diff options
| -rw-r--r-- | erts/doc/src/crash_dump.xml | 88 | ||||
| -rw-r--r-- | erts/emulator/beam/break.c | 128 | ||||
| -rw-r--r-- | erts/emulator/beam/erl_bif_info.c | 7 | ||||
| -rw-r--r-- | erts/emulator/beam/erl_process.c | 215 | ||||
| -rw-r--r-- | erts/emulator/beam/erl_process.h | 13 | ||||
| -rw-r--r-- | erts/emulator/beam/erl_process_dump.c | 174 | ||||
| -rw-r--r-- | erts/emulator/beam/erl_thr_progress.c | 42 | ||||
| -rw-r--r-- | erts/emulator/beam/erl_thr_progress.h | 4 | ||||
| -rw-r--r-- | erts/emulator/beam/sys.h | 2 | ||||
| -rw-r--r-- | erts/emulator/sys/unix/erl_unix_sys.h | 26 | ||||
| -rw-r--r-- | erts/emulator/sys/unix/sys.c | 31 | ||||
| -rw-r--r-- | erts/emulator/sys/win32/erl_win_sys.h | 12 | ||||
| -rw-r--r-- | erts/emulator/sys/win32/sys.c | 21 | ||||
| -rw-r--r-- | erts/emulator/test/bif_SUITE.erl | 33 |
14 files changed, 731 insertions, 65 deletions
diff --git a/erts/doc/src/crash_dump.xml b/erts/doc/src/crash_dump.xml index 2b5fc877c3..8291bf38b7 100644 --- a/erts/doc/src/crash_dump.xml +++ b/erts/doc/src/crash_dump.xml @@ -55,10 +55,12 @@ emulator or the operating system can be reconfigured to avoid the crash, which is why interpreting the crash dump correctly is important.</p> + <p>On systems that support OS signals, it is also possible to stop + the runtime system and generate a crash dump by sending the SIGUSR1.</p> <p>The erlang crash dump is a readable text file, but it might not be very easy to read. Using the Crashdump Viewer tool in the <c><![CDATA[observer]]></c> application will simplify the task. This is an - HTML based tool for browsing Erlang crash dumps.</p> + wx-widget based tool for browsing Erlang crash dumps.</p> <section> <marker id="general_info"></marker> @@ -66,8 +68,9 @@ <p>The first part of the dump shows the creation time for the dump, a slogan indicating the reason for the dump, the system version, of the node from which the dump originates, the compile time of - the emulator running the originating node and the number of - atoms in the atom table. + the emulator running the originating node, the number of + atoms in the atom table and the runtime system thread that caused + the crash dump to happen. </p> <section> @@ -170,6 +173,60 @@ </section> <section> + <marker id="scheduler"></marker> + <title>Scheduler information</title> + <p>Under the tag <em>=scheduler</em> information about the current state + and statistics of the schedulers in the runtime system is displayed. + On OSs that do allow instant suspension of other threads, the data within + this section will reflect what the runtime system looks like at the moment + when the crash happens.</p> + <p>The following fields can exist for a process:</p> + <taglist> + <tag><em>=scheduler:id</em></tag> + <item>Header, states the scheduler identifier.</item> + <tag><em>Scheduler Sleep Info Flags</em></tag> + <item>If empty the scheduler was doing some work. + If not empty the scheduler is either in some state of sleep, + or suspended. This entry is only present in a SMP enabled emulator</item> + <tag><em>Scheduler Sleep Info Aux Work</em></tag> + <item>If not empty, a scheduler internal auxiliary work is scheduled + to be done.</item> + <tag><em>Current Port</em></tag> + <item>The port identifier of the port that is currently being + executed by the scheduler.</item> + <tag><em>Current Process</em></tag> + <item>The process identifier of the process that is currently being + executed by the scheduler. If there is such a process this entry is + followed by the <em>State</em>,<em>Internal State</em>, + <em>Program Counter</em>, <em>CP</em> of that same process. See + <seealso marker="#processes">Process Information</seealso> for a + description what the different entries mean. Keep in mind that + this is a snapshot of what the entries are exactly when the crash + dump is starting to be generated. Therefore they will most likely + be different (and more telling) then the entries for the same + processes found in the <em>=proc</em> section. If there is no currently + running process, only the <em>Current Process</em> entry will be printed. + </item> + <tag><em>Current Process Limited Stack Trace</em></tag> + <item>This entry only shows up if there is a current process. It is very + similar to <seealso marker="#proc_data"><em>=proc_stack</em></seealso>, + except that only the function frames are printed (i.e. the stack variables + are omited). It is also limited to only print the top and bottom part + of the stack. If the stack is small (less that 512 slots) then the + entire stack will be printed. If not, an entry stating + <code>skipping ## slots</code> will be printed where ## is + replaced by the number of slots that has been skipped.</item> + <tag><em>Run Queue</em></tag> + <item>Displays statistics about how many processes and ports + of different priorities are scheduled on this scheduler.