*/
#include <stdio.h>
#include <stdlib.h>
-#include <libgen.h>
#include <assert.h>
#include <sys/types.h>
#include <sys/stat.h>
#include "flist.h"
#include "fio.h"
-#include "verify.h"
#include "trim.h"
#include "filelock.h"
#include "smalloc.h"
#include "blktrace.h"
+#include "pshared.h"
+#include "lib/roundup.h"
+
+#include <netinet/in.h>
+#include <netinet/tcp.h>
+#include <arpa/inet.h>
+#include <sys/stat.h>
+#include <sys/socket.h>
+#include <sys/un.h>
static int iolog_flush(struct io_log *log);
if (!td->o.write_iolog_file)
return;
- fprintf(td->iolog_f, "%s %s %llu %lu\n", io_u->file->file_name,
+ fprintf(td->iolog_f, "%s %s %llu %llu\n", io_u->file->file_name,
io_ddir_name(io_u->ddir),
io_u->offset, io_u->buflen);
}
static void iolog_delay(struct thread_data *td, unsigned long delay)
{
uint64_t usec = utime_since_now(&td->last_issue);
+ unsigned long orig_delay = delay;
uint64_t this_delay;
struct timespec ts;
}
usec = utime_since_now(&ts);
- if (usec > delay)
- td->time_offset = usec - delay;
+ if (usec > orig_delay)
+ td->time_offset = usec - orig_delay;
else
td->time_offset = 0;
}
return 1;
}
+static bool read_iolog2(struct thread_data *td);
+
int read_iolog_get(struct thread_data *td, struct io_u *io_u)
{
struct io_piece *ipo;
while (!flist_empty(&td->io_log_list)) {
int ret;
+ if (!td->io_log_blktrace && td->o.read_iolog_chunked) {
+ if (td->io_log_checkmark == td->io_log_current) {
+ if (!read_iolog2(td))
+ return 1;
+ }
+ td->io_log_current--;
+ }
ipo = flist_first_entry(&td->io_log_list, struct io_piece, list);
flist_del(&ipo->list);
remove_trim_entry(td, ipo);
io_u->ddir = ipo->ddir;
if (ipo->ddir != DDIR_WAIT) {
io_u->offset = ipo->offset;
+ io_u->verify_offset = ipo->offset;
io_u->buflen = ipo->len;
io_u->file = td->files[ipo->fileno];
get_file(io_u->file);
- dprint(FD_IO, "iolog: get %llu/%lu/%s\n", io_u->offset,
+ dprint(FD_IO, "iolog: get %llu/%llu/%s\n", io_u->offset,
io_u->buflen, io_u->file->file_name);
if (ipo->delay)
iolog_delay(td, ipo->delay);
struct fio_rb_node **p, *parent;
struct io_piece *ipo, *__ipo;
- ipo = malloc(sizeof(struct io_piece));
+ ipo = calloc(1, sizeof(struct io_piece));
init_ipo(ipo);
ipo->file = io_u->file;
ipo->offset = io_u->offset;
}
/*
- * We don't need to sort the entries if we only performed sequential
- * writes. In this case, just reading back data in the order we wrote
- * it out is the faster but still safe.
- *
- * One exception is if we don't have a random map in which case we need
+ * Only sort writes if we don't have a random map in which case we need
* to check for duplicate blocks and drop the old one, which we rely on
* the rb insert/lookup for handling.
*/
- if (((!td->o.verifysort) || !td_random(td)) &&
- file_randommap(td, ipo->file)) {
+ if (file_randommap(td, ipo->file)) {
INIT_FLIST_HEAD(&ipo->list);
flist_add_tail(&ipo->list, &td->io_hist_list);
ipo->flags |= IP_F_ONLIST;
td->io_hist_len--;
}
-void trim_io_piece(struct thread_data *td, const struct io_u *io_u)
+void trim_io_piece(const struct io_u *io_u)
{
struct io_piece *ipo = io_u->ipo;
void write_iolog_close(struct thread_data *td)
{
+ if (!td->iolog_f)
+ return;
+
fflush(td->iolog_f);
fclose(td->iolog_f);
free(td->iolog_buf);
td->iolog_buf = NULL;
}
+static int64_t iolog_items_to_fetch(struct thread_data *td)
+{
+ struct timespec now;
+ uint64_t elapsed;
+ uint64_t for_1s;
+ int64_t items_to_fetch;
+
+ if (!td->io_log_highmark)
+ return 10;
+
+
+ fio_gettime(&now, NULL);
+ elapsed = ntime_since(&td->io_log_highmark_time, &now);
+ if (elapsed) {
+ for_1s = (td->io_log_highmark - td->io_log_current) * 1000000000 / elapsed;
+ items_to_fetch = for_1s - td->io_log_current;
+ if (items_to_fetch < 0)
+ items_to_fetch = 0;
+ } else
+ items_to_fetch = 0;
+
+ td->io_log_highmark = td->io_log_current + items_to_fetch;
+ td->io_log_checkmark = (td->io_log_highmark + 1) / 2;
+ fio_gettime(&td->io_log_highmark_time, NULL);
+
+ return items_to_fetch;
+}
+
/*
* Read version 2 iolog data. It is enhanced to include per-file logging,
* syncs, etc.
