diff options
Diffstat (limited to 'erts/emulator')
42 files changed, 647 insertions, 293 deletions
diff --git a/erts/emulator/beam/beam_emu.c b/erts/emulator/beam/beam_emu.c index 73bf443372..ea01ce597d 100644 --- a/erts/emulator/beam/beam_emu.c +++ b/erts/emulator/beam/beam_emu.c @@ -1480,7 +1480,16 @@ next_catch(Process* c_p, Eterm *reg) { ptr = prev = c_p->stop; ASSERT(ptr <= STACK_START(c_p)); - if (ptr == STACK_START(c_p)) return NULL; + + /* This function is only called if we have active catch tags or have + * previously called a function that was exception-traced. As the exception + * trace flag isn't cleared after the traced function returns (and the + * catch tag inserted by it is gone), it's possible to land here with an + * empty stack, and the process should simply die when that happens. */ + if (ptr == STACK_START(c_p)) { + ASSERT(!active_catches && IS_TRACED_FL(c_p, F_EXCEPTION_TRACE)); + return NULL; + } /* * Better safe than sorry here. In debug builds, produce a core @@ -3258,20 +3267,23 @@ erts_is_builtin(Eterm Mod, Eterm Name, int arity) /* - * Return the current number of reductions for the given process. + * Return the current number of reductions consumed by the given process. * To get the total number of reductions, p->reds must be added. */ Uint -erts_current_reductions(Process *current, Process *p) +erts_current_reductions(Process *c_p, Process *p) { - if (current != p) { + Sint reds_left; + if (c_p != p || !(erts_atomic32_read_nob(&c_p->state) + & ERTS_PSFLG_RUNNING)) { return 0; - } else if (current->fcalls < 0 && ERTS_PROC_GET_SAVED_CALLS_BUF(current)) { - return current->fcalls + CONTEXT_REDS; + } else if (c_p->fcalls < 0 && ERTS_PROC_GET_SAVED_CALLS_BUF(c_p)) { + reds_left = c_p->fcalls + CONTEXT_REDS; } else { - return REDS_IN(current) - current->fcalls; + reds_left = c_p->fcalls; } + return REDS_IN(c_p) - reds_left; } int diff --git a/erts/emulator/beam/beam_load.c b/erts/emulator/beam/beam_load.c index 21740caa2c..941c3ebbbe 100644 --- a/erts/emulator/beam/beam_load.c +++ b/erts/emulator/beam/beam_load.c @@ -3145,6 +3145,13 @@ is_killed(LoaderState* stp, GenOpArg Reg, GenOpArg Live) Live.val <= Reg.val; } +static int +is_killed_by_call_fun(LoaderState* stp, GenOpArg Reg, GenOpArg Live) +{ + return Reg.type == TAG_x && Live.type == TAG_u && + Live.val+1 <= Reg.val; +} + /* * Test whether register Reg is killed by make_fun instruction that * creates the fun given by index idx. @@ -3347,19 +3354,12 @@ gen_get_binary2(LoaderState* stp, GenOpArg Fail, GenOpArg Ms, GenOpArg Live, NATIVE_ENDIAN(Flags); if (Size.type == TAG_a && Size.val == am_all) { - if (Ms.type == Dst.type && Ms.val == Dst.val) { - GENOP_NAME_ARITY(op, i_bs_get_binary_all_reuse, 3); - op->a[0] = Ms; - op->a[1] = Fail; - op->a[2] = Unit; - } else { - GENOP_NAME_ARITY(op, i_bs_get_binary_all2, 5); - op->a[0] = Ms; - op->a[1] = Fail; - op->a[2] = Live; - op->a[3] = Unit; - op->a[4] = Dst; - } + GENOP_NAME_ARITY(op, i_bs_get_binary_all2, 5); + op->a[0] = Ms; + op->a[1] = Fail; + op->a[2] = Live; + op->a[3] = Unit; + op->a[4] = Dst; } else if (Size.type == TAG_i) { GENOP_NAME_ARITY(op, i_bs_get_binary_imm2, 6); op->a[0] = Ms; diff --git a/erts/emulator/beam/break.c b/erts/emulator/beam/break.c index 27bf2187c2..06d8c7cda8 100644 --- a/erts/emulator/beam/break.c +++ b/erts/emulator/beam/break.c @@ -39,6 +39,7 @@ #include "erl_hl_timer.h" #include "erl_thr_progress.h" #include "erl_proc_sig_queue.h" +#include "dist.h" /* Forward declarations -- should really appear somewhere else */ static void process_killer(void); @@ -74,6 +75,7 @@ port_info(fmtfn_t to, void *to_arg) void process_info(fmtfn_t to, void *to_arg) { + ErtsSchedulerData *esdp = erts_get_scheduler_data(); int i, max = erts_ptab_max(&erts_proc); for (i = 0; i < max; i++) { Process *p = erts_pix2proc(i); @@ -81,11 +83,35 @@ process_info(fmtfn_t to, void *to_arg) /* Do not include processes with no heap, * they are most likely just created and has invalid data */ - if (!ERTS_PROC_IS_EXITING(p) && p->heap != NULL) - print_process_info(to, to_arg, p, 0); + if (p->heap != NULL) { + ErtsProcLocks locks = (p == esdp->current_process || + p == esdp->free_process) ? ERTS_PROC_LOCK_MAIN : 0; + print_process_info(to, to_arg, p, locks); + } } } + /* Look for FREE processes in the run-queues and dist entries. + These have been removed from the ptab but we still want them + in the crash dump for debugging. */ + + /* First loop through all run-queues */ + for (i = 0; i < erts_no_schedulers + ERTS_NUM_DIRTY_RUNQS; i++) { + ErtsRunQueue *rq = ERTS_RUNQ_IX(i); + int j; + for (j = 0; j < ERTS_NO_PROC_PRIO_QUEUES; j++) { + Process *p = rq->procs.prio[j].first; + while (p) { + if (ERTS_PSFLG_FREE & erts_atomic32_read_acqb(&p->state)) + print_process_info(to, to_arg, p, 0); + p = p->next; + } + } + } + + /* Then check all dist entries */ + erts_dist_print_procs_suspended_on_de(to, to_arg); + port_info(to, to_arg); } @@ -199,13 +225,14 @@ static int doit_print_monitor(ErtsMonitor *mon, void *vpcontext, Sint reds) } return 1; } - + /* Display info about an individual Erlang process */ void print_process_info(fmtfn_t to, void *to_arg, Process *p, ErtsProcLocks orig_locks) { int garbing = 0; int running = 0; + int exiting = 0; Sint len; struct saved_calls *scb; erts_aint32_t state; @@ -226,9 +253,12 @@ print_process_info(fmtfn_t to, void *to_arg, Process *p, ErtsProcLocks orig_lock | ERTS_PSFLG_DIRTY_RUNNING)) running = 1; + if (state & ERTS_PSFLG_EXITING) + exiting = 1; + if (!(locks & ERTS_PROC_LOCK_MAIN)) { locks |= ERTS_PROC_LOCK_MAIN; - if (ERTS_IS_CRASH_DUMPING && running) { + if (ERTS_IS_CRASH_DUMPING) { if (erts_proc_trylock(p, locks)) { /* crash dumping and main lock taken, this probably means that the process is doing a GC on a dirty-scheduler... so we cannot @@ -246,7 +276,7 @@ print_process_info(fmtfn_t to, void *to_arg, Process *p, ErtsProcLocks orig_lock * If the process is registered as a global process, display the * registered name */ - if (p->common.u.alive.reg) + if (!ERTS_PROC_IS_EXITING(p) && p->common.u.alive.reg) erts_print(to, to_arg, "Name: %T\n", p->common.u.alive.reg->name); /* @@ -332,7 +362,7 @@ print_process_info(fmtfn_t to, void *to_arg, Process *p, ErtsProcLocks orig_lock } /* display the links only if there are any*/ - if (ERTS_P_LINKS(p) || ERTS_P_MONITORS(p) || ERTS_P_LT_MONITORS(p)) { + if (!exiting && (ERTS_P_LINKS(p) || ERTS_P_MONITORS(p) || ERTS_P_LT_MONITORS(p))) { PrintMonitorContext context = {1, to, to_arg}; erts_print(to, to_arg,"Link list: ["); erts_link_tree_foreach(ERTS_P_LINKS(p), doit_print_link, &context); diff --git a/erts/emulator/beam/bs_instrs.tab b/erts/emulator/beam/bs_instrs.tab index 652460a66d..9cad2b03c5 100644 --- a/erts/emulator/beam/bs_instrs.tab +++ b/erts/emulator/beam/bs_instrs.tab @@ -1136,7 +1136,6 @@ i_bs_get_utf16.execute(Fail, Flags, Dst) { } bs_context_to_binary := ctx_to_bin.fetch.execute; -i_bs_get_binary_all_reuse := ctx_to_bin.fetch_bin.execute; ctx_to_bin.head() { Eterm context; @@ -1159,16 +1158,6 @@ ctx_to_bin.fetch(Src) { } } -ctx_to_bin.fetch_bin(Src, Fail, Unit) { - context = $Src; - mb = ms_matchbuffer(context); - size = mb->size - mb->offset; - if (size % $Unit != 0) { - $FAIL($Fail); - } - offs = mb->offset; -} - ctx_to_bin.execute() { Uint hole_size; Uint orig = mb->orig; diff --git a/erts/emulator/beam/dist.c b/erts/emulator/beam/dist.c index a1da1addf9..8bbe6450eb 100644 --- a/erts/emulator/beam/dist.c +++ b/erts/emulator/beam/dist.c @@ -772,77 +772,25 @@ void init_dist(void) #define ErtsDistOutputBuf2Binary(OB) OB->bin -#ifdef DEBUG - -struct obuf_list; -struct obuf_list { - erts_refc_t refc; - struct obuf_list *next; - struct obuf_list *prev; -}; -#define obuf_list_size sizeof(struct obuf_list) -static struct obuf_list *erts_obuf_list = NULL; -static erts_mtx_t erts_obuf_list_mtx; - -static void -insert_obuf(struct obuf_list *obuf, erts_aint_t initial) { - erts_mtx_lock(&erts_obuf_list_mtx); - obuf->next = erts_obuf_list; - obuf->prev = NULL; - erts_refc_init(&obuf->refc, initial); - if (erts_obuf_list) - erts_obuf_list->prev = obuf; - erts_obuf_list = obuf; - erts_mtx_unlock(&erts_obuf_list_mtx); -} - -static void -remove_obuf(struct obuf_list *obuf) { - if (erts_refc_dectest(&obuf->refc, 0) == 0) { - erts_mtx_lock(&erts_obuf_list_mtx); - if (obuf->prev) { - obuf->prev->next = obuf->next; - } else { - erts_obuf_list = obuf->next; - } - if (obuf->next) obuf->next->prev = obuf->prev; - erts_mtx_unlock(&erts_obuf_list_mtx); - } -} - -void check_obuf(void); -void check_obuf(void) { - erts_mtx_lock(&erts_obuf_list_mtx); - ERTS_ASSERT(erts_obuf_list == NULL); - erts_mtx_unlock(&erts_obuf_list_mtx); -} -#else -#define insert_obuf(...) -#define remove_obuf(...) -#define obuf_list_size 0 -#endif - static ERTS_INLINE ErtsDistOutputBuf * alloc_dist_obuf(Uint size, Uint headers) { int i; ErtsDistOutputBuf *obuf; Uint obuf_size = sizeof(ErtsDistOutputBuf)*(headers) + - sizeof(byte)*size + obuf_list_size; + sizeof(byte)*size; Binary *bin = erts_bin_drv_alloc(obuf_size); - size += obuf_list_size; obuf = (ErtsDistOutputBuf *) &bin->orig_bytes[size]; erts_refc_add(&bin->intern.refc, headers - 1, 1); for (i = 0; i < headers; i++) { obuf[i].bin = bin; - obuf[i].extp = (byte *)&bin->orig_bytes[0] + obuf_list_size; + obuf[i].extp = (byte *)&bin->orig_bytes[0]; #ifdef DEBUG obuf[i].dbg_pattern = ERTS_DIST_OUTPUT_BUF_DBG_PATTERN; obuf[i].alloc_endp = obuf->extp + size; ASSERT(bin == ErtsDistOutputBuf2Binary(obuf)); #endif } - insert_obuf((struct obuf_list*)&bin->orig_bytes[0], headers); return obuf; } @@ -851,10 +799,7 @@ free_dist_obuf(ErtsDistOutputBuf *obuf) { Binary *bin = ErtsDistOutputBuf2Binary(obuf); ASSERT(obuf->dbg_pattern == ERTS_DIST_OUTPUT_BUF_DBG_PATTERN); - remove_obuf((struct obuf_list*)&bin->orig_bytes[0]); - if (erts_refc_dectest(&bin->intern.refc, 0) == 0) { - erts_bin_free(bin); - } + erts_bin_release(bin); } static ERTS_INLINE Sint @@ -1968,6 +1913,7 @@ int erts_net_message(Port *prt, goto invalid_message; } reason = tuple[5]; + edep = NULL; } if (is_not_ref(ref)) @@ -2014,12 +1960,14 @@ int erts_net_message(Port *prt, } token = NIL; reason = tuple[4]; + edep = NULL; } else if (type == DOP_EXIT_TT){ if (tuple_arity != 5) { goto invalid_message; } token = tuple[4]; reason = tuple[5]; + edep = NULL; } else if (type == DOP_PAYLOAD_EXIT) { if (tuple_arity != 3) { goto invalid_message; @@ -2067,12 +2015,14 @@ int erts_net_message(Port *prt, } reason = tuple[4]; token = NIL; + edep = NULL; } else if (type == DOP_EXIT2_TT) { if (tuple_arity != 5) { goto invalid_message; } token = tuple[4]; reason = tuple[5]; + edep = NULL; } else if (type == DOP_PAYLOAD_EXIT2) { if (tuple_arity != 3) { goto invalid_message; @@ -2430,8 +2380,8 @@ erts_dsig_send(ErtsDSigSendContext *ctx) case ERTS_DSIG_SEND_PHASE_FIN: { ASSERT(ctx->obuf->extp < ctx->obuf->ext_endp); - ASSERT(((byte*)&ctx->obuf->bin->orig_bytes[0]+obuf_list_size) <= ctx->obuf->extp - ctx->max_finalize_prepend); - ASSERT(ctx->obuf->ext_endp <= ((byte*)ctx->obuf->bin->orig_bytes+obuf_list_size) + ctx->data_size + ctx->dhdr_ext_size); + ASSERT(((byte*)&ctx->obuf->bin->orig_bytes[0]) <= ctx->obuf->extp - ctx->max_finalize_prepend); + ASSERT(ctx->obuf->ext_endp <= ((byte*)ctx->obuf->bin->orig_bytes) + ctx->data_size + ctx->dhdr_ext_size); ctx->data_size = ctx->obuf->ext_endp - ctx->obuf->extp; @@ -2935,7 +2885,9 @@ erts_dist_command(Port *prt, int initial_reds) ob = oq.first; ASSERT(ob); do { + obufsize += size_obuf(ob); reds = erts_encode_ext_dist_header_finalize(ob, dep, flags, reds); + obufsize -= size_obuf(ob); if (reds < 0) break; last_finalized = ob; @@ -2974,7 +2926,9 @@ erts_dist_command(Port *prt, int initial_reds) while (oq.first && !preempt) { ErtsDistOutputBuf *fob; Uint size; + obufsize += size_obuf(oq.first); reds = erts_encode_ext_dist_header_finalize(oq.first, dep, flags, reds); + obufsize -= size_obuf(oq.first); if (reds < 0) { preempt = 1; break; @@ -3407,14 +3361,13 @@ dist_ctrl_get_data_1(BIF_ALIST_1) { DistEntry *dep = ERTS_PROC_GET_DIST_ENTRY(BIF_P); const Sint initial_reds = ERTS_BIF_REDS_LEFT(BIF_P); - Sint reds = initial_reds; + Sint reds = initial_reds, obufsize = 0; ErtsDistOutputBuf *obuf; - Eterm *hp; + Eterm *hp, res; ProcBin *pb; erts_aint_t qsize; Uint32 conn_id, get_size; - Eterm res; - Uint hsz, bin_sz; + Uint hsz = 0, bin_sz; if (!dep) BIF_ERROR(BIF_P, EXC_NOTSUP); @@ -3466,7 +3419,9 @@ dist_ctrl_get_data_1(BIF_ALIST_1) } obuf = dep->tmp_out_queue.first; + obufsize += size_obuf(obuf); reds = erts_encode_ext_dist_header_finalize(obuf, dep, dep->flags, reds); + obufsize -= size_obuf(obuf); if (reds < 0) { erts_de_runlock(dep); ERTS_BIF_YIELD1(bif_export[BIF_dist_ctrl_get_data_1], @@ -3482,8 +3437,7 @@ dist_ctrl_get_data_1(BIF_ALIST_1) erts_de_runlock(dep); - bin_sz = obuf->ext_endp - obuf->extp; - hsz = PROC_BIN_SIZE; + bin_sz = obuf->ext_endp - obuf->extp + obuf->hdr_endp - obuf->hdrp; get_size = dep->opts & ERTS_DIST_CTRL_OPT_GET_SIZE; if (get_size) { @@ -3492,18 +3446,50 @@ dist_ctrl_get_data_1(BIF_ALIST_1) hsz += BIG_UINT_HEAP_SIZE; } - hp = HAlloc(BIF_P, hsz); - pb = (ProcBin *) (char *) hp; - pb->thing_word = HEADER_PROC_BIN; - pb->size = bin_sz; - pb->next = MSO(BIF_P).first; - MSO(BIF_P).first = (struct erl_off_heap_header*) pb; - pb->val = ErtsDistOutputBuf2Binary(obuf); - pb->bytes = (byte*) obuf->extp; - pb->flags = 0; - hp += PROC_BIN_SIZE; + if (!obuf->hdrp) { + hp = HAlloc(BIF_P, PROC_BIN_SIZE + hsz); + pb = (ProcBin *) (char *) hp; + pb->thing_word = HEADER_PROC_BIN; + pb->size = obuf->ext_endp - obuf->extp; + pb->next = MSO(BIF_P).first; + MSO(BIF_P).first = (struct erl_off_heap_header*) pb; + pb->val = ErtsDistOutputBuf2Binary(obuf); + pb->bytes = (byte*) obuf->extp; + pb->flags = 0; + res = make_binary(pb); + } else { + hp = HAlloc(BIF_P, PROC_BIN_SIZE * 2 + 4 + hsz); + pb = (ProcBin *) (char *) hp; + pb->thing_word = HEADER_PROC_BIN; + pb->size = obuf->ext_endp - obuf->extp; + pb->next = MSO(BIF_P).first; + MSO(BIF_P).first = (struct erl_off_heap_header*) pb; + pb->val = ErtsDistOutputBuf2Binary(obuf); + pb->bytes = (byte*) obuf->extp; + pb->flags = 0; + hp += PROC_BIN_SIZE; + + res = CONS(hp, make_binary(pb), NIL); + hp += 2; + + pb = (ProcBin *) (char *) hp; + pb->thing_word = HEADER_PROC_BIN; + pb->size = obuf->hdr_endp - obuf->hdrp; + pb->next = MSO(BIF_P).first; + MSO(BIF_P).first = (struct erl_off_heap_header*) pb; + pb->val = ErtsDistOutputBuf2Binary(obuf); + erts_refc_inc(&pb->val->intern.refc, 1); + pb->bytes = (byte*) obuf->hdrp; + pb->flags = 0; + hp += PROC_BIN_SIZE; + res = CONS(hp, make_binary(pb), res); + hp += 2; + } + + obufsize += size_obuf(obuf); + + qsize = erts_atomic_add_read_nob(&dep->qsize, (erts_aint_t) -obufsize); - qsize = erts_atomic_add_read_nob(&dep->qsize, -size_obuf(obuf)); ASSERT(qsize >= 0); if (qsize < erts_dist_buf_busy_limit/2 @@ -3518,8 +3504,6 @@ dist_ctrl_get_data_1(BIF_ALIST_1) } } - res = make_binary(pb); - if (get_size) { Eterm sz_term; if (IS_USMALL(0, bin_sz)) @@ -4711,10 +4695,6 @@ init_nodes_monitors(void) { erts_mtx_init(&nodes_monitors_mtx, "nodes_monitors", NIL, ERTS_LOCK_FLAGS_PROPERTY_STATIC | ERTS_LOCK_FLAGS_CATEGORY_DISTRIBUTION); -#ifdef DEBUG - erts_mtx_init(&erts_obuf_list_mtx, "sad", NIL, - ERTS_LOCK_FLAGS_PROPERTY_STATIC | ERTS_LOCK_FLAGS_CATEGORY_DISTRIBUTION); -#endif nodes_monitors = NULL; no_nodes_monitors = 0; } @@ -5105,3 +5085,22 @@ erts_processes_monitoring_nodes(Process *c_p) return ctxt.res; } + +static void +print_suspended_on_de(fmtfn_t to, void *to_arg, DistEntry *dep) +{ + for (; dep; dep = dep->next) { + ErtsProcList *curr = erts_proclist_peek_first(dep->suspended); + while (curr) { + if (!is_internal_pid(curr->u.pid)) + print_process_info(to, to_arg, curr->u.p, 0); + curr = erts_proclist_peek_next(dep->suspended, curr); + } + } +} + +void +erts_dist_print_procs_suspended_on_de(fmtfn_t to, void *to_arg) { + print_suspended_on_de(to, to_arg, erts_hidden_dist_entries); + print_suspended_on_de(to, to_arg, erts_visible_dist_entries); +} diff --git a/erts/emulator/beam/dist.h b/erts/emulator/beam/dist.h index 9b5e62ab7e..f953a2ab8c 100644 --- a/erts/emulator/beam/dist.h +++ b/erts/emulator/beam/dist.h @@ -319,5 +319,6 @@ extern int erts_dsig_prepare(ErtsDSigSendContext *, int, int); +void erts_dist_print_procs_suspended_on_de(fmtfn_t to, void *to_arg); int erts_auto_connect(DistEntry* dep, Process *proc, ErtsProcLocks proc_locks); #endif diff --git a/erts/emulator/beam/erl_bif_info.c b/erts/emulator/beam/erl_bif_info.c index 504aa8f943..a7424bbcb8 100644 --- a/erts/emulator/beam/erl_bif_info.c +++ b/erts/emulator/beam/erl_bif_info.c @@ -4442,6 +4442,7 @@ BIF_RETTYPE erts_debug_set_internal_state_2(BIF_ALIST_2) BIF_P->fcalls = reds; else BIF_P->fcalls = reds - CONTEXT_REDS; + BIF_P->scheduler_data->virtual_reds = 0; } BIF_RET(am_true); } diff --git a/erts/emulator/beam/erl_gc.c b/erts/emulator/beam/erl_gc.c index 9317850d96..67a73e4d57 100644 --- a/erts/emulator/beam/erl_gc.c +++ b/erts/emulator/beam/erl_gc.c @@ -577,7 +577,7 @@ force_reschedule: } static ERTS_FORCE_INLINE Uint -young_gen_usage(Process *p) +young_gen_usage(Process *p, Uint *ext_msg_usage) { Uint hsz; Eterm *aheap; @@ -604,7 +604,10 @@ young_gen_usage(Process *p) if (ERTS_SIG_IS_MSG(mp) && mp->data.attached && mp->data.attached != ERTS_MSG_COMBINED_HFRAG) { - hsz += erts_msg_attached_data_size(mp); + Uint sz = erts_msg_attached_data_size(mp); + if (ERTS_SIG_IS_EXTERNAL_MSG(mp)) + *ext_msg_usage += sz; + hsz += sz; } }); } @@ -676,6 +679,7 @@ garbage_collect(Process* p, ErlHeapFragment *live_hf_end, { Uint reclaimed_now = 0; Uint ygen_usage; + Uint ext_msg_usage = 0; Eterm gc_trace_end_tag; int reds; ErtsMonotonicTime start_time; @@ -698,7 +702,7 @@ garbage_collect(Process* p, ErlHeapFragment *live_hf_end, return delay_garbage_collection(p, live_hf_end, need, fcalls); } - ygen_usage = max_young_gen_usage ? max_young_gen_usage : young_gen_usage(p); + ygen_usage = max_young_gen_usage ? max_young_gen_usage : young_gen_usage(p, &ext_msg_usage); if (!ERTS_SCHEDULER_IS_DIRTY(esdp)) { check_for_possibly_long_gc(p, ygen_usage); @@ -739,7 +743,7 @@ garbage_collect(Process* p, ErlHeapFragment *live_hf_end, trace_gc(p, am_gc_minor_start, need, THE_NON_VALUE); } DTRACE2(gc_minor_start, pidbuf, need); - reds = minor_collection(p, live_hf_end, need, objv, nobj, + reds = minor_collection(p, live_hf_end, need + ext_msg_usage, objv, nobj, ygen_usage, &reclaimed_now); DTRACE2(gc_minor_end, pidbuf, reclaimed_now); if (reds == -1) { @@ -764,7 +768,7 @@ do_major_collection: trace_gc(p, am_gc_major_start, need, THE_NON_VALUE); } DTRACE2(gc_major_start, pidbuf, need); - reds = major_collection(p, live_hf_end, need, objv, nobj, + reds = major_collection(p, live_hf_end, need + ext_msg_usage, objv, nobj, ygen_usage, &reclaimed_now); if (ERTS_SCHEDULER_IS_DIRTY(esdp)) p->flags &= ~(F_DIRTY_MAJOR_GC|F_DIRTY_MINOR_GC); @@ -1101,6 +1105,7 @@ erts_garbage_collect_literals(Process* p, Eterm* literals, Eterm* old_htop; Uint n; Uint ygen_usage = 0; + Uint ext_msg_usage = 0; struct erl_off_heap_header** prev = NULL; Sint64 reds; int hibernated = !!(p->flags & F_HIBERNATED); @@ -1118,7 +1123,7 @@ erts_garbage_collect_literals(Process* p, Eterm* literals, p->flags &= ~F_DIRTY_CLA; else { Uint size = byte_lit_size/sizeof(Uint); - ygen_usage = young_gen_usage(p); + ygen_usage = young_gen_usage(p, &ext_msg_usage); if (hibernated) size = size*2 + 3*ygen_usage; else @@ -1130,7 +1135,7 @@ erts_garbage_collect_literals(Process* p, Eterm* literals, } } - reds = (Sint64) garbage_collect(p, ERTS_INVALID_HFRAG_PTR, 0, + reds = (Sint64) garbage_collect(p, ERTS_INVALID_HFRAG_PTR, ext_msg_usage, p->arg_reg, p->arity, fcalls, ygen_usage); if (ERTS_PROC_IS_EXITING(p)) { @@ -1348,6 +1353,9 @@ minor_collection(Process* p, ErlHeapFragment *live_hf_end, Eterm *mature = p->abandoned_heap ? p->abandoned_heap : p->heap; Uint mature_size = p->high_water - mature; Uint size_before = ygen_usage; +#ifdef DEBUG + Uint debug_tmp = 0; +#endif /* * Check if we have gone past the max heap size limit @@ -1484,7 +1492,7 @@ minor_collection(Process* p, ErlHeapFragment *live_hf_end, process from there */ ASSERT(!MAX_HEAP_SIZE_GET(p) || !