</item> + <tag><em>** crashed **</em></tag> + <item>This entry is normally not printed. It signifies that getting + the rest of the information about this scheduler failed for some reason. + </item> + </taglist> + </section> + + <section> <marker id="memory"></marker> <title>Memory information</title> <p>Under the tag <em>=memory</em> you will find information similar @@ -314,6 +371,9 @@ <item>The number of live argument registers. The argument registers, if any are live, will follow. These may contain the arguments of the function if they are not yet moved to the stack.</item> + <item><em>Internal State</em></item> + <item>A more detailed internal represantation of the state of + this process.</item> </taglist> <p>See also the section about <seealso marker="#proc_data">process data</seealso>.</p> </section> @@ -339,18 +399,38 @@ <tag><em>Name</em></tag> <item>The name of the table, regardless of whether it is a <c><![CDATA[named_table]]></c> or not.</item> - <tag><em>Buckets</em></tag> + <tag><em>Hash table, Buckets</em></tag> <item>This occurs if the table is a hash table, i.e. if it is not an <c><![CDATA[ordered_set]]></c>.</item> + <tag><em>Hash table, Chain Length</em></tag> + <item>Only applicable for hash tables. Contains statistics about the + hash table, such as the max, min and avg chain length. Having a max much + larger than the avg, and a std dev much larger that + the expected std dev is a sign that the hashing of the terms is + behaving badly for some reason.</item> <tag><em>Ordered set (AVL tree), Elements</em></tag> <item>This occurs only if the table is an <c><![CDATA[ordered_set]]></c>. (The number of elements is the same as the number of objects in the table.)</item> + <tag><em>Fixed</em></tag> + <item>If the table is fixed using ets:safe_fixtable or some internal + mechanism.</item> <tag><em>Objects</em></tag> <item>The number of objects in the table</item> <tag><em>Words</em></tag> <item>The number of words (usually 4 bytes/word) allocated to data in the table.</item> + <tag><em>Type</em></tag> + <item>The type of the table, i.e. <c>set</c>, <c>bag</c>, + <c>dublicate_bag</c> or <c>ordered_set</c>.</item> + <tag><em>Compressed</em></tag> + <item>If this table was compressed.</item> + <tag><em>Protection</em></tag> + <item>The protection of this table.</item> + <tag><em>Write Concurrency</em></tag> + <item>If write_concurrency was enabled for this table.</item> + <tag><em>Read Concurrency</em></tag> + <item>If read_concurrency was enabled for this table.</item> </taglist> </section> diff --git a/erts/emulator/beam/break.c b/erts/emulator/beam/break.c index 5aee85174f..41765dad12 100644 --- a/erts/emulator/beam/break.c +++ b/erts/emulator/beam/break.c @@ -210,25 +210,12 @@ print_process_info(int to, void *to_arg, Process *p) erts_print(to, to_arg, "State: "); state = erts_smp_atomic32_read_acqb(&p->state); - if (state & ERTS_PSFLG_FREE) - erts_print(to, to_arg, "Non Existing\n"); /* Should never happen */ - else if (state & ERTS_PSFLG_EXITING) - erts_print(to, to_arg, "Exiting\n"); - else if (state & ERTS_PSFLG_GC) { - garbing = 1; - running = 1; - erts_print(to, to_arg, "Garbing\n"); - } - else if (state & ERTS_PSFLG_SUSPENDED) - erts_print(to, to_arg, "Suspended\n"); - else if (state & ERTS_PSFLG_RUNNING) { - running = 1; - erts_print(to, to_arg, "Running\n"); - } - else if (state & ERTS_PSFLG_ACTIVE) - erts_print(to, to_arg, "Scheduled\n"); - else - erts_print(to, to_arg, "Waiting\n"); + erts_dump_process_state(to, to_arg, state); + if (state & ERTS_PSFLG_GC) { + garbing = 1; + running = 1; + } else if (state & ERTS_PSFLG_RUNNING) + running = 1; /* * If the process is registered as a global process, display the @@ -352,6 +339,10 @@ print_process_info(int to, void *to_arg, Process *p) #endif erts_stack_dump(to, to_arg, p); } + + /* Display all states */ + erts_print(to, to_arg, "Internal State: "); + erts_dump_extended_process_state(to, to_arg, state); } static void @@ -671,6 +662,7 @@ erl_crash_dump_v(char *file, int line, char* fmt, va_list args) { #ifdef ERTS_SMP ErtsThrPrgrData tpd_buf; /* in case we aren't a managed thread... */ + int bc; #endif int fd; size_t envsz; @@ -681,27 +673,39 @@ erl_crash_dump_v(char *file, int line, char* fmt, va_list args) char* dumpname; int secs; int env_erl_crash_dump_seconds_set = 1; + int i; if (ERTS_SOMEONE_IS_CRASH_DUMPING) return; #ifdef ERTS_SMP + /* Order all managed threads to block, this has to be done + first to guarantee that this is the only thread to generate + crash dump. */ + bc = erts_thr_progress_fatal_error_block(&tpd_buf); + +#ifdef ERTS_THR_HAVE_SIG_FUNCS /* - * Wait for all managed threads to block. If all threads haven't blocked - * after a minute, we go anyway and hope for the best... - * - * We do not release system again. We expect an exit() or abort() after - * dump has been written. + * We suspend all scheduler threads so that we can dump some + * data about the currently running processes and scheduler data. + * We have to be very very careful when doing this as the schedulers + * could be anywhere. */ - erts_thr_progress_fatal_error_block(60000, &tpd_buf); - /* Either worked or not... */ + for (i = 0; i < erts_no_schedulers; i++) { + erts_tid_t tid = ERTS_SCHEDULER_IX(i)->tid; + if (!erts_equal_tids(tid,erts_thr_self())) + sys_thr_suspend(tid); + } + +#endif /* Allow us to pass certain places without locking... */ erts_smp_atomic32_set_mb(&erts_writing_erl_crash_dump, 1); erts_smp_tsd_set(erts_is_crash_dumping_key, (void *) 1); -#else + +#else /* !ERTS_SMP */ erts_writing_erl_crash_dump = 1; -#endif +#endif /* ERTS_SMP */ envsz = sizeof(env); /* ERL_CRASH_DUMP_SECONDS not set @@ -758,9 +762,8 @@ erl_crash_dump_v(char *file, int line, char* fmt, va_list args) erts_fprintf(stderr,"\nCrash dump is being written to: %s...", dumpname); fd = open(dumpname,O_WRONLY | O_CREAT | O_TRUNC,0640); - if (fd < 0) + if (fd < 0) return; /* Can't create the crash dump, skip it */ - time(&now); erts_fdprintf(fd, "=erl_crash_dump:0.3\n%s", ctime(&now)); @@ -774,9 +777,74 @@ erl_crash_dump_v(char *file, int line, char* fmt, va_list args) erts_fdprintf(fd, "System version: "); erts_print_system_version(fd, NULL, NULL); erts_fdprintf(fd, "%s\n", "Compiled: " ERLANG_COMPILE_DATE); + erts_fdprintf(fd, "Taints: "); erts_print_nif_taints(fd, NULL); erts_fdprintf(fd, "Atoms: %d\n", atom_table_size()); + +#ifdef USE_THREADS + /* We want to note which thread it was that called erl_exit */ + if (erts_get_scheduler_data()) { + erts_fdprintf(fd, "Calling Thread: scheduler:%d\n", + erts_get_scheduler_data()->no); + } else { + if (!erts_thr_getname(erts_thr_self(), dumpnamebuf, MAXPATHLEN)) + erts_fdprintf(fd, "Calling Thread: %s\n", dumpnamebuf); + else + erts_fdprintf(fd, "Calling Thread: %p\n", erts_thr_self()); + } +#else + erts_fdprintf(fd, "Calling Thread: scheduler:1\n"); +#endif + +#if defined(ERTS_HAVE_TRY_CATCH) + + /* + * erts_print_scheduler_info is not guaranteed to be safe to call + * here for all schedulers as we may have suspended a scheduler + * in the middle of updating the STACK_TOP and STACK_START + * variables and thus when scanning the stack we could get + * segmentation faults. We protect against this very unlikely + * scenario by using the ERTS_SYS_TRY_CATCH. + */ + for (i = 0; i < erts_no_schedulers; i++) { + ERTS_SYS_TRY_CATCH( + erts_print_scheduler_info(fd, NULL, ERTS_SCHEDULER_IX(i)), + erts_fdprintf(fd, "** crashed **\n")); + } +#endif + +#ifdef ERTS_SMP + +#if defined(ERTS_THR_HAVE_SIG_FUNCS) + + /* We resume all schedulers so that we are in a known safe state + when we write the rest of the crash dump */ + for (i = 0; i < erts_no_schedulers; i++) { + erts_tid_t tid = ERTS_SCHEDULER_IX(i)->tid; + if (!erts_equal_tids(tid,erts_thr_self())) + sys_thr_resume(tid); + } +#endif + + /* + * Wait for all managed threads to block. If all threads haven't blocked + * after a minute, we go anyway and hope for the best... + * + * We do not release system again. We expect an exit() or abort() after + * dump has been written. + */ + erts_thr_progress_fatal_error_wait(60000); + /* Either worked or not... */ +#endif + +#ifndef ERTS_HAVE_TRY_CATCH + /* This is safe to call here, as all schedulers are blocked */ + for (i = 0; i < erts_no_schedulers; i++) { + erts_print_scheduler_info(fd, NULL, ERTS_SCHEDULER_IX(i)); + } +#endif + info(fd, NULL); /* General system info */ if (erts_ptab_initialized(&erts_proc)) process_info(fd, NULL); /* Info about each process and port */ diff --git a/erts/emulator/beam/erl_bif_info.c b/erts/emulator/beam/erl_bif_info.c index d2ee5e4224..d750e34be3 100644 --- a/erts/emulator/beam/erl_bif_info.c +++ b/erts/emulator/beam/erl_bif_info.c @@ -3900,6 +3900,13 @@ BIF_RETTYPE erts_debug_set_internal_state_2(BIF_ALIST_2) } } } + else if (ERTS_IS_ATOM_STR("broken_halt", BIF_ARG_1)) { + /* Ugly ugly code used by bif_SUITE:erlang_halt/1 */ +#if defined(ERTS_HAVE_TRY_CATCH) + erts_get_scheduler_data()->run_queue = NULL; +#endif + erl_exit(ERTS_DUMP_EXIT, "%T", BIF_ARG_2); + } } BIF_ERROR(BIF_P, BADARG); diff --git a/erts/emulator/beam/erl_process.c b/erts/emulator/beam/erl_process.c index 31dda21b47..1ec931e312 100644 --- a/erts/emulator/beam/erl_process.c +++ b/erts/emulator/beam/erl_process.c @@ -457,8 +457,7 @@ do { \ static void exec_misc_ops(ErtsRunQueue *); static void print_function_from_pc(int to, void *to_arg, BeamInstr* x); -static int stack_element_dump(int to, void *to_arg, Process* p, Eterm* sp, - int yreg); +static int stack_element_dump(int to, void *to_arg, Eterm* sp, int yreg); static void aux_work_timeout(void *unused); static void aux_work_timeout_early_init(int no_schedulers); @@ -12186,7 +12185,7 @@ erts_stack_dump(int to, void *to_arg, Process *p) } erts_program_counter_info(to, to_arg, p); for (sp = p->stop; sp < STACK_START(p); sp++) { - yreg = stack_element_dump(to, to_arg, p, sp, yreg); + yreg = stack_element_dump(to, to_arg, sp, yreg); } } @@ -12243,7 +12242,7 @@ print_function_from_pc(int to, void *to_arg, BeamInstr* x) } static int -stack_element_dump(int to, void *to_arg, Process* p, Eterm* sp, int yreg) +stack_element_dump(int to, void *to_arg, Eterm* sp, int yreg) { Eterm x = *sp; @@ -12272,6 +12271,214 @@ stack_element_dump(int to, void *to_arg, Process* p, Eterm* sp, int yreg) } /* + * Print scheduler information + */ +void +erts_print_scheduler_info(int to, void *to_arg, ErtsSchedulerData *esdp) { + int i; + erts_aint32_t flg; + Process *p; + + erts_print(to, to_arg, "=scheduler:%u\n", esdp->no); + +#ifdef ERTS_SMP + flg = erts_smp_atomic32_read_dirty(&esdp->ssi->flags); + erts_print(to, to_arg, "Scheduler Sleep Info Flags: "); + for (i = 0; i < ERTS_SSI_FLGS_MAX && flg; i++) { + erts_aint32_t chk = (1 << i); + if (flg & chk) { + switch (chk) { + case ERTS_SSI_FLG_SLEEPING: + erts_print(to, to_arg, "SLEEPING"); break; + case ERTS_SSI_FLG_POLL_SLEEPING: + erts_print(to, to_arg, "POLL_SLEEPING"); break; + case ERTS_SSI_FLG_TSE_SLEEPING: + erts_print(to, to_arg, "TSE_SLEEPING"); break; + case ERTS_SSI_FLG_WAITING: + erts_print(to, to_arg, "WAITING"); break; + case ERTS_SSI_FLG_SUSPENDED: + erts_print(to, to_arg, "SUSPENDED"); break; + default: + erts_print(to, to_arg, "UNKNOWN(%d)", flg); break; + } + if (flg > chk) + erts_print(to, to_arg, " | "); + flg -= chk; + } + } + erts_print(to, to_arg, "\n"); +#endif + + flg = erts_atomic32_read_dirty(&esdp->ssi->aux_work); + erts_print(to, to_arg, "Scheduler Sleep Info Aux Work: "); + for (i = 0; i < ERTS_SSI_AUX_WORK_MAX && flg; i++) { + erts_aint32_t chk = (1 << i); + if (flg & chk) { + switch (chk) { + case ERTS_SSI_AUX_WORK_DELAYED_AW_WAKEUP: + erts_print(to, to_arg, "DELAYED_AW_WAKEUP"); break; + case ERTS_SSI_AUX_WORK_DD: + erts_print(to, to_arg, "DELAYED_DEALLOC"); break; + case ERTS_SSI_AUX_WORK_DD_THR_PRGR: + erts_print(to, to_arg, "DELAYED_DEALLOC_THR_PRGR"); break; + case ERTS_SSI_AUX_WORK_FIX_ALLOC_DEALLOC: + erts_print(to, to_arg, "FIX_ALLOC_DEALLOC"); break; + case ERTS_SSI_AUX_WORK_FIX_ALLOC_LOWER_LIM: + erts_print(to, to_arg, "FIX_ALLOC_LOWER_LIM"); break; + case ERTS_SSI_AUX_WORK_THR_PRGR_LATER_OP: + erts_print(to, to_arg, "THR_PRGR_LATER_OP"); break; + case ERTS_SSI_AUX_WORK_ASYNC_READY: + erts_print(to, to_arg, "ASYNC_READY"); break; + case ERTS_SSI_AUX_WORK_ASYNC_READY_CLEAN: + erts_print(to, to_arg, "ASYNC_READY_CLEAN"); break; + case ERTS_SSI_AUX_WORK_MISC_THR_PRGR: + erts_print(to, to_arg, "MISC_THR_PRGR"); break; + case ERTS_SSI_AUX_WORK_MISC: + erts_print(to, to_arg, "MISC"); break; + case ERTS_SSI_AUX_WORK_CHECK_CHILDREN: + erts_print(to, to_arg, "CHECK_CHILDREN"); break; + case ERTS_SSI_AUX_WORK_SET_TMO: + erts_print(to, to_arg, "SET_TMO"); break; + case ERTS_SSI_AUX_WORK_MSEG_CACHE_CHECK: + erts_print(to, to_arg, "MSEG_CACHE_CHECK"); break; + case ERTS_SSI_AUX_WORK_REAP_PORTS: + erts_print(to, to_arg, "REAP_PORTS"); break; + default: + erts_print(to, to_arg, "UNKNOWN(%d)", flg); break; + } + if (flg > chk) + erts_print(to, to_arg, " | "); + flg -= chk; + } + } + erts_print(to, to_arg, "\n"); + + erts_print(to, to_arg, "Current Port: "); + if (esdp->current_port) + erts_print(to, to_arg, "%T", esdp->current_port->common.id); + erts_print(to, to_arg, "\n"); + + p = esdp->current_process; + erts_print(to, to_arg, "Current Process: "); + if (esdp->current_process && !(ERTS_TRACE_FLAGS(p) & F_SENSITIVE)) { + flg = erts_smp_atomic32_read_dirty(&p->state); + erts_print(to, to_arg, "%T\n", p->common.id); + + erts_print(to, to_arg, "Current Process State: "); + erts_dump_process_state(to, to_arg, flg); + + erts_print(to, to_arg, "Current Process Internal State: "); + erts_dump_extended_process_state(to, to_arg, flg); + + erts_print(to, to_arg, "Current Process Program counter: %p (", p->i); + print_function_from_pc(to, to_arg, p->i); + erts_print(to, to_arg, ")\n"); + erts_print(to, to_arg, "Current Process CP: %p (", p->cp); + print_function_from_pc(to, to_arg, p->cp); + erts_print(to, to_arg, ")\n"); + + /* Getting this stacktrace can segfault if we are very very + unlucky if called while a process is being garbage collected. + Therefore we only call this on other schedulers if we either + have protection against segfaults, or we know that the process + is not garbage collecting. It *should* always be safe to call + on a process owned by us, even if it is currently being garbage + collected. + */ + erts_print(to, to_arg, "Current Process Limited Stack Trace:\n"); + erts_limited_stack_trace(to, to_arg, p); + } else + erts_print(to, to_arg, "\n"); + + for (i = 0; i < ERTS_NO_PROC_PRIO_LEVELS; i++) { + erts_print(to, to_arg, "Run Queue "); + switch (i) { + case PRIORITY_MAX: + erts_print(to, to_arg, "Max "); + break; + case PRIORITY_HIGH: + erts_print(to, to_arg, "High "); + break; + case PRIORITY_NORMAL: + erts_print(to, to_arg, "Normal "); + break; + case PRIORITY_LOW: + erts_print(to, to_arg, "Low "); + break; + default: + erts_print(to, to_arg, "Unknown "); + break; + } + erts_print(to, to_arg, "Length: %d\n", + erts_smp_atomic32_read_dirty(&esdp->run_queue->procs.prio_info[i].len)); + } + erts_print(to, to_arg, "Run Queue Port Length: %d\n", + erts_smp_atomic32_read_dirty(&esdp->run_queue->ports.info.len)); + + flg = erts_smp_atomic32_read_dirty(&esdp->run_queue->flags); + erts_print(to, to_arg, "Run