X-Git-Url: https://git.kernel.dk/?p=fio.git;a=blobdiff_plain;f=backend.c;h=d3e9bf8160615b64ae714b47f4524403df272f44;hp=86e689f838a00e752e8268ef6efda71fda156f00;hb=4d832322bf67a8a496bee049bc7dd9a2fdd828a2;hpb=6c8611c79713fe73fddf7458ab3ab36feaeae67b diff --git a/backend.c b/backend.c index 86e689f8..d3e9bf81 100644 --- a/backend.c +++ b/backend.c @@ -18,7 +18,7 @@ * * You should have received a copy of the GNU General Public License * along with this program; if not, write to the Free Software - * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. * */ #include @@ -136,7 +136,7 @@ static void set_sig_handlers(void) /* * Check if we are above the minimum rate given. */ -static bool __check_min_rate(struct thread_data *td, struct timeval *now, +static bool __check_min_rate(struct thread_data *td, struct timespec *now, enum fio_ddir ddir) { unsigned long long bytes = 0; @@ -223,7 +223,7 @@ static bool __check_min_rate(struct thread_data *td, struct timeval *now, return false; } -static bool check_min_rate(struct thread_data *td, struct timeval *now) +static bool check_min_rate(struct thread_data *td, struct timespec *now) { bool ret = false; @@ -335,18 +335,18 @@ static int fio_file_fsync(struct thread_data *td, struct fio_file *f) return ret; } -static inline void __update_tv_cache(struct thread_data *td) +static inline void __update_ts_cache(struct thread_data *td) { - fio_gettime(&td->tv_cache, NULL); + fio_gettime(&td->ts_cache, NULL); } -static inline void update_tv_cache(struct thread_data *td) +static inline void update_ts_cache(struct thread_data *td) { - if ((++td->tv_cache_nr & td->tv_cache_mask) == td->tv_cache_mask) - __update_tv_cache(td); + if ((++td->ts_cache_nr & td->ts_cache_mask) == td->ts_cache_mask) + __update_ts_cache(td); } -static inline bool runtime_exceeded(struct thread_data *td, struct timeval *t) +static inline bool runtime_exceeded(struct thread_data *td, struct timespec *t) { if (in_ramp_time(td)) return false; @@ -430,7 +430,7 @@ static void check_update_rusage(struct thread_data *td) } } -static int wait_for_completions(struct thread_data *td, struct timeval *time) +static int wait_for_completions(struct thread_data *td, struct timespec *time) { const int full = queue_full(td); int min_evts = 0; @@ -462,7 +462,7 @@ static int wait_for_completions(struct thread_data *td, struct timeval *time) int io_queue_event(struct thread_data *td, struct io_u *io_u, int *ret, enum fio_ddir ddir, uint64_t *bytes_issued, int from_verify, - struct timeval *comp_time) + struct timespec *comp_time) { int ret2; @@ -499,7 +499,6 @@ int io_queue_event(struct thread_data *td, struct io_u *io_u, int *ret, if (ddir_rw(io_u->ddir)) td->ts.short_io_u[io_u->ddir]++; - f = io_u->file; if (io_u->offset == f->real_file_size) goto sync_done; @@ -586,6 +585,50 @@ static int unlink_all_files(struct thread_data *td) return ret; } +/* + * Check if io_u will overlap an in-flight IO in the queue + */ +static bool in_flight_overlap(struct io_u_queue *q, struct io_u *io_u) +{ + bool overlap; + struct io_u *check_io_u; + unsigned long long x1, x2, y1, y2; + int i; + + x1 = io_u->offset; + x2 = io_u->offset + io_u->buflen; + overlap = false; + io_u_qiter(q, check_io_u, i) { + if (check_io_u->flags & IO_U_F_FLIGHT) { + y1 = check_io_u->offset; + y2 = check_io_u->offset + check_io_u->buflen; + + if (x1 < y2 && y1 < x2) { + overlap = true; + dprint(FD_IO, "in-flight overlap: %llu/%lu, %llu/%lu\n", + x1, io_u->buflen, + y1, check_io_u->buflen); + break; + } + } + } + + return overlap; +} + +static int io_u_submit(struct thread_data *td, struct io_u *io_u) +{ + /* + * Check