{
uint32_t zone_idx;
- if (f->zbd_info->zone_size_log2)
+ if (f->zbd_info->zone_size_log2 > 0)
zone_idx = offset >> f->zbd_info->zone_size_log2;
else
zone_idx = (offset >> 9) / f->zbd_info->zone_size;
ze = &f->zbd_info->zone_info[zone_idx_e];
for (z = zb; z < ze; z++) {
pthread_mutex_lock(&z->mutex);
+ pthread_mutex_lock(&f->zbd_info->mutex);
+ f->zbd_info->sectors_with_data -= z->wp - z->start;
+ pthread_mutex_unlock(&f->zbd_info->mutex);
z->wp = z->start;
z->verify_block = 0;
pthread_mutex_unlock(&z->mutex);
return res;
}
+/*
+ * Reset zbd_info.write_cnt, the counter that counts down towards the next
+ * zone reset.
+ */
+static void zbd_reset_write_cnt(const struct thread_data *td,
+ const struct fio_file *f)
+{
+ assert(0 <= td->o.zrf.u.f && td->o.zrf.u.f <= 1);
+
+ pthread_mutex_lock(&f->zbd_info->mutex);
+ f->zbd_info->write_cnt = td->o.zrf.u.f ?
+ min(1.0 / td->o.zrf.u.f, 0.0 + UINT_MAX) : UINT_MAX;
+ pthread_mutex_unlock(&f->zbd_info->mutex);
+}
+
+static bool zbd_dec_and_reset_write_cnt(const struct thread_data *td,
+ const struct fio_file *f)
+{
+ uint32_t write_cnt = 0;
+
+ pthread_mutex_lock(&f->zbd_info->mutex);
+ assert(f->zbd_info->write_cnt);
+ if (f->zbd_info->write_cnt)
+ write_cnt = --f->zbd_info->write_cnt;
+ if (write_cnt == 0)
+ zbd_reset_write_cnt(td, f);
+ pthread_mutex_unlock(&f->zbd_info->mutex);
+
+ return write_cnt == 0;
+}
+
+/* Check whether the value of zbd_info.sectors_with_data is correct. */
+static void check_swd(const struct thread_data *td, const struct fio_file *f)
+{
+#if 0
+ struct fio_zone_info *zb, *ze, *z;
+ uint64_t swd;
+
+ zb = &f->zbd_info->zone_info[zbd_zone_idx(f, f->file_offset)];
+ ze = &f->zbd_info->zone_info[zbd_zone_idx(f, f->file_offset +
+ f->io_size)];
+ swd = 0;
+ for (z = zb; z < ze; z++) {
+ pthread_mutex_lock(&z->mutex);
+ swd += z->wp - z->start;
+ }
+ pthread_mutex_lock(&f->zbd_info->mutex);
+ assert(f->zbd_info->sectors_with_data == swd);
+ pthread_mutex_unlock(&f->zbd_info->mutex);
+ for (z = zb; z < ze; z++)
+ pthread_mutex_unlock(&z->mutex);
+#endif
+}
+
void zbd_file_reset(struct thread_data *td, struct fio_file *f)
{
- struct fio_zone_info *zb, *ze;
+ struct fio_zone_info *zb, *ze, *z;
uint32_t zone_idx_e;
+ uint64_t swd = 0;
if (!f->zbd_info)
return;
zb = &f->zbd_info->zone_info[zbd_zone_idx(f, f->file_offset)];
zone_idx_e = zbd_zone_idx(f, f->file_offset + f->io_size);
ze = &f->zbd_info->zone_info[zone_idx_e];
+ for (z = zb ; z < ze; z++) {
+ pthread_mutex_lock(&z->mutex);
+ swd += z->wp - z->start;
+ }
+ pthread_mutex_lock(&f->zbd_info->mutex);
+ f->zbd_info->sectors_with_data = swd;
+ pthread_mutex_unlock(&f->zbd_info->mutex);
+ for (z = zb ; z < ze; z++)
+ pthread_mutex_unlock(&z->mutex);
+ dprint(FD_ZBD, "%s(%s): swd = %ld\n", __func__, f->file_name, swd);
/*
* If data verification is enabled reset the affected zones before
* writing any data to avoid that a zone reset has to be issued while
zbd_reset_zones(td, f, zb, ze, td->o.verify != VERIFY_NONE &&
(td->o.td_ddir & TD_DDIR_WRITE) &&
td->runstate != TD_VERIFYING);
+ zbd_reset_write_cnt(td, f);
}
/* The caller must hold f->zbd_info->mutex. */
switch (io_u->ddir) {
case DDIR_WRITE:
zone_end = min(end, (z + 1)->start);
+ pthread_mutex_lock(&zbd_info->mutex);
+ /*
+ * z->wp > zone_end means that one or more I/O errors
+ * have occurred.
