#include <sys/time.h>
#include <sys/types.h>
#include <sys/stat.h>
+#include <sys/sysmacros.h>
#include <dirent.h>
#include <libgen.h>
#include <math.h>
+#include <assert.h>
#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 flist_head disk_list = FLIST_HEAD_INIT(disk_list);
+static struct fio_mutex *disk_util_mutex;
static struct disk_util *__init_per_file_disk_util(struct thread_data *td,
int majdev, int mindev, char *path);
if (du == last_du)
last_du = NULL;
+ while (!flist_empty(&du->slaves)) {
+ struct disk_util *slave;
+
+ slave = flist_first_entry(&du->slaves, struct disk_util, slavelist);
+ flist_del(&slave->slavelist);
+ slave->users--;
+ }
+
fio_mutex_remove(du->lock);
- sfree(du->name);
+ 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;
dprint(FD_DISKUTIL, "%s: %s", du->path, p);
- ret = sscanf(p, "%u %u %llu %u %u %u %llu %u %u %u %u\n", &dus->ios[0],
- &dus->merges[0], &dus->sectors[0],
- &dus->ticks[0], &dus->ios[1],
- &dus->merges[1], &dus->sectors[1],
- &dus->ticks[1], &in_flight,
- &dus->io_ticks, &dus->time_in_queue);
+ 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);
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;
}
static void update_io_tick_disk(struct disk_util *du)
{
struct disk_util_stat __dus, *dus, *ldus;
- struct timeval t;
+ struct timespec t;
- if (get_io_ticks(du, &__dus))
- return;
if (!du->users)
return;
+ if (get_io_ticks(du, &__dus))
+ return;
dus = &du->dus;
ldus = &du->last_dus;
- dus->sectors[0] += (__dus.sectors[0] - ldus->sectors[0]);
- dus->sectors[1] += (__dus.sectors[1] - ldus->sectors[1]);
- dus->ios[0] += (__dus.ios[0] - ldus->ios[0]);
- dus->ios[1] += (__dus.ios[1] - ldus->ios[1]);
- dus->merges[0] += (__dus.merges[0] - ldus->merges[0]);
- dus->merges[1] += (__dus.merges[1] - ldus->merges[1]);
- dus->ticks[0] += (__dus.ticks[0] - ldus->ticks[0]);
- dus->ticks[1] += (__dus.ticks[1] - ldus->ticks[1]);
- dus->io_ticks += (__dus.io_ticks - ldus->io_ticks);
- dus->time_in_queue += (__dus.time_in_queue - ldus->time_in_queue);
+ dus->s.sectors[0] += (__dus.s.sectors[0] - ldus->s.sectors[0]);
+ dus->s.sectors[1] += (__dus.s.sectors[1] - ldus->s.sectors[1]);
+ dus->s.ios[0] += (__dus.s.ios[0] - ldus->s.ios[0]);
+ 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);
fio_gettime(&t, NULL);
- du->msec += mtime_since(&du->time, &t);
+ dus->s.msec += mtime_since(&du->time, &t);
memcpy(&du->time, &t, sizeof(t));
- memcpy(ldus, &__dus, sizeof(__dus));
+ memcpy(&ldus->s, &__dus.s, sizeof(__dus.s));
}
-void update_io_ticks(void)
+int update_io_ticks(void)
{
struct flist_head *entry;
struct disk_util *du;
+ int ret = 0;
dprint(FD_DISKUTIL, "update io ticks\n");
- flist_for_each(entry, &disk_list) {
- du = flist_entry(entry, struct disk_util, list);
- update_io_tick_disk(du);
- }
+ fio_mutex_down(disk_util_mutex);
+
+ if (!helper_should_exit()) {
+ flist_for_each(entry, &disk_list) {
+ du = flist_entry(entry, struct disk_util, list);
+ update_io_tick_disk(du);
+ }
+ } else
+ ret = 1;
+
+ fio_mutex_up(disk_util_mutex);
+ return ret;
}
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);
+
flist_for_each(entry, &disk_list) {
du = flist_entry(entry, struct disk_util, list);
- if (major == du->major && minor == du->minor)
+ if (major == du->major && minor == du->minor) {
+ fio_mutex_up(disk_util_mutex);
return du;
+ }
}
+ fio_mutex_up(disk_util_mutex);
return NULL;
}
/*
* must be a file, open "." in that path
*/
+ tempname[PATH_MAX - 1] = '\0';
strncpy(tempname, file_name, PATH_MAX - 1);
p = dirname(tempname);
if (stat(p, &st)) {
* are links to the real directories for the slave
* devices?