(MAX_HEAP_SIZE_FLAGS_GET(p) & MAX_HEAP_SIZE_KILL) || - MAX_HEAP_SIZE_GET(p) > (young_gen_usage(p) + + MAX_HEAP_SIZE_GET(p) > (young_gen_usage(p, &debug_tmp) + (OLD_HEND(p) - OLD_HEAP(p)) + (HEAP_END(p) - HEAP_TOP(p)))); diff --git a/erts/emulator/beam/erl_init.c b/erts/emulator/beam/erl_init.c index 82d5140d1c..12750b9aa6 100644 --- a/erts/emulator/beam/erl_init.c +++ b/erts/emulator/beam/erl_init.c @@ -2417,17 +2417,12 @@ erts_exit_vv(int n, int flush_async, char *fmt, va_list args1, va_list args2) erts_exit_epilogue(); } -void check_obuf(void); __decl_noreturn void __noreturn erts_exit_epilogue(void) { int n = erts_exit_code; sys_tty_reset(n); -#ifdef DEBUG - check_obuf(); -#endif - if (n == ERTS_INTR_EXIT) exit(0); else if (n == ERTS_DUMP_EXIT) diff --git a/erts/emulator/beam/erl_lock_check.c b/erts/emulator/beam/erl_lock_check.c index 39eabb6710..3aab4828cc 100644 --- a/erts/emulator/beam/erl_lock_check.c +++ b/erts/emulator/beam/erl_lock_check.c @@ -164,8 +164,7 @@ static erts_lc_lock_order_t erts_lock_order[] = { { "os_monotonic_time", NULL }, { "erts_alloc_hard_debug", NULL }, { "hard_dbg_mseg", NULL }, - { "erts_mmap", NULL }, - { "sad", NULL} + { "erts_mmap", NULL } }; #define ERTS_LOCK_ORDER_SIZE \ diff --git a/erts/emulator/beam/erl_message.c b/erts/emulator/beam/erl_message.c index e350a20339..2a0fb9e2aa 100644 --- a/erts/emulator/beam/erl_message.c +++ b/erts/emulator/beam/erl_message.c @@ -1137,10 +1137,28 @@ change_to_off_heap: return res; } -void erts_factory_proc_init(ErtsHeapFactory* factory, - Process* p) +void erts_factory_proc_init(ErtsHeapFactory* factory, Process* p) { - erts_factory_proc_prealloc_init(factory, p, HEAP_LIMIT(p) - HEAP_TOP(p)); + /* This function does not use HAlloc to allocate on the heap + as we do not want to use INIT_HEAP_MEM on the allocated + heap as that completely destroys the DEBUG emulators + performance. */ + ErlHeapFragment *bp = p->mbuf; + factory->mode = FACTORY_HALLOC; + factory->p = p; + factory->hp_start = HEAP_TOP(p); + factory->hp = factory->hp_start; + factory->hp_end = HEAP_LIMIT(p); + factory->off_heap = &p->off_heap; + factory->message = NULL; + factory->off_heap_saved.first = p->off_heap.first; + factory->off_heap_saved.overhead = p->off_heap.overhead; + factory->heap_frags_saved = bp; + factory->heap_frags_saved_used = bp ? bp->used_size : 0; + factory->heap_frags = NULL; /* not used */ + factory->alloc_type = 0; /* not used */ + + HEAP_TOP(p) = HEAP_LIMIT(p); } void erts_factory_proc_prealloc_init(ErtsHeapFactory* factory, diff --git a/erts/emulator/beam/erl_message.h b/erts/emulator/beam/erl_message.h index e5f623a370..5bd25737a7 100644 --- a/erts/emulator/beam/erl_message.h +++ b/erts/emulator/beam/erl_message.h @@ -22,6 +22,7 @@ #define __ERL_MESSAGE_H__ #include "sys.h" +#include "erl_vm.h" #define ERTS_PROC_SIG_QUEUE_TYPE_ONLY #include "erl_proc_sig_queue.h" #undef ERTS_PROC_SIG_QUEUE_TYPE_ONLY @@ -117,6 +118,7 @@ erts_produce_heap(ErtsHeapFactory* factory, Uint need, Uint xtra) } res = factory->hp; factory->hp += need; + INIT_HEAP_MEM(res, need); return res; } diff --git a/erts/emulator/beam/erl_nif.h b/erts/emulator/beam/erl_nif.h index a599511c78..cc389a093f 100644 --- a/erts/emulator/beam/erl_nif.h +++ b/erts/emulator/beam/erl_nif.h @@ -59,12 +59,17 @@ */ #define ERL_NIF_MAJOR_VERSION 2 #define ERL_NIF_MINOR_VERSION 15 + /* - * WHEN CHANGING INTERFACE VERSION, also replace erts version below - * with ticket syntax like "erts-@OTP-12345@", or a temporary placeholder - * between two @ like "erts-@MyName@", if you don't know what a ticket is. + * WHEN CHANGING INTERFACE VERSION, also replace erts version below with + * a ticket number e.g. "erts-@OTP-12345@". The syntax is the same as for + * runtime dependencies so multiple tickets should be separated with ":", e.g. + * "erts-@OTP-12345:OTP-54321@". + * + * If you're not on the OTP team, you should use a placeholder like + * erts-@MyName@ instead. */ -#define ERL_NIF_MIN_ERTS_VERSION "erts-@OTP-15095 OTP-15640@ (OTP-22)" +#define ERL_NIF_MIN_ERTS_VERSION "erts-@OTP-15095:OTP-15640@" /* * The emulator will refuse to load a nif-lib with a major version diff --git a/erts/emulator/beam/erl_node_tables.c b/erts/emulator/beam/erl_node_tables.c index 215dc6fa71..49dea8919b 100644 --- a/erts/emulator/beam/erl_node_tables.c +++ b/erts/emulator/beam/erl_node_tables.c @@ -2295,7 +2295,7 @@ reference_table_term(Uint **hpp, ErlOffHeap *ohp, Uint *szp) tup = MK_2TUP(AM_ets, drp->id); } else { - ASSERT(!drp->ctrl_ref && drp->node_ref && !drp->signal_ref); + ASSERT(!drp->ctrl_ref && (drp->node_ref || drp->sequence_ref) && !drp->signal_ref); ASSERT(is_atom(drp->id)); tup = MK_2TUP(drp->id, MK_UINT(drp->creation)); tup = MK_2TUP(AM_node, tup); diff --git a/erts/emulator/beam/erl_posix_str.c b/erts/emulator/beam/erl_posix_str.c index 7b3e640d3f..5b515d6e78 100644 --- a/erts/emulator/beam/erl_posix_str.c +++ b/erts/emulator/beam/erl_posix_str.c @@ -171,6 +171,9 @@ erl_errno_id(error) #if defined(EIDRM) && (!defined(EINPROGRESS) || (EIDRM != EINPROGRESS)) case EIDRM: return "eidrm"; #endif +#ifdef EILSEQ + case EILSEQ: return "eilseq"; +#endif #ifdef EINIT case EINIT: return "einit"; #endif diff --git a/erts/emulator/beam/erl_proc_sig_queue.c b/erts/emulator/beam/erl_proc_sig_queue.c index bd59c4afa3..4e9f177e51 100644 --- a/erts/emulator/beam/erl_proc_sig_queue.c +++ b/erts/emulator/beam/erl_proc_sig_queue.c @@ -611,7 +611,8 @@ proc_queue_signal(Process *c_p, Eterm pid, ErtsSignal *sig, int op) #endif return 1; } - ASSERT(esdp->pending_signal.dbg_from == esdp->current_process); + ASSERT(esdp->pending_signal.dbg_from == esdp->current_process || + esdp->pending_signal.dbg_from == esdp->free_process); if (pend_sig != sig) { /* Switch them and send previously pending signal instead */ Eterm pend_to = esdp->pending_signal.to; @@ -3973,6 +3974,9 @@ clear_seq_trace_token(ErtsMessage *sig) case ERTS_MON_TYPE_PROC: case ERTS_MON_TYPE_DIST_PROC: case ERTS_MON_TYPE_NODE: + case ERTS_MON_TYPE_NODES: + case ERTS_MON_TYPE_SUSPEND: + case ERTS_MON_TYPE_TIME_OFFSET: break; default: ERTS_INTERNAL_ERROR("Unexpected sig type"); @@ -3989,6 +3993,11 @@ clear_seq_trace_token(ErtsMessage *sig) case ERTS_SIG_Q_OP_LINK: case ERTS_SIG_Q_OP_UNLINK: case ERTS_SIG_Q_OP_TRACE_CHANGE_STATE: + case ERTS_SIG_Q_OP_GROUP_LEADER: + case ERTS_SIG_Q_OP_IS_ALIVE: + case ERTS_SIG_Q_OP_PROCESS_INFO: + case ERTS_SIG_Q_OP_SYNC_SUSPEND: + case ERTS_SIG_Q_OP_RPC: break; default: diff --git a/erts/emulator/beam/erl_process.c b/erts/emulator/beam/erl_process.c index f34289339f..9e662632b4 100644 --- a/erts/emulator/beam/erl_process.c +++ b/erts/emulator/beam/erl_process.c @@ -12090,9 +12090,11 @@ erts_proc_exit_handle_dist_monitor(ErtsMonitor *mon, void *vctxt, Sint reds) ErtsDSigSendContext ctx; ErtsMonLnkDist *dist; DistEntry *dep; - Eterm watcher; + Eterm watcher, ref, *hp; ErtsMonitorData *mdp = NULL; Eterm watched; + Uint watcher_sz, ref_sz; + ErtsHeapFactory factory; ASSERT(erts_monitor_is_target(mon) && mon->type == ERTS_MON_TYPE_DIST_PROC); @@ -12124,10 +12126,18 @@ erts_proc_exit_handle_dist_monitor(ErtsMonitor *mon, void *vctxt, Sint reds) case ERTS_DSIG_PREP_CONNECTED: if (dist->connection_id != ctx.connection_id) break; + erts_factory_proc_init(&factory, c_p); + watcher_sz = size_object(watcher); + hp = erts_produce_heap(&factory, watcher_sz, 0); + watcher = copy_struct(watcher, watcher_sz, &hp, factory.off_heap); + ref_sz = size_object(mdp->ref); + hp = erts_produce_heap(&factory, ref_sz, 0); + ref = copy_struct(mdp->ref, ref_sz, &hp, factory.off_heap); + erts_factory_close(&factory); code = erts_dsig_send_m_exit(&ctx, watcher, watched, - mdp->ref, + ref, reason); switch (code) { case ERTS_DSIG_SEND_CONTINUE: @@ -12332,13 +12342,15 @@ erts_proc_exit_handle_dist_link(ErtsLink *lnk, void *vctxt, Sint reds) { ErtsProcExitContext *ctxt = (ErtsProcExitContext *) vctxt; Process *c_p = ctxt->c_p; - Eterm reason = ctxt->reason; + Eterm reason = ctxt->reason, item, *hp; + Uint item_sz; int code; ErtsDSigSendContext ctx; ErtsMonLnkDist *dist; DistEntry *dep; ErtsLink *dlnk; ErtsLinkData *ldp = NULL; + ErtsHeapFactory factory; ASSERT(lnk->type == ERTS_LNK_TYPE_DIST_PROC); dlnk = erts_link_to_other(lnk, &ldp); @@ -12365,9 +12377,14 @@ erts_proc_exit_handle_dist_link(ErtsLink *lnk, void *vctxt, Sint reds) case ERTS_DSIG_PREP_CONNECTED: if (dist->connection_id != ctx.connection_id) break; + erts_factory_proc_init(&factory, c_p); + item_sz = size_object(lnk->other.item); + hp = erts_produce_heap(&factory, item_sz, 0); + item = copy_struct(lnk->other.item, item_sz, &hp, factory.off_heap); + erts_factory_close(&factory); code = erts_dsig_send_exit_tt(&ctx, c_p->common.id, - lnk->other.item, + item, reason, SEQ_TRACE_TOKEN(c_p)); switch (code) { @@ -12584,6 +12601,8 @@ erts_continue_exit_process(Process *p) if (p->u.terminate) { trap_state = p->u.terminate; + /* Re-set the reason as it may have been gc:ed */ + trap_state->reason = p->fvalue; } else { trap_state->phase = ERTS_CONTINUE_EXIT_TIMERS; trap_state->reason = p->fvalue; @@ -12661,11 +12680,11 @@ restart: p->flags &= ~F_USING_DB; } - erts_set_gc_state(p, 1); - trap_state->phase = ERTS_CONTINUE_EXIT_CLEAN_SYS_TASKS; case ERTS_CONTINUE_EXIT_CLEAN_SYS_TASKS: - + + /* We enable GC again as it can produce more sys-tasks */ + erts_set_gc_state(p, 1); state = erts_atomic32_read_acqb(&p->state); if ((state & ERTS_PSFLG_SYS_TASKS) || p->dirty_sys_tasks) { reds -= cleanup_sys_tasks(p, state, reds); @@ -12788,6 +12807,9 @@ restart: erts_deref_dist_entry(trap_state->dep); } + /* Disable GC so that reason does not get moved */ + erts_set_gc_state(p, 0); + trap_state->pectxt.c_p = p; trap_state->pectxt.reason = trap_state->reason; trap_state->pectxt.dist_links = NULL; @@ -12872,13 +12894,13 @@ restart: switch (result) { case ERTS_DSIG_SEND_OK: case ERTS_DSIG_SEND_TOO_LRG: /*SEND_SYSTEM_LIMIT*/ - case ERTS_DSIG_SEND_YIELD: /*SEND_YIELD_RETURN*/ break; + case ERTS_DSIG_SEND_YIELD: /*SEND_YIELD_RETURN*/ case ERTS_DSIG_SEND_CONTINUE: { /*SEND_YIELD_CONTINUE*/ goto yield; } } - erts_set_gc_state(p, 1); + trap_state->pectxt.dist_state = NIL; if (reds <= 0) goto yield; @@ -12917,11 +12939,12 @@ restart: * From this point on we are no longer allowed to yield * this process. */ - #ifdef DEBUG yield_allowed = 0; #endif + erts_set_gc_state(p, 1); + /* Set state to not active as we don't want this process to be scheduled in again after this. */ state = erts_atomic32_read_band_relb(&p->state, @@ -12960,7 +12983,7 @@ restart: erts_free(ERTS_ALC_T_CONT_EXIT_TRAP, trap_state); p->u.terminate = NULL; } - + ERTS_CHK_HAVE_ONLY_MAIN_PROC_LOCK(p); if (IS_TRACED_FL(p, F_TRACE_SCHED_EXIT)) diff --git a/erts/emulator/beam/erl_process.h b/erts/emulator/beam/erl_process.h index 4ffa022d5c..711b73417d 100644 --- a/erts/emulator/beam/erl_process.h +++ b/erts/emulator/beam/erl_process.h @@ -1322,9 +1322,6 @@ ERTS_GLB_INLINE void erts_heap_frag_shrink(Process* p, Eterm* hp) #endif /* inline */ Eterm* erts_heap_alloc(Process* p, Uint need, Uint xtra); -#ifdef CHECK_FOR_HOLES -Eterm* erts_set_hole_marker(Eterm* ptr, Uint sz); -#endif extern erts_rwmtx_t erts_cpu_bind_rwmtx; /* If any of the erts_system_monitor_* variables are set (enabled), diff --git a/erts/emulator/beam/erl_process_dump.c b/erts/emulator/beam/erl_process_dump.c index a164ed543e..71262061dd 100644 --- a/erts/emulator/beam/erl_process_dump.c +++ b/erts/emulator/beam/erl_process_dump.c @@ -121,12 +121,14 @@ Uint erts_process_memory(Process *p, int include_sigs_in_transit) size += sizeof(Process); - erts_link_tree_foreach(ERTS_P_LINKS(p), - link_size, (void *) &size); - erts_monitor_tree_foreach(ERTS_P_MONITORS(p), - monitor_size, (void *) &size); - erts_monitor_list_foreach(ERTS_P_LT_MONITORS(p), - monitor_size, (void *) &size); + if ((erts_atomic32_read_nob(&p->state) & ERTS_PSFLG_EXITING) == 0) { + erts_link_tree_foreach(ERTS_P_LINKS(p), + link_size, (void *) &size); + erts_monitor_tree_foreach(ERTS_P_MONITORS(p), + monitor_size, (void *) &size); + erts_monitor_list_foreach(ERTS_P_LT_MONITORS(p), + monitor_size, (void *) &size); + } size += (p->heap_sz + p->mbuf_sz) * sizeof(Eterm); if (p->abandoned_heap) size += (p->hend - p->heap) * sizeof(Eterm); diff --git a/erts/emulator/beam/erl_utils.h b/erts/emulator/beam/erl_utils.h index 880febba8b..430ac305c5 100644 --- a/erts/emulator/beam/erl_utils.h +++ b/erts/emulator/beam/erl_utils.h @@ -69,7 +69,6 @@ int erts_fit_in_bits_int32(Sint32); int erts_fit_in_bits_uint(Uint); Sint erts_list_length(Eterm); int erts_is_builtin(Eterm, Eterm, int); -Uint32 block_hash(byte *, unsigned, Uint32); Uint32 make_hash2(Eterm); Uint32 make_hash(Eterm); Uint32 make_internal_hash(Eterm, Uint32 salt); diff --git a/erts/emulator/beam/erl_vm.h b/erts/emulator/beam/erl_vm.h index 35eae18394..e623148587 100644 --- a/erts/emulator/beam/erl_vm.h +++ b/erts/emulator/beam/erl_vm.h @@ -67,9 +67,10 @@ (unsigned long)HEAP_TOP(p),(sz),__FILE__,__LINE__)), \ */ # ifdef CHECK_FOR_HOLES -# define INIT_HEAP_MEM(p,sz) erts_set_hole_marker(HEAP_TOP(p), (sz)) +Eterm* erts_set_hole_marker(Eterm* ptr, Uint sz); +# define INIT_HEAP_MEM(p,sz) erts_set_hole_marker(p, (sz)) # else -# define INIT_HEAP_MEM(p,sz) sys_memset(HEAP_TOP(p),0x01,(sz)*sizeof(Eterm*)) +# define INIT_HEAP_MEM(p,sz) sys_memset(p,0x01,(sz)*sizeof(Eterm*)) # endif #else # define INIT_HEAP_MEM(p,sz) ((void)0) @@ -91,7 +92,7 @@ ErtsHAllocLockCheck(p), \ (IS_FORCE_HEAP_FRAGS || (((HEAP_LIMIT(p) - HEAP_TOP(p)) < (sz))) \ ? erts_heap_alloc((p),(sz),(xtra)) \ - : (INIT_HEAP_MEM(p,sz), \ + : (INIT_HEAP_MEM(HEAP_TOP(p),sz), \ HEAP_TOP(p) = HEAP_TOP(p) + (sz), HEAP_TOP(p) - (sz)))) #define HAlloc(P, SZ) HAllocX(P,SZ,0) diff --git a/erts/emulator/beam/msg_instrs.tab b/erts/emulator/beam/msg_instrs.tab index 6f8d1469ef..b08466c830 100644 --- a/erts/emulator/beam/msg_instrs.tab +++ b/erts/emulator/beam/msg_instrs.tab @@ -105,6 +105,7 @@ i_loop_rec(Dest) { $SET_CP_I_ABS(I); c_p->arity = 0; c_p->current = NULL; + c_p->fcalls = FCALLS; FCALLS -= erts_proc_sig_receive_helper(c_p, FCALLS, neg_o_reds, &msgp, &get_out); SWAPIN; diff --git a/erts/emulator/beam/ops.tab b/erts/emulator/beam/ops.tab index 6832e65b1b..10ca74cd60 100644 --- a/erts/emulator/beam/ops.tab +++ b/erts/emulator/beam/ops.tab @@ -338,7 +338,7 @@ swap_temp R1 R2 Tmp | line Loc | apply Live | is_killed_apply(Tmp, Live) => \ swap_temp R1 R2 Tmp | line Loc | apply_last Live D | is_killed_apply(Tmp, Live) => \ swap R1 R2 | line Loc | apply_last Live D -swap_temp R1 R2 Tmp | line Loc | call_fun Live | is_killed(Tmp, Live) => \ +swap_temp R1 R2 Tmp | line Loc | call_fun Live | is_killed_by_call_fun(Tmp, Live) => \ swap R1 R2 | line Loc | call_fun Live swap_temp R1 R2 Tmp | make_fun2 OldIndex=u | is_killed_by_make_fun(Tmp, OldIndex) => \ swap R1 R2 | make_fun2 OldIndex @@ -1262,7 +1262,6 @@ bs_get_binary2 Fail=f Ms=xy Live=u Sz=sq Unit=u Flags=u Dst=d => \ i_bs_get_binary_imm2 xy f? t W t d i_bs_get_binary2 xy f t? s t d i_bs_get_binary_all2 xy f? t t d -i_bs_get_binary_all_reuse xy f? t # Fetching float from binaries. bs_get_float2 Fail=f Ms=xy Live=u Sz=s Unit=u Flags=u Dst=d => \ @@ -1691,9 +1690,14 @@ i_plus S1=c S2=c Fail Dst => move S1 x | i_plus x S2 Fail Dst i_plus xy xyc j? d -i_minus x x j? d -i_minus c x j? d -i_minus s s j? d +# A minus instruction with a constant right operand will be +# converted to an i_increment instruction, except in guards or +# when the negated value of the constant won't fit in a guard. +# Therefore, it very rare. +i_minus S1 S2=c Fail Dst => move S2 x | i_minus S1 x Fail Dst + +i_minus xy xy j? d +i_minus c xy j? d i_times j? s s d diff --git a/erts/emulator/beam/utils.c b/erts/emulator/beam/utils.c index 36cfe0548e..0bbae65e28 100644 --- a/erts/emulator/beam/utils.c +++ b/erts/emulator/beam/utils.c @@ -1069,11 +1069,11 @@ do { \ #define HCONST 0x9e3779b9UL /* the golden ratio; an arbitrary value */ -Uint32 -block_hash(byte *k, unsigned length, Uint32 initval) +static Uint32 +block_hash(byte *k, Uint length, Uint32 initval) { Uint32 a,b,c; - unsigned len; + Uint len; /* Set up the internal state */ len = length; @@ -1749,7 +1749,7 @@ make_internal_hash(Eterm term, Uint32 salt) case SUB_BINARY_SUBTAG: { byte* bptr; - unsigned sz = binary_size(term); + Uint sz = binary_size(term); Uint32 con = HCONST_13 + hash; Uint bitoffs; Uint bitsize; diff --git a/erts/emulator/internal_doc/CarrierMigration.md b/erts/emulator/internal_doc/CarrierMigration.md index bb3d8aac28..40f6031ca8 100644 --- a/erts/emulator/internal_doc/CarrierMigration.md +++ b/erts/emulator/internal_doc/CarrierMigration.md @@ -1,6 +1,9 @@ Carrier Migration ================= +Introduction +------------ + The ERTS memory allocators manage memory blocks in two types of raw memory chunks. We call these chunks of raw memory *carriers*. Single-block carriers which only contain one large block, @@ -141,11 +144,11 @@ Since the carrier has been unlinked from the data structure of available free blocks, no more allocations will be made in the carrier. -The allocator instance that created a carrier is called its **owner**. +The allocator instance that created a carrier is called its *owner*. Ownership never changes. The allocator instance that has the responsibility to perform deallocations in a -carrier is called its **employer**. The employer may also perform allocations if +carrier is called its *employer*. The employer may also perform allocations if the carrier is not in the pool. Employment may change when a carrier is fetched from or inserted into the pool. @@ -153,14 +156,14 @@ Deallocations in a carrier, while it remains in the pool, is always performed the owner. That is, all pooled carriers are employed by their owners. Each carrier has an atomic word containing a pointer to the employing allocator -instance and three bit flags; IN_POOL, BUSY and HOMECOMING. +instance and three bit flags; IN\_POOL, BUSY and HOMECOMING. When fetching a carrier from the pool, employment may change and further deallocations in the carrier will be redirected to the new employer using the delayed dealloc functionality. When a foreign allocator instance abandons a carrier back into the pool, it will -also pass it back to its **owner** using the delayed dealloc queue. When doing +also pass it back to its *owner* using the delayed dealloc queue. When doing this it will set the HOMECOMING bit flag to mark it as "enqueued". The owner will later clear the HOMECOMING bit when the carrier is dequeued. This mechanism prevents a carrier from being enqueued again before it has been dequeued. @@ -180,14 +183,14 @@ back to the owner for deallocation using the delayed dealloc functionality. In short: -* The allocator instance that created a carrier **owns** it. -* An empty carrier is always deallocated by its **owner**. -* **Ownership** never changes. -* The allocator instance that uses a carrier **employs** it. -* An **employer** can abandon a carrier into the pool. +* The allocator instance that created a carrier *owns* it. +* An empty carrier is always deallocated by its *owner*. +* *Ownership* never changes. +* The allocator instance that uses a carrier *employs* it. +* An *employer* can abandon a carrier into the pool. * Pooled carriers are not allocated from. -* Pooled carriers are always **employed** by their **owner**. -* **Employment** can only change from **owner** to a foreign allocator +* Pooled carriers are always *employed* by their *owner*. +* *Employment* can only change from *owner* to a foreign allocator when a carrier is fetched from the pool. @@ -229,7 +232,7 @@ carrier. When the cluster gets to the same size as the search limit, all searches will essentially fail. To counter the "bad cluster" problem and also ease the contention, the -search will now always start by first looking at the allocators **own** +search will now always start by first looking at the allocators *own* carriers. That is, carriers that were initially created by the allocator itself and later had been abandoned to the pool. If none of our own abandoned carrier would do, then the search continues into the diff --git a/erts/emulator/internal_doc/CodeLoading.md b/erts/emulator/internal_doc/CodeLoading.md index 151b9cd57c..0b2e3070e7 100644 --- a/erts/emulator/internal_doc/CodeLoading.md +++ b/erts/emulator/internal_doc/CodeLoading.md @@ -45,7 +45,7 @@ free to schedule other work while the second loader is waiting. (See `erts_release_code_write_permission`). The ability to prepare several modules in parallel is not currently -used as almost all code loading is serialized by the code_server +used as almost all code loading is serialized by the code\_server process. The BIF interface is however prepared for this. erlang:prepare_loading(Module, Code) -> LoaderState @@ -71,8 +71,8 @@ structures. These *code access structures* are * Export table. One entry for every exported function. * Module table. One entry for each loaded module. -* "beam_catches". Identifies jump destinations for catch instructions. -* "beam_ranges". Map code address to function and line in source file. +* "beam\_catches". Identifies jump destinations for catch instructions. +* "beam\_ranges". Map code address to function and line in source file. The most frequently used of these structures is the export table that is accessed in run time for every executed external function call to diff --git a/erts/emulator/internal_doc/GarbageCollection.md b/erts/emulator/internal_doc/GarbageCollection.md index 1d9e3f4160..a1627b3233 100644 --- a/erts/emulator/internal_doc/GarbageCollection.md +++ b/erts/emulator/internal_doc/GarbageCollection.md @@ -1,6 +1,6 @@ # Erlang Garbage Collector -Erlang manages dynamic memory with a [tracing garbage collector](https://en.wikipedia.org/wiki/Tracing_garbage_collection). More precisely a per process generational semi-space copying collector using [Cheney's](#cheney) copy collection algorithm together with a global large object space. +Erlang manages dynamic memory with a [tracing garbage collector](https://en.wikipedia.org/wiki/Tracing_garbage_collection). More precisely a per process generational semi-space copying collector using Cheney's copy collection algorithm together with a global large object space. (See C. J. Cheney in [References](#references).) ## Overview @@ -12,12 +12,11 @@ Terms are created on the heap by evaluating expressions. There are two major typ Let's look at an example that returns a tuple with the newly created data. -```erlang -data(Foo) -> - Cons = [42|Foo], - Literal = {text, "hello world!"