/* $OpenBSD: scheduler_ramqueue.c,v 1.43 2017/01/09 09:53:23 reyk Exp $ */ /* * Copyright (c) 2012 Gilles Chehade * Copyright (c) 2012 Eric Faurot * * Permission to use, copy, modify, and distribute this software for any * purpose with or without fee is hereby granted, provided that the above * copyright notice and this permission notice appear in all copies. * * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "smtpd.h" #include "log.h" TAILQ_HEAD(evplist, rq_envelope); struct rq_message { uint32_t msgid; struct tree envelopes; }; struct rq_envelope { TAILQ_ENTRY(rq_envelope) entry; SPLAY_ENTRY(rq_envelope) t_entry; uint64_t evpid; uint64_t holdq; enum delivery_type type; #define RQ_EVPSTATE_PENDING 0 #define RQ_EVPSTATE_SCHEDULED 1 #define RQ_EVPSTATE_INFLIGHT 2 #define RQ_EVPSTATE_HELD 3 uint8_t state; #define RQ_ENVELOPE_EXPIRED 0x01 #define RQ_ENVELOPE_REMOVED 0x02 #define RQ_ENVELOPE_SUSPEND 0x04 #define RQ_ENVELOPE_UPDATE 0x08 #define RQ_ENVELOPE_OVERFLOW 0x10 uint8_t flags; time_t ctime; time_t sched; time_t expire; struct rq_message *message; time_t t_inflight; time_t t_scheduled; }; struct rq_holdq { struct evplist q; size_t count; }; struct rq_queue { size_t evpcount; struct tree messages; SPLAY_HEAD(prioqtree, rq_envelope) q_priotree; struct evplist q_pending; struct evplist q_inflight; struct evplist q_mta; struct evplist q_mda; struct evplist q_bounce; struct evplist q_update; struct evplist q_expired; struct evplist q_removed; }; static int rq_envelope_cmp(struct rq_envelope *, struct rq_envelope *); SPLAY_PROTOTYPE(prioqtree, rq_envelope, t_entry, rq_envelope_cmp); static int scheduler_ram_init(const char *); static int scheduler_ram_insert(struct scheduler_info *); static size_t scheduler_ram_commit(uint32_t); static size_t scheduler_ram_rollback(uint32_t); static int scheduler_ram_update(struct scheduler_info *); static int scheduler_ram_delete(uint64_t); static int scheduler_ram_hold(uint64_t, uint64_t); static int scheduler_ram_release(int, uint64_t, int); static int scheduler_ram_batch(int, int *, size_t *, uint64_t *, int *); static size_t scheduler_ram_messages(uint32_t, uint32_t *, size_t); static size_t scheduler_ram_envelopes(uint64_t, struct evpstate *, size_t); static int scheduler_ram_schedule(uint64_t); static int scheduler_ram_remove(uint64_t); static int scheduler_ram_suspend(uint64_t); static int scheduler_ram_resume(uint64_t); static int scheduler_ram_query(uint64_t); static void sorted_insert(struct rq_queue *, struct rq_envelope *); static void rq_queue_init(struct rq_queue *); static void rq_queue_merge(struct rq_queue *, struct rq_queue *); static void rq_queue_dump(struct rq_queue *, const char *); static void rq_queue_schedule(struct rq_queue *rq); static struct evplist *rq_envelope_list(struct rq_queue *, struct rq_envelope *); static void rq_envelope_schedule(struct rq_queue *, struct rq_envelope *); static int rq_envelope_remove(struct rq_queue *, struct rq_envelope *); static int rq_envelope_suspend(struct rq_queue *, struct rq_envelope *); static int rq_envelope_resume(struct rq_queue *, struct rq_envelope *); static void rq_envelope_delete(struct rq_queue *, struct rq_envelope *); static const char *rq_envelope_to_text(struct rq_envelope *); struct scheduler_backend scheduler_backend_ramqueue = { scheduler_ram_init, scheduler_ram_insert, scheduler_ram_commit, scheduler_ram_rollback, scheduler_ram_update, scheduler_ram_delete, scheduler_ram_hold, scheduler_ram_release, scheduler_ram_batch, scheduler_ram_messages, scheduler_ram_envelopes, scheduler_ram_schedule, scheduler_ram_remove, scheduler_ram_suspend, scheduler_ram_resume, scheduler_ram_query, }; static struct rq_queue ramqueue; static struct tree updates; static struct tree holdqs[3]; /* delivery type */ static time_t currtime; #define BACKOFF_TRANSFER 400 #define BACKOFF_DELIVERY 10 #define BACKOFF_OVERFLOW 3 static time_t scheduler_backoff(time_t t0, time_t base, uint32_t step) { return (t0 + base * step * step); } static time_t scheduler_next(time_t t0, time_t base, uint32_t step) { time_t t; /* XXX be more efficient */ while ((t = scheduler_backoff(t0, base, step)) <= currtime) step++; return (t); } static int scheduler_ram_init(const char *arg) { rq_queue_init(&ramqueue); tree_init(&updates); tree_init(&holdqs[D_MDA]); tree_init(&holdqs[D_MTA]); tree_init(&holdqs[D_BOUNCE]); return (1); } static int scheduler_ram_insert(struct scheduler_info *si) { struct rq_queue *update; struct rq_message *message; struct rq_envelope *envelope; uint32_t msgid; currtime = time(NULL); msgid = evpid_to_msgid(si->evpid); /* find/prepare a ramqueue update */ if ((update = tree_get(&updates, msgid)) == NULL) { update = xcalloc(1, sizeof *update, "scheduler_insert"); stat_increment("scheduler.ramqueue.update", 1); rq_queue_init(update); tree_xset(&updates, msgid, update); } /* find/prepare the msgtree message in ramqueue update */ if ((message = tree_get(&update->messages, msgid)) == NULL) { message = xcalloc(1, sizeof *message, "scheduler_insert"); message->msgid = msgid; tree_init(&message->envelopes); tree_xset(&update->messages, msgid, message); stat_increment("scheduler.ramqueue.message", 1); } /* create envelope in ramqueue message */ envelope = xcalloc(1, sizeof *envelope, "scheduler_insert"); envelope->evpid = si->evpid; envelope->type = si->type; envelope->message = message; envelope->ctime = si->creation; envelope->expire = si->creation + si->expire; envelope->sched = scheduler_backoff(si->creation, (si->type == D_MTA) ? BACKOFF_TRANSFER : BACKOFF_DELIVERY, si->retry); tree_xset(&message->envelopes, envelope->evpid, envelope); update->evpcount++; stat_increment("scheduler.ramqueue.envelope", 1); envelope->state = RQ_EVPSTATE_PENDING; TAILQ_INSERT_TAIL(&update->q_pending, envelope, entry); si->nexttry = envelope->sched; return (1); } static size_t scheduler_ram_commit(uint32_t msgid) { struct rq_queue *update; size_t r; currtime = time(NULL); update = tree_xpop(&updates, msgid); r = update->evpcount; if (tracing & TRACE_SCHEDULER) rq_queue_dump(update, "update to commit"); rq_queue_merge(&ramqueue, update); if (tracing & TRACE_SCHEDULER) rq_queue_dump(&ramqueue, "resulting queue"); rq_queue_schedule(&ramqueue); free(update); stat_decrement("scheduler.ramqueue.update", 1); return (r); } static size_t scheduler_ram_rollback(uint32_t msgid) { struct rq_queue *update; struct rq_envelope *evp; size_t r; currtime = time(NULL); if ((update = tree_pop(&updates, msgid)) == NULL) return (0); r = update->evpcount; while ((evp = TAILQ_FIRST(&update->q_pending))) { TAILQ_REMOVE(&update->q_pending, evp, entry); rq_envelope_delete(update, evp); } free(update); stat_decrement("scheduler.ramqueue.update", 1); return (r); } static int scheduler_ram_update(struct scheduler_info *si) { struct rq_message *msg; struct rq_envelope *evp; uint32_t msgid; currtime = time(NULL); msgid = evpid_to_msgid(si->evpid); msg = tree_xget(&ramqueue.messages, msgid); evp = tree_xget(&msg->envelopes, si->evpid); /* it *must* be in-flight */ if (evp->state != RQ_EVPSTATE_INFLIGHT) errx(1, "evp:%016" PRIx64 " not in-flight", si->evpid); TAILQ_REMOVE(&ramqueue.q_inflight, evp, entry); /* * If the envelope was removed while inflight, schedule it for * removal immediately. */ if (evp->flags & RQ_ENVELOPE_REMOVED) { TAILQ_INSERT_TAIL(&ramqueue.q_removed, evp, entry); evp->state = RQ_EVPSTATE_SCHEDULED; evp->t_scheduled = currtime; return (1); } evp->sched = scheduler_next(evp->ctime, (si->type == D_MTA) ? BACKOFF_TRANSFER : BACKOFF_DELIVERY, si->retry); evp->state = RQ_EVPSTATE_PENDING; if (!(evp->flags & RQ_ENVELOPE_SUSPEND)) sorted_insert(&ramqueue, evp); si->nexttry = evp->sched; return (1); } static int scheduler_ram_delete(uint64_t evpid) { struct rq_message *msg; struct rq_envelope *evp; uint32_t msgid; currtime = time(NULL); msgid = evpid_to_msgid(evpid); msg = tree_xget(&ramqueue.messages, msgid); evp = tree_xget(&msg->envelopes, evpid); /* it *must* be in-flight */ if (evp->state != RQ_EVPSTATE_INFLIGHT) errx(1, "evp:%016" PRIx64 " not in-flight", evpid); TAILQ_REMOVE(&ramqueue.q_inflight, evp, entry); rq_envelope_delete(&ramqueue, evp); return (1); } #define HOLDQ_MAXSIZE 1000 static int scheduler_ram_hold(uint64_t evpid, uint64_t holdq) { struct rq_holdq *hq; struct rq_message *msg; struct rq_envelope *evp; uint32_t msgid; currtime = time(NULL); msgid = evpid_to_msgid(evpid); msg = tree_xget(&ramqueue.messages, msgid); evp = tree_xget(&msg->envelopes, evpid); /* it *must* be in-flight */ if (evp->state != RQ_EVPSTATE_INFLIGHT) errx(1, "evp:%016" PRIx64 " not in-flight", evpid); TAILQ_REMOVE(&ramqueue.q_inflight, evp, entry); /* If the envelope is suspended, just mark it as pending */ if (evp->flags & RQ_ENVELOPE_SUSPEND) { evp->state = RQ_EVPSTATE_PENDING; return (0); } hq = tree_get(&holdqs[evp->type], holdq); if (hq == NULL) { hq = xcalloc(1, sizeof(*hq), "scheduler_hold"); TAILQ_INIT(&hq->q); tree_xset(&holdqs[evp->type], holdq, hq); stat_increment("scheduler.ramqueue.holdq", 1); } /* If the holdq is full, just "tempfail" the envelope */ if (hq->count >= HOLDQ_MAXSIZE) { evp->state = RQ_EVPSTATE_PENDING; evp->flags |= RQ_ENVELOPE_UPDATE; evp->flags |= RQ_ENVELOPE_OVERFLOW; sorted_insert(&ramqueue, evp); stat_increment("scheduler.ramqueue.hold-overflow", 1); return (0); } evp->state = RQ_EVPSTATE_HELD; evp->holdq = holdq; /* This is an optimization: upon release, the envelopes will be * inserted in the pending queue from the first element to the last. * Since elements already in the queue were received first, they * were scheduled first, so they will be reinserted before the * current element. */ TAILQ_INSERT_HEAD(&hq->q, evp, entry); hq->count += 1; stat_increment("scheduler.ramqueue.hold", 