for overlap if the user asked us to, and we have + * at least one IO in flight besides this one. + */ + if (td->o.serialize_overlap && td->cur_depth > 1 && + in_flight_overlap(&td->io_u_all, io_u)) + return FIO_Q_BUSY; + + return td_io_queue(td, io_u); +} + /* * The main verify engine. Runs over the writes we previously submitted, * reads the blocks back in, and checks the crc/md5 of the data. @@ -633,12 +676,12 @@ static void do_verify(struct thread_data *td, uint64_t verify_bytes) enum fio_ddir ddir; int full; - update_tv_cache(td); + update_ts_cache(td); check_update_rusage(td); - if (runtime_exceeded(td, &td->tv_cache)) { - __update_tv_cache(td); - if (runtime_exceeded(td, &td->tv_cache)) { + if (runtime_exceeded(td, &td->ts_cache)) { + __update_ts_cache(td); + if (runtime_exceeded(td, &td->ts_cache)) { fio_mark_td_terminate(td); break; } @@ -716,7 +759,7 @@ static void do_verify(struct thread_data *td, uint64_t verify_bytes) if (!td->o.disable_slat) fio_gettime(&io_u->start_time, NULL); - ret = td_io_queue(td, io_u); + ret = io_u_submit(td, io_u); if (io_queue_event(td, io_u, &ret, ddir, NULL, 1, NULL)) break; @@ -813,8 +856,9 @@ static long long usec_for_io(struct thread_data *td, enum fio_ddir ddir) val = (int64_t) (1000000 / iops) * -logf(__rand_0_1(&td->poisson_state[ddir])); if (val) { - dprint(FD_RATE, "poisson rate iops=%llu\n", - (unsigned long long) 1000000 / val); + dprint(FD_RATE, "poisson rate iops=%llu, ddir=%d\n", + (unsigned long long) 1000000 / val, + ddir); } td->last_usec[ddir] += val; return td->last_usec[ddir]; @@ -873,7 +917,7 @@ static void do_io(struct thread_data *td, uint64_t *bytes_done) while ((td->o.read_iolog_file && !flist_empty(&td->io_log_list)) || (!flist_empty(&td->trim_list)) || !io_issue_bytes_exceeded(td) || td->o.time_based) { - struct timeval comp_time; + struct timespec comp_time; struct io_u *io_u; int full; enum fio_ddir ddir; @@ -883,11 +927,11 @@ static void do_io(struct thread_data *td, uint64_t *bytes_done) if (td->terminate || td->done) break; - update_tv_cache(td); + update_ts_cache(td); - if (runtime_exceeded(td, &td->tv_cache)) { - __update_tv_cache(td); - if (runtime_exceeded(td, &td->tv_cache)) { + if (runtime_exceeded(td, &td->ts_cache)) { + __update_ts_cache(td); + if (runtime_exceeded(td, &td->ts_cache)) { fio_mark_td_terminate(td); break; } @@ -982,7 +1026,7 @@ static void do_io(struct thread_data *td, uint64_t *bytes_done) td->rate_next_io_time[ddir] = usec_for_io(td, ddir); } else { - ret = td_io_queue(td, io_u); + ret = io_u_submit(td, io_u); if (should_check_rate(td)) td->rate_next_io_time[ddir] = usec_for_io(td, ddir); @@ -1346,6 +1390,8 @@ static bool keep_running(struct thread_data *td) if (td->done) return false; + if (td->terminate) + return false; if (td->o.time_based) return true; if (td->o.loops) { @@ -1456,8 +1502,9 @@ static void *thread_main(void *data) struct thread_data *td = fd->td; struct thread_options *o = &td->o; struct sk_out *sk_out = fd->sk_out; + uint64_t bytes_done[DDIR_RWDIR_CNT]; int deadlock_loop_cnt; - int clear_state; + bool clear_state, did_some_io; int ret; sk_out_assign(sk_out); @@ -1645,16 +1692,14 @@ static void *thread_main(void *data) if (td_io_init(td)) goto err; - if (init_random_map(td)) + if (!init_random_map(td)) goto err; if (o->exec_prerun && exec_string(o, o->exec_prerun, (const char *)"prerun")) goto err; - if (o->pre_read) { - if (pre_read_files(td) < 0) - goto err; - } + if (o->pre_read && !pre_read_files(td)) + goto err; fio_verify_init(td); @@ -1677,12 +1722,15 @@ static void *thread_main(void *data) sizeof(td->bw_sample_time)); } - clear_state = 