Queue Flags: "); + for (i = 0; i < ERTS_RUNQ_FLG_MAX && flg; i++) { + erts_aint32_t chk = (1 << i); + if (flg & chk) { + switch (chk) { + case (1 << PRIORITY_MAX): + erts_print(to, to_arg, "NONEMPTY_MAX"); break; + case (1 << PRIORITY_HIGH): + erts_print(to, to_arg, "NONEMPTY_HIGH"); break; + case (1 << PRIORITY_NORMAL): + erts_print(to, to_arg, "NONEMPTY_NORMAL"); break; + case (1 << PRIORITY_LOW): + erts_print(to, to_arg, "NONEMPTY_LOW"); break; + case (1 << (PRIORITY_MAX + ERTS_RUNQ_FLGS_EMIGRATE_SHFT)): + erts_print(to, to_arg, "EMIGRATE_MAX"); break; + case (1 << (PRIORITY_HIGH + ERTS_RUNQ_FLGS_EMIGRATE_SHFT)): + erts_print(to, to_arg, "EMIGRATE_HIGH"); break; + case (1 << (PRIORITY_NORMAL + ERTS_RUNQ_FLGS_EMIGRATE_SHFT)): + erts_print(to, to_arg, "EMIGRATE_NORMAL"); break; + case (1 << (PRIORITY_LOW + ERTS_RUNQ_FLGS_EMIGRATE_SHFT)): + erts_print(to, to_arg, "EMIGRATE_LOW"); break; + case (1 << (PRIORITY_MAX + ERTS_RUNQ_FLGS_IMMIGRATE_SHFT)): + erts_print(to, to_arg, "IMMIGRATE_MAX"); break; + case (1 << (PRIORITY_HIGH + ERTS_RUNQ_FLGS_IMMIGRATE_SHFT)): + erts_print(to, to_arg, "IMMIGRATE_HIGH"); break; + case (1 << (PRIORITY_NORMAL + ERTS_RUNQ_FLGS_IMMIGRATE_SHFT)): + erts_print(to, to_arg, "IMMIGRATE_NORMAL"); break; + case (1 << (PRIORITY_LOW + ERTS_RUNQ_FLGS_IMMIGRATE_SHFT)): + erts_print(to, to_arg, "IMMIGRATE_LOW"); break; + case (1 << (PRIORITY_MAX + ERTS_RUNQ_FLGS_EVACUATE_SHFT)): + erts_print(to, to_arg, "EVACUATE_MAX"); break; + case (1 << (PRIORITY_HIGH + ERTS_RUNQ_FLGS_EVACUATE_SHFT)): + erts_print(to, to_arg, "EVACUATE_HIGH"); break; + case (1 << (PRIORITY_NORMAL + ERTS_RUNQ_FLGS_EVACUATE_SHFT)): + erts_print(to, to_arg, "EVACUATE_NORMAL"); break; + case (1 << (PRIORITY_LOW + ERTS_RUNQ_FLGS_EVACUATE_SHFT)): + erts_print(to, to_arg, "EVACUATE_LOW"); break; + case ERTS_RUNQ_FLG_OUT_OF_WORK: + erts_print(to, to_arg, "OUT_OF_WORK"); break; + case ERTS_RUNQ_FLG_HALFTIME_OUT_OF_WORK: + erts_print(to, to_arg, "HALFTIME_OUT_OF_WORK"); break; + case ERTS_RUNQ_FLG_SUSPENDED: + erts_print(to, to_arg, "SUSPENDED"); break; + case ERTS_RUNQ_FLG_CHK_CPU_BIND: + erts_print(to, to_arg, "CHK_CPU_BIND"); break; + case ERTS_RUNQ_FLG_INACTIVE: + erts_print(to, to_arg, "INACTIVE"); break; + case ERTS_RUNQ_FLG_NONEMPTY: + erts_print(to, to_arg, "NONEMPTY"); break; + case ERTS_RUNQ_FLG_PROTECTED: + erts_print(to, to_arg, "PROTECTED"); break; + default: + erts_print(to, to_arg, "UNKNOWN(%d)", flg); break; + } + if (flg > chk) + erts_print(to, to_arg, " | "); + flg -= chk; + } + } + erts_print(to, to_arg, "\n"); +} + +/* * A nice system halt closing all open port goes as follows: * 1) This function schedules the aux work ERTS_SSI_AUX_WORK_REAP_PORTS * on all schedulers, then schedules itself out. diff --git a/erts/emulator/beam/erl_process.h b/erts/emulator/beam/erl_process.h index 3d08be25ff..a12b577656 100644 --- a/erts/emulator/beam/erl_process.h +++ b/erts/emulator/beam/erl_process.h @@ -170,6 +170,8 @@ extern int erts_sched_thread_suggested_stack_size; #define ERTS_RUNQ_FLG_PROTECTED \ (((Uint32) 1) << (ERTS_RUNQ_FLG_BASE2 + 6)) +#define ERTS_RUNQ_FLG_MAX (ERTS_RUNQ_FLG_BASE2 + 7) + #define ERTS_RUNQ_FLGS_MIGRATION_QMASKS \ (ERTS_RUNQ_FLGS_EMIGRATE_QMASK \ | ERTS_RUNQ_FLGS_IMMIGRATE_QMASK \ @@ -252,6 +254,8 @@ typedef enum { #define ERTS_SSI_FLG_WAITING (((erts_aint32_t) 1) << 3) #define ERTS_SSI_FLG_SUSPENDED (((erts_aint32_t) 1) << 4) +#define ERTS_SSI_FLGS_MAX 5 + #define ERTS_SSI_FLGS_SLEEP_TYPE \ (ERTS_SSI_FLG_TSE_SLEEPING|ERTS_SSI_FLG_POLL_SLEEPING) @@ -283,6 +287,8 @@ typedef enum { #define ERTS_SSI_AUX_WORK_MSEG_CACHE_CHECK (((erts_aint32_t) 1) << 12) #define ERTS_SSI_AUX_WORK_REAP_PORTS (((erts_aint32_t) 1) << 13) +#define ERTS_SSI_AUX_WORK_MAX 14 + typedef struct ErtsSchedulerSleepInfo_ ErtsSchedulerSleepInfo; #ifdef ERTS_DIRTY_SCHEDULERS @@ -1081,6 +1087,9 @@ void erts_check_for_holes(Process* p); #define ERTS_PSFLG_DIRTY_IO_PROC ERTS_PSFLG_BIT(19) #define ERTS_PSFLG_DIRTY_CPU_PROC_IN_Q ERTS_PSFLG_BIT(20) #define ERTS_PSFLG_DIRTY_IO_PROC_IN_Q ERTS_PSFLG_BIT(21) +#define ERTS_PSFLG_MAX (ERTS_PSFLGS_ZERO_BIT_OFFSET + 22) +#else +#define ERTS_PSFLG_MAX (ERTS_PSFLGS_ZERO_BIT_OFFSET + 18) #endif #define ERTS_PSFLGS_IN_PRQ_MASK (ERTS_PSFLG_IN_PRQ_MAX \ @@ -1616,7 +1625,11 @@ void erts_cleanup_empty_process(Process* p); void erts_debug_verify_clean_empty_process(Process* p); #endif void erts_stack_dump(int to, void *to_arg, Process *); +void erts_limited_stack_trace(int to, void *to_arg, Process *); void erts_program_counter_info(int to, void *to_arg, Process *); +void erts_print_scheduler_info(int to, void *to_arg, ErtsSchedulerData *esdp); +void erts_dump_extended_process_state(int to, void *to_arg, erts_aint32_t psflg); +void erts_dump_process_state(int to, void *to_arg, erts_aint32_t psflg); Eterm erts_get_process_priority(Process *p); Eterm erts_set_process_priority(Process *p, Eterm prio); diff --git a/erts/emulator/beam/erl_process_dump.c b/erts/emulator/beam/erl_process_dump.c index 2f3cf23b00..36bb6b2f0e 