1); return (1); } static int scheduler_ram_release(int type, uint64_t holdq, int n) { struct rq_holdq *hq; struct rq_envelope *evp; int i, update; currtime = time(NULL); hq = tree_get(&holdqs[type], holdq); if (hq == NULL) return (0); if (n == -1) { n = 0; update = 1; } else update = 0; for (i = 0; n == 0 || i < n; i++) { evp = TAILQ_FIRST(&hq->q); if (evp == NULL) break; TAILQ_REMOVE(&hq->q, evp, entry); hq->count -= 1; evp->holdq = 0; /* When released, all envelopes are put in the pending queue * and will be rescheduled immediately. As an optimization, * we could just schedule them directly. */ evp->state = RQ_EVPSTATE_PENDING; if (update) evp->flags |= RQ_ENVELOPE_UPDATE; sorted_insert(&ramqueue, evp); } if (TAILQ_EMPTY(&hq->q)) { tree_xpop(&holdqs[type], holdq); free(hq); stat_decrement("scheduler.ramqueue.holdq", 1); } stat_decrement("scheduler.ramqueue.hold", i); return (i); } static int scheduler_ram_batch(int mask, int *delay, size_t *count, uint64_t *evpids, int *types) { struct rq_envelope *evp; size_t i, n; time_t t; currtime = time(NULL); rq_queue_schedule(&ramqueue); if (tracing & TRACE_SCHEDULER) rq_queue_dump(&ramqueue, "scheduler_ram_batch()"); i = 0; n = 0; for (;;) { if (mask & SCHED_REMOVE && (evp = TAILQ_FIRST(&ramqueue.q_removed))) { TAILQ_REMOVE(&ramqueue.q_removed, evp, entry); types[i] = SCHED_REMOVE; evpids[i] = evp->evpid; rq_envelope_delete(&ramqueue, evp); if (++i == *count) break; } if (mask & SCHED_EXPIRE && (evp = TAILQ_FIRST(&ramqueue.q_expired))) { TAILQ_REMOVE(&ramqueue.q_expired, evp, entry); types[i] = SCHED_EXPIRE; evpids[i] = evp->evpid; rq_envelope_delete(&ramqueue, evp); if (++i == *count) break; } if (mask & SCHED_UPDATE && (evp = TAILQ_FIRST(&ramqueue.q_update))) { TAILQ_REMOVE(&ramqueue.q_update, evp, entry); types[i] = SCHED_UPDATE; evpids[i] = evp->evpid; if (evp->flags & RQ_ENVELOPE_OVERFLOW) t = BACKOFF_OVERFLOW; else if (evp->type == D_MTA) t = BACKOFF_TRANSFER; else t = BACKOFF_DELIVERY; evp->sched = scheduler_next(evp->ctime, t, 0); evp->flags &= ~(RQ_ENVELOPE_UPDATE|RQ_ENVELOPE_OVERFLOW); evp->state = RQ_EVPSTATE_PENDING; if (!(evp->flags & RQ_ENVELOPE_SUSPEND)) sorted_insert(&ramqueue, evp); if (++i == *count) break; } if (mask & SCHED_BOUNCE && (evp = TAILQ_FIRST(&ramqueue.q_bounce))) { TAILQ_REMOVE(&ramqueue.q_bounce, evp, entry); types[i] = SCHED_BOUNCE; evpids[i] = evp->evpid; TAILQ_INSERT_TAIL(&ramqueue.q_inflight, evp, entry); evp->state = RQ_EVPSTATE_INFLIGHT; evp->t_inflight = currtime; if (++i == *count) break; } if (mask & SCHED_MDA && (evp = TAILQ_FIRST(&ramqueue.q_mda))) { TAILQ_REMOVE(&ramqueue.q_mda, evp, entry); types[i] = SCHED_MDA; evpids[i] = evp->evpid; TAILQ_INSERT_TAIL(&ramqueue.q_inflight, evp, entry); evp->state = RQ_EVPSTATE_INFLIGHT; evp->t_inflight = currtime; if (++i == *count) break; } if (mask & SCHED_MTA && (evp = TAILQ_FIRST(&ramqueue.q_mta))) { TAILQ_REMOVE(&ramqueue.q_mta, evp, entry); types[i] = SCHED_MTA; evpids[i] = evp->evpid; TAILQ_INSERT_TAIL(&ramqueue.q_inflight, evp, entry); evp->state = RQ_EVPSTATE_INFLIGHT; evp->t_inflight = currtime; if (++i == *count) break; } /* nothing seen this round */ if (i == n) break; n = i; } if (i) { *count = i; return (1); } if ((evp = TAILQ_FIRST(&ramqueue.q_pending))) { if (evp->sched < evp->expire) t = evp->sched; else t = evp->expire; *delay = (t < currtime) ? 