/*- * Copyright (c) 2016-2018 Netflix Inc. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * */ #include __FBSDID("$FreeBSD$"); #include "opt_inet.h" #include "opt_inet6.h" #include "opt_ipsec.h" #include "opt_tcpdebug.h" /** * Some notes about usage. * * The tcp_hpts system is designed to provide a high precision timer * system for tcp. Its main purpose is to provide a mechanism for * pacing packets out onto the wire. It can be used in two ways * by a given TCP stack (and those two methods can be used simultaneously). * * First, and probably the main thing its used by Rack and BBR for, it can * be used to call tcp_output() of a transport stack at some time in the future. * The normal way this is done is that tcp_output() of the stack schedules * itself to be called again by calling tcp_hpts_insert(tcpcb, slot). The * slot is the time from now that the stack wants to be called but it * must be converted to tcp_hpts's notion of slot. This is done with * one of the macros HPTS_MS_TO_SLOTS or HPTS_USEC_TO_SLOTS. So a typical * call from the tcp_output() routine might look like: * * tcp_hpts_insert(tp, HPTS_USEC_TO_SLOTS(550)); * * The above would schedule tcp_ouput() to be called in 550 useconds. * Note that if using this mechanism the stack will want to add near * its top a check to prevent unwanted calls (from user land or the * arrival of incoming ack's). So it would add something like: * * if (inp->inp_in_hpts) * return; * * to prevent output processing until the time alotted has gone by. * Of course this is a bare bones example and the stack will probably * have more consideration then just the above. * * Now the tcp_hpts system will call tcp_output in one of two forms, * it will first check to see if the stack as defined a * tfb_tcp_output_wtime() function, if so that is the routine it * will call, if that function is not defined then it will call the * tfb_tcp_output() function. The only difference between these * two calls is that the former passes the time in to the function * so the function does not have to access the time (which tcp_hpts * already has). What these functions do is of course totally up * to the individual tcp stack. * * Now the second function (actually two functions I guess :D) * the tcp_hpts system provides is the ability to either abort * a connection (later) or process input on a connection. * Why would you want to do this? To keep processor locality. * * So in order to use the input redirection function the * stack changes its tcp_do_segment() routine to instead * of process the data call the function: * * tcp_queue_pkt_to_input() * * You will note that the arguments to this function look * a lot like tcp_do_segments's arguments. This function * will assure that the tcp_hpts system will * call the functions tfb_tcp_hpts_do_segment() from the * correct CPU. Note that multiple calls can get pushed * into the tcp_hpts system this will be indicated by * the next to last argument to tfb_tcp_hpts_do_segment() * (nxt_pkt). If nxt_pkt is a 1 then another packet is * coming. If nxt_pkt is a 0 then this is the last call * that the tcp_hpts system has available for the tcp stack. * * The other point of the input system is to be able to safely * drop a tcp connection without worrying about the recursive * locking that may be occuring on the INP_WLOCK. So if * a stack wants to drop a connection it calls: * * tcp_set_inp_to_drop(tp, ETIMEDOUT) * * To schedule the tcp_hpts system to call * * tcp_drop(tp, drop_reason) * * at a future point. This is quite handy to prevent locking * issues when dropping connections. * */ #include #include #include #include #include #include #include #include #include /* for proc0 declaration */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #define TCPSTATES /* for logging */ #include #include #include #include #include /* required for icmp_var.h */ #include /* for ICMP_BANDLIM */ #include #include #include #include #define TCPOUTFLAGS #include #include #include #include #include #include #include #include #ifdef tcpdebug #include #endif /* tcpdebug */ #ifdef tcp_offload #include #endif #ifdef ipsec #include #include #endif /* ipsec */ #include "opt_rss.h" MALLOC_DEFINE(M_TCPHPTS, "tcp_hpts", "TCP hpts"); #ifdef RSS static int tcp_bind_threads = 1; #else static int tcp_bind_threads = 0; #endif TUNABLE_INT("net.inet.tcp.bind_hptss", &tcp_bind_threads); static uint32_t tcp_hpts_logging_size = DEFAULT_HPTS_LOG; TUNABLE_INT("net.inet.tcp.hpts_logging_sz", &tcp_hpts_logging_size); static struct tcp_hptsi tcp_pace; static int tcp_hptsi_lock_inpinfo(struct inpcb *inp, struct tcpcb **tp); static void tcp_wakehpts(struct tcp_hpts_entry *p); static void tcp_wakeinput(struct tcp_hpts_entry *p); static void tcp_input_data(struct tcp_hpts_entry *hpts, struct timeval *tv); static void tcp_hptsi(struct tcp_hpts_entry *hpts, struct timeval *ctick); static void tcp_hpts_thread(void *ctx); static void tcp_init_hptsi(void *st); int32_t tcp_min_hptsi_time = DEFAULT_MIN_SLEEP; static int32_t tcp_hpts_callout_skip_swi = 0; SYSCTL_DECL(_net_inet_tcp); SYSCTL_NODE(_net_inet_tcp, OID_AUTO, hpts, CTLFLAG_RW, 0, "TCP Hpts controls"); #define timersub(tvp, uvp, vvp) \ do { \ (vvp)->tv_sec = (tvp)->tv_sec - (uvp)->tv_sec; \ (vvp)->tv_usec = (tvp)->tv_usec - (uvp)->tv_usec; \ if ((vvp)->tv_usec < 0) { \ (vvp)->tv_sec--; \ (vvp)->tv_usec += 1000000; \ } \ } while (0) static int32_t logging_on = 0; static int32_t hpts_sleep_max = (NUM_OF_HPTSI_SLOTS - 2); static int32_t tcp_hpts_precision = 120; SYSCTL_INT(_net_inet_tcp_hpts, OID_AUTO, precision, CTLFLAG_RW, &tcp_hpts_precision, 120, "Value for PRE() precision of callout"); SYSCTL_INT(_net_inet_tcp_hpts, OID_AUTO, logging, CTLFLAG_RW, &logging_on, 0, "Turn on logging if compiled in"); counter_u64_t hpts_loops; SYSCTL_COUNTER_U64(_net_inet_tcp_hpts, OID_AUTO, loops, CTLFLAG_RD, &hpts_loops, "Number of times hpts had to loop to catch up"); counter_u64_t back_tosleep; SYSCTL_COUNTER_U64(_net_inet_tcp_hpts, OID_AUTO, no_tcbsfound, CTLFLAG_RD, &back_tosleep, "Number of times hpts found no tcbs"); static int32_t in_newts_every_tcb = 