}, - {tag, Cons, Literal}. -``` + + data(Foo) -> + Cons = [42|Foo], + Literal = {text, "hello world!"}, + {tag, Cons, Literal}. In this example we first create a new cons cell with an integer and a tuple with some text. Then a tuple of size three wrapping the other values with an atom tag is created and returned. @@ -25,7 +24,6 @@ On the heap tuples require a word size for each of its elements as well as for t Compiling this code to beam assembly (`erlc -S`) shows exactly what is happening. -```erlang ... {test_heap,6,1}. {put_list,{integer,42},{x,0},{x,1}}. @@ -34,9 +32,8 @@ Compiling this code to beam assembly (`erlc -S`) shows exactly what is happening {put,{x,1}}. {put,{literal,{text,"hello world!"}}}. return. -``` -Looking at the assembler code we can see three things; The heap requirement in this function turns out to be only six words, as seen by the `{test_heap,6,1}` instruction. All the allocations are combined to a single instruction. The bulk of the data `{text, "hello world!"}` is a *literal*. Literals, sometimes referred to as constants, are not allocated in the function since they are a part of the module and allocated at load time. +Looking at the assembler code we can see three things: The heap requirement in this function turns out to be only six words, as seen by the `{test_heap,6,1}` instruction. All the allocations are combined to a single instruction. The bulk of the data `{text, "hello world!"}` is a *literal*. Literals, sometimes referred to as constants, are not allocated in the function since they are a part of the module and allocated at load time. If there is not enough space available on the heap to satisfy the `test_heap` instructions request for memory, then a garbage collection is initiated. It may happen immediately in the `test_heap` instruction, or it can be delayed until a later time depending on what state the process is in. If the garbage collection is delayed, any memory needed will be allocated in heap fragments. Heap fragments are extra memory blocks that are a part of the young heap, but are not allocated in the contigious area where terms normally reside. See [The young heap](#the-young-heap) for more details. @@ -50,11 +47,9 @@ It follows all the pointers from the root-set to the heap and copies each term w After the header word has been copied a [*move marker*](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_gc.h#L45-L46) is destructively placed in it pointing to the term in the *to space*. Any other term that points to the already moved term will [see this move marker](https://github.com/erlang/otp/blob/OTP-18.0/erts/emulator/beam/erl_gc.c#L1125) and copy the referring pointer instead. For example, if the have the following Erlang code: -```erlang -foo(Arg) -> - T = {test, Arg}, - {wrapper, T, T, T}. -``` + foo(Arg) -> + T = {test, Arg}, + {wrapper, T, T, T}. Only one copy of T exists on the heap and during the garbage collection only the first time T is encountered will it be copied. @@ -86,15 +81,15 @@ In the next garbage collection, any pointers to the old heap will be ignored and Generational garbage collection aims to increase performance at the expense of memory. This is achieved because only the young, smaller, heap is considered in most garbage collections. -The generational [hypothesis](#ungar) predicts that most terms tend to die young, and for an immutable language such as Erlang, young terms die even faster than in other languages. So for most usage patterns the data in the new heap will die very soon after it is allocated. This is good because it limits the amount of data copied to the old heap and also because the garbage collection algorithm used is proportional to the amount of live data on the heap. +The generational hypothesis predicts that most terms tend to die young (see D. Ungar in [References](#references)), and for an immutable language such as Erlang, young terms die even faster than in other languages. So for most usage patterns the data in the new heap will die very soon after it is allocated. This is good because it limits the amount of data copied to the old heap and also because the garbage collection algorithm used is proportional to the amount of live data on the heap. One critical issue to note here is that any term on the young heap can reference terms on the old heap but *no* term on the old heap may refer to a term on the young heap. This is due to the nature of the copy algorithm. Anything referenced by an old heap term is not included in the reference tree, root-set and its followers, and hence is not copied. If it was, the data would be lost, fire and brimstone would rise to cover the earth. Fortunately, this comes naturally for Erlang because the terms are immutable and thus there can be no pointers modified on the old heap to point to the young heap. -To reclaim data from the old heap, both young and old heaps are included during the collection and copied to a common *to space*. Both the *from space* of the young and old heap are then deallocated and the procedure will start over from the beginning. This type of garbage collection is called a full sweep and is triggered when the size of the area under the high-watermark is larger than the size of the free area of the old heap. It can also be triggered by doing a manual call to [erlang:garbage_collect()](http://erlang.org/doc/man/erlang.html#garbage_collect-0), or by running into the young garbage collection limit set by [spawn_opt(fun(),[{fullsweep_after, N}])](http://erlang.org/doc/man/erlang.html#spawn_opt-4) where N is the number of young garbage collections to do before forcing a garbage collection of both young and old heap. +To reclaim data from the old heap, both young and old heaps are included during the collection and copied to a common *to space*. Both the *from space* of the young and old heap are then deallocated and the procedure will start over from the beginning. This type of garbage collection is called a full sweep and is triggered when the size of the area under the high-watermark is larger than the size of the free area of the old heap. It can also be triggered by doing a manual call to [erlang:garbage_collect()](http://erlang.org/doc/man/erlang.html#garbage_collect-0), or by running into the young garbage collection limit set by [spawn\_opt(fun(),[{fullsweep\_after, N}\])](http://erlang.org/doc/man/erlang.html#spawn_opt-4) where N is the number of young garbage collections to do before forcing a garbage collection of both young and old heap. ## The young heap -The young heap, or the allocation heap, consists of the stack and heap as described in the Overview. However, it also includes any heap fragments that are attached to the heap. All of the heap fragments are considered to be above the high-watermark and part of the young generation. Heap fragments contain terms that either did not fit on the heap, or were created by another process and then attached to the heap. For instance if the bif binary_to_term created a term which does not fit on the current heap without doing a garbage collection, it will create a heap-fragment for the term and then schedule a garbage collection for later. Also if a message is sent to the process, the payload may be placed in a heap-fragment and that fragment is added to young heap when the message is matched in a receive clause. +The young heap, or the allocation heap, consists of the stack and heap as described in the Overview. However, it also includes any heap fragments that are attached to the heap. All of the heap fragments are considered to be above the high-watermark and part of the young generation. Heap fragments contain terms that either did not fit on the heap, or were created by another process and then attached to the heap. For instance if the bif `binary_to_term/1` created a term which does not fit on the current heap without doing a garbage collection, it will create a heap-fragment for the term and then schedule a garbage collection for later. Also if a message is sent to the process, the payload may be placed in a heap-fragment and that fragment is added to young heap when the message is matched in a receive clause. This procedure differs from how it worked prior to Erlang/OTP 19.0. Before 19.0, only a contiguous memory block where the young heap and stack resided was considered to be part of the young heap. Heap fragments and messages were immediately copied into the young heap before they could be inspected by the Erlang program. The behaviour introduced in 19.0 is superior in many ways - most significantly it reduces the number of necessary copy operations and the root set for garbage collection. @@ -118,21 +113,19 @@ The old heap is always one step ahead in the heap growth stages than the young h When garbage collecting a heap (young or old) all literals are left in place and not copied. To figure out if a term should be copied or not when doing a garbage collection the following pseudo code is used: -```c -if (erts_is_literal(ptr) || (on_old_heap(ptr) && !fullsweep)) { - /* literal or non fullsweep - do not copy */ -} else { - copy(ptr); -} -``` + if (erts_is_literal(ptr) || (on_old_heap(ptr) && !fullsweep)) { + /* literal