0 : (t - currtime); } else *delay = -1; return (0); } static size_t scheduler_ram_messages(uint32_t from, uint32_t *dst, size_t size) { uint64_t id; size_t n; void *i; for (n = 0, i = NULL; n < size; n++) { if (tree_iterfrom(&ramqueue.messages, &i, from, &id, NULL) == 0) break; dst[n] = id; } return (n); } static size_t scheduler_ram_envelopes(uint64_t from, struct evpstate *dst, size_t size) { struct rq_message *msg; struct rq_envelope *evp; void *i; size_t n; if ((msg = tree_get(&ramqueue.messages, evpid_to_msgid(from))) == NULL) return (0); for (n = 0, i = NULL; n < size; ) { if (tree_iterfrom(&msg->envelopes, &i, from, NULL, (void**)&evp) == 0) break; if (evp->flags & (RQ_ENVELOPE_REMOVED | RQ_ENVELOPE_EXPIRED)) continue; dst[n].evpid = evp->evpid; dst[n].flags = 0; dst[n].retry = 0; dst[n].time = 0; if (evp->state == RQ_EVPSTATE_PENDING) { dst[n].time = evp->sched; dst[n].flags = EF_PENDING; } else if (evp->state == RQ_EVPSTATE_SCHEDULED) { dst[n].time = evp->t_scheduled; dst[n].flags = EF_PENDING; } else if (evp->state == RQ_EVPSTATE_INFLIGHT) { dst[n].time = evp->t_inflight; dst[n].flags = EF_INFLIGHT; } else if (evp->state == RQ_EVPSTATE_HELD) { /* same as scheduled */ dst[n].time = evp->t_scheduled; dst[n].flags = EF_PENDING; dst[n].flags |= EF_HOLD; } if (evp->flags & RQ_ENVELOPE_SUSPEND) dst[n].flags |= EF_SUSPEND; n++; } return (n); } static int scheduler_ram_schedule(uint64_t evpid) { struct rq_message *msg; struct rq_envelope *evp; uint32_t msgid; void *i; int r; currtime = time(NULL); if (evpid > 0xffffffff) { msgid = evpid_to_msgid(evpid); if ((msg = tree_get(&ramqueue.messages, msgid)) == NULL) return (0); if ((evp = tree_get(&msg->envelopes, evpid)) == NULL) return (0); if (evp->state == RQ_EVPSTATE_INFLIGHT) return (0); rq_envelope_schedule(&ramqueue, evp); return (1); } else { msgid = evpid; if ((msg = tree_get(&ramqueue.messages, msgid)) == NULL) return (0); i = NULL; r = 0; while (tree_iter(&msg->envelopes, &i, NULL, (void*)(&evp))) { if (evp->state == RQ_EVPSTATE_INFLIGHT) continue; rq_envelope_schedule(&ramqueue, evp); r++; } return (r); } } static int scheduler_ram_remove(uint64_t evpid) { struct rq_message *msg; struct rq_envelope *evp; uint32_t msgid; void *i; int r; currtime = time(NULL); if (evpid > 0xffffffff) { msgid = evpid_to_msgid(evpid); if ((msg = tree_get(&ramqueue.messages, msgid)) == NULL) return (0); if ((evp = tree_get(&msg->envelopes, evpid)) == NULL) return (0); if (rq_envelope_remove(&ramqueue, evp)) return (1); return (0); } else { msgid = evpid; if ((msg = tree_get(&ramqueue.messages, msgid)) == NULL) return (0); i = NULL; r = 0; while (tree_iter(&msg->envelopes, &i, NULL, (void*)(&evp))) if (rq_envelope_remove(&ramqueue, evp)) r++; return (r); } } static int scheduler_ram_suspend(uint64_t evpid) { struct rq_message *msg; struct rq_envelope *evp; uint32_t msgid; void *i; int r; currtime = time(NULL); if (evpid > 0xffffffff) { msgid = evpid_to_msgid(evpid); if ((msg = tree_get(&ramqueue.messages, msgid)) == NULL) return (0); if ((evp = tree_get(&msg->envelopes, evpid)) == NULL) return (0); if (rq_envelope_suspend(&ramqueue, evp)) return (1); return (0); } else { msgid = evpid; if ((msg = tree_get(&ramqueue.messages, msgid)) == NULL) return (0); i = NULL; r = 0; while (tree_iter(&msg->envelopes, &i, NULL, (void*)(&evp))) if (rq_envelope_suspend(&ramqueue, evp)) r++; return (r); } } static int scheduler_ram_resume(uint64_t evpid) { struct rq_message *msg; struct rq_envelope *evp; uint32_t msgid; void *i; int r; currtime = time(NULL); if (evpid > 0xffffffff) { msgid = evpid_to_msgid(evpid); if ((msg = tree_get(&ramqueue.messages, msgid)) == NULL) return (0); if ((evp = tree_get(&msg->envelopes, evpid)) == NULL) return (0); if (rq_envelope_resume(&ramqueue, evp)) return (1); return (0); } else { msgid = evpid; if ((msg = tree_get(&ramqueue.messages, msgid)) == NULL) return (0); i = NULL; r = 0; while (tree_iter(&msg->envelopes, &i, NULL, (void*)(&evp))) if (rq_envelope_resume(&ramqueue, evp)) r++; return (r); } } static int scheduler_ram_query(uint64_t evpid) { uint32_t msgid; if (evpid > 0xffffffff) msgid = evpid_to_msgid(evpid); else msgid = evpid; if (tree_get(&ramqueue.messages, msgid) == NULL) return (0); return (1); } static void sorted_insert(struct rq_queue *rq, struct rq_envelope *evp) { struct rq_envelope *evp2; SPLAY_INSERT(prioqtree, &rq->q_priotree, evp); evp2 = SPLAY_NEXT(prioqtree, &rq->q_priotree, evp); if (evp2) TAILQ_INSERT_BEFORE(evp2, evp, entry); else TAILQ_INSERT_TAIL(&rq->q_pending, evp, entry); } static void rq_queue_init(struct rq_queue *rq) { memset(rq, 0, sizeof *rq); tree_init(&rq->messages); TAILQ_INIT(&rq->q_pending); TAILQ_INIT(&rq->q_inflight); TAILQ_INIT(&rq->q_mta); TAILQ_INIT(&rq->q_mda); TAILQ_INIT(&rq->q_bounce); TAILQ_INIT(&rq->q_update); TAILQ_INIT(&rq->q_expired); TAILQ_INIT(&rq->q_removed); SPLAY_INIT(&rq->q_priotree); } static void rq_queue_merge(struct rq_queue *rq, struct rq_queue *update) { struct rq_message *message, *tomessage; struct rq_envelope *envelope; uint64_t id; void *i; while (tree_poproot(&update->messages, &id, (void*)&message)) { if ((tomessage = tree_get(&rq->messages, id)) == NULL) { /* message does not exist. re-use structure */ tree_xset(&rq->messages, id, message); continue; } /* need to re-link all envelopes before merging them */ i = NULL; while ((tree_iter(&message->envelopes, &i, &id, (void*)&envelope))) envelope->message = tomessage; tree_merge(&tomessage->envelopes, &message->envelopes); free(message); stat_decrement("scheduler.ramqueue.message", 1); } /* Sorted insert in the pending queue */ while ((envelope = TAILQ_FIRST(&update->q_pending))) { TAILQ_REMOVE(&update->q_pending, envelope, entry); sorted_insert(rq, envelope); } rq->evpcount += update->evpcount; } #define SCHEDULEMAX 1024 static void rq_queue_schedule(struct rq_queue *rq) { struct rq_envelope *evp; size_t n; n = 0; while ((evp = TAILQ_FIRST(&rq->q_pending))) { if (evp->sched > currtime && evp->expire > currtime) break; if (n == SCHEDULEMAX) break; if (evp->state != RQ_EVPSTATE_PENDING) errx(1, "evp:%016" PRIx64 " flags=0x%x", evp->evpid, evp->flags); if (evp->expire <= currtime) { TAILQ_REMOVE(&rq->q_pending, evp, entry); SPLAY_REMOVE(prioqtree, &rq->q_priotree, evp); TAILQ_INSERT_TAIL(&rq->q_expired, evp, entry); evp->state = RQ_EVPSTATE_SCHEDULED; evp->flags |= RQ_ENVELOPE_EXPIRED; evp->t_scheduled = currtime; continue; } rq_envelope_schedule(rq, evp); n += 1; } } static struct evplist * rq_envelope_list(struct rq_queue *rq, struct rq_envelope *evp) { switch (evp->state) { case RQ_EVPSTATE_PENDING: return &rq->q_pending; case RQ_EVPSTATE_SCHEDULED: if (evp->flags & RQ_ENVELOPE_EXPIRED) return &rq->q_expired; if (evp->flags & RQ_ENVELOPE_REMOVED) return &rq->q_removed; if (evp->flags & RQ_ENVELOPE_UPDATE) return &rq->q_update; if (evp->type == D_MTA) return &rq->q_mta; if (evp->type == D_MDA) return &rq->q_mda; if (evp->type == D_BOUNCE) return &rq->q_bounce; errx(1, "%016" PRIx64 " bad evp type %d", evp->evpid, evp->type); case RQ_EVPSTATE_INFLIGHT: return &rq->q_inflight; case RQ_EVPSTATE_HELD: return (NULL); } errx(1, "%016" PRIx64 " bad state %d", evp->evpid, evp->state); return (NULL); } static void rq_envelope_schedule(struct rq_queue *rq, struct rq_envelope *evp) { struct rq_holdq *hq; struct evplist *q = NULL; switch (evp->type) { case D_MTA: q = &rq->q_mta; break; case D_MDA: q = &rq->q_mda; break; case D_BOUNCE: q = &rq->q_bounce; break; } if (evp->flags & RQ_ENVELOPE_UPDATE) q = &rq->q_update; if (evp->state == RQ_EVPSTATE_HELD) { hq = tree_xget(&holdqs[evp->type], evp->holdq); TAILQ_REMOVE(&hq->q, evp, entry); hq->count -= 1; if (TAILQ_EMPTY(&hq->q)) { tree_xpop(&holdqs[evp->type], evp->holdq); free(hq); } evp->holdq = 0; stat_decrement("scheduler.ramqueue.hold", 1); } else if (!(evp->flags & RQ_ENVELOPE_SUSPEND)) { TAILQ_REMOVE(&rq->q_pending, evp, entry); SPLAY_REMOVE(prioqtree, &rq->q_priotree, evp); } TAILQ_INSERT_TAIL(q, evp, entry); evp->state = RQ_EVPSTATE_SCHEDULED; evp->t_scheduled = currtime; } static int rq_envelope_remove(struct rq_queue *rq, struct rq_envelope *evp) { struct rq_holdq *hq; struct evplist *evl; if (evp->flags & (RQ_ENVELOPE_REMOVED | RQ_ENVELOPE_EXPIRED)) return (0); /* * If envelope is inflight, mark it envelope for removal. */ if (evp->state == RQ_EVPSTATE_INFLIGHT) { evp->flags |= RQ_ENVELOPE_REMOVED; return (1); } if (evp->state == RQ_EVPSTATE_HELD) { hq = tree_xget(&holdqs[evp->type], evp->holdq); TAILQ_REMOVE(&hq->q, evp, entry); hq->count -= 1; if (TAILQ_EMPTY(&hq->q)) { tree_xpop(&holdqs[evp->type], evp->holdq); free(hq); } evp->holdq = 0; stat_decrement("scheduler.ramqueue.hold", 1); } else if (!(evp->flags & RQ_ENVELOPE_SUSPEND)) { evl = rq_envelope_list(rq, evp); TAILQ_REMOVE(evl, evp, entry); if (evl == &rq->q_pending) SPLAY_REMOVE(prioqtree, &rq->q_priotree, evp); } TAILQ_INSERT_TAIL(&rq->q_removed, evp, entry); evp->state = RQ_EVPSTATE_SCHEDULED; evp->flags |= RQ_ENVELOPE_REMOVED; evp->t_scheduled = currtime; return (1); } static int rq_envelope_suspend(struct rq_queue *rq, struct rq_envelope *evp) { struct rq_holdq *hq; struct evplist *evl; if (evp->flags & RQ_ENVELOPE_SUSPEND) return (0); if (evp->state == RQ_EVPSTATE_HELD) { hq = tree_xget(&holdqs[evp->type], evp->holdq); TAILQ_REMOVE(&hq->q, evp, entry); hq->count -= 1; if (TAILQ_EMPTY(&hq->q)) { tree_xpop(&holdqs[evp->type], evp->holdq); free(hq); } evp->holdq = 0; evp->state = RQ_EVPSTATE_PENDING; stat_decrement("scheduler.ramqueue.hold", 1); } else if (evp->state != RQ_EVPSTATE_INFLIGHT) { evl = rq_envelope_list(rq, evp); TAILQ_REMOVE(evl, evp, entry); if (evl == &rq->q_pending) SPLAY_REMOVE(prioqtree, &rq->q_priotree, evp); } evp->flags |= RQ_ENVELOPE_SUSPEND; return (1); } static int rq_envelope_resume(struct rq_queue *rq, struct rq_envelope *evp) { struct evplist *evl; if (!