*/
-static int read_iolog2(struct thread_data *td, FILE *f)
+static bool read_iolog2(struct thread_data *td)
{
unsigned long long offset;
unsigned int bytes;
char *rfname, *fname, *act;
char *str, *p;
enum fio_ddir rw;
+ bool realloc = false;
+ int64_t items_to_fetch = 0;
- free_release_files(td);
+ if (td->o.read_iolog_chunked) {
+ items_to_fetch = iolog_items_to_fetch(td);
+ if (!items_to_fetch)
+ return true;
+ }
/*
* Read in the read iolog and store it, reuse the infrastructure
act = malloc(256+16);
reads = writes = waits = 0;
- while ((p = fgets(str, 4096, f)) != NULL) {
+ while ((p = fgets(str, 4096, td->io_log_rfile)) != NULL) {
struct io_piece *ipo;
int r;
dprint(FD_FILE, "iolog: ignoring"
" re-add of file %s\n", fname);
} else {
- fileno = add_file(td, fname, 0, 1);
+ fileno = add_file(td, fname, td->subjob_number, 1);
file_action = FIO_LOG_ADD_FILE;
}
continue;
/*
* Make note of file
*/
- ipo = malloc(sizeof(*ipo));
+ ipo = calloc(1, sizeof(*ipo));
init_ipo(ipo);
ipo->ddir = rw;
if (rw == DDIR_WAIT) {
ipo_bytes_align(td->o.replay_align, ipo);
ipo->len = bytes;
- if (rw != DDIR_INVAL && bytes > td->o.max_bs[rw])
+ if (rw != DDIR_INVAL && bytes > td->o.max_bs[rw]) {
+ realloc = true;
td->o.max_bs[rw] = bytes;
+ }
ipo->fileno = fileno;
ipo->file_action = file_action;
td->o.size += bytes;
}
queue_io_piece(td, ipo);
+
+ if (td->o.read_iolog_chunked) {
+ td->io_log_current++;
+ items_to_fetch--;
+ if (items_to_fetch == 0)
+ break;
+ }
}
free(str);
free(act);
free(rfname);
+ if (td->o.read_iolog_chunked) {
+ td->io_log_highmark = td->io_log_current;
+ td->io_log_checkmark = (td->io_log_highmark + 1) / 2;
+ fio_gettime(&td->io_log_highmark_time, NULL);
+ }
+
if (writes && read_only) {
log_err("fio: <%s> skips replay of %d writes due to"
" read-only\n", td->o.name, writes);
writes = 0;
}
+ if (td->o.read_iolog_chunked) {
+ if (td->io_log_current == 0) {
+ return false;
+ }
+ td->o.td_ddir = TD_DDIR_RW;
+ if (realloc && td->orig_buffer)
+ {
+ io_u_quiesce(td);
+ free_io_mem(td);
+ init_io_u_buffers(td);
+ }
+ return true;
+ }
+
if (!reads && !writes && !waits)
- return 1;
+ return false;
else if (reads && !writes)
td->o.td_ddir = TD_DDIR_READ;
else if (!reads && writes)
else
td->o.td_ddir = TD_DDIR_RW;
- return 0;
+ return true;
+}
+
+static bool is_socket(const char *path)
+{
+ struct stat buf;
+ int r;
+
+ r = stat(path, &buf);
+ if (r == -1)
+ return false;
+
+ return S_ISSOCK(buf.st_mode);
+}
+
+static int open_socket(const char *path)
+{
+ struct sockaddr_un addr;
+ int ret, fd;