100644 --- a/erts/emulator/beam/erl_process_dump.c +++ b/erts/emulator/beam/erl_process_dump.c @@ -43,8 +43,9 @@ static void dump_process_info(int to, void *to_arg, Process *p); static void dump_element(int to, void *to_arg, Eterm x); static void dump_dist_ext(int to, void *to_arg, ErtsDistExternal *edep); static void dump_element_nl(int to, void *to_arg, Eterm x); -static int stack_element_dump(int to, void *to_arg, Process* p, Eterm* sp, +static int stack_element_dump(int to, void *to_arg, Eterm* sp, int yreg); +static void stack_trace_dump(int to, void *to_arg, Eterm* sp); static void print_function_from_pc(int to, void *to_arg, BeamInstr* x); static void heap_dump(int to, void *to_arg, Eterm x); static void dump_binaries(int to, void *to_arg, Binary* root); @@ -148,7 +149,7 @@ dump_process_info(int to, void *to_arg, Process *p) if ((ERTS_TRACE_FLAGS(p) & F_SENSITIVE) == 0) { erts_print(to, to_arg, "=proc_stack:%T\n", p->common.id); for (sp = p->stop; sp < STACK_START(p); sp++) { - yreg = stack_element_dump(to, to_arg, p, sp, yreg); + yreg = stack_element_dump(to, to_arg, sp, yreg); } erts_print(to, to_arg, "=proc_heap:%T\n", p->common.id); @@ -243,9 +244,65 @@ dump_element_nl(int to, void *to_arg, Eterm x) erts_putc(to, to_arg, '\n'); } +static void +stack_trace_dump(int to, void *to_arg, Eterm *sp) { + Eterm x = *sp; + if (is_CP(x)) { + erts_print(to, to_arg, "%p:", sp); + erts_print(to, to_arg, "SReturn addr 0x%X (", cp_val(x)); + print_function_from_pc(to, to_arg, cp_val(x)); + erts_print(to, to_arg, ")\n"); + } +} + +void +erts_limited_stack_trace(int to, void *to_arg, Process *p) +{ + Eterm* sp; + + + if (ERTS_TRACE_FLAGS(p) & F_SENSITIVE) { + return; + } + + if (STACK_START(p) < STACK_TOP(p)) { + return; + } + + if ((STACK_START(p) - STACK_TOP(p)) < 512) { + if (erts_sys_is_area_readable((char*)STACK_TOP(p), + (char*)STACK_START(p))) + for (sp = STACK_TOP(p); sp < STACK_START(p); sp++) + stack_trace_dump(to, to_arg, sp); + else + erts_print(to, to_arg, "Could not read from stack memory: %p - %p\n", + STACK_TOP(p), STACK_START(p)); + } else { + sp = STACK_TOP(p); + if (erts_sys_is_area_readable((char*)STACK_TOP(p), + (char*)(STACK_TOP(p) + 25))) + for (; sp < (STACK_TOP(p) + 256); sp++) + stack_trace_dump(to, to_arg, sp); + else + erts_print(to, to_arg, "Could not read from stack memory: %p - %p\n", + STACK_TOP(p), STACK_TOP(p) + 256); + + erts_print(to, to_arg, "%p: skipping %d frames\n", + sp, STACK_START(p) - STACK_TOP(p) - 512); + + if (erts_sys_is_area_readable((char*)(STACK_START(p) - 256), + (char*)STACK_START(p))) + for (sp = STACK_START(p) - 256; sp < STACK_START(p); sp++) + stack_trace_dump(to, to_arg, sp); + else + erts_print(to, to_arg, "Could not read from stack memory: %p - %p\n", + STACK_START(p) - 256, STACK_START(p)); + } + +} static int -stack_element_dump(int to, void *to_arg, Process* p, Eterm* sp, int yreg) +stack_element_dump(int to, void *to_arg, Eterm* sp, int yreg) { Eterm x = *sp; @@ -508,3 +565,114 @@ dump_externally(int to, void *to_arg, Eterm term) erts_print(to, to_arg, "%02X", *s++); } } + +void erts_dump_process_state(int to, void *to_arg, erts_aint32_t psflg) { + if (psflg & ERTS_PSFLG_FREE) + erts_print(to, to_arg, "Non Existing\n"); /* Should never happen */ + else if (psflg & ERTS_PSFLG_EXITING) + erts_print(to, to_arg, "Exiting\n"); + else if (psflg & ERTS_PSFLG_GC) { + erts_print(to, to_arg, "Garbing\n"); + } + else if (psflg & ERTS_PSFLG_SUSPENDED) + erts_print(to, to_arg, "Suspended\n"); + else if (psflg & ERTS_PSFLG_RUNNING) { + erts_print(to, to_arg, "Running\n"); + } + else if (psflg & ERTS_PSFLG_ACTIVE) + erts_print(to, to_arg, "Scheduled\n"); + else + erts_print(to, to_arg, "Waiting\n"); +} + +void +erts_dump_extended_process_state(int to, void *to_arg, erts_aint32_t psflg) { + + int i; + + switch (ERTS_PSFLGS_GET_ACT_PRIO(psflg)) { + case PRIORITY_MAX: erts_print(to, to_arg, "ACT_PRIO_MAX | "); break; + case PRIORITY_HIGH: erts_print(to, to_arg, "ACT_PRIO_HIGH | "); break; + case PRIORITY_NORMAL: erts_print(to, to_arg, "ACT_PRIO_NORMAL | "); break; + case PRIORITY_LOW: erts_print(to, to_arg, "ACT_PRIO_LOW | "); break; + } + switch (ERTS_PSFLGS_GET_USR_PRIO(psflg)) { + case PRIORITY_MAX: erts_print(to, to_arg, "USR_PRIO_MAX | "); break; + case PRIORITY_HIGH: erts_print(to, to_arg, "USR_PRIO_HIGH | "); break; + case PRIORITY_NORMAL: erts_print(to, to_arg, "USR_PRIO_NORMAL | "); break; + case PRIORITY_LOW: erts_print(to, to_arg, "USR_PRIO_LOW | "); break; + } + switch (ERTS_PSFLGS_GET_PRQ_PRIO(psflg)) { + case PRIORITY_MAX: erts_print(to, to_arg, "PRQ_PRIO_MAX"); break; + case PRIORITY_HIGH: erts_print(to, to_arg, "PRQ_PRIO_HIGH"); break; + case PRIORITY_NORMAL: erts_print(to, to_arg, "PRQ_PRIO_NORMAL"); break; + case PRIORITY_LOW: erts_print(to, to_arg, "PRQ_PRIO_LOW"); break; + } + + psflg &= ~(ERTS_PSFLGS_ACT_PRIO_MASK | + ERTS_PSFLGS_USR_PRIO_MASK | + ERTS_PSFLGS_PRQ_PRIO_MASK); + + if (psflg) + erts_print(to, to_arg, " | "); + + for (i = 0; i < ERTS_PSFLG_MAX && psflg; i++) { + erts_aint32_t