0; SYSCTL_INT(_net_inet_tcp_hpts, OID_AUTO, in_tsperpcb, CTLFLAG_RW, &in_newts_every_tcb, 0, "Do we have a new cts every tcb we process for input"); static int32_t in_ts_percision = 0; SYSCTL_INT(_net_inet_tcp_hpts, OID_AUTO, in_tspercision, CTLFLAG_RW, &in_ts_percision, 0, "Do we use percise timestamp for clients on input"); static int32_t out_newts_every_tcb = 0; SYSCTL_INT(_net_inet_tcp_hpts, OID_AUTO, out_tsperpcb, CTLFLAG_RW, &out_newts_every_tcb, 0, "Do we have a new cts every tcb we process for output"); static int32_t out_ts_percision = 0; SYSCTL_INT(_net_inet_tcp_hpts, OID_AUTO, out_tspercision, CTLFLAG_RW, &out_ts_percision, 0, "Do we use a percise timestamp for every output cts"); SYSCTL_INT(_net_inet_tcp_hpts, OID_AUTO, maxsleep, CTLFLAG_RW, &hpts_sleep_max, 0, "The maximum time the hpts will sleep <1 - 254>"); SYSCTL_INT(_net_inet_tcp_hpts, OID_AUTO, minsleep, CTLFLAG_RW, &tcp_min_hptsi_time, 0, "The minimum time the hpts must sleep before processing more slots"); SYSCTL_INT(_net_inet_tcp_hpts, OID_AUTO, skip_swi, CTLFLAG_RW, &tcp_hpts_callout_skip_swi, 0, "Do we have the callout call directly to the hpts?"); static void __tcp_hpts_log_it(struct tcp_hpts_entry *hpts, struct inpcb *inp, int event, uint32_t slot, uint32_t ticknow, int32_t line) { struct hpts_log *pl; HPTS_MTX_ASSERT(hpts); if (hpts->p_log == NULL) return; pl = &hpts->p_log[hpts->p_log_at]; hpts->p_log_at++; if (hpts->p_log_at >= hpts->p_logsize) { hpts->p_log_at = 0; hpts->p_log_wrapped = 1; } pl->inp = inp; if (inp) { pl->t_paceslot = inp->inp_hptsslot; pl->t_hptsreq = inp->inp_hpts_request; pl->p_onhpts = inp->inp_in_hpts; pl->p_oninput = inp->inp_in_input; } else { pl->t_paceslot = 0; pl->t_hptsreq = 0; pl->p_onhpts = 0; pl->p_oninput = 0; } pl->is_notempty = 1; pl->event = event; pl->line = line; pl->cts = tcp_get_usecs(NULL); pl->p_curtick = hpts->p_curtick; pl->p_prevtick = hpts->p_prevtick; pl->p_on_queue_cnt = hpts->p_on_queue_cnt; pl->ticknow = ticknow; pl->slot_req = slot; pl->p_nxt_slot = hpts->p_nxt_slot; pl->p_cur_slot = hpts->p_cur_slot; pl->p_hpts_sleep_time = hpts->p_hpts_sleep_time; pl->p_flags = (hpts->p_cpu & 0x7f); pl->p_flags <<= 7; pl->p_flags |= (hpts->p_num & 0x7f); pl->p_flags <<= 2; if (hpts->p_hpts_active) { pl->p_flags |= HPTS_HPTS_ACTIVE; } } #define tcp_hpts_log_it(a, b, c, d, e) __tcp_hpts_log_it(a, b, c, d, e, __LINE__) static void hpts_timeout_swi(void *arg) { struct tcp_hpts_entry *hpts; hpts = (struct tcp_hpts_entry *)arg; swi_sched(hpts->ie_cookie, 0); } static void hpts_timeout_dir(void *arg) { tcp_hpts_thread(arg); } static inline void hpts_sane_pace_remove(struct tcp_hpts_entry *hpts, struct inpcb *inp, struct hptsh *head, int clear) { #ifdef INVARIANTS if (mtx_owned(&hpts->p_mtx) == 0) { /* We don't own the mutex? */ panic("%s: hpts:%p inp:%p no hpts mutex", __FUNCTION__, hpts, inp); } if (hpts->p_cpu != inp->inp_hpts_cpu) { /* It is not the right cpu/mutex? */ panic("%s: hpts:%p inp:%p incorrect CPU", __FUNCTION__, hpts, inp); } if (inp->inp_in_hpts == 0) { /* We are not on the hpts? */ panic("%s: hpts:%p inp:%p not on the hpts?", __FUNCTION__, hpts, inp); } if (TAILQ_EMPTY(head) && (hpts->p_on_queue_cnt != 0)) { /* We should not be empty with a queue count */ panic("%s hpts:%p hpts bucket empty but cnt:%d", __FUNCTION__, hpts, hpts->p_on_queue_cnt); } #endif TAILQ_REMOVE(head, inp, inp_hpts); hpts->p_on_queue_cnt--; if (hpts->p_on_queue_cnt < 0) { /* Count should not go negative .. */ #ifdef INVARIANTS panic("Hpts goes negative inp:%p hpts:%p", inp, hpts); #endif hpts->p_on_queue_cnt = 0; } if (clear) { inp->inp_hpts_request = 0; inp->inp_in_hpts = 0; } } static inline void hpts_sane_pace_insert(struct tcp_hpts_entry *hpts, struct inpcb *inp, struct hptsh *head, int line, int noref) { #ifdef INVARIANTS if (mtx_owned(&hpts->p_mtx) == 0) { /* We don't own the mutex? */ panic("%s: hpts:%p inp:%p no hpts mutex", __FUNCTION__, hpts, inp); } if (hpts->p_cpu != inp->inp_hpts_cpu) { /* It is not the right cpu/mutex? */ panic("%s: hpts:%p inp:%p incorrect CPU", __FUNCTION__, hpts, inp); } if ((noref == 0) && (inp->inp_in_hpts == 1)) { /* We are already on the hpts? */ panic("%s: hpts:%p inp:%p already on the hpts?", __FUNCTION__, hpts, inp); } #endif TAILQ_INSERT_TAIL(head, inp, inp_hpts); inp->inp_in_hpts = 1; hpts->p_on_queue_cnt++; if (noref == 0) { in_pcbref(inp); } } static inline void hpts_sane_input_remove(struct tcp_hpts_entry *hpts, struct inpcb *inp, int clear) { #ifdef INVARIANTS if (mtx_owned(&hpts->p_mtx) == 0) { /* We don't own the mutex? */ panic("%s: hpts:%p inp:%p no hpts mutex", __FUNCTION__, hpts, inp); } if (hpts->p_cpu != inp->inp_input_cpu) { /* It is not the right cpu/mutex? */ panic("%s: hpts:%p inp:%p incorrect CPU", __FUNCTION__, hpts, inp); } if (inp->inp_in_input == 0) { /* We are not on the input hpts? */ panic("%s: hpts:%p inp:%p not on the input hpts?", __FUNCTION__, hpts, inp); } #endif TAILQ_REMOVE(&hpts->p_input, inp, inp_input); hpts->p_on_inqueue_cnt--; if (hpts->p_on_inqueue_cnt < 0) { #ifdef INVARIANTS panic("Hpts in goes negative inp:%p hpts:%p", inp, hpts); #endif hpts->p_on_inqueue_cnt = 0; } #ifdef INVARIANTS if (TAILQ_EMPTY(&hpts->p_input) && (hpts->p_on_inqueue_cnt != 0)) { /* We should not be empty with a queue count */ panic("%s hpts:%p in_hpts input empty but cnt:%d", __FUNCTION__, hpts, hpts->p_on_inqueue_cnt); } #endif if (clear) inp->inp_in_input = 0; } static inline void hpts_sane_input_insert(struct tcp_hpts_entry *hpts, struct inpcb *inp, int line) { #ifdef INVARIANTS if (mtx_owned(&hpts->p_mtx) == 0) { /* We don't own the mutex? */ panic("%s: hpts:%p inp:%p no hpts mutex", __FUNCTION__, hpts, inp); } if (hpts->p_cpu != inp->inp_input_cpu) { /* It is not the right cpu/mutex? */ panic("%s: hpts:%p inp:%p incorrect CPU", __FUNCTION__, hpts, inp); } if (inp->inp_in_input == 1) { /* We are already on the input hpts? */ panic("%s: hpts:%p inp:%p already on the input hpts?", __FUNCTION__, hpts, inp); } #endif TAILQ_INSERT_TAIL(&hpts->p_input, inp, inp_input); inp->inp_in_input = 1; hpts->p_on_inqueue_cnt++; in_pcbref(inp); } static int sysctl_tcp_hpts_log(SYSCTL_HANDLER_ARGS) { struct tcp_hpts_entry *hpts; size_t sz; int32_t