+
+ fd = socket(AF_UNIX, SOCK_STREAM, 0);
+ if (fd < 0)
+ return fd;
+
+ addr.sun_family = AF_UNIX;
+ if (snprintf(addr.sun_path, sizeof(addr.sun_path), "%s", path) >=
+ sizeof(addr.sun_path)) {
+ log_err("%s: path name %s is too long for a Unix socket\n",
+ __func__, path);
+ }
+
+ ret = connect(fd, (const struct sockaddr *)&addr, strlen(path) + sizeof(addr.sun_family));
+ if (!ret)
+ return fd;
+
+ close(fd);
+ return -1;
}
/*
* open iolog, check version, and call appropriate parser
*/
-static int init_iolog_read(struct thread_data *td)
+static bool init_iolog_read(struct thread_data *td, char *fname)
{
char buffer[256], *p;
- FILE *f;
- int ret;
+ FILE *f = NULL;
+
+ dprint(FD_IO, "iolog: name=%s\n", fname);
+
+ if (is_socket(fname)) {
+ int fd;
+
+ fd = open_socket(fname);
+ if (fd >= 0)
+ f = fdopen(fd, "r");
+ } else if (!strcmp(fname, "-")) {
+ f = stdin;
+ } else
+ f = fopen(fname, "r");
+
+ free(fname);
- f = fopen(td->o.read_iolog_file, "r");
if (!f) {
perror("fopen read iolog");
- return 1;
+ return false;
}
p = fgets(buffer, sizeof(buffer), f);
td_verror(td, errno, "iolog read");
log_err("fio: unable to read iolog\n");
fclose(f);
- return 1;
+ return false;
}
/*
* version 2 of the iolog stores a specific string as the
* first line, check for that
*/
- if (!strncmp(iolog_ver2, buffer, strlen(iolog_ver2)))
- ret = read_iolog2(td, f);
- else {
- log_err("fio: iolog version 1 is no longer supported\n");
- ret = 1;
+ if (!strncmp(iolog_ver2, buffer, strlen(iolog_ver2))) {
+ free_release_files(td);
+ td->io_log_rfile = f;
+ return read_iolog2(td);
}
+ log_err("fio: iolog version 1 is no longer supported\n");
fclose(f);
- return ret;
+ return false;
}
/*
* Set up a log for storing io patterns.
*/
-static int init_iolog_write(struct thread_data *td)
+static bool init_iolog_write(struct thread_data *td)
{
struct fio_file *ff;
FILE *f;
f = fopen(td->o.write_iolog_file, "a");
if (!f) {
perror("fopen write iolog");
- return 1;
+ return false;
}
/*
*/
if (fprintf(f, "%s\n", iolog_ver2) < 0) {
perror("iolog init\n");
- return 1;
+ return false;
}
/*
for_each_file(td, ff, i)
log_file(td, ff, FIO_LOG_ADD_FILE);
- return 0;
+ return true;
}
-int init_iolog(struct thread_data *td)
+bool init_iolog(struct thread_data *td)
{
- int ret = 0;
+ bool ret;
if (td->o.read_iolog_file) {
int need_swap;
+ char * fname = get_name_by_idx(td->o.read_iolog_file, td->subjob_number);
/*
* Check if it's a blktrace file and load that if possible.
* Otherwise assume it's a normal log file and load that.
*/
- if (is_blktrace(td->o.read_iolog_file, &need_swap))
- ret = load_blktrace(td, td->o.read_iolog_file, need_swap);
- else
- ret = init_iolog_read(td);
+ if (is_blktrace(fname, &need_swap)) {
+ td->io_log_blktrace = 1;
+ ret = load_blktrace(td, fname, need_swap);
+ } else {
+ td->io_log_blktrace = 0;
+ ret = init_iolog_read(td, fname);
+ }
} else if (td->o.write_iolog_file)
ret = init_iolog_write(td);
+ else
+ ret = true;
- if (ret)
+ if (!ret)
td_verror(td, EINVAL, "failed initializing iolog");
return ret;
INIT_FLIST_HEAD(&l->io_logs);
l->log_type = p->log_type;
l->log_offset = p->log_offset;
+ l->log_prio = p->log_prio;
l->log_gz = p->log_gz;
l->log_gz_store = p->log_gz_store;
l->avg_msec = p->avg_msec;
}
if (l->td && l->td->o.io_submit_mode != IO_MODE_OFFLOAD) {
- struct io_logs *p;
-
- p = calloc(1, sizeof(*l->pending));
- p->max_samples = DEF_LOG_ENTRIES;
- p->log = calloc(p->max_samples, log_entry_sz(l));
- l->pending = p;
+ unsigned int def_samples = DEF_LOG_ENTRIES;
+ struct io_logs *__p;
+
+ __p = calloc(1, sizeof(*l->pending));
+ if (l->td->o.iodepth > DEF_LOG_ENTRIES)
+ def_samples = roundup_pow2(l->td->o.iodepth);
+ __p->max_samples = def_samples;
+ __p->log = calloc(__p->max_samples, log_entry_sz(l));
+ l->pending = __p;
}
if (l->log_offset)
l->log_ddir_mask = LOG_OFFSET_SAMPLE_BIT;
+ if (l->log_prio)
+ l->log_ddir_mask |= LOG_PRIO_SAMPLE_BIT;
INIT_FLIST_HEAD(&l->chunk_list);
entry_before = flist_first_entry(&entry->list, struct io_u_plat_entry, list);
io_u_plat_before = entry_before->io_u_plat;
- fprintf(f, "%lu, %u, %u, ", (unsigned long) s->time,
- io_sample_ddir(s), s->bs);
+ fprintf(f, "%lu, %u, %llu, ", (unsigned long) s->time,
+ io_sample_ddir(s), (unsigned long long) s->bs);
for (j = 0; j < FIO_IO_U_PLAT_NR - stride; j += stride) {
fprintf(f, "%llu, ", (unsigned long long)
hist_sum(j, stride, io_u_plat, io_u_plat_before));
void flush_samples(FILE *f, void *samples, uint64_t sample_size)
{
struct io_sample *s;
- int log_offset;
+ int log_offset, log_prio;
uint64_t i, nr_samples;
+ unsigned int prio_val;
+ const char *fmt;
if (!sample_size)
return;
s = __get_sample(samples, 0, 0);
log_offset = (s->__ddir & LOG_OFFSET_SAMPLE_BIT) != 0;
+ log_prio = (s->__ddir & LOG_PRIO_SAMPLE_BIT) != 0;
+
+ if (log_offset) {
+ if (log_prio)
+ fmt = "%lu, %" PRId64 ", %u, %llu, %llu, 0x%04x\n";
+ else
+ fmt = "%lu, %" PRId64 ", %u, %llu, %llu, %u\n";
+ } else {
+ if (log_prio)
+ fmt = "%lu, %" PRId64 ", %u, %llu, 0x%04x\n";
+ else
+ fmt = "%lu, %" PRId64 ", %u, %llu, %u\n";
+ }
nr_samples = sample_size / __log_entry_sz(log_offset);
for (i = 0; i < nr_samples; i++) {
s = __get_sample(samples, log_offset, i);
+ if (log_prio)
+ prio_val = s->priority;
+ else
+ prio_val = ioprio_value_is_class_rt(s->priority);
+
if (!log_offset) {
- fprintf(f, "%lu, %" PRId64 ", %u, %u\n",
- (unsigned long) s->time,
- s->data.val,
- io_sample_ddir(s), s->bs);
+ fprintf(f, fmt,
+ (unsigned long) s->time,
+ s->data.val,
+ io_sample_ddir(s), (unsigned long long) s->bs,
+ prio_val);
} else {
struct io_sample_offset *so = (void *) s;
- fprintf(f, "%lu, %" PRId64 ", %u, %u, %llu\n",
- (unsigned long) s->time,
- s->data.val,
- io_sample_ddir(s), s->bs,
- (unsigned long long) so->offset);
+ fprintf(f, fmt,
+ (unsigned long) s->time,
+ s->data.val,
+ io_sample_ddir(s), (unsigned long long) s->bs,
+ (unsigned long long) so->offset,
+ prio_val);
}
}
}
struct iolog_compress ic;
z_stream stream;
struct stat sb;
- ssize_t ret;
+ size_t ret;
size_t total;
void *buf;
FILE *f;
ic.seq = 1;
ret = fread(ic.buf, ic.len, 1, f);
- if (ret < 0) {
+ if (ret == 0 && ferror(f)) {
perror("fread");
fclose(f);
free(buf);
return 1;
- } else if (ret != 1) {
+ } else if (ferror(f) || (!feof(f) && ret != 1)) {
log_err("fio: short read on reading log\n");
fclose(f);
free(buf);