X-Git-Url: https://git.kernel.dk/?p=fio.git;a=blobdiff_plain;f=diskutil.c;h=ace7af3d5b5bb7368584a1369a0872cbe79c92ee;hp=52d87f6433d85132b2ae2fd3e6bfdc127fc64393;hb=HEAD;hpb=67a176fc7fa918cc91880530fac07cc3aa864c71 diff --git a/diskutil.c b/diskutil.c index 52d87f64..69b3dd26 100644 --- a/diskutil.c +++ b/diskutil.c @@ -1,23 +1,26 @@ +#include #include #include -#include #include #include +#include #include #include -#include -#include +#ifdef CONFIG_VALGRIND_DEV +#include +#else +#define DRD_IGNORE_VAR(x) do { } while (0) +#endif #include "fio.h" #include "smalloc.h" #include "diskutil.h" +#include "helper_thread.h" static int last_majdev, last_mindev; static struct disk_util *last_du; -static struct fio_mutex *disk_util_mutex; - -FLIST_HEAD(disk_list); +static struct fio_sem *disk_util_sem; static struct disk_util *__init_per_file_disk_util(struct thread_data *td, int majdev, int mindev, char *path); @@ -35,14 +38,13 @@ static void disk_util_free(struct disk_util *du) slave->users--; } - fio_mutex_remove(du->lock); + fio_sem_remove(du->lock); + free(du->sysfs_root); sfree(du); } static int get_io_ticks(struct disk_util *du, struct disk_util_stat *dus) { - unsigned in_flight; - unsigned long long sectors[2]; char line[256]; FILE *f; char *p; @@ -62,29 +64,40 @@ static int get_io_ticks(struct disk_util *du, struct disk_util_stat *dus) dprint(FD_DISKUTIL, "%s: %s", du->path, p); - ret = sscanf(p, "%llu %llu %llu %llu %llu %llu %llu %llu %u %llu %llu\n", - (unsigned long long *) &dus->s.ios[0], - (unsigned long long *) &dus->s.merges[0], - §ors[0], - (unsigned long long *) &dus->s.ticks[0], - (unsigned long long *) &dus->s.ios[1], - (unsigned long long *) &dus->s.merges[1], - §ors[1], - (unsigned long long *) &dus->s.ticks[1], - &in_flight, - (unsigned long long *) &dus->s.io_ticks, - (unsigned long long *) &dus->s.time_in_queue); + ret = sscanf(p, "%"SCNu64" %"SCNu64" %"SCNu64" %"SCNu64" " + "%"SCNu64" %"SCNu64" %"SCNu64" %"SCNu64" " + "%*u %"SCNu64" %"SCNu64"\n", + &dus->s.ios[0], &dus->s.merges[0], &dus->s.sectors[0], + &dus->s.ticks[0], + &dus->s.ios[1], &dus->s.merges[1], &dus->s.sectors[1], + &dus->s.ticks[1], + &dus->s.io_ticks, &dus->s.time_in_queue); fclose(f); - dprint(FD_DISKUTIL, "%s: stat read ok? %d\n", du->path, ret == 1); - dus->s.sectors[0] = sectors[0]; - dus->s.sectors[1] = sectors[1]; - return ret != 11; + dprint(FD_DISKUTIL, "%s: stat read ok? %d\n", du->path, ret == 10); + return ret != 10; +} + +static uint64_t safe_32bit_diff(uint64_t nval, uint64_t oval) +{ + /* Linux kernel prints some of the stat fields as 32-bit integers. It is + * possible that the value overflows, but since fio uses unsigned 64-bit + * arithmetic in update_io_tick_disk(), it instead results in a huge + * bogus value being added to the respective accumulating field. Just + * in case Linux starts reporting these metrics as 64-bit values in the + * future, check that overflow actually happens around the 32-bit + * unsigned boundary; assume overflow only happens once between + * successive polls. + */ + if (oval <= nval || oval >= (1ull << 32)) + return nval - oval; + else + return (1ull << 32) + nval - oval; } static void update_io_tick_disk(struct disk_util *du) { struct disk_util_stat __dus, *dus, *ldus; - struct timeval t; + struct timespec t; if (!du->users) return; @@ -100,15 +113,16 @@ static void update_io_tick_disk(struct disk_util *du) dus->s.ios[1] += (__dus.s.ios[1] - ldus->s.ios[1]); dus->s.merges[0] += (__dus.s.merges[0] - ldus->s.merges[0]); dus->s.merges[1] += (__dus.s.merges[1] - ldus->s.merges[1]); - dus->s.ticks[0] += (__dus.s.ticks[0] - ldus->s.ticks[0]); - dus->s.ticks[1] += (__dus.s.ticks[1] - ldus->s.ticks[1]); - dus->s.io_ticks += (__dus.s.io_ticks - ldus->s.io_ticks); - dus->s.time_in_queue += (__dus.s.time_in_queue - ldus->s.time_in_queue); + dus->s.ticks[0] += safe_32bit_diff(__dus.s.ticks[0], ldus->s.ticks[0]); + dus->s.ticks[1] += safe_32bit_diff(__dus.s.ticks[1], ldus->s.ticks[1]); + dus->s.io_ticks += safe_32bit_diff(__dus.s.io_ticks, ldus->s.io_ticks); + dus->s.time_in_queue += + safe_32bit_diff(__dus.s.time_in_queue, ldus->s.time_in_queue); fio_gettime(&t, NULL); dus->s.msec += mtime_since(&du->time, &t); - memcpy(&du->time, &t, sizeof(t)); - memcpy(&ldus->s, &__dus.s, sizeof(__dus.s)); + du->time = t; + ldus->s = __dus.s; } int update_io_ticks(void) @@ -119,9 +133,9 @@ int update_io_ticks(void) dprint(FD_DISKUTIL, "update io ticks\n"); - fio_mutex_down(disk_util_mutex); + fio_sem_down(disk_util_sem); - if (!helper_exit) { + if (!helper_should_exit()) { flist_for_each(entry, &disk_list) { du = flist_entry(entry, struct disk_util, list); update_io_tick_disk(du); @@ -129,7 +143,7 @@ int update_io_ticks(void) } else ret = 1; - fio_mutex_up(disk_util_mutex); + fio_sem_up(disk_util_sem); return ret; } @@ -138,18 +152,18 @@ static struct disk_util *disk_util_exists(int major, int minor) struct flist_head *entry; struct disk_util *du; - fio_mutex_down(disk_util_mutex); + fio_sem_down(disk_util_sem); flist_for_each(entry, &disk_list) { du = flist_entry(entry, struct disk_util, list); if (major == du->major && minor == du->minor) { - fio_mutex_up(disk_util_mutex); + fio_sem_up(disk_util_sem); return du; } } - fio_mutex_up(disk_util_mutex); + fio_sem_up(disk_util_sem); return NULL; } @@ -163,14 +177,10 @@ static int get_device_numbers(char *file_name, int *maj, int *min) if (S_ISBLK(st.st_mode)) { majdev = major(st.st_rdev); mindev = minor(st.st_rdev); - } else if (S_ISCHR(st.st_mode)) { - majdev = major(st.st_rdev); - mindev = minor(st.st_rdev); - if (fio_lookup_raw(st.st_rdev, &majdev, &mindev)) - return -1; - } else if (S_ISFIFO(st.st_mode)) + } else if (S_ISCHR(st.st_mode) || + S_ISFIFO(st.st_mode)) { return -1; - else { + } else { majdev = major(st.st_dev); mindev = minor(st.st_dev); } @@ -178,7 +188,7 @@ static int get_device_numbers(char *file_name, int *maj, int *min) /* * must be a file, open "." in that path */ - strncpy(tempname, file_name, PATH_MAX - 1); + snprintf(tempname, FIO_ARRAY_SIZE(tempname), "%s", file_name); p = dirname(tempname); if (stat(p, &st)) { perror("disk util stat"); @@ -238,22 +248,28 @@ static void find_add_disk_slaves(struct thread_data *td, char *path, !strcmp(dirent->d_name, "..")) continue; - sprintf(temppath, "%s%s%s", slavesdir, FIO_OS_PATH_SEPARATOR, dirent->d_name); + nowarn_snprintf(temppath, sizeof(temppath), "%s/%s", slavesdir, + dirent->d_name); /* Can we always assume that the slaves device entries * are links to the real directories for the slave * devices? */ linklen = readlink(temppath, slavepath, PATH_MAX - 1); - if (linklen < 0) { + if (linklen < 0) { perror("readlink() for slave device."); closedir(dirhandle); return; } slavepath[linklen] = '\0'; - sprintf(temppath, "%s/%s/dev", slavesdir, slavepath); + nowarn_snprintf(temppath, sizeof(temppath), "%s/%s/dev", + slavesdir, slavepath); + if (access(temppath, F_OK) != 0) + nowarn_snprintf(temppath, sizeof(temppath), + "%s/%s/device/dev", slavesdir, + slavepath); if (read_block_dev_entry(temppath, &majdev, &mindev)) { - perror("Error getting slave device numbers."); + perror("Error getting slave device numbers"); closedir(dirhandle); return; } @@ -265,7 +281,8 @@ static void find_add_disk_slaves(struct thread_data *td, char *path, if (slavedu) continue; - sprintf(temppath, "%s%s%s", slavesdir, FIO_OS_PATH_SEPARATOR, slavepath); + nowarn_snprintf(temppath, sizeof(temppath), "%s/%s", slavesdir, + slavepath); __init_per_file_disk_util(td, majdev, mindev, temppath); slavedu = disk_util_exists(majdev, mindev); @@ -290,11 +307,10 @@ static struct disk_util *disk_util_add(struct thread_data *td, int majdev, dprint(FD_DISKUTIL, "add maj/min %d/%d: %s\n", majdev, mindev, path); du = smalloc(sizeof(*du)); - if (!du) { - log_err("fio: smalloc() pool exhausted\n"); + if (!du) return NULL; - } + DRD_IGNORE_VAR(du->users); memset(du, 0, sizeof(*du)); INIT_FLIST_HEAD(&du->list); l = snprintf(du->path, sizeof(du->path), "%s/stat", path); @@ -304,16 +320,17 @@ static struct disk_util *disk_util_add(struct thread_data *td, int majdev, sfree(du); return NULL; } - strncpy((char *) du->dus.name, basename(path), FIO_DU_NAME_SZ - 1); - du->sysfs_root = path; + snprintf((char *) du->dus.name, FIO_ARRAY_SIZE(du->dus.name), "%s", + basename(path)); + du->sysfs_root = strdup(path); du->major = majdev; du->minor = mindev; INIT_FLIST_HEAD(&du->slavelist); INIT_FLIST_HEAD(&du->slaves); - du->lock = fio_mutex_init(FIO_MUTEX_UNLOCKED); + du->lock = fio_sem_init(FIO_SEM_UNLOCKED); du->users = 0; - fio_mutex_down(disk_util_mutex); + fio_sem_down(disk_util_sem); flist_for_each(entry, &disk_list) { __du = flist_entry(entry, struct disk_util, list); @@ -322,7 +339,7 @@ static struct disk_util *disk_util_add(struct thread_data *td, int majdev, if (!strcmp((char *) du->dus.name, (char *) __du->dus.name)) { disk_util_free(du); - fio_mutex_up(disk_util_mutex); + fio_sem_up(disk_util_sem); return __du; } } @@ -333,7 +350,7 @@ static struct disk_util *disk_util_add(struct thread_data *td, int majdev, get_io_ticks(du, &du->last_dus); flist_add_tail(&du->list, &disk_list); - fio_mutex_up(disk_util_mutex); + fio_sem_up(disk_util_sem); find_add_disk_slaves(td, path, du); return du; @@ -364,12 +381,12 @@ static int find_block_dir(int majdev, int mindev, char *path, int link_ok) return 0; while ((dir = readdir(D)) != NULL) { - char full_path[256]; + char full_path[257]; if (!strcmp(dir->d_name, ".") || !strcmp(dir->d_name, "..")) continue; - sprintf(full_path, "%s%s%s", path, FIO_OS_PATH_SEPARATOR, dir->d_name); + sprintf(full_path, "%s/%s", path, dir->d_name); if (!strcmp(dir->d_name, "dev")) { if (!check_dev_match(majdev, mindev, full_path)) { @@ -425,13 +442,10 @@ static struct disk_util *__init_per_file_disk_util(struct thread_data *td, log_err("unknown sysfs layout\n"); return NULL; } - strncpy(tmp, p, PATH_MAX - 1); + snprintf(tmp, FIO_ARRAY_SIZE(tmp), "%s", p); sprintf(path, "%s", tmp); } - if (td->o.ioscheduler && !td->sysfs_root) - td->sysfs_root = strdup(path); - return disk_util_add(td, majdev, mindev, path); } @@ -450,12 +464,8 @@ static struct disk_util *init_per_file_disk_util(struct thread_data *td, mindev); du = disk_util_exists(majdev, mindev); - if (du) { - if (td->o.ioscheduler && !td->sysfs_root) - td->sysfs_root = strdup(du->sysfs_root); - + if (du) return du; - } /* * for an fs without a device, we will repeatedly stat through @@ -488,81 +498,16 @@ void init_disk_util(struct thread_data *td) unsigned int i; if (!td->o.do_disk_util || - (td->io_ops->flags & (FIO_DISKLESSIO | FIO_NODISKUTIL))) + td_ioengine_flagged(td, FIO_DISKLESSIO | FIO_NODISKUTIL)) return; for_each_file(td, f, i) f->du = __init_disk_util(td, f); } -static void show_agg_stats(struct disk_util_agg *agg, int terse) -{ - if (!agg->slavecount) - return; - - if (!terse) { - log_info(", aggrios=%llu/%llu, aggrmerge=%llu/%llu, " - "aggrticks=%llu/%llu, aggrin_queue=%llu, " - "aggrutil=%3.2f%%", - (unsigned long long) agg->ios[0] / agg->slavecount, - (unsigned long long) agg->ios[1] / agg->slavecount, - (unsigned long long) agg->merges[0] / agg->slavecount, - (unsigned long long) agg->merges[1] / agg->slavecount, - (unsigned long long) agg->ticks[0] / agg->slavecount, - (unsigned long long) agg->ticks[1] / agg->slavecount, - (unsigned long long) agg->time_in_queue / agg->slavecount, - agg->max_util.u.f); - } else { - log_info(";slaves;%llu;%llu;%llu;%llu;%llu;%llu;%llu;%3.2f%%", - (unsigned long long) agg->ios[0] / agg->slavecount, - (unsigned long long) agg->ios[1] / agg->slavecount, - (unsigned long long) agg->merges[0] / agg->slavecount, - (unsigned long long) agg->merges[1] / agg->slavecount, - (unsigned long long) agg->ticks[0] / agg->slavecount, - (unsigned long long) agg->ticks[1] / agg->slavecount, - (unsigned long long) agg->time_in_queue / agg->slavecount, - agg->max_util.u.f); - } -} - -static void aggregate_slaves_stats(struct disk_util *masterdu) -{ - struct disk_util_agg *agg = &masterdu->agg; - struct disk_util_stat *dus; - struct flist_head *entry; - struct disk_util *slavedu; - double util; - - flist_for_each(entry, &masterdu->slaves) { - slavedu = flist_entry(entry, struct disk_util, slavelist); - dus = &slavedu->dus; - agg->ios[0] += dus->s.ios[0]; - agg->ios[1] += dus->s.ios[1]; - agg->merges[0] += dus->s.merges[0]; - agg->merges[1] += dus->s.merges[1]; - agg->sectors[0] += dus->s.sectors[0]; - agg->sectors[1] += dus->s.sectors[1]; - agg->ticks[0] += dus->s.ticks[0]; - agg->ticks[1] += dus->s.ticks[1]; - agg->time_in_queue += dus->s.time_in_queue; - agg->slavecount++; - - util = (double) (100 * dus->s.io_ticks / (double) slavedu->dus.s.msec); - /* System utilization is the utilization of the - * component with the highest utilization. - */ - if (util > agg->max_util.u.f) - agg->max_util.u.f = util; - - } - - if (agg->max_util.u.f > 100.0) - agg->max_util.u.f = 100.0; -} - void disk_util_prune_entries(void) { - fio_mutex_down(disk_util_mutex); + fio_sem_down(disk_util_sem); while (!flist_empty(&disk_list)) { struct disk_util *du; @@ -573,157 +518,11 @@ void disk_util_prune_entries(void) } last_majdev = last_mindev = -1; - fio_mutex_up(disk_util_mutex); - fio_mutex_remove(disk_util_mutex); -} - -void print_disk_util(struct disk_util_stat *dus, struct disk_util_agg *agg, - int terse) -{ - double util = 0; - - if (dus->s.msec) - util = (double) 100 * dus->s.io_ticks / (double) dus->s.msec; - if (util > 100.0) - util = 100.0; - - if (!terse) { - if (agg->slavecount) - log_info(" "); - - log_info(" %s: ios=%llu/%llu, merge=%llu/%llu, " - "ticks=%llu/%llu, in_queue=%llu, util=%3.2f%%", - dus->name, - (unsigned long long) dus->s.ios[0], - (unsigned long long) dus->s.ios[1], - (unsigned long long) dus->s.merges[0], - (unsigned long long) dus->s.merges[1], - (unsigned long long) dus->s.ticks[0], - (unsigned long long) dus->s.ticks[1], - (unsigned long long) dus->s.time_in_queue, - util); - } else { - log_info(";%s;%llu;%llu;%llu;%llu;%llu;%llu;%llu;%3.2f%%", - dus->name, - (unsigned long long) dus->s.ios[0], - (unsigned long