or non fullsweep - do not copy */ + } else { + copy(ptr); + } The [`erts_is_literal`](https://github.com/erlang/otp/blob/OTP-19.0/erts/emulator/beam/global.h#L1452-L1465) check works differently on different architectures and operating systems. -On 64 bit systems that allow mapping of unreserved virtual memory areas (most operating systems except Windows), an area of size 1 GB (by default) is mapped and then all literals are placed within that area. Then all that has to be done to determine if something is a literal or not is [two quick pointer checks](https://github.com/erlang/otp/blob/OTP-19.0/erts/emulator/beam/erl_alloc.h#L322-L324). This system relies on the fact that a memory page that has not been touched yet does not take any actual space. So even if 1 GB of virtual memory is mapped, only the memory which is actually needed for literals is allocated in ram. The size of the literal area is configurable through the +MIscs erts_alloc option. +On 64 bit systems that allow mapping of unreserved virtual memory areas (most operating systems except Windows), an area of size 1 GB (by default) is mapped and then all literals are placed within that area. Then all that has to be done to determine if something is a literal or not is [two quick pointer checks](https://github.com/erlang/otp/blob/OTP-19.0/erts/emulator/beam/erl_alloc.h#L322-L324). This system relies on the fact that a memory page that has not been touched yet does not take any actual space. So even if 1 GB of virtual memory is mapped, only the memory which is actually needed for literals is allocated in ram. The size of the literal area is configurable through the +MIscs erts\_alloc option. On 32 bit systems, there is not enough virtual memory space to allocate 1 GB for just literals, so instead small 256 KB sized literal regions are created on demand and a card mark bit-array of the entire 32 bit memory space is then used to determine if a term is a literal or not. Since the total memory space is only 32 bits, the card mark bit-array is only 256 words large. On a 64 bit system the same bit-array would have to be 1 tera words large, so this technique is only viable on 32 bit systems. Doing [lookups in the array](https://github.com/erlang/otp/blob/OTP-19.0/erts/emulator/beam/erl_alloc.h#L316-L319) is a little more expensive then just doing the pointer checks that can be done in 64 bit systems, but not extremely so. -On 64 bit windows, on which erts_alloc cannot do unreserved virtual memory mappings, a [special tag](https://github.com/erlang/otp/blob/OTP-19.0/erts/emulator/beam/erl_term.h#L59) within the Erlang term object is used to determine if something [is a literal or not](https://github.com/erlang/otp/blob/OTP-19.0/erts/emulator/beam/erl_term.h#L248-L252). This is very cheap, however, the tag is only available on 64 bit machines, and it is possible to do a great deal of other nice optimizations with this tag in the future (like for instance a more compact list implementation) so it is not used on operating systems where it is not needed. +On 64 bit windows, on which erts\_alloc cannot do unreserved virtual memory mappings, a [special tag](https://github.com/erlang/otp/blob/OTP-19.0/erts/emulator/beam/erl_term.h#L59) within the Erlang term object is used to determine if something [is a literal or not](https://github.com/erlang/otp/blob/OTP-19.0/erts/emulator/beam/erl_term.h#L248-L252). This is very cheap, however, the tag is only available on 64 bit machines, and it is possible to do a great deal of other nice optimizations with this tag in the future (like for instance a more compact list implementation) so it is not used on operating systems where it is not needed. This behaviour is different from how it worked prior to Erlang/OTP 19.0. Before 19.0 the literal check was done by checking if the pointer pointed to the young or old heap block. If it did not, then it was considered a literal. This lead to considerable overhead and strange memory usage scenarios, so it was removed in 19.0. @@ -182,6 +175,8 @@ Using `on_heap` will force all messages to be part of on the young heap which wi Which one of these strategies is best depends a lot on what the process is doing and how it interacts with other processes. So, as always, profile the application and see how it behaves with the different options. - <a name="cheney">[1]</a>: C. J. Cheney. A nonrecursive list compacting algorithm. Commun. ACM, 13(11):677–678, Nov. 1970. +## References + +C. J. Cheney. A nonrecursive list compacting algorithm. Commun. ACM, 13(11):677–678, Nov. 1970. - <a name="ungar">[2]</a>: D. Ungar. Generation scavenging: A non-disruptive high performance storage reclamation algorithm. SIGSOFT Softw. Eng. Notes, 9(3):157–167, Apr. 1984. +D. Ungar. Generation scavenging: A non-disruptive high performance storage reclamation algorithm. SIGSOFT Softw. Eng. Notes, 9(3):157–167, Apr. 1984. diff --git a/erts/emulator/internal_doc/PTables.md b/erts/emulator/internal_doc/PTables.md index 6fe0e7665d..ef61963a40 100644 --- a/erts/emulator/internal_doc/PTables.md +++ b/erts/emulator/internal_doc/PTables.md @@ -85,13 +85,13 @@ following: 3. Depending on use, issue appropriate memory barrier. A common barrier used is a barrier with acquire semantics. On - x86/x86_64 this maps to a compiler barrier preventing the compiler + x86/x86\_64 this maps to a compiler barrier preventing the compiler to reorder instructions, but on other hardware often some kind of light weight hardware memory barrier is also needed. When comparing with a locked approach, at least one heavy weight memory barrier will be issued when locking the lock on most, if - not all, hardware architectures (including x86/x86_64), and often + not all, hardware architectures (including x86/x86\_64), and often some kind of light weight memory barrier will be issued when unlocking the lock. diff --git a/erts/emulator/internal_doc/SuperCarrier.md b/erts/emulator/internal_doc/SuperCarrier.md index acf722ea37..f52c6613d5 100644 --- a/erts/emulator/internal_doc/SuperCarrier.md +++ b/erts/emulator/internal_doc/SuperCarrier.md @@ -5,7 +5,7 @@ A super carrier is large memory area, allocated at VM start, which can be used during runtime to allocate normal carriers from. The super carrier feature was introduced in OTP R16B03. It is -enabled with command line option +MMscs <size in Mb> +enabled with command line option +MMscs <size in Mb> and can be configured with other options. Problem @@ -65,7 +65,7 @@ carrier is full. ### Implementation ### -The entire super carrier implementation is kept in erl_mmap.c. The +The entire super carrier implementation is kept in erl\_mmap.c. The name suggest that it can be viewed as our own mmap implementation. A super carrier needs to satisfy two slightly different kinds of @@ -98,8 +98,8 @@ other. ### Data structures ### -The MBC area is called **sa** as in super aligned and the SBC area is -called **sua** as in super un-aligned. +The MBC area is called *sa* as in super aligned and the SBC area is +called *sua* as in super un-aligned. Note that the "super" in super alignment and the "super" in super carrier has nothing to do with each other. We could have choosen @@ -128,7 +128,7 @@ down or up. We need to keep track of all the free segments in order to reuse them for new carrier allocations. One initial idea was to use the same mechanism that is used to keep track of free blocks within MBCs -(alloc_util and the different strategies). However, that would not be +(alloc\_util and the different strategies). However, that would not be as straight forward as one can think and can also waste quite a lot of memory as it uses prepended block headers. The granularity of the super carrier is one memory page (usually 4kb). We want to allocate diff --git a/erts/emulator/internal_doc/Tracing.md b/erts/emulator/internal_doc/Tracing.md index 7f97f64765..196ae0dd4e 100644 --- a/erts/emulator/internal_doc/Tracing.md +++ b/erts/emulator/internal_doc/Tracing.md @@ -37,6 +37,7 @@ what different type of break actions that are enabled. Same Same but Different ----------------------- + Even though `trace_pattern` use the same technique as the non-blocking code loading with replicated generations of data structures and an atomic switch, the implementations are quite separate from each @@ -72,6 +73,7 @@ aligned write operation on all hardware architectures we use. Adding a new Breakpoint ----------------------- + This is a simplified sequence describing what `trace_pattern` goes through when adding a new breakpoint. @@ -82,7 +84,7 @@ through when adding a new breakpoint. instruction word in the breakpoint. 3. Write a pointer to the breakpoint at offset -4 from the first - instruction "func_info" header. + instruction "func\_info" header. 4. Set the staging part of the breakpoint as enabled with specified breakpoint data. @@ -139,7 +141,7 @@ and removing breakpoints. 2. Allocate new breakpoint structures with a disabled active part and the original beam instruction. Write a pointer to the breakpoint in - "func_info" header at offset -4. + "func\_info" header at offset -4. 