X-Git-Url: https://git.kernel.dk/?a=blobdiff_plain;f=zbd.c;h=68cd58e1b9db99dfa883dab39bc06daea6b972e8;hb=c91fd13d479dc38bbb7ef6995256ad098ebbceb2;hp=84165ddc50b9d08933db2d45a3fb1ea0eb180ad6;hpb=b423c4b0ebd539338bdce1babc0cf4c3f3e5d513;p=fio.git diff --git a/zbd.c b/zbd.c index 84165ddc..68cd58e1 100644 --- a/zbd.c +++ b/zbd.c @@ -113,6 +113,34 @@ int zbd_reset_wp(struct thread_data *td, struct fio_file *f, return ret; } +/** + * zbd_get_max_open_zones - Get the maximum number of open zones + * @td: FIO thread data + * @f: FIO file for which to get max open zones + * @max_open_zones: Upon success, result will be stored here. + * + * A @max_open_zones value set to zero means no limit. + * + * Returns 0 upon success and a negative error code upon failure. + */ +int zbd_get_max_open_zones(struct thread_data *td, struct fio_file *f, + unsigned int *max_open_zones) +{ + int ret; + + if (td->io_ops && td->io_ops->get_max_open_zones) + ret = td->io_ops->get_max_open_zones(td, f, max_open_zones); + else + ret = blkzoned_get_max_open_zones(td, f, max_open_zones); + if (ret < 0) { + td_verror(td, errno, "get max open zones failed"); + log_err("%s: get max open zones failed (%d).\n", + f->file_name, errno); + } + + return ret; +} + /** * zbd_zone_idx - convert an offset into a zone number * @f: file pointer. @@ -166,7 +194,8 @@ static bool zbd_zone_full(const struct fio_file *f, struct fio_zone_info *z, z->wp + required > zbd_zone_capacity_end(z); } -static void zone_lock(struct thread_data *td, struct fio_file *f, struct fio_zone_info *z) +static void zone_lock(struct thread_data *td, const struct fio_file *f, + struct fio_zone_info *z) { struct zoned_block_device_info *zbd = f->zbd_info; uint32_t nz = z - zbd->zone_info; @@ -284,9 +313,7 @@ static bool zbd_verify_sizes(void) return false; } - if (td->o.zone_skip && - (td->o.zone_skip < td->o.zone_size || - td->o.zone_skip % td->o.zone_size)) { + if (td->o.zone_skip % td->o.zone_size) { log_err("%s: zoneskip %llu is not a multiple of the device zone size %llu.\n", f->file_name, (unsigned long long) td->o.zone_skip, (unsigned long long) td->o.zone_size); @@ -334,20 +361,21 @@ static bool zbd_verify_bs(void) { struct thread_data *td; struct fio_file *f; - uint32_t zone_size; int i, j, k; for_each_td(td, i) { for_each_file(td, f, j) { + uint64_t zone_size; + if (!f->zbd_info) continue; zone_size = f->zbd_info->zone_size; for (k = 0; k < FIO_ARRAY_SIZE(td->o.bs); k++) { if (td->o.verify != VERIFY_NONE && zone_size % td->o.bs[k] != 0) { - log_info("%s: block size %llu is not a divisor of the zone size %d\n", + log_info("%s: block size %llu is not a divisor of the zone size %llu\n", f->file_name, td->o.bs[k], - zone_size); + (unsigned long long)zone_size); return false; } } @@ -554,6 +582,51 @@ out: return ret; } +static int zbd_set_max_open_zones(struct thread_data *td, struct fio_file *f) +{ + struct zoned_block_device_info *zbd = f->zbd_info; + unsigned int max_open_zones; + int ret; + + if (zbd->model != ZBD_HOST_MANAGED) { + /* Only host-managed devices have a max open limit */ + zbd->max_open_zones = td->o.max_open_zones; + goto out; + } + + /* If host-managed, get the max open limit */ + ret = zbd_get_max_open_zones(td, f, &max_open_zones); + if (ret) + return ret; + + if (!max_open_zones) { + /* No device limit */ + zbd->max_open_zones = td->o.max_open_zones; + } else if (!td->o.max_open_zones) { + /* No user limit. Set limit to device limit */ + zbd->max_open_zones = max_open_zones; + } else if (td->o.max_open_zones <= max_open_zones) { + /* Both user limit and dev limit. User limit not too large */ + zbd->max_open_zones = td->o.max_open_zones; + } else { + /* Both user limit and dev limit. User limit too large */ + td_verror(td, EINVAL, + "Specified --max_open_zones is too large"); + log_err("Specified --max_open_zones (%d) is larger than max (%u)\n", + td->o.max_open_zones, max_open_zones); + return -EINVAL; + } + +out: + /* Ensure that the limit is not larger than FIO's internal limit */ + zbd->max_open_zones = min_not_zero(zbd->max_open_zones, + (uint32_t) ZBD_MAX_OPEN_ZONES); + dprint(FD_ZBD, "%s: using max open zones limit: %"PRIu32"\n", + f->file_name, zbd->max_open_zones); + + return 0; +} + /* * Allocate zone information and store it into f->zbd_info if zonemode=zbd. * @@ -576,9 +649,13 @@ static int zbd_create_zone_info(struct thread_data *td, struct fio_file *f) case ZBD_HOST_AWARE: case ZBD_HOST_MANAGED: ret = parse_zone_info(td, f); + if (ret) + return ret; break; case ZBD_NONE: ret = init_zone_info(td, f); + if (ret) + return ret; break; default: td_verror(td, EINVAL, "Unsupported zoned model"); @@ -586,11 +663,15 @@ static int zbd_create_zone_info(struct thread_data *td, struct fio_file *f) return -EINVAL; } - if (ret == 0) { - f->zbd_info->model = zbd_model; - f->zbd_info->max_open_zones = td->o.max_open_zones; + f->zbd_info->model = zbd_model; + + ret = zbd_set_max_open_zones(td, f); + if (ret) { + zbd_free_zone_info(f); + return ret; } - return ret; + + return 0; } void zbd_free_zone_info(struct fio_file *f) @@ -647,7 +728,7 @@ static bool zbd_open_zone(struct thread_data *td, const struct fio_file *f, static int zbd_reset_zone(struct thread_data *td, struct fio_file *f, struct fio_zone_info *z); -int zbd_setup_files(struct thread_data *td) +int zbd_init_files(struct thread_data *td) { struct fio_file *f; int i; @@ -656,6 +737,44 @@ int zbd_setup_files(struct thread_data *td) if (zbd_init_zone_info(td, f)) return 1; } + return 0; +} + +void zbd_recalc_options_with_zone_granularity(struct thread_data *td) +{ + struct fio_file *f; + int i; + + for_each_file(td, f, i) { + struct zoned_block_device_info *zbd = f->zbd_info; + // zonemode=strided doesn't get per-file zone size. + uint64_t zone_size = zbd ? zbd->zone_size : td->o.zone_size; + + if (zone_size == 0) + continue; + + if (td->o.size_nz > 0) { + td->o.size = td->o.size_nz * zone_size; + } + if (td->o.io_size_nz > 0) { + td->o.io_size = td->o.io_size_nz * zone_size; + } + if (td->o.start_offset_nz > 0) { + td->o.start_offset = td->o.start_offset_nz * zone_size; + } + if (td->o.offset_increment_nz > 0) { + td->o.offset_increment = td->o.offset_increment_nz * zone_size; + } + if (td->o.zone_skip_nz > 0) { + td->o.zone_skip = td->o.zone_skip_nz * zone_size; + } + } +} + +int zbd_setup_files(struct thread_data *td) +{ + struct fio_file *f; + int i; if (!zbd_using_direct_io()) { log_err("Using direct I/O is mandatory for writing to ZBD drives\n\n"); @@ -688,8 +807,6 @@ int zbd_setup_files(struct thread_data *td) if (zbd_is_seq_job(f)) assert(f->min_zone < f->max_zone); - zbd->max_open_zones = zbd->max_open_zones ?: ZBD_MAX_OPEN_ZONES; - if (td->o.max_open_zones > 0 && zbd->max_open_zones != td->o.max_open_zones) { log_err("Different 'max_open_zones' values\n"); @@ -785,11 +902,8 @@ static void zbd_close_zone(struct thread_data *td, const struct fio_file *f, if (f->zbd_info->open_zones[open_zone_idx] == zone_idx) break; } - if (open_zone_idx == f->zbd_info->num_open_zones) { - dprint(FD_ZBD, "%s: zone %d is not open\n", - f->file_name, zone_idx); + if (open_zone_idx == f->zbd_info->num_open_zones) return; - } dprint(FD_ZBD, "%s: closing zone %d\n", f->file_name, zone_idx); memmove(f->zbd_info->open_zones + open_zone_idx, @@ -807,16 +921,13 @@ static void zbd_close_zone(struct thread_data *td, const struct fio_file *f, * @f: fio file for which to reset zones * @zb: first zone to reset. * @ze: first zone not to reset. - * @all_zones: whether to reset all zones or only those zones for which the - * write pointer is not a multiple of td->o.min_bs[DDIR_WRITE]. */ static int zbd_reset_zones(struct thread_data *td, struct fio_file *f, struct fio_zone_info *const zb, - struct fio_zone_info *const ze, bool all_zones) + struct fio_zone_info *const ze) { struct fio_zone_info *z; const uint32_t min_bs = td->o.min_bs[DDIR_WRITE]; - bool reset_wp; int res = 0; assert(min_bs); @@ -829,16 +940,10 @@ static int zbd_reset_zones(struct thread_data *td, struct fio_file *f, if (!z->has_wp) continue; zone_lock(td, f, z); - if (all_zones) { - pthread_mutex_lock(&f->zbd_info->mutex); - zbd_close_zone(td, f, nz); - pthread_mutex_unlock(&f->zbd_info->mutex); - - reset_wp = z->wp != z->start; - } else { - reset_wp = z->wp % min_bs != 0; - } - if (reset_wp) { + pthread_mutex_lock(&f->zbd_info->mutex); + zbd_close_zone(td, f, nz); + pthread_mutex_unlock(&f->zbd_info->mutex); + if (z->wp != z->start) { dprint(FD_ZBD, "%s: resetting zone %u\n", f->file_name, zbd_zone_nr(f, z)); if (zbd_reset_zone(td, f, z) < 0) @@ -893,7 +998,8 @@ enum swd_action { }; /* Calculate the number of sectors with data (swd) and perform action 'a' */ -static uint64_t zbd_process_swd(const struct fio_file *f, enum swd_action a) +static uint64_t zbd_process_swd(struct thread_data *td, + const struct fio_file *f, enum swd_action a) { struct fio_zone_info *zb, *ze, *z; uint64_t swd = 0; @@ -903,7 +1009,7 @@ static uint64_t zbd_process_swd(const struct fio_file *f, enum swd_action a) ze = get_zone(f, f->max_zone); for (z = zb; z < ze; z++) { if (z->has_wp) { - pthread_mutex_lock(&z->mutex); + zone_lock(td, f, z); wp_swd += z->wp - z->start; } swd += z->wp - z->start; @@ -934,40 +1040,34 @@ static uint64_t zbd_process_swd(const struct fio_file *f, enum swd_action a) static const bool enable_check_swd = false; /* Check whether the values of zbd_info.*sectors_with_data are correct. */ -static void zbd_check_swd(const struct fio_file *f) +static void zbd_check_swd(struct thread_data *td, const struct fio_file *f) { if (!enable_check_swd) return; - zbd_process_swd(f, CHECK_SWD); -} - -static void zbd_init_swd(struct fio_file *f) -{ - uint64_t swd; - - swd = zbd_process_swd(f, SET_SWD); - dprint(FD_ZBD, "%s(%s): swd = %" PRIu64 "\n", __func__, f->file_name, - swd); + zbd_process_swd(td, f, CHECK_SWD); } void zbd_file_reset(struct thread_data *td, struct fio_file *f) { struct fio_zone_info *zb, *ze; + uint64_t swd; if (!f->zbd_info || !td_write(td)) return; zb = get_zone(f, f->min_zone); ze = get_zone(f, f->max_zone); - zbd_init_swd(f); + swd = zbd_process_swd(td, f, SET_SWD); + dprint(FD_ZBD, "%s(%s): swd = %" PRIu64 "\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 * writing data, which causes data loss. */ - zbd_reset_zones(td, f, zb, ze, td->o.verify != VERIFY_NONE && - td->runstate != TD_VERIFYING); + if (td->o.verify != VERIFY_NONE && td->runstate != TD_VERIFYING) + zbd_reset_zones(td, f, zb, ze); zbd_reset_write_cnt(td, f); } @@ -1094,16 +1194,18 @@ static struct fio_zone_info *zbd_convert_to_open_zone(struct thread_data *td, uint32_t tmp_idx; z = get_zone(f, zone_idx); - - zone_lock(td, f, z); + if (z->has_wp) + zone_lock(td, f, z); pthread_mutex_lock(&f->zbd_info->mutex); - if (z->cond != ZBD_ZONE_COND_OFFLINE && - td->o.max_open_zones == 0 && td->o.job_max_open_zones == 0) - goto examine_zone; - if (f->zbd_info->num_open_zones == 0) { - dprint(FD_ZBD, "%s(%s): no zones are open\n", - __func__, f->file_name); - goto open_other_zone; + if (z->has_wp) { + if (z->cond != ZBD_ZONE_COND_OFFLINE && + td->o.max_open_zones == 0 && td->o.job_max_open_zones == 0) + goto examine_zone; + if (f->zbd_info->num_open_zones == 0) { + dprint(FD_ZBD, "%s(%s): no zones are open\n", + __func__, f->file_name); + goto open_other_zone; + } } /* @@ -1112,7 +1214,8 @@ static struct fio_zone_info *zbd_convert_to_open_zone(struct thread_data *td, * Ignore zones which don't belong to thread's offset/size area. */ open_zone_idx = pick_random_zone_idx(f, io_u); - assert(open_zone_idx < f->zbd_info->num_open_zones); + assert(!open_zone_idx || + open_zone_idx < f->zbd_info->num_open_zones); tmp_idx = open_zone_idx; for (i = 0; i < f->zbd_info->num_open_zones; i++) { uint32_t tmpz; @@ -1131,7 +1234,8 @@ static struct fio_zone_info *zbd_convert_to_open_zone(struct thread_data *td, dprint(FD_ZBD, "%s(%s): no candidate zone\n", __func__, f->file_name); pthread_mutex_unlock(&f->zbd_info->mutex); - zone_unlock(z); + if (z->has_wp) + zone_unlock(z); return NULL; found_candidate_zone: @@ -1140,7 +1244,8 @@ found_candidate_zone: break; zone_idx = new_zone_idx; pthread_mutex_unlock(&f->zbd_info->mutex); - zone_unlock(z); + if (z->has_wp) + zone_unlock(z); } /* Both z->mutex and f->zbd_info->mutex are held. */ @@ -1242,10 +1347,23 @@ static struct fio_zone_info *zbd_replay_write_order(struct thread_data *td, assert(z); } - if (z->verify_block * min_bs >= z->capacity) + if (z->verify_block * min_bs >= z->capacity) { log_err("%s: %d * %d >= %llu\n", f->file_name, z->verify_block, min_bs, (unsigned long long)z->capacity); - io_u->offset = z->start + z->verify_block++ * min_bs; + /* + * If the assertion below fails during a test run, adding + * "--experimental_verify=1" to the command line may help. + */ + assert(false); + } + io_u->offset = z->start + z->verify_block * min_bs; + if (io_u->offset + io_u->buflen >= zbd_zone_capacity_end(z)) { + log_err("%s: %llu + %llu >= %llu\n", f->file_name, io_u->offset, + io_u->buflen, (unsigned long long) zbd_zone_capacity_end(z)); + assert(false); + } + z->verify_block += io_u->buflen / min_bs; + return z; } @@ -1413,7 +1531,7 @@ static void zbd_put_io(struct thread_data *td, const struct io_u *io_u) zbd_end_zone_io(td, io_u, z); zone_unlock(z); - zbd_check_swd(f); + zbd_check_swd(td, f); } /* @@ -1567,9 +1685,31 @@ enum io_u_action zbd_adjust_block(struct thread_data *td, struct io_u *io_u) zb = get_zone(f, zone_idx_b); orig_zb = zb; - /* Accept the I/O offset for conventional zones. */ - if (!zb->has_wp) + if (!zb->has_wp) { + /* Accept non-write I/Os for conventional zones. */ + if (io_u->ddir != DDIR_WRITE) + return io_u_accept; + /* + * Make sure that writes to conventional zones + * don't cross over to any sequential zones. + */ + if (!