*/
- linklen = readlink(temppath, slavepath, PATH_MAX - 0);
+ linklen = readlink(temppath, slavepath, PATH_MAX - 1);
if (linklen < 0) {
perror("readlink() for slave device.");
+ closedir(dirhandle);
return;
}
slavepath[linklen] = '\0';
sprintf(temppath, "%s/%s/dev", slavesdir, slavepath);
if (read_block_dev_entry(temppath, &majdev, &mindev)) {
perror("Error getting slave device numbers.");
+ closedir(dirhandle);
return;
}
/* Should probably use an assert here. slavedu should
* always be present at this point. */
- if (slavedu)
+ if (slavedu) {
+ slavedu->users++;
flist_add_tail(&slavedu->slavelist, &masterdu->slaves);
+ }
}
closedir(dirhandle);
}
-static struct disk_util *disk_util_add(struct thread_data * td, int majdev,
+static struct disk_util *disk_util_add(struct thread_data *td, int majdev,
int mindev, char *path)
{
struct disk_util *du, *__du;
struct flist_head *entry;
+ int l;
dprint(FD_DISKUTIL, "add maj/min %d/%d: %s\n", majdev, mindev, path);
du = smalloc(sizeof(*du));
+ if (!du)
+ return NULL;
+
memset(du, 0, sizeof(*du));
INIT_FLIST_HEAD(&du->list);
- sprintf(du->path, "%s/stat", path);
- du->name = smalloc_strdup(basename(path));
- du->sysfs_root = path;
+ l = snprintf(du->path, sizeof(du->path), "%s/stat", path);
+ if (l < 0 || l >= sizeof(du->path)) {
+ log_err("constructed path \"%.100s[...]/stat\" larger than buffer (%zu bytes)\n",
+ path, sizeof(du->path) - 1);
+ sfree(du);
+ return NULL;
+ }
+ strncpy((char *) du->dus.name, basename(path), FIO_DU_NAME_SZ - 1);
+ 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(1);
+ du->lock = fio_mutex_init(FIO_MUTEX_UNLOCKED);
du->users = 0;
+ fio_mutex_down(disk_util_mutex);
+
flist_for_each(entry, &disk_list) {
__du = flist_entry(entry, struct disk_util, list);
- dprint(FD_DISKUTIL, "found %s in list\n", __du->name);
+ dprint(FD_DISKUTIL, "found %s in list\n", __du->dus.name);
- if (!strcmp(du->name, __du->name)) {
+ if (!strcmp((char *) du->dus.name, (char *) __du->dus.name)) {
disk_util_free(du);
+ fio_mutex_up(disk_util_mutex);
return __du;
}
}
- dprint(FD_DISKUTIL, "add %s to list\n", du->name);
+ dprint(FD_DISKUTIL, "add %s to list\n", du->dus.name);
fio_gettime(&du->time, NULL);
get_io_ticks(du, &du->last_dus);
flist_add_tail(&du->list, &disk_list);
+ fio_mutex_up(disk_util_mutex);
+
find_add_disk_slaves(td, path, du);
return du;
}
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;
log_err("unknown sysfs layout\n");
return NULL;
}
+ tmp[PATH_MAX - 1] = '\0';
strncpy(tmp, p, PATH_MAX - 1);
sprintf(path, "%s", tmp);
}
- if (td->o.ioscheduler && !td->sysfs_root)
- td->sysfs_root = strdup(path);
-
return disk_util_add(td, majdev, mindev, path);
}
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
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,
+ struct buf_output *out)
+{
+ if (!agg->slavecount)
+ return;
+
+ if (!terse) {
+ log_buf(out, ", 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_buf(out, ";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, max_util = 0;
- int slavecount = 0;
-
- unsigned merges[2] = { 0, };
- unsigned ticks[2] = { 0, };
- unsigned time_in_queue = { 0, };
- unsigned long long sectors[2] = { 0, };
- unsigned ios[2] = { 0, };
+ double util;
flist_for_each(entry, &masterdu->slaves) {
slavedu = flist_entry(entry, struct disk_util, slavelist);
dus = &slavedu->dus;
- ios[0] += dus->ios[0];
- ios[1] += dus->ios[1];
- merges[0] += dus->merges[0];
- merges[1] += dus->merges[1];
- sectors[0] += dus->sectors[0];
- sectors[1] += dus->sectors[1];
- ticks[0] += dus->ticks[0];
- ticks[1] += dus->ticks[1];
- time_in_queue += dus->time_in_queue;
- ++slavecount;
-
- util = (double) (100 * dus->io_ticks / (double) slavedu->msec);
+ 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 > max_util)
- max_util = util;
+ if (util > agg->max_util.u.f)
+ agg->max_util.u.f = util;
}
- if (max_util > 100.0)
- max_util = 100.0;
+ 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);
- log_info(", aggrios=%u/%u, aggrmerge=%u/%u, aggrticks=%u/%u,"
- " aggrin_queue=%u, aggrutil=%3.2f%%",
- ios[0]/slavecount, ios[1]/slavecount,
- merges[0]/slavecount, merges[1]/slavecount,
- ticks[0]/slavecount, ticks[1]/slavecount,
- time_in_queue/slavecount, max_util);
+ while (!flist_empty(&disk_list)) {
+ struct disk_util *du;
+ du = flist_first_entry(&disk_list, struct disk_util, list);
+ flist_del(&du->list);
+ disk_util_free(du);
+ }
+
+ last_majdev = last_mindev = -1;
+ fio_mutex_up(disk_util_mutex);
+ fio_mutex_remove(disk_util_mutex);
}
-void show_disk_util(void)
+void print_disk_util(struct disk_util_stat *dus, struct disk_util_agg *agg,
+ int terse, struct buf_output *out)
{
- struct disk_util_stat *dus;
- struct flist_head *entry, *next;
- struct disk_util *du;
- double util;
+ 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_buf(out, " ");
+
+ log_buf(out, " %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_buf(out, ";%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, out);
- if (flist_empty(&disk_list))
+ if (!terse)
+ log_buf(out, "\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;
- log_info("\nDisk stats (read/write):\n");
+ json_object_add_value_array(obj, "disk_util", array);
- flist_for_each(entry, &disk_list) {
+ flist_for_each(entry, head) {
du = flist_entry(entry, struct disk_util, list);
- dus = &du->dus;
-
- util = (double) 100 * du->dus.io_ticks / (double) du->msec;
- if (util > 100.0)
- util = 100.0;
-
- /* If this node is the slave of a master device, as
- * happens in case of software RAIDs, inward-indent
- * this stats line to reflect a master-slave
- * relationship. Because the master device gets added
- * before the slave devices, we can safely assume that
- * the master's stats line has been displayed in a
- * previous iteration of this loop.
- */
- if (!flist_empty(&du->slavelist))
- log_info(" ");
-
- log_info(" %s: ios=%u/%u, merge=%u/%u, ticks=%u/%u, "
- "in_queue=%u, util=%3.2f%%", du->name,
- dus->ios[0], dus->ios[1],
- dus->merges[0], dus->merges[1],
- dus->ticks[0], dus->ticks[1],
- dus->time_in_queue, util);
-
- /* If the device has slaves, aggregate the stats for
- * those slave devices also.
- */
- if (!flist_empty(&du->slaves))
- aggregate_slaves_stats(du);
- log_info("\n");
+ aggregate_slaves_stats(du);
+ json_array_add_disk_util(&du->dus, &du->agg, array);
}
+}
- /*
- * now free the list
- */
- flist_for_each_safe(entry, next, &disk_list) {
- flist_del(entry);
- du = flist_entry(entry, struct disk_util, list);
- disk_util_free(du);
+void show_disk_util(int terse, struct json_object *parent,
+ struct buf_output *out)
+{
+ struct flist_head *entry;
+ struct disk_util *du;
+ bool do_json;
+
+ 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) && parent)
+ do_json = true;
+ else
+ do_json = false;
+
+ if (!terse && !do_json)
+ log_buf(out, "\nDisk stats (read/write):\n");
+
+ if (do_json)
+ json_object_add_disk_utils(parent, &disk_list);
+ else if (output_format & ~(FIO_OUTPUT_JSON | FIO_OUTPUT_JSON_PLUS)) {
+ 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, out);
+ }
+ }
+
+ fio_mutex_up(disk_util_mutex);
+}
+
+void setup_disk_util(void)
+{
+ disk_util_mutex = fio_mutex_init(FIO_MUTEX_UNLOCKED);
}