3. Update the staging part of all affected breakpoints. Disable breakpoints that are to be removed. diff --git a/erts/emulator/internal_doc/beam_makeops.md b/erts/emulator/internal_doc/beam_makeops.md index 1da8d2ab05..2880099b70 100644 --- a/erts/emulator/internal_doc/beam_makeops.md +++ b/erts/emulator/internal_doc/beam_makeops.md @@ -403,7 +403,7 @@ A line with `//` is also a comment. It is recommended to only use this style of comments in files that define implementations of instructions. -A long line can be broken into shorter lines by a placing a`\` before +A long line can be broken into shorter lines by a placing a `\` before the newline. ### Variable definitions ### @@ -1159,7 +1159,6 @@ implementation of `gen_element()`: return op; } -} ### Defining the implementation ### @@ -1452,7 +1451,7 @@ optionally additional heap space. ##### The NEXT_INSTRUCTION pre-bound variable ##### -The NEXT_INSTRUCTION is a pre-bound variable that is available in +The NEXT\_INSTRUCTION is a pre-bound variable that is available in all instructions. It expands to the address of the next instruction. Here is an example: @@ -1545,7 +1544,7 @@ register, the pointer will no longer be valid. (Y registers are stored on the stack.) In those circumstances, `$REFRESH_GEN_DEST()` must be invoked -to set up the pointer again. **beam\_makeops** will notice +to set up the pointer again. **beam\_makeops** will notice if there is a call to a function that does a garbage collection and `$REFRESH_GEN_DEST()` is not called. diff --git a/erts/emulator/sys/common/erl_poll.c b/erts/emulator/sys/common/erl_poll.c index c71d23f58c..9662996039 100644 --- a/erts/emulator/sys/common/erl_poll.c +++ b/erts/emulator/sys/common/erl_poll.c @@ -2354,6 +2354,7 @@ uint32_t epoll_events(int kp_fd, int fd) fprintf(stderr,"failed to parse file %s on line %d, errno = %d\n", fname, line, errno); + fclose(f); return 0; } if (fd == ev_fd) { @@ -2408,6 +2409,7 @@ ERTS_POLL_EXPORT(erts_poll_get_selected_events)(ErtsPollSet *ps, if (fscanf(f,"pos:\t%x\nflags:\t%x", &pos, &flags) != 2) { fprintf(stderr,"failed to parse file %s, errno = %d\n", fname, errno); ASSERT(0); + fclose(f); return; } if (fscanf(f,"\nmnt_id:\t%x\n", &mnt_id)); @@ -2422,6 +2424,7 @@ ERTS_POLL_EXPORT(erts_poll_get_selected_events)(ErtsPollSet *ps, fprintf(stderr,"failed to parse file %s on line %d, errno = %d\n", fname, line, errno); ASSERT(0); + fclose(f); return; } if (fd == ps->wake_fds[0] || fd == ps->wake_fds[1]) @@ -2437,6 +2440,7 @@ ERTS_POLL_EXPORT(erts_poll_get_selected_events)(ErtsPollSet *ps, ev[fd] = (ERTS_POLL_EV_IN|ERTS_POLL_EV_OUT) & ERTS_POLL_EV_N2E(events); line++; } + fclose(f); #else for (fd = 0; fd < len; fd++) ev[fd] = ERTS_POLL_EV_NONE; diff --git a/erts/emulator/test/binary_SUITE.erl b/erts/emulator/test/binary_SUITE.erl index 1406ddc9dc..563d60cc3f 100644 --- a/erts/emulator/test/binary_SUITE.erl +++ b/erts/emulator/test/binary_SUITE.erl @@ -1703,13 +1703,16 @@ error_after_yield_bad_ext_term() -> BadAtomExt]). %% Invalid atom at the end cmp_old_impl(Config) when is_list(Config) -> - %% Compare results from new yielding implementations with - %% old non yielding implementations + %% This test was originally a comparison with the non yielding + %% implementation in R16B. Since OTP 22 we can't talk distribution with such + %% old nodes (< 19). The test case it kept but compares with previous major + %% version for semantic regression test. Cookie = atom_to_list(erlang:get_cookie()), - Rel = "r16b_latest", + Rel = (integer_to_list(list_to_integer(erlang:system_info(otp_release)) - 1) + ++ "_latest"), case test_server:is_release_available(Rel) of false -> - {skipped, "No "++Rel++" available"}; + {skipped, "No OTP "++Rel++" available"}; true -> {ok, Node} = test_server:start_node(list_to_atom(atom_to_list(?MODULE)++"_"++Rel), peer, diff --git a/erts/emulator/test/bs_match_misc_SUITE.erl b/erts/emulator/test/bs_match_misc_SUITE.erl index 17759d78f3..cae4eb54d2 100644 --- a/erts/emulator/test/bs_match_misc_SUITE.erl +++ b/erts/emulator/test/bs_match_misc_SUITE.erl @@ -24,7 +24,7 @@ kenneth/1,encode_binary/1,native/1,happi/1, size_var/1,wiger/1,x0_context/1,huge_float_field/1, writable_binary_matched/1,otp_7198/1,unordered_bindings/1, - float_middle_endian/1]). + float_middle_endian/1,unsafe_get_binary_reuse/1]). -include_lib("common_test/include/ct.hrl"). @@ -36,7 +36,8 @@ all() -> [bound_var, bound_tail, t_float, little_float, sean, kenneth, encode_binary, native, happi, size_var, wiger, x0_context, huge_float_field, writable_binary_matched, - otp_7198, unordered_bindings, float_middle_endian]. + otp_7198, unordered_bindings, float_middle_endian, + unsafe_get_binary_reuse]. %% Test matching of bound variables. @@ -556,5 +557,21 @@ unordered_bindings(CompressedLength, HashSize, PadLength, T) -> Padding:PadLength/binary,PadLength>> = T, {Content,Mac,Padding}. +%% ERL-901: A load-time optimization assumed that match contexts had no further +%% uses when a bs_get_binary2 overwrote the match context's register, and +%% figured it would be safe to reuse the match context's memory for the +%% resulting binary. +%% +%% This is no longer safe as of OTP 22, as a match context may be reused after +%% being passed to another function. +unsafe_get_binary_reuse(Config) when is_list(Config) -> + <<_First, Rest/binary>> = <<"hello">>, + ubgr_1(Rest), + <<Second,_/bits>> = Rest, + $e = Second, + ok. + +ubgr_1(<<_CP/utf8, Rest/binary>>) -> id(Rest); +ubgr_1(_) -> false. id(I) -> I. diff --git a/erts/emulator/test/distribution_SUITE.erl b/erts/emulator/test/distribution_SUITE.erl index 4f70b51aa0..449821e5ad 100644 --- a/erts/emulator/test/distribution_SUITE.erl +++ b/erts/emulator/test/distribution_SUITE.erl @@ -744,6 +744,7 @@ link_to_dead_new_node(Config) when is_list(Config) -> {'EXIT', Pid, noproc} -> ok; Other -> + stop_node(Node), ct:fail({unexpected_message, Other}) after 5000 -> ct:fail(nothing_received) @@ -1421,7 +1422,7 @@ message_latency_large_exit(Nodename, ReasonFun) -> FlushTrace = fun F() -> receive - {trace, Pid, _, _} = M -> + {trace, Pid, _, _} -> F() after 0 -> ok @@ -1455,8 +1456,14 @@ measure_latency_large_message(Nodename, DataFun) -> Echo = spawn(N, fun F() -> receive {From, Msg} -> From ! Msg, F() end end), - %% Test 32 MB and 320 MB and test the latency difference of sent messages - Payloads = [{I, <<0:(I * 32 * 1024 * 1024 * 8)>>} || I <- [1,10]], + case erlang:system_info(build_type) of + debug -> + %% Test 3.2 MB and 32 MB and test the latency difference of sent messages + Payloads = [{I, <<0:(I * 32 * 1024 * 8)>>} || I <- [1,10]]; + _ -> + %% Test 32 MB and 320 MB and test the latency difference of sent messages + Payloads = [{I, <<0:(I * 32 * 1024 * 1024 * 8)>>} || I <- [1,10]] + end, IndexTimes = [{I, measure_latency(DataFun, Dropper, Echo, P)} || {I, P} <- Payloads], @@ -1465,6 +1472,8 @@ measure_latency_large_message(Nodename, DataFun) -> ct:pal("~p",[IndexTimes]), + stop_node(N), + case {lists:max(Times), lists:min(Times)} of {Max, Min} when Max * 0.25 > Min -> ct:fail({incorrect_latency, IndexTimes}); @@ -1487,7 +1496,7 @@ measure_latency(DataFun, Dropper, Echo, Payload) -> end) || _ <- lists:seq(1,2)], [receive - {monitor, _Sender, busy_dist_port, _Info} = M -> + {monitor, _Sender, busy_dist_port, _Info} -> ok end || _ <- lists:seq(1,10)], @@ -1733,7 +1742,7 @@ bad_dist_fragments(Config) when is_list(Config) -> start_monitor(Offender,P), Exit2Victim = spawn(Victim, fun() -> receive ok -> ok end end), - send_bad_fragments(Offender, Victim, P,{?DOP_PAYLOAD_EXIT2,P,ExitVictim},2, + send_bad_fragments(Offender, Victim, P,{?DOP_PAYLOAD_EXIT2,P,Exit2Victim},2, [{hdr, 1, [132]}]), start_monitor(Offender,P), @@ -2428,17 +2437,22 @@ stop_node(Node) -> verify_nc(Node) -> P = self(), Ref = make_ref(), - spawn(Node, - fun() -> - R = erts_test_utils:check_node_dist(fun(E) -> E end), - P ! {Ref, R} - end), + Pid = spawn(Node, + fun() -> + R = erts_test_utils:check_node_dist(fun(E) -> E end), + P ! {Ref, R} + end), + MonRef = monitor(process, Pid), receive {Ref, ok} -> + demonitor(MonRef,[flush]), ok; {Ref, Error} -> - ct:log("~s",[Error]), - ct:fail(failed_nc_refc_check) + ct:log("~p",[Error]), + ct:fail(failed_nc_refc_check); + {'DOWN', MonRef, _, _, _} = Down -> + ct:log("~p",[Down]), + ct:fail(crashed_nc_refc_check) end. freeze_node(Node, MS) -> diff --git a/erts/emulator/test/dump_SUITE.erl b/erts/emulator/test/dump_SUITE.erl index d0237b78cc..3b860ebdf6 100644 --- a/erts/emulator/test/dump_SUITE.erl +++ b/erts/emulator/test/dump_SUITE.erl @@ -24,7 +24,7 @@ -export([all/0, suite/0, init_per_testcase/2, end_per_testcase/2]). --export([signal_abort/1]). +-export([signal_abort/1, exiting_dump/1, free_dump/1]). -export([load/0]). @@ -35,7 +35,7 @@ suite() -> {timetrap, {minutes, 2}}]. all() -> - [signal_abort]. + [signal_abort, exiting_dump, free_dump]. init_per_testcase(signal_abort, Config) -> SO = erlang:system_info(schedulers_online), @@ -48,7 +48,10 @@ init_per_testcase(signal_abort, Config) -> {skip, "the platform does not support scheduler dump"}; Dump -> Config - end. + end; +init_per_testcase(_, Config) -> + Config. + end_per_testcase(_, Config) -> Config. @@ -79,8 +82,6 @@ signal_abort(Config) -> {ok, Bin} = get_dump_when_done(Dump), - ct:log("~s",[Bin]), - {match, Matches} = re:run(Bin,"Current Process: <",[global]), ct:log("Found ~p",[Matches]), @@ -91,6 +92,85 @@ signal_abort(Config) -> ok. +load() -> + lists:seq(1,10000), + load(). + + +%% Test that crash dumping when a process is in the state EXITING works +exiting_dump(Config) when is_list(Config) -> + Dump = filename:join(proplists:get_value(priv_dir, Config),"signal_abort.dump"), + + {ok, Node} = start_node(Config), + + Self = self(), + + Pid = spawn_link(Node, + fun() -> + [begin + T = ets:new(hej,[]), + [ets:insert(T,{I,I}) || I <- lists:seq(1,1000)] + end || _ <- lists:seq(1,1000)], + Self ! ready, + receive ok -> ok end + end), + + true = rpc:call(Node, os, putenv, ["ERL_CRASH_DUMP",Dump]), + + receive ready -> unlink(Pid), Pid ! ok end, + + rpc:call(Node, erlang, halt, ["dump"]), + + {ok, Bin} = get_dump_when_done(Dump), + + {match, Matches} = re:run(Bin,"^State: Exiting", [global, multiline]), + + ct:log("Found ~p",[Matches]), + + true = length(Matches) == 1, + + file:delete(Dump), + + ok. + +%% Test that crash dumping when a process is in the state FREE works +free_dump(Config) when is_list(Config) -> + Dump = filename:join(proplists:get_value(priv_dir, Config),"signal_abort.dump"), + + {ok, Node} = start_node(Config), + + Self = self(), + + Pid = spawn_link(Node, + fun() -> + Self ! ready, + receive + ok -> + unlink(Self), + exit(lists:duplicate(1000,1000)) + end + end), + + true = rpc:call(Node, os, putenv, ["ERL_CRASH_DUMP",Dump]), + + [erlang:monitor(process, Pid) || _ <- lists:seq(1,10000)], + receive ready -> unlink(Pid), Pid ! ok end, + + rpc:call(Node, erlang, halt, ["dump"]), + + {ok, Bin} = get_dump_when_done(Dump), + + {match, Matches} = re:run(Bin,"^State: Non Existing", [global, multiline]), + + ct:log("Found ~p",[Matches]), + + true = length(Matches) == 1, + + file:delete(Dump), + + ok. + + get_dump_when_done(Dump) -> case file:read_file_info(Dump) of {ok, #file_info{ size = Sz }} -> @@ -104,15 +184,13 @@ get_dump_when_done(Dump, Sz) -> timer:sleep(1000), case file:read_file_info(Dump) of {ok, #file_info{ size = Sz }} -> - file:read_file(Dump); + {ok, Bin} = file:read_file(Dump), + ct:log("~s",[Bin]), + {ok, Bin}; {ok, #file_info{ size = NewSz }} -> get_dump_when_done(Dump, NewSz) end. -load() -> - lists:seq(1,10000), - load(). - start_node(Config) when is_list(Config) -> Pa = filename:dirname(code:which(?MODULE)), Name = list_to_atom(atom_to_list(?MODULE) diff --git a/erts/emulator/test/efile_SUITE.erl b/erts/emulator/test/efile_SUITE.erl index 7dcf302742..045b351e02 100644 --- a/erts/emulator/test/efile_SUITE.erl +++ b/erts/emulator/test/efile_SUITE.erl @@ -45,7 +45,12 @@ iter_max_files(Config) when is_list(Config) -> iter_max_files_1(Config) -> DataDir = proplists:get_value(data_dir,Config), TestFile = filename:join(DataDir, "existing_file"), - N = 10, + case erlang:system_info(debug_compiled) of + true -> + N = 5; + false -> + N = 10 + end, %% Run on a different node in order to make the test more stable. Dir = filename:dirname(code:which(?MODULE)), {ok,Node} = test_server:start_node(test_iter_max_files,slave, @@ -100,34 +105,27 @@ open_files(Name) -> %% a /proc directory), let's read some zero sized files 500 times each, while %% ensuring that response isn't empty << >> proc_zero_sized_files(Config) when is_list(Config) -> - {Type, Flavor} = os:type(), - %% Some files which exist on Linux but might be missing on other systems - Inputs = ["/proc/cpuinfo", - "/proc/meminfo", - "/proc/partitions", - "/proc/swaps", - "/proc/version", - "/proc/uptime", - %% curproc is present on freebsd - "/proc/curproc/cmdline"], - case filelib:is_dir("/proc") of - false -> {skip, "/proc not found"}; % skip the test if no /proc - _ when Type =:= unix andalso Flavor =:= sunos -> - %% SunOS has a /proc, but no zero sized special files - {skip, "sunos does not have any zero sized special files"}; - true -> - %% Take away files which do not exist in proc - Inputs1 = lists:filter(fun filelib:is_file/1, Inputs), - - %% Fail if none of mentioned files exist in /proc, did we just get - %% a normal /proc directory without any special files? - ?assertNotEqual([], Inputs1), - + TestFiles0 = [%% Some files which exist on Linux but might be missing on + %% other systems + "/proc/cpuinfo", + "/proc/meminfo", + "/proc/partitions", + "/proc/swaps", + "/proc/version", + "/proc/uptime", + %% curproc is present on FreeBSD + "/proc/curproc/cmdline"], + + TestFiles = [F || F <- TestFiles0, filelib:is_file(F)], + + case TestFiles of + [_|_] -> %% For 6 inputs and 500 attempts each this do run anywhere %% between 500 and 3000 function calls. - lists:foreach( - fun(Filename) -> do_proc_zero_sized(Filename, 500) end, - Inputs1) + [do_proc_zero_sized(F, 500) || F <- TestFiles], + ok; + [] -> + {skip, "Failed to find any known zero-sized files"} end. %% @doc Test one file N times to also trigger possible leaking fds and memory diff --git a/erts/emulator/test/erts_test_utils.erl b/erts/emulator/test/erts_test_utils.erl index 0c3ef3e0fc..e4e00a0a16 100644 --- a/erts/emulator/test/erts_test_utils.erl +++ b/erts/emulator/test/erts_test_utils.erl @@ -19,6 +19,7 @@ %% -module(erts_test_utils). +-compile(r16). %% %% THIS MODULE IS ALSO USED BY *OTHER* APPLICATIONS TEST CODE diff --git a/erts/emulator/test/fun_SUITE.erl b/erts/emulator/test/fun_SUITE.erl index 4042b58ff2..7f6caa08f1 100644 --- a/erts/emulator/test/fun_SUITE.erl +++ b/erts/emulator/test/fun_SUITE.erl @@ -592,7 +592,8 @@ refc_dist(Config) when is_list(Config) -> 3 = fun_refc(F2), true = erlang:garbage_collect(), 2 = fun_refc(F), - refc_dist_send(Node, F). + refc_dist_send(Node, F), + test_server:stop_node(Node). refc_dist_send(Node, F) -> Pid = spawn_link(Node, fun() -> receive @@ -682,6 +683,7 @@ t_arity(Config) when is_list(Config) -> 43 = spawn_call(Node, fun(X, Y) -> A+X+Y end), 1 = spawn_call(Node, fun(X, Y) -> X+Y end), 45 = spawn_call(Node, fun(X, Y, Z) -> A+X+Y+Z end), + test_server:stop_node(Node), ok. t_is_function2(Config) when is_list(Config) -> diff --git a/erts/emulator/test/lcnt_SUITE.erl b/erts/emulator/test/lcnt_SUITE.erl index 87b97037d6..2dbaec9942 100644 --- a/erts/emulator/test/lcnt_SUITE.erl +++ b/erts/emulator/test/lcnt_SUITE.erl @@ -187,5 +187,8 @@ remove_untoggleable_locks([]) -> []; remove_untoggleable_locks([{resource_monitors, _, _, _} | T]) -> remove_untoggleable_locks(T); +remove_untoggleable_locks([{'socket[gcnt]', _, _, _} | T]) -> + %% Global lock used by socket NIF + remove_untoggleable_locks(T); remove_untoggleable_locks([H | T]) -> [H | remove_untoggleable_locks(T)]. diff --git a/erts/emulator/test/map_SUITE.erl b/erts/emulator/test/map_SUITE.erl index d0a6763fe5..9ea59e1084 100644 --- a/erts/emulator/test/map_SUITE.erl +++ b/erts/emulator/test/map_SUITE.erl @@ -17,7 +17,7 @@ %% %CopyrightEnd% %% -module(map_SUITE). --export([all/0, suite/0]). +-export([all/0, suite/0, init_per_suite/1, end_per_suite/1]). -export([t_build_and_match_literals/1, t_build_and_match_literals_large/1, t_update_literals/1, t_update_literals_large/1, @@ -84,7 +84,10 @@ %% instruction-level tests t_has_map_fields/1, y_regs/1, - badmap_17/1]). + badmap_17/1, + + %%Bugs + t_large_unequal_bins_same_hash_bug/1]). -include_lib("stdlib/include/ms_transform.hrl"). @@ -149,7 +152,26 @@ all() -> [t_build_and_match_literals, t_build_and_match_literals_large, %% instruction-level tests t_has_map_fields, y_regs, - badmap_17]. + badmap_17, + + %% Bugs + t_large_unequal_bins_same_hash_bug]. + +init_per_suite(Config) -> + A0 = case application:start(sasl) of + ok -> [sasl]; + _ -> [] + end, + A = case application:start(os_mon) of + ok -> [os_mon|A0]; + _ -> A0 + end, + [{started_apps, A}|Config]. + +end_per_suite(Config) -> + As = proplists:get_value(started_apps, Config), + lists:foreach(fun (A) -> application:stop(A) end, As), + Config. %% tests @@ -3374,3 +3396,67 @@ fannerl() -> 104,2,97,9,97,16,70,63,184,100,97,32,0,0,0,104,2,97,10,97,16,70,63,169,174, 254,64,0,0,0,104,2,97,11,97,16,70,191,119,121,234,0,0,0,0,104,2,97,12,97, 16,70,63,149,12,170,128,0,0,0,104,2,97,13,97,16,70,191,144,193,191,0,0,0,0>>. + +%% This test case checks that the bug with ticket number OTP-15707 is +%% fixed. The bug could cause a crash or memory usage to grow until +%% the machine ran out of memory. +t_large_unequal_bins_same_hash_bug(Config) when is_list(Config) -> + run_when_enough_resources( + fun() -> + K1 = get_4GB_bin(1), + K2 = get_4GB_bin(2), + Map = make_map(500), + Map2 = maps:put(K1, 42, Map), + %% The map needed to contain at least 32 key-value pairs + %% at this point to get the crash or out of memory + %% problem on the next line + Map3 = maps:put(K2, 43, Map2), + %% The following line should avoid that the compiler + %% optimizes away the above + io:format("~p ~p~n", [erlang:phash2(Map3), maps:size(Map3)]) + end). + +make_map(0) -> + #{}; +make_map(Size) -> + maps:put(Size, Size, make_map(Size-1)). + +get_4GB_bin(Value) -> + List = lists:duplicate(65536, Value), + Bin = erlang:iolist_to_binary(List), + IOList4GB = duplicate_iolist(Bin, 16), + Bin4GB = erlang:iolist_to_binary(IOList4GB), + 4294967296 = size(Bin4GB), + Bin4GB. + +duplicate_iolist(IOList, 0) -> + IOList; +duplicate_iolist(IOList, NrOfTimes) -> + duplicate_iolist([IOList, IOList], NrOfTimes - 1). + +run_when_enough_resources(Fun) -> + case {total_memory(), erlang:system_info(wordsize)} of + {Mem, 8} when is_integer(Mem) andalso Mem >= 31 -> + Fun(); + {Mem, WordSize} -> + {skipped, + io_lib:format("Not enough resources (System Memory >= ~p, Word Size = ~p)", + [Mem, WordSize])} + end. + +total_memory() -> + %% Total memory in GB. + try + MemoryData = memsup:get_system_memory_data(), + case lists:keysearch(total_memory, 1, MemoryData) of + {value, {total_memory, TM}} -> + TM div (1024*1024*1024); + false -> + {value, {system_total_memory, STM}} = + lists:keysearch(system_total_memory, 1, MemoryData), + STM div (1024*1024*1024) + end + catch + _ : _ -> + undefined + end. diff --git a/erts/emulator/test/process_SUITE.erl b/erts/emulator/test/process_SUITE.erl index edf08ce0bd..b530ced566 100644 --- a/erts/emulator/test/process_SUITE.erl +++ b/erts/emulator/test/process_SUITE.erl @@ -44,6 +44,7 @@ process_info_garbage_collection/1, process_info_smoke_all/1, process_info_status_handled_signal/1, + process_info_reductions/1, bump_reductions/1, low_prio/1, binary_owner/1, yield/1, yield2/1, otp_4725/1, bad_register/1, garbage_collect/1, otp_6237/1, process_info_messages/1, process_flag_badarg/1, process_flag_heap_size/1, @@ -84,6 +85,7 @@ all() -> process_info_garbage_collection, process_info_smoke_all, process_info_status_handled_signal, + process_info_reductions, bump_reductions, low_prio, yield, yield2, otp_4725, bad_register, garbage_collect, process_info_messages, process_flag_badarg, process_flag_heap_size, @@ -133,6 +135,15 @@ init_per_group(_GroupName, Config) -> end_per_group(_GroupName, Config) -> Config. +init_per_testcase(Func, Config) + when Func =:= processes_default_tab; + Func =:= processes_this_tab -> + case erlang:system_info(debug_compiled) of + true -> + {skip, "Don't run in debug"}; + false -> + [{testcase, Func} | Config] + end; init_per_testcase(Func, Config) when is_atom(Func), is_list(Config) -> [{testcase, Func}|Config]. @@ -1084,6 +1095,46 @@ process_info_status_handled_signal(Config) when is_list(Config) -> false = erlang:is_process_alive(P), ok. +%% OTP-15709 +%% Provoke a bug where process_info(reductions) returned wrong result +%% because REDS_IN (def_arg_reg[5]) is read when the process in not running. +process_info_reductions(Config) when is_list(Config) -> + pi_reductions_tester(spawn_link(fun() -> pi_reductions_spinnloop() end)), + pi_reductions_tester(spawn_link(fun() -> pi_reductions_recvloop() end)), + ok. + +pi_reductions_tester(Pid) -> + {_, DiffList} = + lists:foldl(fun(_, {Prev, Acc}) -> + %% Add another item that force sending the request + %% as a signal, like 'current_function'. + PI = process_info(Pid, [reductions, current_function]), + [{reductions,Reds}, {current_function,_}] = PI, + Diff = Reds - Prev, + {Diff, true} = {Diff, (Diff >= 0)}, + {Diff, true} = {Diff, (Diff =< 1000*1000)}, + {Reds, [Diff | Acc]} + end, + {0, []}, + lists:seq(1,10)), + unlink(Pid), + exit(Pid,kill), + io:format("Reduction diffs: ~p\n", [DiffList]), + ok. + +pi_reductions_spinnloop() -> + %% 6 args to make use of def_arg_reg[5] which is also used as REDS_IN + pi_reductions_spinnloop(1, atom, "hej", self(), make_ref(), 3.14). + +pi_reductions_spinnloop(A,B,C,D,E,F) -> + pi_reductions_spinnloop(B,C,D,E,F,A). + +pi_reductions_recvloop() -> + receive + "a free lunch" -> false + end. + + %% Tests erlang:bump_reductions/1. bump_reductions(Config) when is_list(Config) -> erlang:garbage_collect(), |