logging_was, i; int32_t error = 0; /* * HACK: Turn off logging so no locks are required this really needs * a memory barrier :) */ logging_was = logging_on; logging_on = 0; if (!req->oldptr) { /* How much? */ sz = 0; for (i = 0; i < tcp_pace.rp_num_hptss; i++) { hpts = tcp_pace.rp_ent[i]; if (hpts->p_log == NULL) continue; sz += (sizeof(struct hpts_log) * hpts->p_logsize); } error = SYSCTL_OUT(req, 0, sz); } else { for (i = 0; i < tcp_pace.rp_num_hptss; i++) { hpts = tcp_pace.rp_ent[i]; if (hpts->p_log == NULL) continue; if (hpts->p_log_wrapped) sz = (sizeof(struct hpts_log) * hpts->p_logsize); else sz = (sizeof(struct hpts_log) * hpts->p_log_at); error = SYSCTL_OUT(req, hpts->p_log, sz); } } logging_on = logging_was; return error; } SYSCTL_PROC(_net_inet_tcp_hpts, OID_AUTO, log, CTLTYPE_STRING | CTLFLAG_RD | CTLFLAG_MPSAFE, 0, 0, sysctl_tcp_hpts_log, "A", "tcp hptsi log"); /* * Try to get the INP_INFO lock. * * This function always succeeds in getting the lock. It will clear * *tpp and return (1) if something critical changed while the inpcb * was unlocked. Otherwise, it will leave *tpp unchanged and return (0). * * This function relies on the fact that the hpts always holds a * reference on the inpcb while the segment is on the hptsi wheel and * in the input queue. * */ static int tcp_hptsi_lock_inpinfo(struct inpcb *inp, struct tcpcb **tpp) { struct tcp_function_block *tfb; struct tcpcb *tp; void *ptr; /* Try the easy way. */ if (INP_INFO_TRY_RLOCK(&V_tcbinfo)) return (0); /* * OK, let's try the hard way. We'll save the function pointer block * to make sure that doesn't change while we aren't holding the * lock. */ tp = *tpp; tfb = tp->t_fb; ptr = tp->t_fb_ptr; INP_WUNLOCK(inp); INP_INFO_RLOCK(&V_tcbinfo); INP_WLOCK(inp); /* If the session went away, return an error. */ if ((inp->inp_flags & (INP_TIMEWAIT | INP_DROPPED)) || (inp->inp_flags2 & INP_FREED)) { *tpp = NULL; return (1); } /* * If the function block or stack-specific data block changed, * report an error. */ tp = intotcpcb(inp); if ((tp->t_fb != tfb) && (tp->t_fb_ptr != ptr)) { *tpp = NULL; return (1); } return (0); } static void tcp_wakehpts(struct tcp_hpts_entry *hpts) { HPTS_MTX_ASSERT(hpts); swi_sched(hpts->ie_cookie, 0); if (hpts->p_hpts_active == 2) { /* Rare sleeping on a ENOBUF */ wakeup_one(hpts); } } static void tcp_wakeinput(struct tcp_hpts_entry *hpts) { HPTS_MTX_ASSERT(hpts); swi_sched(hpts->ie_cookie, 0); if (hpts->p_hpts_active == 2) { /* Rare sleeping on a ENOBUF */ wakeup_one(hpts); } } struct tcp_hpts_entry * tcp_cur_hpts(struct inpcb *inp) { int32_t hpts_num; struct tcp_hpts_entry *hpts; hpts_num = inp->inp_hpts_cpu; hpts = tcp_pace.rp_ent[hpts_num]; return (hpts); } struct tcp_hpts_entry * tcp_hpts_lock(struct inpcb *inp) { struct tcp_hpts_entry *hpts; int32_t hpts_num; again: hpts_num = inp->inp_hpts_cpu; hpts = tcp_pace.rp_ent[hpts_num]; #ifdef INVARIANTS if (mtx_owned(&hpts->p_mtx)) { panic("Hpts:%p owns mtx prior-to lock line:%d", hpts, __LINE__); } #endif mtx_lock(&hpts->p_mtx); if (hpts_num != inp->inp_hpts_cpu) { mtx_unlock(&hpts->p_mtx); goto again; } return (hpts); } struct tcp_hpts_entry * tcp_input_lock(struct inpcb *inp) { struct tcp_hpts_entry *hpts; int32_t hpts_num; again: hpts_num = inp->inp_input_cpu; hpts = tcp_pace.rp_ent[hpts_num]; #ifdef INVARIANTS if (mtx_owned(&hpts->p_mtx)) { panic("Hpts:%p owns mtx prior-to lock line:%d", hpts, __LINE__); } #endif mtx_lock(&hpts->p_mtx); if (hpts_num != inp->inp_input_cpu) { mtx_unlock(&hpts->p_mtx); goto again; } return (hpts); } static void tcp_remove_hpts_ref(struct inpcb *inp, struct tcp_hpts_entry *hpts, int line) { int32_t add_freed; if (inp->inp_flags2 & INP_FREED) { /* * Need to play a special trick so that in_pcbrele_wlocked * does not return 1 when it really should have returned 0. */ add_freed = 1; inp->inp_flags2 &= ~INP_FREED; } else { add_freed = 0; } #ifndef INP_REF_DEBUG if (in_pcbrele_wlocked(inp)) { /* * This should not happen. We have the inpcb referred to by * the main socket (why we are called) and the hpts. It * should always return 0. */ panic("inpcb:%p release ret 1", inp); } #else if (__in_pcbrele_wlocked(inp, line)) { /* * This should not happen. We have the inpcb referred to by * the main socket (why we are called) and the hpts. It * should always return 0. */ panic("inpcb:%p release ret 1", inp); } #endif if (add_freed) { inp->inp_flags2 |= INP_FREED; } } static void tcp_hpts_remove_locked_output(struct tcp_hpts_entry *hpts, struct inpcb *inp, int32_t flags, int32_t line) { if (inp->inp_in_hpts) { hpts_sane_pace_remove(hpts, inp, &hpts->p_hptss[inp->inp_hptsslot], 1); tcp_remove_hpts_ref(inp, hpts, line); } } static void tcp_hpts_remove_locked_input(struct tcp_hpts_entry *hpts, struct inpcb *inp, int32_t flags, int32_t line) { HPTS_MTX_ASSERT(hpts); if (inp->inp_in_input) { hpts_sane_input_remove(hpts, inp, 1); tcp_remove_hpts_ref(inp, hpts, line); } } /* * Called normally with the INP_LOCKED but it * does not matter, the hpts lock is the key * but the lock order allows us to hold the * INP lock and then get the hpts lock. * * Valid values in the flags are * HPTS_REMOVE_OUTPUT - remove from the output of the hpts. * HPTS_REMOVE_INPUT - remove from the input of the hpts. * Note that you can or both values together and get two * actions. */ void __tcp_hpts_remove(struct inpcb *inp, int32_t flags, int32_t line) { struct tcp_hpts_entry *hpts; INP_WLOCK_ASSERT(inp); if (flags & HPTS_REMOVE_OUTPUT) { hpts = tcp_hpts_lock(inp); tcp_hpts_remove_locked_output(hpts, inp, flags, line); mtx_unlock(&hpts->p_mtx); } if (flags & HPTS_REMOVE_INPUT) { hpts = tcp_input_lock(inp); tcp_hpts_remove_locked_input(hpts, inp, flags, line); mtx_unlock(&hpts->p_mtx); } } static inline int hpts_tick(struct tcp_hpts_entry *hpts, int32_t plus) { return ((hpts->p_prevtick + plus) % NUM_OF_HPTSI_SLOTS); } static int tcp_queue_to_hpts_immediate_locked(struct inpcb *inp, struct tcp_hpts_entry *hpts, int32_t line, int32_t noref) { int32_t need_wake = 0; uint32_t ticknow = 0; HPTS_MTX_ASSERT(hpts); if (inp->inp_in_hpts == 0) { /* Ok we need to set it on the hpts in the current slot */ if (hpts->p_hpts_active == 