+ */
+ if (z->wp <= zone_end)
+ zbd_info->sectors_with_data += zone_end - z->wp;
+ pthread_mutex_unlock(&zbd_info->mutex);
z->wp = zone_end;
break;
case DDIR_TRIM:
{
const struct fio_file *f = io_u->file;
uint32_t zone_idx_b;
- struct fio_zone_info *zb, *zl;
+ struct fio_zone_info *zb, *zl, *orig_zb;
uint32_t orig_len = io_u->buflen;
uint32_t min_bs = td->o.min_bs[io_u->ddir];
uint64_t new_len;
assert(io_u->buflen);
zone_idx_b = zbd_zone_idx(f, io_u->offset);
zb = &f->zbd_info->zone_info[zone_idx_b];
+ orig_zb = zb;
/* Accept the I/O offset for conventional zones. */
if (zb->type == BLK_ZONE_TYPE_CONVENTIONAL)
goto accept;
}
/*
- * Avoid reads past the write pointer because such reads do not
- * hit the medium.
+ * Check that there is enough written data in the zone to do an
+ * I/O of at least min_bs B. If there isn't, find a new zone for
+ * the I/O.
*/
range = zb->cond != BLK_ZONE_COND_OFFLINE ?
- ((zb->wp - zb->start) << 9) - io_u->buflen : 0;
- if (td_random(td) && range >= 0) {
- io_u->offset = (zb->start << 9) +
- ((io_u->offset - (zb->start << 9)) %
- (range + 1)) / min_bs * min_bs;
- assert(zb->start << 9 <= io_u->offset);
- assert(io_u->offset + io_u->buflen <= zb->wp << 9);
- goto accept;
- }
- if (zb->cond == BLK_ZONE_COND_OFFLINE ||
- (io_u->offset + io_u->buflen) >> 9 > zb->wp) {
+ (zb->wp - zb->start) << 9 : 0;
+ if (range < min_bs ||
+ ((!td_random(td)) && (io_u->offset + min_bs > zb->wp << 9))) {
pthread_mutex_unlock(&zb->mutex);
zl = &f->zbd_info->zone_info[zbd_zone_idx(f,
f->file_offset + f->io_size)];
io_u->buflen);
goto eof;
}
- io_u->offset = zb->start << 9;
+ /*
+ * zbd_find_zone() returned a zone with a range of at
+ * least min_bs.
+ */
+ range = (zb->wp - zb->start) << 9;
+ assert(range >= min_bs);
+
+ if (!td_random(td))
+ io_u->offset = zb->start << 9;
}
- if ((io_u->offset + io_u->buflen) >> 9 > zb->wp) {
- dprint(FD_ZBD, "%s: %lld + %lld > %" PRIu64 "\n",
- f->file_name, io_u->offset, io_u->buflen,
- zb->wp);
- goto eof;
+ /*
+ * Make sure the I/O is within the zone valid data range while
+ * maximizing the I/O size and preserving randomness.
+ */
+ if (range <= io_u->buflen)
+ io_u->offset = zb->start << 9;
+ else if (td_random(td))
+ io_u->offset = (zb->start << 9) +
+ ((io_u->offset - (orig_zb->start << 9)) %
+ (range - io_u->buflen)) / min_bs * min_bs;
+ /*
+ * Make sure the I/O does not cross over the zone wp position.
+ */
+ new_len = min((unsigned long long)io_u->buflen,
+ (unsigned long long)(zb->wp << 9) - io_u->offset);
+ new_len = new_len / min_bs * min_bs;
+ if (new_len < io_u->buflen) {
+ io_u->buflen = new_len;
+ dprint(FD_IO, "Changed length from %u into %llu\n",
+ orig_len, io_u->buflen);
}
+ assert(zb->start << 9 <= io_u->offset);
+ assert(io_u->offset + io_u->buflen <= zb->wp << 9);
goto accept;
case DDIR_WRITE:
if (io_u->buflen > (f->zbd_info->zone_size << 9))
goto eof;
zone_idx_b = zb - f->zbd_info->zone_info;
}
+ /* Check whether the zone reset threshold has been exceeded */
+ if (td->o.zrf.u.f) {
+ check_swd(td, f);
+ if ((f->zbd_info->sectors_with_data << 9) >=
+ f->io_size * td->o.zrt.u.f &&
+ zbd_dec_and_reset_write_cnt(td, f)) {
+ zb->reset_zone = 1;
+ }
+ }
/* Reset the zone pointer if necessary */
if (zb->reset_zone || zbd_zone_full(f, zb, min_bs)) {
assert(td->o.verify == VERIFY_NONE);
zb->reset_zone = 0;
if (zbd_reset_zone(td, f, zb) < 0)
goto eof;
+ check_swd(td, f);
}
/* Make writes occur at the write pointer */
assert(!zbd_zone_full(f, zb, min_bs));
projects / fio.git / blobdiff