hint) +{ + if (so->so_flags & SOF_KNOTE) + KNOTE(&so->so_klist, hint); + + soevupcall(so, hint); + + /* + * Don't post an event if this a subflow socket or + * the app has opted out of using cellular interface + */ + if ((hint & SO_FILT_HINT_IFDENIED) && + !(so->so_flags & SOF_MP_SUBFLOW) && + !(so->so_restrictions & SO_RESTRICT_DENY_CELLULAR) && + !(so->so_restrictions & SO_RESTRICT_DENY_EXPENSIVE)) + soevent_ifdenied(so); +} + +void +soevupcall(struct socket *so, u_int32_t hint) +{ + if (so->so_event != NULL) { + caddr_t so_eventarg = so->so_eventarg; + int locked = hint & SO_FILT_HINT_LOCKED; + + hint &= so->so_eventmask; + if (hint != 0) { + if (locked) + socket_unlock(so, 0); + + so->so_event(so, so_eventarg, hint); + + if (locked) + socket_lock(so, 0); + } + } +} + +static void +soevent_ifdenied(struct socket *so) +{ + struct kev_netpolicy_ifdenied ev_ifdenied; + + bzero(&ev_ifdenied, sizeof (ev_ifdenied)); + /* + * The event consumer is interested about the effective {upid,pid,uuid} + * info which can be different than the those related to the process + * that recently performed a system call on the socket, i.e. when the + * socket is delegated. + */ + if (so->so_flags & SOF_DELEGATED) { + ev_ifdenied.ev_data.eupid = so->e_upid; + ev_ifdenied.ev_data.epid = so->e_pid; + uuid_copy(ev_ifdenied.ev_data.euuid, so->e_uuid); + } else { + ev_ifdenied.ev_data.eupid = so->last_upid; + ev_ifdenied.ev_data.epid = so->last_pid; + uuid_copy(ev_ifdenied.ev_data.euuid, so->last_uuid); + } + + if (++so->so_ifdenied_notifies > 1) { + /* + * Allow for at most one kernel event to be generated per + * socket; so_ifdenied_notifies is reset upon changes in + * the UUID policy. See comments in inp_update_policy. + */ + if (net_io_policy_log) { + uuid_string_t buf; + + uuid_unparse(ev_ifdenied.ev_data.euuid, buf); + log(LOG_DEBUG, "%s[%d]: so 0x%llx [%d,%d] epid %d " + "euuid %s%s has %d redundant events supressed\n", + __func__, so->last_pid, + (uint64_t)VM_KERNEL_ADDRPERM(so), SOCK_DOM(so), + SOCK_TYPE(so), ev_ifdenied.ev_data.epid, buf, + ((so->so_flags & SOF_DELEGATED) ? + " [delegated]" : ""), so->so_ifdenied_notifies); + } + } else { + if (net_io_policy_log) { + uuid_string_t buf; + + uuid_unparse(ev_ifdenied.ev_data.euuid, buf); + log(LOG_DEBUG, "%s[%d]: so 0x%llx [%d,%d] epid %d " + "euuid %s%s event posted\n", __func__, + so->last_pid, (uint64_t)VM_KERNEL_ADDRPERM(so), + SOCK_DOM(so), SOCK_TYPE(so), + ev_ifdenied.ev_data.epid, buf, + ((so->so_flags & SOF_DELEGATED) ? + " [delegated]" : "")); + } + netpolicy_post_msg(KEV_NETPOLICY_IFDENIED, &ev_ifdenied.ev_data, + sizeof (ev_ifdenied)); + } +} /* * Make a copy of a sockaddr in a malloced buffer of type M_SONAME. @@ -1746,15 +2775,45 @@ void sotoxsocket(struct socket *so, struct xsocket *xso) { xso->xso_len = sizeof (*xso); - xso->xso_so = so; + xso->xso_so = (_XSOCKET_PTR(struct socket *))VM_KERNEL_ADDRPERM(so); + xso->so_type = so->so_type; + xso->so_options = (short)(so->so_options & 0xffff); + xso->so_linger = so->so_linger; + xso->so_state = so->so_state; + xso->so_pcb = (_XSOCKET_PTR(caddr_t))VM_KERNEL_ADDRPERM(so->so_pcb); + if (so->so_proto) { + xso->xso_protocol = SOCK_PROTO(so); + xso->xso_family = SOCK_DOM(so); + } else { + xso->xso_protocol = xso->xso_family = 0; + } + xso->so_qlen = so->so_qlen; + xso->so_incqlen = so->so_incqlen; + xso->so_qlimit = so->so_qlimit; + xso->so_timeo = so->so_timeo; + xso->so_error = so->so_error; + xso->so_pgid = so->so_pgid; + xso->so_oobmark = so->so_oobmark; + sbtoxsockbuf(&so->so_snd, &xso->so_snd); + sbtoxsockbuf(&so->so_rcv, &xso->so_rcv); + xso->so_uid = kauth_cred_getuid(so->so_cred); +} + + + +void +sotoxsocket64(struct socket *so, struct xsocket64 *xso) +{ + xso->xso_len = sizeof (*xso); + xso->xso_so = (u_int64_t)VM_KERNEL_ADDRPERM(so); xso->so_type = so->so_type; - xso->so_options = so->so_options; + xso->so_options = (short)(so->so_options & 0xffff); xso->so_linger = so->so_linger; xso->so_state = so->so_state; - xso->so_pcb = so->so_pcb; + xso->so_pcb = (u_int64_t)VM_KERNEL_ADDRPERM(so->so_pcb); if (so->so_proto) { - xso->xso_protocol = so->so_proto->pr_protocol; - xso->xso_family = so->so_proto->pr_domain->dom_family; + xso->xso_protocol = SOCK_PROTO(so); + xso->xso_family = SOCK_DOM(so); } else { xso->xso_protocol = xso->xso_family = 0; } @@ -1767,9 +2826,10 @@ sotoxsocket(struct socket *so, struct xsocket *xso) xso->so_oobmark = so->so_oobmark; sbtoxsockbuf(&so->so_snd, &xso->so_snd); sbtoxsockbuf(&so->so_rcv, &xso->so_rcv); - xso->so_uid = so->so_uid; + xso->so_uid = kauth_cred_getuid(so->so_cred); } + /* * This does the same for sockbufs. Note that the xsockbuf structure, * since it is always embedded in a socket, does not include a self @@ -1785,35 +2845,142 @@ sbtoxsockbuf(struct sockbuf *sb, struct xsockbuf *xsb) xsb->sb_mbmax = sb->sb_mbmax; xsb->sb_lowat = sb->sb_lowat; xsb->sb_flags = sb->sb_flags; - xsb->sb_timeo = (u_long) + xsb->sb_timeo = (short) (sb->sb_timeo.tv_sec * hz) + sb->sb_timeo.tv_usec / tick; if (xsb->sb_timeo == 0 && sb->sb_timeo.tv_usec != 0) xsb->sb_timeo = 1; } +/* + * Based on the policy set by an all knowing decison maker, throttle sockets + * that either have been marked as belonging to "background" process. + */ +inline int +soisthrottled(struct socket *so) +{ + /* + * On non-embedded, we rely on implicit throttling by the + * application, as we're missing the system wide "decision maker" + */ + return ( + (so->so_traffic_mgt_flags & TRAFFIC_MGT_SO_BACKGROUND)); +} + +inline int +soisprivilegedtraffic(struct socket *so) +{ + return ((so->so_flags & SOF_PRIVILEGED_TRAFFIC_CLASS) ? 1 : 0); +} + +inline int +soissrcbackground(struct socket *so) +{ + return ((so->so_traffic_mgt_flags & TRAFFIC_MGT_SO_BACKGROUND) || + IS_SO_TC_BACKGROUND(so->so_traffic_class)); +} + +inline int +soissrcrealtime(struct socket *so) +{ + return (so->so_traffic_class >= SO_TC_AV && + so->so_traffic_class <= SO_TC_VO); +} + +inline int +soissrcbesteffort(struct socket *so) +{ + return (so->so_traffic_class == SO_TC_BE || + so->so_traffic_class == SO_TC_RD || + so->so_traffic_class == SO_TC_OAM); +} + +void +sonullevent(struct socket *so, void *arg, uint32_t hint) +{ +#pragma unused(so, arg, hint) +} + /* * Here is the definition of some of the basic objects in the kern.ipc * branch of the MIB. */ -SYSCTL_NODE(_kern, KERN_IPC, ipc, CTLFLAG_RW|CTLFLAG_LOCKED, 0, "IPC"); - -/* This takes the place of kern.maxsockbuf, which moved to kern.ipc. */ -static int dummy; -SYSCTL_INT(_kern, KERN_DUMMY, dummy, CTLFLAG_RW, &dummy, 0, ""); - -SYSCTL_INT(_kern_ipc, KIPC_MAXSOCKBUF, maxsockbuf, CTLFLAG_RW, - &sb_max, 