0) { /* A sleeping hpts we want in next slot to run */ if (logging_on) { tcp_hpts_log_it(hpts, inp, HPTSLOG_INSERT_SLEEPER, 0, hpts_tick(hpts, 1)); } inp->inp_hptsslot = hpts_tick(hpts, 1); inp->inp_hpts_request = 0; if (logging_on) { tcp_hpts_log_it(hpts, inp, HPTSLOG_SLEEP_BEFORE, 1, ticknow); } need_wake = 1; } else if ((void *)inp == hpts->p_inp) { /* * We can't allow you to go into the same slot we * are in. We must put you out. */ inp->inp_hptsslot = hpts->p_nxt_slot; } else inp->inp_hptsslot = hpts->p_cur_slot; hpts_sane_pace_insert(hpts, inp, &hpts->p_hptss[inp->inp_hptsslot], line, noref); inp->inp_hpts_request = 0; if (logging_on) { tcp_hpts_log_it(hpts, inp, HPTSLOG_IMMEDIATE, 0, 0); } if (need_wake) { /* * Activate the hpts if it is sleeping and its * timeout is not 1. */ if (logging_on) { tcp_hpts_log_it(hpts, inp, HPTSLOG_WAKEUP_HPTS, 0, ticknow); } hpts->p_direct_wake = 1; tcp_wakehpts(hpts); } } return (need_wake); } int __tcp_queue_to_hpts_immediate(struct inpcb *inp, int32_t line) { int32_t ret; struct tcp_hpts_entry *hpts; INP_WLOCK_ASSERT(inp); hpts = tcp_hpts_lock(inp); ret = tcp_queue_to_hpts_immediate_locked(inp, hpts, line, 0); mtx_unlock(&hpts->p_mtx); return (ret); } static void tcp_hpts_insert_locked(struct tcp_hpts_entry *hpts, struct inpcb *inp, uint32_t slot, uint32_t cts, int32_t line, struct hpts_diag *diag, int32_t noref) { int32_t need_new_to = 0; int32_t need_wakeup = 0; uint32_t largest_slot; uint32_t ticknow = 0; uint32_t slot_calc; HPTS_MTX_ASSERT(hpts); if (diag) { memset(diag, 0, sizeof(struct hpts_diag)); diag->p_hpts_active = hpts->p_hpts_active; diag->p_nxt_slot = hpts->p_nxt_slot; diag->p_cur_slot = hpts->p_cur_slot; diag->slot_req = slot; } if ((inp->inp_in_hpts == 0) || noref) { inp->inp_hpts_request = slot; if (slot == 0) { /* Immediate */ tcp_queue_to_hpts_immediate_locked(inp, hpts, line, noref); return; } if (hpts->p_hpts_active) { /* * Its slot - 1 since nxt_slot is the next tick that * will go off since the hpts is awake */ if (logging_on) { tcp_hpts_log_it(hpts, inp, HPTSLOG_INSERT_NORMAL, slot, 0); } /* * We want to make sure that we don't place a inp in * the range of p_cur_slot <-> p_nxt_slot. If we * take from p_nxt_slot to the end, plus p_cur_slot * and then take away 2, we will know how many is * the max slots we can use. */ if (hpts->p_nxt_slot > hpts->p_cur_slot) { /* * Non-wrap case nxt_slot <-> cur_slot we * don't want to land in. So the diff gives * us what is taken away from the number of * slots. */ largest_slot = NUM_OF_HPTSI_SLOTS - (hpts->p_nxt_slot - hpts->p_cur_slot); } else if (hpts->p_nxt_slot == hpts->p_cur_slot) { largest_slot = NUM_OF_HPTSI_SLOTS - 2; } else { /* * Wrap case so the diff gives us the number * of slots that we can land in. */ largest_slot = hpts->p_cur_slot - hpts->p_nxt_slot; } /* * We take away two so we never have a problem (20 * usec's) out of 1024000 usecs */ largest_slot -= 2; if (inp->inp_hpts_request > largest_slot) { /* * Restrict max jump of slots and remember * leftover */ slot = largest_slot; inp->inp_hpts_request -= largest_slot; } else { /* This one will run when we hit it */ inp->inp_hpts_request = 0; } if (hpts->p_nxt_slot == hpts->p_cur_slot) slot_calc = (hpts->p_nxt_slot + slot) % NUM_OF_HPTSI_SLOTS; else slot_calc = (hpts->p_nxt_slot + slot - 1) % NUM_OF_HPTSI_SLOTS; if (slot_calc == hpts->p_cur_slot) { #ifdef INVARIANTS /* TSNH */ panic("Hpts:%p impossible slot calculation slot_calc:%u slot:%u largest:%u\n", hpts, slot_calc, slot, largest_slot); #endif if (slot_calc) slot_calc--; else slot_calc = NUM_OF_HPTSI_SLOTS - 1; } inp->inp_hptsslot = slot_calc; if (diag) { diag->inp_hptsslot = inp->inp_hptsslot; } } else { /* * The hpts is sleeping, we need to figure out where * it will wake up at and if we need to reschedule * its time-out. */ uint32_t have_slept, yet_to_sleep; uint32_t slot_now; struct timeval tv; ticknow = tcp_gethptstick(&tv); slot_now = ticknow % NUM_OF_HPTSI_SLOTS; /* * The user wants to be inserted at (slot_now + * slot) % NUM_OF_HPTSI_SLOTS, so lets set that up. */ largest_slot = NUM_OF_HPTSI_SLOTS - 2; if (inp->inp_hpts_request > largest_slot) { /* Adjust the residual in inp_hpts_request */ slot = largest_slot; inp->inp_hpts_request -= largest_slot; } else { /* No residual it all fits */ inp->inp_hpts_request = 0; } inp->inp_hptsslot = (slot_now + slot) % NUM_OF_HPTSI_SLOTS; if (diag) { diag->slot_now = slot_now; diag->inp_hptsslot = inp->inp_hptsslot; diag->p_on_min_sleep = hpts->p_on_min_sleep; } if (logging_on) { tcp_hpts_log_it(hpts, inp, HPTSLOG_INSERT_SLEEPER, slot, ticknow); } /* Now do we need to restart the hpts's timer? */ if (TSTMP_GT(ticknow, hpts->p_curtick)) have_slept = ticknow - hpts->p_curtick; else have_slept = 0; if (have_slept < hpts->p_hpts_sleep_time) { /* This should be what happens */ yet_to_sleep = hpts->p_hpts_sleep_time - have_slept; } else { /* We are over-due */ yet_to_sleep = 0; need_wakeup = 1; } if (diag) { diag->have_slept = have_slept; diag->yet_to_sleep = yet_to_sleep; diag->hpts_sleep_time = hpts->p_hpts_sleep_time; } if ((hpts->p_on_min_sleep == 0) && (yet_to_sleep > slot)) { /* * We need to reschedule the hptss time-out. */ hpts->p_hpts_sleep_time = slot; need_new_to = slot * HPTS_TICKS_PER_USEC; } } hpts_sane_pace_insert(hpts, inp, &hpts->p_hptss[inp->inp_hptsslot], line, noref); if (logging_on) { tcp_hpts_log_it(hpts, inp, HPTSLOG_INSERTED, slot, ticknow); } /* * Now how far is the hpts sleeping to? if active is 1, its * up and ticking we do nothing, otherwise we may need to * reschedule its callout if need_new_to is set from above. */ if (need_wakeup) { if (logging_on) { tcp_hpts_log_it(hpts, inp, HPTSLOG_RESCHEDULE, 1, 0); } hpts->p_direct_wake = 1; tcp_wakehpts(hpts); if (diag) { diag->need_new_to = 0; diag->co_ret = 0xffff0000; } } else if (need_new_to) { int32_t co_ret; struct timeval tv; sbintime_t sb; tv.tv_sec = 0; tv.tv_usec = 0; while (need_new_to > HPTS_USEC_IN_SEC) { tv.tv_sec++; need_new_to -= HPTS_USEC_IN_SEC; } tv.tv_usec = need_new_to; sb = tvtosbt(tv); if (tcp_hpts_callout_skip_swi == 0) { co_ret = callout_reset_sbt_on(&hpts->co, sb, 