0; + memset(bytes_done, 0, sizeof(bytes_done)); + clear_state = false; + did_some_io = false; + while (keep_running(td)) { uint64_t verify_bytes; fio_gettime(&td->start, NULL); - memcpy(&td->tv_cache, &td->start, sizeof(td->start)); + memcpy(&td->ts_cache, &td->start, sizeof(td->start)); if (clear_state) { clear_io_state(td, 0); @@ -1696,8 +1744,6 @@ static void *thread_main(void *data) if (td->o.verify_only && td_write(td)) verify_bytes = do_dry_run(td); else { - uint64_t bytes_done[DDIR_RWDIR_CNT]; - do_io(td, bytes_done); if (!ddir_rw_sum(bytes_done)) { @@ -1717,7 +1763,7 @@ static void *thread_main(void *data) if (td->runstate >= TD_EXITED) break; - clear_state = 1; + clear_state = true; /* * Make sure we've successfully updated the rusage stats @@ -1756,6 +1802,9 @@ static void *thread_main(void *data) td_ioengine_flagged(td, FIO_UNIDIR)) continue; + if (ddir_rw_sum(bytes_done)) + did_some_io = true; + clear_io_state(td, 0); fio_gettime(&td->start, NULL); @@ -1776,6 +1825,19 @@ static void *thread_main(void *data) break; } + /* + * If td ended up with no I/O when it should have had, + * then something went wrong unless FIO_NOIO or FIO_DISKLESSIO. + * (Are we not missing other flags that can be ignored ?) + */ + if ((td->o.size || td->o.io_size) && !ddir_rw_sum(bytes_done) && + !did_some_io && !td->o.create_only && + !(td_ioengine_flagged(td, FIO_NOIO) || + td_ioengine_flagged(td, FIO_DISKLESSIO))) + log_err("%s: No I/O performed by %s, " + "perhaps try --debug=io option for details?\n", + td->o.name, td->io_ops->name); + td_set_runstate(td, TD_FINISHING); update_rusage_stat(td); @@ -1848,14 +1910,6 @@ err: return (void *) (uintptr_t) td->error; } -static void dump_td_info(struct thread_data *td) -{ - log_err("fio: job '%s' (state=%d) hasn't exited in %lu seconds, it " - "appears to be stuck. Doing forceful exit of this job.\n", - td->o.name, td->runstate, - (unsigned long) time_since_now(&td->terminate_time)); -} - /* * Run over the job map and reap the threads that have exited, if any. */ @@ -1873,11 +1927,7 @@ static void reap_threads(unsigned int *nr_running, uint64_t *t_rate, for_each_td(td, i) { int flags = 0; - /* - * ->io_ops is NULL for a thread that has closed its - * io engine - */ - if (td->io_ops && !strcmp(td->io_ops->name, "cpuio")) + if (!strcmp(td->o.ioengine, "cpuio")) cputhreads++; else realthreads++; @@ -1940,7 +1990,11 @@ static void reap_threads(unsigned int *nr_running, uint64_t *t_rate, if (td->terminate && td->runstate < TD_FSYNCING && time_since_now(&td->terminate_time) >= FIO_REAP_TIMEOUT) { - dump_td_info(td); + log_err("fio: job '%s' (state=%d) hasn't exited in " + "%lu seconds, it appears to be stuck. Doing " + "forceful exit of this job.\n", + td->o.name, td->runstate, + (unsigned long) time_since_now(&td->terminate_time)); td_set_runstate(td, TD_REAPED); goto reaped; } @@ -1988,7 +2042,10 @@ static bool __check_trigger_file(void) static bool trigger_timedout(void) { if (trigger_timeout) - return time_since_genesis() >= trigger_timeout; + if (time_since_genesis() >= trigger_timeout) { + trigger_timeout = 0; + return true; + } return false; } @@ -1997,7 +2054,7 @@ void exec_trigger(const char *cmd) { int ret; - if (!cmd) + if (!cmd || cmd[0] == '\0') return; ret = system(cmd); @@ -2192,7 +2249,7 @@ reap: while (todo) { struct thread_data *map[REAL_MAX_JOBS]; - struct timeval this_start; + struct timespec this_start; int this_jobs = 0, left; struct fork_data *fd; @@ -2283,6 +2340,7 @@ reap: fio_terminate_threads(TERMINATE_ALL); fio_abort = 1; nr_started--; + free(fd); break; } dprint(FD_MUTEX, "done waiting on startup_mutex\n");