0, "Maximum socket buffer size"); -SYSCTL_INT(_kern_ipc, OID_AUTO, maxsockets, CTLFLAG_RD, - &maxsockets, 0, "Maximum number of sockets avaliable"); -SYSCTL_INT(_kern_ipc, KIPC_SOCKBUF_WASTE, sockbuf_waste_factor, CTLFLAG_RW, - &sb_efficiency, 0, ""); -SYSCTL_INT(_kern_ipc, OID_AUTO, sbspace_factor, CTLFLAG_RW, - &sbspace_factor, 0, "Ratio of mbuf/cluster use for socket layers"); -SYSCTL_INT(_kern_ipc, KIPC_NMBCLUSTERS, nmbclusters, CTLFLAG_RD, - &nmbclusters, 0, ""); -SYSCTL_INT(_kern_ipc, OID_AUTO, njcl, CTLFLAG_RD, &njcl, 0, ""); -SYSCTL_INT(_kern_ipc, OID_AUTO, njclbytes, CTLFLAG_RD, &njclbytes, 0, ""); -SYSCTL_INT(_kern_ipc, KIPC_SOQLIMITCOMPAT, soqlimitcompat, CTLFLAG_RW, - &soqlimitcompat, 1, "Enable socket queue limit compatibility"); -SYSCTL_INT(_kern_ipc, OID_AUTO, soqlencomp, CTLFLAG_RW, - &soqlencomp, 0, "Listen backlog represents only complete queue"); +SYSCTL_NODE(_kern, KERN_IPC, ipc, + CTLFLAG_RW|CTLFLAG_LOCKED|CTLFLAG_ANYBODY, 0, "IPC"); + +/* Check that the maximum socket buffer size is within a range */ + +static int +sysctl_sb_max SYSCTL_HANDLER_ARGS +{ +#pragma unused(oidp, arg1, arg2) + u_int32_t new_value; + int changed = 0; + int error = sysctl_io_number(req, sb_max, sizeof (u_int32_t), + &new_value, &changed); + if (!error && changed) { + if (new_value > LOW_SB_MAX && new_value <= high_sb_max) { + sb_max = new_value; + } else { + error = ERANGE; + } + } + return (error); +} + +static int +sysctl_io_policy_throttled SYSCTL_HANDLER_ARGS +{ +#pragma unused(arg1, arg2) + int i, err; + + i = net_io_policy_throttled; + + err = sysctl_handle_int(oidp, &i, 0, req); + if (err != 0 || req->newptr == USER_ADDR_NULL) + return (err); + + if (i != net_io_policy_throttled) + SOTHROTTLELOG(("throttle: network IO policy throttling is " + "now %s\n", i ? "ON" : "OFF")); + + net_io_policy_throttled = i; + + return (err); +} + +SYSCTL_PROC(_kern_ipc, KIPC_MAXSOCKBUF, maxsockbuf, + CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_LOCKED, + &sb_max, 0, &sysctl_sb_max, "IU", "Maximum socket buffer size"); + +SYSCTL_INT(_kern_ipc, KIPC_SOCKBUF_WASTE, sockbuf_waste_factor, + CTLFLAG_RW | CTLFLAG_LOCKED, &sb_efficiency, 0, ""); + +SYSCTL_INT(_kern_ipc, KIPC_NMBCLUSTERS, nmbclusters, + CTLFLAG_RD | CTLFLAG_LOCKED, &nmbclusters, 0, ""); + +SYSCTL_INT(_kern_ipc, OID_AUTO, njcl, + CTLFLAG_RD | CTLFLAG_LOCKED, &njcl, 0, ""); + +SYSCTL_INT(_kern_ipc, OID_AUTO, njclbytes, + CTLFLAG_RD | CTLFLAG_LOCKED, &njclbytes, 0, ""); + +SYSCTL_INT(_kern_ipc, KIPC_SOQLIMITCOMPAT, soqlimitcompat, + CTLFLAG_RW | CTLFLAG_LOCKED, &soqlimitcompat, 1, + "Enable socket queue limit compatibility"); + +SYSCTL_INT(_kern_ipc, OID_AUTO, soqlencomp, CTLFLAG_RW | CTLFLAG_LOCKED, + &soqlencomp, 0, "Listen backlog represents only complete queue"); + +SYSCTL_NODE(_kern_ipc, OID_AUTO, io_policy, CTLFLAG_RW, 0, "network IO policy"); + +SYSCTL_PROC(_kern_ipc_io_policy, OID_AUTO, throttled, + CTLTYPE_INT | CTLFLAG_RW | CTLFLAG_LOCKED, &net_io_policy_throttled, 0, + sysctl_io_policy_throttled, "I", ""); + +SYSCTL_INT(_kern_ipc_io_policy, OID_AUTO, log, CTLFLAG_RW | CTLFLAG_LOCKED, + &net_io_policy_log, 0, ""); + +#if CONFIG_PROC_UUID_POLICY +SYSCTL_INT(_kern_ipc_io_policy, OID_AUTO, uuid, CTLFLAG_RW | CTLFLAG_LOCKED, + &net_io_policy_uuid, 0, ""); +#endif /* CONFIG_PROC_UUID_POLICY */