0, hpts_timeout_swi, hpts, hpts->p_cpu, (C_DIRECT_EXEC | C_PREL(tcp_hpts_precision))); } else { co_ret = callout_reset_sbt_on(&hpts->co, sb, 0, hpts_timeout_dir, hpts, hpts->p_cpu, C_PREL(tcp_hpts_precision)); } if (diag) { diag->need_new_to = need_new_to; diag->co_ret = co_ret; } } } else { #ifdef INVARIANTS panic("Hpts:%p tp:%p already on hpts and add?", hpts, inp); #endif } } uint32_t tcp_hpts_insert_diag(struct inpcb *inp, uint32_t slot, int32_t line, struct hpts_diag *diag){ struct tcp_hpts_entry *hpts; uint32_t slot_on, cts; struct timeval tv; /* * We now return the next-slot the hpts will be on, beyond its * current run (if up) or where it was when it stopped if it is * sleeping. */ INP_WLOCK_ASSERT(inp); hpts = tcp_hpts_lock(inp); if (in_ts_percision) microuptime(&tv); else getmicrouptime(&tv); cts = tcp_tv_to_usectick(&tv); tcp_hpts_insert_locked(hpts, inp, slot, cts, line, diag, 0); slot_on = hpts->p_nxt_slot; mtx_unlock(&hpts->p_mtx); return (slot_on); } uint32_t __tcp_hpts_insert(struct inpcb *inp, uint32_t slot, int32_t line){ return (tcp_hpts_insert_diag(inp, slot, line, NULL)); } int __tcp_queue_to_input_locked(struct inpcb *inp, struct tcp_hpts_entry *hpts, int32_t line) { int32_t retval = 0; HPTS_MTX_ASSERT(hpts); if (inp->inp_in_input == 0) { /* Ok we need to set it on the hpts in the current slot */ hpts_sane_input_insert(hpts, inp, line); retval = 1; if (hpts->p_hpts_active == 0) { /* * Activate the hpts if it is sleeping. */ if (logging_on) { tcp_hpts_log_it(hpts, inp, HPTSLOG_WAKEUP_INPUT, 0, 0); } retval = 2; hpts->p_direct_wake = 1; tcp_wakeinput(hpts); } } else if (hpts->p_hpts_active == 0) { retval = 4; hpts->p_direct_wake = 1; tcp_wakeinput(hpts); } return (retval); } void tcp_queue_pkt_to_input(struct tcpcb *tp, struct mbuf *m, struct tcphdr *th, int32_t tlen, int32_t drop_hdrlen, uint8_t iptos, uint8_t ti_locked) { /* Setup packet for input first */ INP_WLOCK_ASSERT(tp->t_inpcb); m->m_pkthdr.pace_thoff = (uint16_t) ((caddr_t)th - mtod(m, caddr_t)); m->m_pkthdr.pace_tlen = (uint16_t) tlen; m->m_pkthdr.pace_drphdrlen = drop_hdrlen; m->m_pkthdr.pace_tos = iptos; m->m_pkthdr.pace_lock = (uint8_t) ti_locked; if (tp->t_in_pkt == NULL) { tp->t_in_pkt = m; tp->t_tail_pkt = m; } else { tp->t_tail_pkt->m_nextpkt = m; tp->t_tail_pkt = m; } } int32_t __tcp_queue_to_input(struct tcpcb *tp, struct mbuf *m, struct tcphdr *th, int32_t tlen, int32_t drop_hdrlen, uint8_t iptos, uint8_t ti_locked, int32_t line){ struct tcp_hpts_entry *hpts; int32_t ret; tcp_queue_pkt_to_input(tp, m, th, tlen, drop_hdrlen, iptos, ti_locked); hpts = tcp_input_lock(tp->t_inpcb); ret = __tcp_queue_to_input_locked(tp->t_inpcb, hpts, line); mtx_unlock(&hpts->p_mtx); return (ret); } void __tcp_set_inp_to_drop(struct inpcb *inp, uint16_t reason, int32_t line) { struct tcp_hpts_entry *hpts; struct tcpcb *tp; tp = intotcpcb(inp); hpts = tcp_input_lock(tp->t_inpcb); if (inp->inp_in_input == 0) { /* Ok we need to set it on the hpts in the current slot */ hpts_sane_input_insert(hpts, inp, line); if (hpts->p_hpts_active == 0) { /* * Activate the hpts if it is sleeping. */ hpts->p_direct_wake = 1; tcp_wakeinput(hpts); } } else if (hpts->p_hpts_active == 0) { hpts->p_direct_wake = 1; tcp_wakeinput(hpts); } inp->inp_hpts_drop_reas = reason; mtx_unlock(&hpts->p_mtx); } static uint16_t hpts_random_cpu(struct inpcb *inp){ /* * No flow type set distribute the load randomly. */ uint16_t cpuid; uint32_t ran; /* * If one has been set use it i.e. we want both in and out on the * same hpts. */ if (inp->inp_input_cpu_set) { return (inp->inp_input_cpu); } else if (inp->inp_hpts_cpu_set) { return (inp->inp_hpts_cpu); } /* Nothing set use a random number */ ran = arc4random(); cpuid = (ran & 0xffff) % mp_ncpus; return (cpuid); } static uint16_t hpts_cpuid(struct inpcb *inp){ uint16_t cpuid; /* * If one has been set use it i.e. we want both in and out on the * same hpts. */ if (inp->inp_input_cpu_set) { return (inp->inp_input_cpu); } else if (inp->inp_hpts_cpu_set) { return (inp->inp_hpts_cpu); } /* If one is set the other must be the same */ #ifdef RSS cpuid = rss_hash2cpuid(inp->inp_flowid, inp->inp_flowtype); if (cpuid == NETISR_CPUID_NONE) return (hpts_random_cpu(inp)); else return (cpuid); #else /* * We don't have a flowid -> cpuid mapping, so cheat and just map * unknown cpuids to curcpu. Not the best, but apparently better * than defaulting to swi 0. */ if (inp->inp_flowtype != M_HASHTYPE_NONE) { cpuid = inp->inp_flowid % mp_ncpus; return (cpuid); } cpuid = hpts_random_cpu(inp); return (cpuid); #endif } /* * Do NOT try to optimize the processing of inp's * by first pulling off all the inp's into a temporary * list (e.g. TAILQ_CONCAT). If you do that the subtle * interactions of switching CPU's will kill because of * problems in the linked list manipulation. Basically * you would switch cpu's with the hpts mutex locked * but then while you were processing one of the inp's * some other one that you switch will get a new * packet on the different CPU. It will insert it * on the new hptss input list. Creating a temporary * link in the inp will not fix it either, since * the other hpts will be doing the same thing and * you will both end up using the temporary link. * * You will die in an ASSERT for tailq corruption if you * run INVARIANTS or you will die horribly without * INVARIANTS in some unknown way with a corrupt linked * list. */ static void tcp_input_data(struct tcp_hpts_entry *hpts, struct timeval *tv) { struct mbuf *m, *n; struct tcpcb *tp; struct inpcb *inp; uint16_t drop_reason; int16_t set_cpu; uint32_t did_prefetch = 0; int32_t ti_locked = TI_UNLOCKED; HPTS_MTX_ASSERT(hpts); while ((inp = TAILQ_FIRST(&hpts->p_input)) != NULL) { HPTS_MTX_ASSERT(hpts); hpts_sane_input_remove(hpts, inp, 0); if (inp->inp_input_cpu_set == 0) { set_cpu = 1; } else { set_cpu = 0; } hpts->p_inp = inp; drop_reason = inp->inp_hpts_drop_reas; inp->inp_in_input = 0; mtx_unlock(&hpts->p_mtx); if (drop_reason) { INP_INFO_RLOCK(&V_tcbinfo); ti_locked = TI_RLOCKED; } else { ti_locked = TI_UNLOCKED; } INP_WLOCK(inp); if ((inp->inp_flags & (INP_TIMEWAIT | INP_DROPPED)) || (inp->inp_flags2 & INP_FREED)) { out: hpts->p_inp = NULL; if (ti_locked == TI_RLOCKED) { INP_INFO_RUNLOCK(&V_tcbinfo); } if (in_pcbrele_wlocked(inp) == 0) { INP_WUNLOCK(inp); } ti_locked = TI_UNLOCKED; mtx_lock(&hpts->p_mtx); continue; } tp = intotcpcb(inp); if ((tp == NULL) || (tp->t_inpcb == NULL)) { goto out; } if (drop_reason) { /* This tcb is being destroyed for drop_reason */ m = tp->t_in_pkt; if (m) n = m->m_nextpkt; else n = NULL; tp->t_in_pkt = NULL; while (m) { m_freem(m); m = n; if (m) n = m->m_nextpkt; } tp = tcp_drop(tp, drop_reason); INP_INFO_RUNLOCK(&V_tcbinfo); if (tp == NULL) { INP_WLOCK(inp); } if (in_pcbrele_wlocked(inp) == 0) INP_WUNLOCK(inp); mtx_lock(&hpts->p_mtx); continue; } if (set_cpu) { /* * Setup so the next time we will move to the right * CPU. This should be a rare event. It will * sometimes happens when we are the client side * (usually not the server). Somehow tcp_output() * gets called before the tcp_do_segment() sets the * intial state. This means the r_cpu and r_hpts_cpu * is 0. We get on the hpts, and then tcp_input() * gets called setting up the r_cpu to the correct * value. The hpts goes off and sees the mis-match. * We simply correct it here and the CPU will switch * to the new hpts nextime the tcb gets added to the * the hpts (not this time) :-) */ tcp_set_hpts(inp); } CURVNET_SET(tp->t_vnet); m = tp->t_in_pkt; n = NULL; if (m != NULL && (m->m_pkthdr.pace_lock == TI_RLOCKED || tp->t_state != TCPS_ESTABLISHED)) { ti_locked = TI_RLOCKED; if (tcp_hptsi_lock_inpinfo(inp, &tp)) { CURVNET_RESTORE(); goto out; } m = tp->t_in_pkt; } if (in_newts_every_tcb) { if (in_ts_percision) microuptime(tv); else getmicrouptime(tv); } if (tp->t_fb_ptr != NULL) { kern_prefetch(tp->t_fb_ptr, &did_prefetch); did_prefetch = 1; } /* Any input work to do, if so do it first */ if ((m != NULL) && (m == tp->t_in_pkt)) { struct tcphdr *th; int32_t tlen, drop_hdrlen, nxt_pkt; uint8_t iptos; n = m->m_nextpkt; tp->t_in_pkt = tp->t_tail_pkt = NULL; while (m) { th = (struct tcphdr *)(mtod(m, caddr_t)+m->m_pkthdr.pace_thoff); tlen = m->m_pkthdr.pace_tlen; drop_hdrlen = m->m_pkthdr.pace_drphdrlen; iptos = m->m_pkthdr.pace_tos; m->m_nextpkt = NULL; if (n) nxt_pkt = 1; else nxt_pkt = 0; inp->inp_input_calls = 1; if (tp->t_fb->tfb_tcp_hpts_do_segment) { /* Use the hpts specific do_segment */ (*tp->t_fb->tfb_tcp_hpts_do_segment) (m, th, inp->inp_socket, tp, drop_hdrlen, tlen, iptos, ti_locked, nxt_pkt, tv); } else { /* Use the default do_segment */ (*tp->t_fb->tfb_tcp_do_segment) (m, th, inp->inp_socket, tp, drop_hdrlen, tlen, iptos, ti_locked); } /* * Do segment returns unlocked we need the * lock again but we also need some kasserts * here. */ INP_INFO_UNLOCK_ASSERT(&V_tcbinfo); INP_UNLOCK_ASSERT(inp); m = n; if (m) n = m->m_nextpkt; if (m != NULL && m->m_pkthdr.pace_lock == TI_RLOCKED) { INP_INFO_RLOCK(&V_tcbinfo); ti_locked = TI_RLOCKED; } else ti_locked = TI_UNLOCKED; INP_WLOCK(inp); /* * Since we have an opening here we must * re-check if the tcb went away while we * were getting the lock(s). */ if ((inp->inp_flags & (INP_TIMEWAIT | INP_DROPPED)) || (inp->inp_flags2 & INP_FREED)) { out_free: while (m) { m_freem(m); m = n; if (m) n = m->m_nextpkt; } CURVNET_RESTORE(); goto out; } /* * Now that we hold the INP lock, check if * we need to upgrade our lock. */ if (ti_locked == TI_UNLOCKED && (tp->t_state != TCPS_ESTABLISHED)) { ti_locked = TI_RLOCKED; if (tcp_hptsi_lock_inpinfo(inp, &tp)) goto out_free; } } /** end while(m) */ } /** end if ((m != NULL) && (m == tp->t_in_pkt)) */ if (in_pcbrele_wlocked(inp) == 0) INP_WUNLOCK(inp); if (ti_locked == TI_RLOCKED) INP_INFO_RUNLOCK(&V_tcbinfo); INP_INFO_UNLOCK_ASSERT(&V_tcbinfo); INP_UNLOCK_ASSERT(inp); ti_locked = TI_UNLOCKED; mtx_lock(&hpts->p_mtx); hpts->p_inp = NULL; CURVNET_RESTORE(); } } static int tcp_hpts_est_run(struct tcp_hpts_entry *hpts) { int32_t ticks_to_run; if (hpts->p_prevtick && (SEQ_GT(hpts->p_curtick, hpts->p_prevtick))) { ticks_to_run = hpts->p_curtick - hpts->p_prevtick; if (ticks_to_run >= (NUM_OF_HPTSI_SLOTS - 1)) { ticks_to_run = NUM_OF_HPTSI_SLOTS - 2; } } else { if (hpts->p_prevtick == hpts->p_curtick) { /* This happens when we get woken up right away */ return (-1); } ticks_to_run = 1; } /* Set in where we will be when we catch up */ hpts->p_nxt_slot = (hpts->p_cur_slot + ticks_to_run) % NUM_OF_HPTSI_SLOTS; if (hpts->p_nxt_slot == hpts->p_cur_slot) { panic("Impossible math -- hpts:%p p_nxt_slot:%d p_cur_slot:%d ticks_to_run:%d", hpts, hpts->p_nxt_slot, hpts->p_cur_slot, ticks_to_run); } return (ticks_to_run); } static void tcp_hptsi(struct tcp_hpts_entry *hpts, struct timeval *ctick) { struct tcpcb *tp; struct inpcb *inp = NULL, *ninp; struct timeval tv; int32_t ticks_to_run, i, error, tick_now, interum_tick; int32_t paced_cnt = 0; int32_t did_prefetch = 0; int32_t prefetch_ninp = 0; int32_t prefetch_tp = 0; uint32_t cts; int16_t set_cpu; HPTS_MTX_ASSERT(hpts); hpts->p_curtick = tcp_tv_to_hptstick(ctick); cts = tcp_tv_to_usectick(ctick); memcpy(&tv, ctick, sizeof(struct timeval)); hpts->p_cur_slot = hpts_tick(hpts, 1); /* Figure out if we had missed ticks */ again: HPTS_MTX_ASSERT(hpts); ticks_to_run = tcp_hpts_est_run(hpts); if (!TAILQ_EMPTY(&hpts->p_input)) { tcp_input_data(hpts, &tv); } #ifdef INVARIANTS if (TAILQ_EMPTY(&hpts->p_input) && (hpts->p_on_inqueue_cnt != 0)) { panic("tp:%p in_hpts input empty but cnt:%d", hpts, hpts->p_on_inqueue_cnt); } #endif HPTS_MTX_ASSERT(hpts); /* Reset the ticks to run and time if we need too */ interum_tick = tcp_gethptstick(&tv); if (interum_tick != hpts->p_curtick) { /* Save off the new time we execute to */ *ctick = tv; hpts->p_curtick = interum_tick; cts = tcp_tv_to_usectick(&tv); hpts->p_cur_slot = hpts_tick(hpts, 1); ticks_to_run = tcp_hpts_est_run(hpts); } if (ticks_to_run == -1) { goto no_run; } if (logging_on) { tcp_hpts_log_it(hpts, inp, HPTSLOG_SETTORUN, ticks_to_run, 0); } if (hpts->p_on_queue_cnt == 0) { goto no_one; } HPTS_MTX_ASSERT(hpts); for (i = 0; i < ticks_to_run; i++) { /* * Calculate our delay, if there are no extra ticks there * was not any */ hpts->p_delayed_by = (ticks_to_run - (i + 1)) * HPTS_TICKS_PER_USEC; HPTS_MTX_ASSERT(hpts); while ((inp = TAILQ_FIRST(&hpts->p_hptss[hpts->p_cur_slot])) != NULL) { /* For debugging */ if (logging_on) { tcp_hpts_log_it(hpts, inp, HPTSLOG_HPTSI, ticks_to_run, i); } hpts->p_inp = inp; paced_cnt++; if (hpts->p_cur_slot != inp->inp_hptsslot) { panic("Hpts:%p inp:%p slot mis-aligned %u vs %u", hpts, inp, hpts->p_cur_slot, inp->inp_hptsslot); } /* Now pull it */ if (inp->inp_hpts_cpu_set == 0) { set_cpu = 1; } else { set_cpu = 0; } hpts_sane_pace_remove(hpts, inp, &hpts->p_hptss[hpts->p_cur_slot], 0); if ((ninp = TAILQ_FIRST(&hpts->p_hptss[hpts->p_cur_slot])) != NULL) { /* We prefetch the next inp if possible */ kern_prefetch(ninp, &prefetch_ninp); prefetch_ninp = 1; } if (inp->inp_hpts_request) { /* * This guy is deferred out further in time * then our wheel had on it. Push him back * on the wheel. */ int32_t remaining_slots; remaining_slots = ticks_to_run - (i + 1); if (inp->inp_hpts_request > remaining_slots) { /* * Keep INVARIANTS happy by clearing * the flag */ tcp_hpts_insert_locked(hpts, inp, inp->inp_hpts_request, cts, __LINE__, NULL, 1); hpts->p_inp = NULL; continue; } inp->inp_hpts_request = 0; } /* * We clear the hpts flag here after dealing with * remaining slots. This way anyone looking with the * TCB lock will see its on the hpts until just * before we unlock. */ inp->inp_in_hpts = 0; mtx_unlock(&hpts->p_mtx); INP_WLOCK(inp); if (in_pcbrele_wlocked(inp)) { mtx_lock(&hpts->p_mtx); if (logging_on) tcp_hpts_log_it(hpts, hpts->p_inp, HPTSLOG_INP_DONE, 0, 1); hpts->p_inp = NULL; continue; } if (inp->inp_flags & (INP_TIMEWAIT | INP_DROPPED)) { out_now: #ifdef INVARIANTS if (mtx_owned(&hpts->p_mtx)) { panic("Hpts:%p owns mtx prior-to lock line:%d", hpts, __LINE__); } #endif INP_WUNLOCK(inp); mtx_lock(&hpts->p_mtx); if (logging_on) tcp_hpts_log_it(hpts, hpts->p_inp, HPTSLOG_INP_DONE, 0, 3); hpts->p_inp = NULL; continue; } tp = intotcpcb(inp); if ((tp == NULL) || (tp->t_inpcb == NULL)) { goto out_now; } if (set_cpu) { /* * Setup so the next time we will move to * the right CPU. This should be a rare * event. It will sometimes happens when we * are the client side (usually not the * server). Somehow tcp_output() gets called * before the tcp_do_segment() sets the * intial state. This means the r_cpu and * r_hpts_cpu is 0. We get on the hpts, and * then tcp_input() gets called setting up * the r_cpu to the correct value. The hpts * goes off and sees the mis-match. We * simply correct it here and the CPU will * switch to the new hpts nextime the tcb * gets added to the the hpts (not this one) * :-) */ tcp_set_hpts(inp); } if (out_newts_every_tcb) { struct timeval sv; if (out_ts_percision) microuptime(&sv); else getmicrouptime(&sv); cts = tcp_tv_to_usectick(&sv); } CURVNET_SET(tp->t_vnet); /* * There is a hole here, we get the refcnt on the * inp so it will still be preserved but to make * sure we can get the INP we need to hold the p_mtx * above while we pull out the tp/inp, as long as * fini gets the lock first we are assured of having * a sane INP we can lock and test. */ #ifdef INVARIANTS if (mtx_owned(&hpts->p_mtx)) { panic("Hpts:%p owns mtx before tcp-output:%d", hpts, __LINE__); } #endif if (tp->t_fb_ptr != NULL) { kern_prefetch(tp->t_fb_ptr, &did_prefetch); did_prefetch = 1; } inp->inp_hpts_calls = 1; if (tp->t_fb->tfb_tcp_output_wtime != NULL) { error = (*tp->t_fb->tfb_tcp_output_wtime) (tp, &tv); } else { error = tp->t_fb->tfb_tcp_output(tp); } if (ninp && ninp->inp_ppcb) { /* * If we have a nxt inp, see if we can * prefetch its ppcb. Note this may seem * "risky" since we have no locks (other * than the previous inp) and there no * assurance that ninp was not pulled while * we were processing inp and freed. If this * occured it could mean that either: * * a) Its NULL (which is fine we won't go * here) b) Its valid (which is cool we * will prefetch it) c) The inp got * freed back to the slab which was * reallocated. Then the piece of memory was * re-used and something else (not an * address) is in inp_ppcb. If that occurs * we don't crash, but take a TLB shootdown * performance hit (same as if it was NULL * and we tried to pre-fetch it). * * Considering that the likelyhood of is * quite rare we will take a risk on doing * this. If performance drops after testing * we can always take this out. NB: the * kern_prefetch on amd64 actually has * protection against a bad address now via * the DMAP_() tests. This will prevent the * TLB hit, and instead if occurs just * cause us to load cache with a useless * address (to us). */ kern_prefetch(ninp->inp_ppcb, &prefetch_tp); prefetch_tp = 1; } INP_WUNLOCK(inp); INP_UNLOCK_ASSERT(inp); CURVNET_RESTORE(); #ifdef INVARIANTS if (mtx_owned(&hpts->p_mtx)) { panic("Hpts:%p owns mtx prior-to lock line:%d", hpts, __LINE__); } #endif mtx_lock(&hpts->p_mtx); if (logging_on) tcp_hpts_log_it(hpts, hpts->p_inp, HPTSLOG_INP_DONE, 0, 4); hpts->p_inp = NULL; } HPTS_MTX_ASSERT(hpts); hpts->p_inp = NULL; hpts->p_cur_slot++; if (hpts->p_cur_slot >= NUM_OF_HPTSI_SLOTS) { hpts->p_cur_slot = 0; } } no_one: HPTS_MTX_ASSERT(hpts); hpts->p_prevtick = hpts->p_curtick; hpts->p_delayed_by = 0; /* * Check to see if we took an excess amount of time and need to run * more ticks (if we did not hit eno-bufs). */ /* Re-run any input that may be there */ (void)tcp_gethptstick(&tv); if (!TAILQ_EMPTY(&hpts->p_input)) { tcp_input_data(hpts, &tv); } #ifdef INVARIANTS if (TAILQ_EMPTY(&hpts->p_input) && (hpts->p_on_inqueue_cnt != 0)) { panic("tp:%p in_hpts input empty but cnt:%d", hpts, hpts->p_on_inqueue_cnt); } #endif tick_now = tcp_gethptstick(&tv); if (SEQ_GT(tick_now, hpts->p_prevtick)) { struct timeval res; /* Did we really spend a full tick or more in here? */ timersub(&tv, ctick, &res); if (res.tv_sec || (res.tv_usec >= HPTS_TICKS_PER_USEC)) { counter_u64_add(hpts_loops, 1); if (logging_on) { tcp_hpts_log_it(hpts, inp, HPTSLOG_TOLONG, (uint32_t) res.tv_usec, tick_now); } *ctick = res; hpts->p_curtick = tick_now; goto again; } } no_run: { uint32_t t = 0, i, fnd = 0; if (hpts->p_on_queue_cnt) { /* * Find next slot that is occupied and use that to * be the sleep time. */ for (i = 1, t = hpts->p_nxt_slot; i < NUM_OF_HPTSI_SLOTS; i++) { if (TAILQ_EMPTY(&hpts->p_hptss[t]) == 0) { fnd = 1; break; } t = (t + 1) % NUM_OF_HPTSI_SLOTS; } if (fnd) { hpts->p_hpts_sleep_time = i; } else { counter_u64_add(back_tosleep, 1); #ifdef INVARIANTS panic("Hpts:%p cnt:%d but non found", hpts, hpts->p_on_queue_cnt); #endif hpts->p_on_queue_cnt = 0; goto non_found; } t++; } else { /* No one on the wheel sleep for all but 2 slots */ non_found: if (hpts_sleep_max == 0) hpts_sleep_max = 1; hpts->p_hpts_sleep_time = min((NUM_OF_HPTSI_SLOTS - 2), hpts_sleep_max); t = 0; } if (logging_on) { tcp_hpts_log_it(hpts, inp, HPTSLOG_SLEEPSET, t, (hpts->p_hpts_sleep_time * HPTS_TICKS_PER_USEC)); } } } void __tcp_set_hpts(struct inpcb *inp, int32_t line) { struct tcp_hpts_entry *hpts; INP_WLOCK_ASSERT(inp); hpts = tcp_hpts_lock(inp); if ((inp->inp_in_hpts == 0) && (inp->inp_hpts_cpu_set == 0)) { inp->inp_hpts_cpu = hpts_cpuid(inp); inp->inp_hpts_cpu_set = 1; } mtx_unlock(&hpts->p_mtx); hpts = tcp_input_lock(inp); if ((inp->inp_input_cpu_set == 0) && (inp->inp_in_input == 0)) { inp->inp_input_cpu = hpts_cpuid(inp); inp->inp_input_cpu_set = 1; } mtx_unlock(&hpts->p_mtx); } uint16_t tcp_hpts_delayedby(struct inpcb *inp){ return (tcp_pace.rp_ent[inp->inp_hpts_cpu]->p_delayed_by); } static void tcp_hpts_thread(void *ctx) { struct tcp_hpts_entry *hpts; struct timeval tv; sbintime_t sb; hpts = (struct tcp_hpts_entry *)ctx; mtx_lock(&hpts->p_mtx); if (hpts->p_direct_wake) { /* Signaled by input */ if (logging_on) tcp_hpts_log_it(hpts, NULL, HPTSLOG_AWAKE, 1, 1); callout_stop(&hpts->co); } else { /* Timed out */ if (callout_pending(&hpts->co) || !callout_active(&hpts->co)) { if (logging_on) tcp_hpts_log_it(hpts, NULL, HPTSLOG_AWAKE, 2, 2); mtx_unlock(&hpts->p_mtx); return; } callout_deactivate(&hpts->co); if (logging_on) tcp_hpts_log_it(hpts, NULL, HPTSLOG_AWAKE, 3, 3); } hpts->p_hpts_active = 1; (void)tcp_gethptstick(&tv); tcp_hptsi(hpts, &tv); HPTS_MTX_ASSERT(hpts); tv.tv_sec = 0; tv.tv_usec = hpts->p_hpts_sleep_time * HPTS_TICKS_PER_USEC; if (tcp_min_hptsi_time && (tv.tv_usec < tcp_min_hptsi_time)) { tv.tv_usec = tcp_min_hptsi_time; hpts->p_on_min_sleep = 1; } else { /* Clear the min sleep flag */ hpts->p_on_min_sleep = 0; } hpts->p_hpts_active = 0; sb = tvtosbt(tv); if (tcp_hpts_callout_skip_swi == 0) { callout_reset_sbt_on(&hpts->co, sb, 0, hpts_timeout_swi, hpts, hpts->p_cpu, (C_DIRECT_EXEC | C_PREL(tcp_hpts_precision))); } else { callout_reset_sbt_on(&hpts->co, sb, 0, hpts_timeout_dir, hpts, hpts->p_cpu, C_PREL(tcp_hpts_precision)); } hpts->p_direct_wake = 0; mtx_unlock(&hpts->p_mtx); } #undef timersub static void tcp_init_hptsi(void *st) { int32_t i, j, error, bound = 0, created = 0; size_t sz, asz; struct timeval tv; sbintime_t sb; struct tcp_hpts_entry *hpts; char unit[16]; uint32_t ncpus = mp_ncpus ? mp_ncpus : MAXCPU; tcp_pace.rp_proc = NULL; tcp_pace.rp_num_hptss = ncpus; hpts_loops = counter_u64_alloc(M_WAITOK); back_tosleep = counter_u64_alloc(M_WAITOK); sz = (tcp_pace.rp_num_hptss * sizeof(struct tcp_hpts_entry *)); tcp_pace.rp_ent = malloc(sz, M_TCPHPTS, M_WAITOK | M_ZERO); asz = sizeof(struct hptsh) * NUM_OF_HPTSI_SLOTS; for (i = 0; i < tcp_pace.rp_num_hptss; i++) { tcp_pace.rp_ent[i] = malloc(sizeof(struct tcp_hpts_entry), M_TCPHPTS, M_WAITOK | M_ZERO); tcp_pace.rp_ent[i]->p_hptss = malloc(asz, M_TCPHPTS, M_WAITOK); hpts = tcp_pace.rp_ent[i]; /* * Init all the hpts structures that are not specifically * zero'd by the allocations. Also lets attach them to the * appropriate sysctl block as well. */ mtx_init(&hpts->p_mtx, "tcp_hpts_lck", "hpts", MTX_DEF | MTX_DUPOK); TAILQ_INIT(&hpts->p_input); for (j = 0; j < NUM_OF_HPTSI_SLOTS; j++) { TAILQ_INIT(&hpts->p_hptss[j]); } sysctl_ctx_init(&hpts->hpts_ctx); sprintf(unit, "%d", i); hpts->hpts_root = SYSCTL_ADD_NODE(&hpts->hpts_ctx, SYSCTL_STATIC_CHILDREN(_net_inet_tcp_hpts), OID_AUTO, unit, CTLFLAG_RW, 0, ""); SYSCTL_ADD_INT(&hpts->hpts_ctx, SYSCTL_CHILDREN(hpts->hpts_root), OID_AUTO, "in_qcnt", CTLFLAG_RD, &hpts->p_on_inqueue_cnt, 0, "Count TCB's awaiting input processing"); SYSCTL_ADD_INT(&hpts->hpts_ctx, SYSCTL_CHILDREN(hpts->hpts_root), OID_AUTO, "out_qcnt", CTLFLAG_RD, &hpts->p_on_queue_cnt, 0, "Count TCB's awaiting output processing"); SYSCTL_ADD_UINT(&hpts->hpts_ctx, SYSCTL_CHILDREN(hpts->hpts_root), OID_AUTO, "active", CTLFLAG_RD, &hpts->p_hpts_active, 0, "Is the hpts active"); SYSCTL_ADD_UINT(&hpts->hpts_ctx, SYSCTL_CHILDREN(hpts->hpts_root), OID_AUTO, "curslot", CTLFLAG_RD, &hpts->p_cur_slot, 0, "What the current slot is if active"); SYSCTL_ADD_UINT(&hpts->hpts_ctx, SYSCTL_CHILDREN(hpts->hpts_root), OID_AUTO, "curtick", CTLFLAG_RD, &hpts->p_curtick, 0, "What the current tick on if active"); SYSCTL_ADD_UINT(&hpts->hpts_ctx, SYSCTL_CHILDREN(hpts->hpts_root), OID_AUTO, "logsize", CTLFLAG_RD, &hpts->p_logsize, 0, "Hpts logging buffer size"); hpts->p_hpts_sleep_time = NUM_OF_HPTSI_SLOTS - 2; hpts->p_num = i; hpts->p_prevtick = hpts->p_curtick = tcp_gethptstick(&tv); hpts->p_prevtick -= 1; hpts->p_prevtick %= NUM_OF_HPTSI_SLOTS; hpts->p_cpu = 0xffff; hpts->p_nxt_slot = 1; hpts->p_logsize = tcp_hpts_logging_size; if (hpts->p_logsize) { sz = (sizeof(struct hpts_log) * hpts->p_logsize); hpts->p_log = malloc(sz, M_TCPHPTS, M_WAITOK | M_ZERO); } callout_init(&hpts->co, 1); } /* * Now lets start ithreads to handle the hptss. */ CPU_FOREACH(i) { hpts = tcp_pace.rp_ent[i]; hpts->p_cpu = i; error = swi_add(&hpts->ie, "hpts", tcp_hpts_thread, (void *)hpts, SWI_NET, INTR_MPSAFE, &hpts->ie_cookie); if (error) { panic("Can't add hpts:%p i:%d err:%d", hpts, i, error); } created++; if (tcp_bind_threads) { if (intr_event_bind(hpts->ie, i) == 0) bound++; } tv.tv_sec = 0; tv.tv_usec = hpts->p_hpts_sleep_time * HPTS_TICKS_PER_USEC; sb = tvtosbt(tv); if (tcp_hpts_callout_skip_swi == 0) { callout_reset_sbt_on(&hpts->co, sb, 0, hpts_timeout_swi, hpts, hpts->p_cpu, (C_DIRECT_EXEC | C_PREL(tcp_hpts_precision))); } else { callout_reset_sbt_on(&hpts->co, sb, 0, hpts_timeout_dir, hpts, hpts->p_cpu, C_PREL(tcp_hpts_precision)); } } printf("TCP Hpts created %d swi interrupt thread and bound %d\n", created, bound); return; } SYSINIT(tcphptsi, SI_SUB_KTHREAD_IDLE, SI_ORDER_ANY, tcp_init_hptsi, NULL);