chk = (1 << i); + if (psflg & chk) { + switch (chk) { + case ERTS_PSFLG_IN_PRQ_MAX: + erts_print(to, to_arg, "IN_PRQ_MAX"); break; + case ERTS_PSFLG_IN_PRQ_HIGH: + erts_print(to, to_arg, "IN_PRQ_HIGH"); break; + case ERTS_PSFLG_IN_PRQ_NORMAL: + erts_print(to, to_arg, "IN_PRQ_NORMAL"); break; + case ERTS_PSFLG_IN_PRQ_LOW: + erts_print(to, to_arg, "IN_PRQ_LOW"); break; + case ERTS_PSFLG_FREE: + erts_print(to, to_arg, "FREE"); break; + case ERTS_PSFLG_EXITING: + erts_print(to, to_arg, "EXITING"); break; + case ERTS_PSFLG_PENDING_EXIT: + erts_print(to, to_arg, "PENDING_EXIT"); break; + case ERTS_PSFLG_ACTIVE: + erts_print(to, to_arg, "ACTIVE"); break; + case ERTS_PSFLG_IN_RUNQ: + erts_print(to, to_arg, "IN_RUNQ"); break; + case ERTS_PSFLG_RUNNING: + erts_print(to, to_arg, "RUNNING"); break; + case ERTS_PSFLG_SUSPENDED: + erts_print(to, to_arg, "SUSPENDED"); break; + case ERTS_PSFLG_GC: + erts_print(to, to_arg, "GC"); break; + case ERTS_PSFLG_BOUND: + erts_print(to, to_arg, "BOUND"); break; + case ERTS_PSFLG_TRAP_EXIT: + erts_print(to, to_arg, "TRAP_EXIT"); break; + case ERTS_PSFLG_ACTIVE_SYS: + erts_print(to, to_arg, "ACTIVE_SYS"); break; + case ERTS_PSFLG_RUNNING_SYS: + erts_print(to, to_arg, "RUNNING_SYS"); break; + case ERTS_PSFLG_PROXY: + erts_print(to, to_arg, "PROXY"); break; + case ERTS_PSFLG_DELAYED_SYS: + erts_print(to, to_arg, "DELAYED_SYS"); break; +#ifdef ERTS_DIRTY_SCHEDULERS + case ERTS_PSFLG_DIRTY_CPU_PROC: + erts_print(to, to_arg, "DIRTY_CPU_PROC"); break; + case ERTS_PSFLG_DIRTY_IO_PROC: + erts_print(to, to_arg, "DIRTY_IO_PROC"); break; + case ERTS_PSFLG_DIRTY_CPU_PROC_IN_Q: + erts_print(to, to_arg, "DIRTY_CPU_PROC_IN_Q"); break; + case ERTS_PSFLG_DIRTY_IO_PROC_IN_Q: + erts_print(to, to_arg, "DIRTY_IO_PROC_IN_Q"); break; +#endif + default: + erts_print(to, to_arg, "UNKNOWN(%d)", chk); break; + } + if (psflg > chk) + erts_print(to, to_arg, " | "); + psflg -= chk; + } + } + erts_print(to, to_arg, "\n"); +} diff --git a/erts/emulator/beam/erl_thr_progress.c b/erts/emulator/beam/erl_thr_progress.c index 545a0343d0..3dca2aba10 100644 --- a/erts/emulator/beam/erl_thr_progress.c +++ b/erts/emulator/beam/erl_thr_progress.c @@ -1381,25 +1381,10 @@ erts_thr_progress_block(void) thr_progress_block(tmp_thr_prgr_data(NULL), 1); } -void -erts_thr_progress_fatal_error_block(SWord timeout, - ErtsThrPrgrData *tmp_tpd_bufp) +int +erts_thr_progress_fatal_error_block(ErtsThrPrgrData *tmp_tpd_bufp) { ErtsThrPrgrData *tpd = perhaps_thr_prgr_data(NULL); - erts_aint32_t bc; - SWord time_left = timeout; - SysTimeval to; - - /* - * Counting poll intervals may give us a too long timeout - * if cpu is busy. If we got tolerant time of day we use it - * to prevent this. - */ - if (!erts_disable_tolerant_timeofday) { - erts_get_timeval(&to); - to.tv_sec += timeout / 1000; - to.tv_sec += timeout % 1000; - } if (!tpd) { /* @@ -1412,9 +1397,26 @@ erts_thr_progress_fatal_error_block(SWord timeout, init_tmp_thr_prgr_data(tpd); } - bc = thr_progress_block(tpd, 0); - if (bc == 0) - return; /* Succefully blocked all managed threads */ + /* Returns number of threads that have not yes been blocked */ + return thr_progress_block(tpd, 0); +} + +void +erts_thr_progress_fatal_error_wait(SWord timeout) { + erts_aint32_t bc; + SWord time_left = timeout; + SysTimeval to; + + /* + * Counting poll intervals may give us a too long timeout + * if cpu is busy. If we got tolerant time of day we use it + * to prevent this. + */ + if (!erts_disable_tolerant_timeofday) { + erts_get_timeval(&to); + to.tv_sec += timeout / 1000; + to.tv_sec += timeout % 1000; + } while (1) { if (erts_milli_sleep(ERTS_THR_PRGR_FTL_ERR_BLCK_POLL_INTERVAL) == 0) diff --git a/erts/emulator/beam/erl_thr_progress.h b/erts/emulator/beam/erl_thr_progress.h index 5f392944c2..85b4eeb0eb 100644 --- a/erts/emulator/beam/erl_thr_progress.h +++ b/erts/emulator/beam/erl_thr_progress.h @@ -83,8 +83,8 @@ typedef struct { ErtsThrPrgrLeaderState leader_state; } ErtsThrPrgrData; -void erts_thr_progress_fatal_error_block(SWord timeout, - ErtsThrPrgrData *tmp_tpd_bufp); +int erts_thr_progress_fatal_error_block(ErtsThrPrgrData *tmp_tpd_bufp); +void erts_thr_progress_fatal_error_wait(SWord timeout); #endif /* ERTS_SMP */ diff --git a/erts/emulator/beam/sys.h b/erts/emulator/beam/sys.h index bdc5237815..9227ce74a2 100644 --- a/erts/emulator/beam/sys.h +++ b/erts/emulator/beam/sys.h @@ -756,6 +756,8 @@ typedef struct { } ErtsCheckIoDebugInfo; int erts_check_io_debug(ErtsCheckIoDebugInfo *ip); +int erts_sys_is_area_readable(char *start, char *stop); + /* xxxP */ #define SYS_DEFAULT_FLOAT_DECIMALS 20 void init_sys_float(void); diff --git a/erts/emulator/sys/unix/erl_unix_sys.h b/erts/emulator/sys/unix/erl_unix_sys.h index bc55fab8fb..1932e97742 100644 --- a/erts/emulator/sys/unix/erl_unix_sys.h +++ b/erts/emulator/sys/unix/erl_unix_sys.h @@ -45,7 +45,7 @@ #include <fcntl.h> #include "erl_errno.h" #include <signal.h> - +#include <setjmp.h> #if HAVE_SYS_SOCKETIO_H # include <sys/socketio.h> @@ -349,4 +349,28 @@ extern