long) dus->s.ios[1], - (unsigned long long) dus->s.merges[0], - (unsigned long long) dus->s.merges[1], - (unsigned long long) dus->s.ticks[0], - (unsigned long long) dus->s.ticks[1], - (unsigned long long) dus->s.time_in_queue, - util); - } - - /* - * If the device has slaves, aggregate the stats for - * those slave devices also. - */ - show_agg_stats(agg, terse); - - if (!terse) - log_info("\n"); -} - -void json_array_add_disk_util(struct disk_util_stat *dus, - struct disk_util_agg *agg, struct json_array *array) -{ - struct json_object *obj; - double util = 0; - - if (dus->s.msec) - util = (double) 100 * dus->s.io_ticks / (double) dus->s.msec; - if (util > 100.0) - util = 100.0; - - obj = json_create_object(); - json_array_add_value_object(array, obj); - - json_object_add_value_string(obj, "name", dus->name); - json_object_add_value_int(obj, "read_ios", dus->s.ios[0]); - json_object_add_value_int(obj, "write_ios", dus->s.ios[1]); - json_object_add_value_int(obj, "read_merges", dus->s.merges[0]); - json_object_add_value_int(obj, "write_merges", dus->s.merges[1]); - json_object_add_value_int(obj, "read_ticks", dus->s.ticks[0]); - json_object_add_value_int(obj, "write_ticks", dus->s.ticks[1]); - json_object_add_value_int(obj, "in_queue", dus->s.time_in_queue); - json_object_add_value_float(obj, "util", util); - - /* - * If the device has slaves, aggregate the stats for - * those slave devices also. - */ - if (!agg->slavecount) - return; - json_object_add_value_int(obj, "aggr_read_ios", - agg->ios[0] / agg->slavecount); - json_object_add_value_int(obj, "aggr_write_ios", - agg->ios[1] / agg->slavecount); - json_object_add_value_int(obj, "aggr_read_merges", - agg->merges[0] / agg->slavecount); - json_object_add_value_int(obj, "aggr_write_merge", - agg->merges[1] / agg->slavecount); - json_object_add_value_int(obj, "aggr_read_ticks", - agg->ticks[0] / agg->slavecount); - json_object_add_value_int(obj, "aggr_write_ticks", - agg->ticks[1] / agg->slavecount); - json_object_add_value_int(obj, "aggr_in_queue", - agg->time_in_queue / agg->slavecount); - json_object_add_value_float(obj, "aggr_util", agg->max_util.u.f); -} - -static void json_object_add_disk_utils(struct json_object *obj, - struct flist_head *head) -{ - struct json_array *array = json_create_array(); - struct flist_head *entry; - struct disk_util *du; - - json_object_add_value_array(obj, "disk_util", array); - - flist_for_each(entry, head) { - du = flist_entry(entry, struct disk_util, list); - - aggregate_slaves_stats(du); - json_array_add_disk_util(&du->dus, &du->agg, array); - } -} - -void show_disk_util(int terse, struct json_object *parent) -{ - struct flist_head *entry; - struct disk_util *du; - - if (!disk_util_mutex) - return; - - fio_mutex_down(disk_util_mutex); - - if (flist_empty(&disk_list)) { - fio_mutex_up(disk_util_mutex); - return; - } - - if (output_format == FIO_OUTPUT_JSON) - assert(parent); - - if (!terse && output_format != FIO_OUTPUT_JSON) - log_info("\nDisk stats (read/write):\n"); - - if (output_format == FIO_OUTPUT_JSON) - json_object_add_disk_utils(parent, &disk_list); - else - flist_for_each(entry, &disk_list) { - du = flist_entry(entry, struct disk_util, list); - - aggregate_slaves_stats(du); - print_disk_util(&du->dus, &du->agg, terse); - } - - fio_mutex_up(disk_util_mutex); + fio_sem_up(disk_util_sem); + fio_sem_remove(disk_util_sem); } void setup_disk_util(void) { - disk_util_mutex = fio_mutex_init(FIO_MUTEX_UNLOCKED); + disk_util_sem = fio_sem_init(FIO_SEM_UNLOCKED); }