(evp->flags & RQ_ENVELOPE_SUSPEND)) return (0); if (evp->state != RQ_EVPSTATE_INFLIGHT) { evl = rq_envelope_list(rq, evp); if (evl == &rq->q_pending) sorted_insert(rq, evp); else TAILQ_INSERT_TAIL(evl, evp, entry); } evp->flags &= ~RQ_ENVELOPE_SUSPEND; return (1); } static void rq_envelope_delete(struct rq_queue *rq, struct rq_envelope *evp) { tree_xpop(&evp->message->envelopes, evp->evpid); if (tree_empty(&evp->message->envelopes)) { tree_xpop(&rq->messages, evp->message->msgid); free(evp->message); stat_decrement("scheduler.ramqueue.message", 1); } free(evp); rq->evpcount--; stat_decrement("scheduler.ramqueue.envelope", 1); } static const char * rq_envelope_to_text(struct rq_envelope *e) { static char buf[256]; char t[64]; (void)snprintf(buf, sizeof buf, "evp:%016" PRIx64 " [", e->evpid); if (e->type == D_BOUNCE) (void)strlcat(buf, "bounce", sizeof buf); else if (e->type == D_MDA) (void)strlcat(buf, "mda", sizeof buf); else if (e->type == D_MTA) (void)strlcat(buf, "mta", sizeof buf); (void)snprintf(t, sizeof t, ",expire=%s", duration_to_text(e->expire - currtime)); (void)strlcat(buf, t, sizeof buf); switch (e->state) { case RQ_EVPSTATE_PENDING: (void)snprintf(t, sizeof t, ",pending=%s", duration_to_text(e->sched - currtime)); (void)strlcat(buf, t, sizeof buf); break; case RQ_EVPSTATE_SCHEDULED: (void)snprintf(t, sizeof t, ",scheduled=%s", duration_to_text(currtime - e->t_scheduled)); (void)strlcat(buf, t, sizeof buf); break; case RQ_EVPSTATE_INFLIGHT: (void)snprintf(t, sizeof t, ",inflight=%s", duration_to_text(currtime - e->t_inflight)); (void)strlcat(buf, t, sizeof buf); break; case RQ_EVPSTATE_HELD: (void)snprintf(t, sizeof t, ",held=%s", duration_to_text(currtime - e->t_inflight)); (void)strlcat(buf, t, sizeof buf); break; default: errx(1, "%016" PRIx64 " bad state %d", e->evpid, e->state); } if (e->flags & RQ_ENVELOPE_REMOVED) (void)strlcat(buf, ",removed", sizeof buf); if (e->flags & RQ_ENVELOPE_EXPIRED) (void)strlcat(buf, ",expired", sizeof buf); if (e->flags & RQ_ENVELOPE_SUSPEND) (void)strlcat(buf, ",suspended", sizeof buf); (void)strlcat(buf, "]", sizeof buf); return (buf); } static void rq_queue_dump(struct rq_queue *rq, const char * name) { struct rq_message *message; struct rq_envelope *envelope; void *i, *j; uint64_t id; log_debug("debug: /--- ramqueue: %s", name); i = NULL; while ((tree_iter(&rq->messages, &i, &id, (void*)&message))) { log_debug("debug: | msg:%08" PRIx32, message->msgid); j = NULL; while ((tree_iter(&message->envelopes, &j, &id, (void*)&envelope))) log_debug("debug: | %s", rq_envelope_to_text(envelope)); } log_debug("debug: \\---"); } static int rq_envelope_cmp(struct rq_envelope *e1, struct rq_envelope *e2) { time_t ref1, ref2; ref1 = (e1->sched < e1->expire) ? e1->sched : e1->expire; ref2 = (e2->sched < e2->expire) ? e2->sched : e2->expire; if (ref1 != ref2) return (ref1 < ref2) ? -1 : 1; if (e1->evpid != e2->evpid) return (e1->evpid < e2->evpid) ? -1 : 1; return 0; } SPLAY_GENERATE(prioqtree, rq_envelope, t_entry, rq_envelope_cmp);