int exit_async(void); #define ERTS_EXIT_AFTER_DUMP _exit +#if !defined(__APPLE__) && !defined(__MACH__) +/* Some OS X versions do not allow (ab)using signal handlers like this */ +#define ERTS_HAVE_TRY_CATCH 1 + +/* We try to simulate a try catch in C with the help of signal handlers. + * Only use this as a very last resort, as it is not very portable and + * quite unstable. It is also not thread safe, so make sure that only + * one thread can call this at a time! + */ +extern void erts_sys_sigsegv_handler(int); +extern jmp_buf erts_sys_sigsegv_jmp; +#define ERTS_SYS_TRY_CATCH(EXPR,CATCH) \ + do { \ + SIGFUNC prev_handler = sys_signal(SIGSEGV, \ + erts_sys_sigsegv_handler); \ + if (!setjmp(erts_sys_sigsegv_jmp)) { \ + EXPR; \ + } else { \ + CATCH; \ + } \ + sys_signal(SIGSEGV,prev_handler); \ + } while(0) +#endif + #endif /* #ifndef _ERL_UNIX_SYS_H */ diff --git a/erts/emulator/sys/unix/sys.c b/erts/emulator/sys/unix/sys.c index 74d67cbd57..57f9fbef53 100644 --- a/erts/emulator/sys/unix/sys.c +++ b/erts/emulator/sys/unix/sys.c @@ -223,6 +223,8 @@ static int sig_suspend_fds[2] = {-1, -1}; static int async_fd[2]; #endif +jmp_buf erts_sys_sigsegv_jmp; + #if CHLDWTHR || defined(ERTS_SMP) erts_mtx_t chld_stat_mtx; #endif @@ -683,6 +685,35 @@ void sys_sigrelease(int sig) sigprocmask(SIG_UNBLOCK, &mask, (sigset_t *)NULL); } +void erts_sys_sigsegv_handler(int signo) { + if (signo == SIGSEGV) { + longjmp(erts_sys_sigsegv_jmp, 1); + } +} + +/* + * Function returns 1 if we can read from all values in between + * start and stop. + */ +int +erts_sys_is_area_readable(char *start, char *stop) { + int fds[2]; + if (!pipe(fds)) { + /* We let write try to figure out if the pointers are readable */ + int res = write(fds[1], start, (char*)stop - (char*)start); + if (res == -1) { + close(fds[0]); + close(fds[1]); + return 0; + } + close(fds[0]); + close(fds[1]); + return 1; + } + return 0; + +} + static ERTS_INLINE int prepare_crash_dump(int secs) { diff --git a/erts/emulator/sys/win32/erl_win_sys.h b/erts/emulator/sys/win32/erl_win_sys.h index 838f0c61eb..fde32c8684 100644 --- a/erts/emulator/sys/win32/erl_win_sys.h +++ b/erts/emulator/sys/win32/erl_win_sys.h @@ -236,4 +236,16 @@ typedef long ssize_t; int init_async(int); int exit_async(void); #endif + +#define ERTS_HAVE_TRY_CATCH 1 + +#define ERTS_SYS_TRY_CATCH(EXPR,CATCH) \ + __try { \ + EXPR; \ + } \ + __except(GetExceptionCode() == EXCEPTION_ACCESS_VIOLATION ? EXCEPTION_EXECUTE_HANDLER : EXCEPTION_CONTINUE_SEARCH) \ + { \ + CATCH; \ + } + #endif diff --git a/erts/emulator/sys/win32/sys.c b/erts/emulator/sys/win32/sys.c index 164ef95629..5d51659b4e 100644 --- a/erts/emulator/sys/win32/sys.c +++ b/erts/emulator/sys/win32/sys.c @@ -247,6 +247,27 @@ void erl_sys_args(int* argc, char** argv) #endif } +/* + * Function returns 1 if we can read from all values in between + * start and stop. + */ +int +erts_sys_is_area_readable(char *start, char *stop) { + volatile char tmp; + __try + { + while(start < stop) { + tmp = *start; + start++; + } + } + __except(EXCEPTION_EXECUTE_HANDLER) + { + return 0; + } + return 1; +} + int erts_sys_prepare_crash_dump(int secs) { Port *heart_port; diff --git a/erts/emulator/test/bif_SUITE.erl b/erts/emulator/test/bif_SUITE.erl index fbc229bc53..fc9bdae0a0 100644 --- a/erts/emulator/test/bif_SUITE.erl +++ b/erts/emulator/test/bif_SUITE.erl @@ -20,6 +20,7 @@ -module(bif_SUITE). -include_lib("test_server/include/test_server.hrl"). +-include_lib("kernel/include/file.hrl"). -export([all/0, suite/0,groups/0,init_per_suite/1, end_per_suite/1, init_per_group/2,end_per_group/2, @@ -681,8 +682,38 @@ erlang_halt(Config) when is_list(Config) -> {badrpc,nodedown} = rpc:call(N2, erlang, halt, [0]), {ok,N3} = slave:start(H, halt_node3), {badrpc,nodedown} = rpc:call(N3, erlang, halt, [0,[]]), - ok. + % This test triggers a segfault when dumping a crash dump + % to make sure that we can handle it properly. + {ok,N4} = slave:start(H, halt_node4), + CrashDump = filename:join(proplists:get_value(priv_dir,Config), + "segfault_erl_crash.dump"), + true = rpc:call(N4, os, putenv, ["ERL_CRASH_DUMP",CrashDump]), + false = rpc:call(N4, erts_debug, set_internal_state, + [available_internal_state, true]), + {badrpc,nodedown} = rpc:call(N4, erts_debug, set_internal_state, + [broken_halt, "Validate correct crash dump"]), + ok = wait_until_stable_size(CrashDump,-1), + {ok, Bin} = file:read_file(CrashDump), + case {string:str(binary_to_list(Bin),"\n=end\n"), + string:str(binary_to_list(Bin),"\r\n=end\r\n")} of + {0,0} -> ct:fail("Could not find end marker in crash dump"); + _ -> ok + end. + +wait_until_stable_size(_File,-10) -> + {error,enoent}; +wait_until_stable_size(File,PrevSz) -> + timer:sleep(250), + case file:read_file_info(File) of + {error,enoent} -> + wait_until_stable_size(File,PrevSz-1); + {ok,#file_info{size = PrevSz }} when PrevSz /= -1 -> + io:format("Crashdump file size was: ~p (~s)~n",[PrevSz,File]), + ok; + {ok,#file_info{size = NewSz }} -> + wait_until_stable_size(File,NewSz) + end. %% Helpers |