(zb + 1)->has_wp || + io_u->offset + io_u->buflen <= (zb + 1)->start) + return io_u_accept; + + if (io_u->offset + min_bs > (zb + 1)->start) { + dprint(FD_IO, + "%s: off=%llu + min_bs=%u > next zone %llu\n", + f->file_name, io_u->offset, + min_bs, (unsigned long long) (zb + 1)->start); + io_u->offset = zb->start + (zb + 1)->start - io_u->offset; + new_len = min(io_u->buflen, (zb + 1)->start - io_u->offset); + } else { + new_len = (zb + 1)->start - io_u->offset; + } + io_u->buflen = new_len / min_bs * min_bs; return io_u_accept; + } /* * Accept the I/O offset for reads if reading beyond the write pointer @@ -1579,7 +1719,7 @@ enum io_u_action zbd_adjust_block(struct thread_data *td, struct io_u *io_u) io_u->ddir == DDIR_READ && td->o.read_beyond_wp) return io_u_accept; - zbd_check_swd(f); + zbd_check_swd(td, f); zone_lock(td, f, zb); @@ -1587,12 +1727,6 @@ enum io_u_action zbd_adjust_block(struct thread_data *td, struct io_u *io_u) case DDIR_READ: if (td->runstate == TD_VERIFYING && td_write(td)) { zb = zbd_replay_write_order(td, io_u, zb); - /* - * Since we return with the zone lock still held, - * add an annotation to let Coverity know that it - * is intentional. - */ - /* coverity[missing_unlock] */ goto accept; } /* @@ -1655,13 +1789,22 @@ enum io_u_action zbd_adjust_block(struct thread_data *td, struct io_u *io_u) assert(io_u->offset + io_u->buflen <= zb->wp); goto accept; case DDIR_WRITE: - if (io_u->buflen > f->zbd_info->zone_size) + if (io_u->buflen > f->zbd_info->zone_size) { + td_verror(td, EINVAL, "I/O buflen exceeds zone size"); + dprint(FD_IO, + "%s: I/O buflen %llu exceeds zone size %llu\n", + f->file_name, io_u->buflen, + (unsigned long long) f->zbd_info->zone_size); goto eof; + } if (!zbd_open_zone(td, f, zone_idx_b)) { zone_unlock(zb); zb = zbd_convert_to_open_zone(td, io_u); - if (!zb) + if (!zb) { + dprint(FD_IO, "%s: can't convert to open zone", + f->file_name); goto eof; + } zone_idx_b = zbd_zone_nr(f, zb); } /* Check whether the zone reset threshold has been exceeded */ @@ -1688,6 +1831,7 @@ enum io_u_action zbd_adjust_block(struct thread_data *td, struct io_u *io_u) goto eof; if (zb->capacity < min_bs) { + td_verror(td, EINVAL, "ZCAP is less min_bs"); log_err("zone capacity %llu smaller than minimum block size %d\n", (unsigned long long)zb->capacity, min_bs); @@ -1698,8 +1842,9 @@ enum io_u_action zbd_adjust_block(struct thread_data *td, struct io_u *io_u) assert(!zbd_zone_full(f, zb, min_bs)); io_u->offset = zb->wp; if (!is_valid_offset(f, io_u->offset)) { - dprint(FD_ZBD, "Dropped request with offset %llu\n", - io_u->offset); + td_verror(td, EINVAL, "invalid WP value"); + dprint(FD_ZBD, "%s: dropped request with offset %llu\n", + f->file_name, io_u->offset); goto eof; } /* @@ -1718,9 +1863,9 @@ enum io_u_action zbd_adjust_block(struct thread_data *td, struct io_u *io_u) orig_len, io_u->buflen); goto accept; } - log_err("Zone remainder %lld smaller than minimum block size %d\n", - (zbd_zone_capacity_end(zb) - io_u->offset), - min_bs); + td_verror(td, EIO, "zone remainder too small"); + log_err("zone remainder %lld smaller than min block size %d\n", + (zbd_zone_capacity_end(zb) - io_u->offset), min_bs); goto eof; case DDIR_TRIM: /* fall-through */ @@ -1742,6 +1887,12 @@ accept: assert(!io_u->zbd_put_io); io_u->zbd_queue_io = zbd_queue_io; io_u->zbd_put_io = zbd_put_io; + /* + * Since we return with the zone lock still held, + * add an annotation to let Coverity know that it + * is intentional. + */ + /* coverity[missing_unlock] */ return io_u_accept; eof: