X-Git-Url: https://git.proxmox.com/?a=blobdiff_plain;f=block.c;h=3f874899377c785c45171bde8feac19b1e27c6db;hb=637d640fbbce498bd43a58b7a20b4190a1534751;hp=4a90dd1a78175d5612a24fc3dada5e002fe18971;hpb=029d091e4975af60ff9622717af19c5910f2f4e9;p=qemu.git diff --git a/block.c b/block.c index 4a90dd1a7..3f8748993 100644 --- a/block.c +++ b/block.c @@ -140,8 +140,6 @@ void bdrv_io_limits_disable(BlockDriverState *bs) bs->slice_start = 0; bs->slice_end = 0; - bs->slice_time = 0; - memset(&bs->io_base, 0, sizeof(bs->io_base)); } static void bdrv_block_timer(void *opaque) @@ -580,6 +578,26 @@ static int refresh_total_sectors(BlockDriverState *bs, int64_t hint) return 0; } +/** + * Set open flags for a given discard mode + * + * Return 0 on success, -1 if the discard mode was invalid. + */ +int bdrv_parse_discard_flags(const char *mode, int *flags) +{ + *flags &= ~BDRV_O_UNMAP; + + if (!strcmp(mode, "off") || !strcmp(mode, "ignore")) { + /* do nothing */ + } else if (!strcmp(mode, "on") || !strcmp(mode, "unmap")) { + *flags |= BDRV_O_UNMAP; + } else { + return -1; + } + + return 0; +} + /** * Set open flags for a given cache mode * @@ -645,17 +663,38 @@ static int bdrv_open_flags(BlockDriverState *bs, int flags) /* * Common part for opening disk images and files + * + * Removes all processed options from *options. */ static int bdrv_open_common(BlockDriverState *bs, BlockDriverState *file, - const char *filename, - int flags, BlockDriver *drv) + QDict *options, int flags, BlockDriver *drv) { int ret, open_flags; + const char *filename; assert(drv != NULL); assert(bs->file == NULL); + assert(options != NULL && bs->options != options); - trace_bdrv_open_common(bs, filename, flags, drv->format_name); + if (file != NULL) { + filename = file->filename; + } else { + filename = qdict_get_try_str(options, "filename"); + } + + trace_bdrv_open_common(bs, filename ?: "", flags, drv->format_name); + + if (use_bdrv_whitelist && !bdrv_is_whitelisted(drv)) { + return -ENOTSUP; + } + + /* bdrv_open() with directly using a protocol as drv. This layer is already + * opened, so assign it to bs (while file becomes a closed BlockDriverState) + * and return immediately. */ + if (file != NULL && drv->bdrv_file_open) { + bdrv_swap(file, bs); + return 0; + } bs->open_flags = flags; bs->buffer_alignment = 512; @@ -665,10 +704,10 @@ static int bdrv_open_common(BlockDriverState *bs, BlockDriverState *file, bdrv_enable_copy_on_read(bs); } - pstrcpy(bs->filename, sizeof(bs->filename), filename); - - if (use_bdrv_whitelist && !bdrv_is_whitelisted(drv)) { - return -ENOTSUP; + if (filename != NULL) { + pstrcpy(bs->filename, sizeof(bs->filename), filename); + } else { + bs->filename[0] = '\0'; } bs->drv = drv; @@ -681,16 +720,20 @@ static int bdrv_open_common(BlockDriverState *bs, BlockDriverState *file, /* Open the image, either directly or using a protocol */ if (drv->bdrv_file_open) { - if (file != NULL) { - bdrv_swap(file, bs); - ret = 0; - } else { - ret = drv->bdrv_file_open(bs, filename, open_flags); - } + assert(file == NULL); + assert(drv->bdrv_parse_filename || filename != NULL); + ret = drv->bdrv_file_open(bs, options, open_flags); } else { + if (file == NULL) { + qerror_report(ERROR_CLASS_GENERIC_ERROR, "Can't use '%s' as a " + "block driver for the protocol level", + drv->format_name); + ret = -EINVAL; + goto free_and_fail; + } assert(file != NULL); bs->file = file; - ret = drv->bdrv_open(bs, open_flags); + ret = drv->bdrv_open(bs, options, open_flags); } if (ret < 0) { @@ -704,6 +747,7 @@ static int bdrv_open_common(BlockDriverState *bs, BlockDriverState *file, #ifndef _WIN32 if (bs->is_temporary) { + assert(filename != NULL); unlink(filename); } #endif @@ -719,41 +763,136 @@ free_and_fail: /* * Opens a file using a protocol (file, host_device, nbd, ...) + * + * options is a QDict of options to pass to the block drivers, or NULL for an + * empty set of options. The reference to the QDict belongs to the block layer + * after the call (even on failure), so if the caller intends to reuse the + * dictionary, it needs to use QINCREF() before calling bdrv_file_open. */ -int bdrv_file_open(BlockDriverState **pbs, const char *filename, int flags) +int bdrv_file_open(BlockDriverState **pbs, const char *filename, + QDict *options, int flags) { BlockDriverState *bs; BlockDriver *drv; + const char *drvname; int ret; - drv = bdrv_find_protocol(filename); - if (!drv) { - return -ENOENT; + /* NULL means an empty set of options */ + if (options == NULL) { + options = qdict_new(); } bs = bdrv_new(""); - ret = bdrv_open_common(bs, NULL, filename, flags, drv); + bs->options = options; + options = qdict_clone_shallow(options); + + /* Fetch the file name from the options QDict if necessary */ + if (!filename) { + filename = qdict_get_try_str(options, "filename"); + } else if (filename && !qdict_haskey(options, "filename")) { + qdict_put(options, "filename", qstring_from_str(filename)); + } else { + qerror_report(ERROR_CLASS_GENERIC_ERROR, "Can't specify 'file' and " + "'filename' options at the same time"); + ret = -EINVAL; + goto fail; + } + + /* Find the right block driver */ + drvname = qdict_get_try_str(options, "driver"); + if (drvname) { + drv = bdrv_find_whitelisted_format(drvname); + qdict_del(options, "driver"); + } else if (filename) { + drv = bdrv_find_protocol(filename); + } else { + qerror_report(ERROR_CLASS_GENERIC_ERROR, + "Must specify either driver or file"); + drv = NULL; + } + + if (!drv) { + ret = -ENOENT; + goto fail; + } + + /* Parse the filename and open it */ + if (drv->bdrv_parse_filename && filename) { + Error *local_err = NULL; + drv->bdrv_parse_filename(filename, options, &local_err); + if (error_is_set(&local_err)) { + qerror_report_err(local_err); + error_free(local_err); + ret = -EINVAL; + goto fail; + } + qdict_del(options, "filename"); + } else if (!drv->bdrv_parse_filename && !filename) { + qerror_report(ERROR_CLASS_GENERIC_ERROR, + "The '%s' block driver requires a file name", + drv->format_name); + ret = -EINVAL; + goto fail; + } + + ret = bdrv_open_common(bs, NULL, options, flags, drv); if (ret < 0) { - bdrv_delete(bs); - return ret; + goto fail; } + + /* Check if any unknown options were used */ + if (qdict_size(options) != 0) { + const QDictEntry *entry = qdict_first(options); + qerror_report(ERROR_CLASS_GENERIC_ERROR, "Block protocol '%s' doesn't " + "support the option '%s'", + drv->format_name, entry->key); + ret = -EINVAL; + goto fail; + } + QDECREF(options); + bs->growable = 1; *pbs = bs; return 0; + +fail: + QDECREF(options); + if (!bs->drv) { + QDECREF(bs->options); + } + bdrv_delete(bs); + return ret; } -int bdrv_open_backing_file(BlockDriverState *bs) +/* + * Opens the backing file for a BlockDriverState if not yet open + * + * options is a QDict of options to pass to the block drivers, or NULL for an + * empty set of options. The reference to the QDict is transferred to this + * function (even on failure), so if the caller intends to reuse the dictionary, + * it needs to use QINCREF() before calling bdrv_file_open. + */ +int bdrv_open_backing_file(BlockDriverState *bs, QDict *options) { char backing_filename[PATH_MAX]; int back_flags, ret; BlockDriver *back_drv = NULL; if (bs->backing_hd != NULL) { + QDECREF(options); return 0; } + /* NULL means an empty set of options */ + if (options == NULL) { + options = qdict_new(); + } + bs->open_flags &= ~BDRV_O_NO_BACKING; - if (bs->backing_file[0] == '\0') { + if (qdict_haskey(options, "file.filename")) { + backing_filename[0] = '\0'; + } else if (bs->backing_file[0] == '\0' && qdict_size(options) == 0) { + QDECREF(options); return 0; } @@ -768,7 +907,9 @@ int bdrv_open_backing_file(BlockDriverState *bs) /* backing files always opened read-only */ back_flags = bs->open_flags & ~(BDRV_O_RDWR | BDRV_O_SNAPSHOT); - ret = bdrv_open(bs->backing_hd, backing_filename, back_flags, back_drv); + ret = bdrv_open(bs->backing_hd, + *backing_filename ? backing_filename : NULL, options, + back_flags, back_drv); if (ret < 0) { bdrv_delete(bs->backing_hd); bs->backing_hd = NULL; @@ -778,68 +919,109 @@ int bdrv_open_backing_file(BlockDriverState *bs) return 0; } +static void extract_subqdict(QDict *src, QDict **dst, const char *start) +{ + const QDictEntry *entry, *next; + const char *p; + + *dst = qdict_new(); + entry = qdict_first(src); + + while (entry != NULL) { + next = qdict_next(src, entry); + if (strstart(entry->key, start, &p)) { + qobject_incref(entry->value); + qdict_put_obj(*dst, p, entry->value); + qdict_del(src, entry->key); + } + entry = next; + } +} + /* * Opens a disk image (raw, qcow2, vmdk, ...) + * + * options is a QDict of options to pass to the block drivers, or NULL for an + * empty set of options. The reference to the QDict belongs to the block layer + * after the call (even on failure), so if the caller intends to reuse the + * dictionary, it needs to use QINCREF() before calling bdrv_open. */ -int bdrv_open(BlockDriverState *bs, const char *filename, int flags, - BlockDriver *drv) +int bdrv_open(BlockDriverState *bs, const char *filename, QDict *options, + int flags, BlockDriver *drv) { int ret; /* TODO: extra byte is a hack to ensure MAX_PATH space on Windows. */ char tmp_filename[PATH_MAX + 1]; BlockDriverState *file = NULL; + QDict *file_options = NULL; + + /* NULL means an empty set of options */ + if (options == NULL) { + options = qdict_new(); + } + bs->options = options; + options = qdict_clone_shallow(options); + + /* For snapshot=on, create a temporary qcow2 overlay */ if (flags & BDRV_O_SNAPSHOT) { BlockDriverState *bs1; int64_t total_size; - int is_protocol = 0; BlockDriver *bdrv_qcow2; - QEMUOptionParameter *options; + QEMUOptionParameter *create_options; char backing_filename[PATH_MAX]; + if (qdict_size(options) != 0) { + error_report("Can't use snapshot=on with driver-specific options"); + ret = -EINVAL; + goto fail; + } + assert(filename != NULL); + /* if snapshot, we create a temporary backing file and open it instead of opening 'filename' directly */ /* if there is a backing file, use it */ bs1 = bdrv_new(""); - ret = bdrv_open(bs1, filename, 0, drv); + ret = bdrv_open(bs1, filename, NULL, 0, drv); if (ret < 0) { bdrv_delete(bs1); - return ret; + goto fail; } total_size = bdrv_getlength(bs1) & BDRV_SECTOR_MASK; - if (bs1->drv && bs1->drv->protocol_name) - is_protocol = 1; - bdrv_delete(bs1); ret = get_tmp_filename(tmp_filename, sizeof(tmp_filename)); if (ret < 0) { - return ret; + goto fail; } /* Real path is meaningless for protocols */ - if (is_protocol) + if (path_has_protocol(filename)) { snprintf(backing_filename, sizeof(backing_filename), "%s", filename); - else if (!realpath(filename, backing_filename)) - return -errno; + } else if (!realpath(filename, backing_filename)) { + ret = -errno; + goto fail; + } bdrv_qcow2 = bdrv_find_format("qcow2"); - options = parse_option_parameters("", bdrv_qcow2->create_options, NULL); + create_options = parse_option_parameters("", bdrv_qcow2->create_options, + NULL); - set_option_parameter_int(options, BLOCK_OPT_SIZE, total_size); - set_option_parameter(options, BLOCK_OPT_BACKING_FILE, backing_filename); + set_option_parameter_int(create_options, BLOCK_OPT_SIZE, total_size); + set_option_parameter(create_options, BLOCK_OPT_BACKING_FILE, + backing_filename); if (drv) { - set_option_parameter(options, BLOCK_OPT_BACKING_FMT, + set_option_parameter(create_options, BLOCK_OPT_BACKING_FMT, drv->format_name); } - ret = bdrv_create(bdrv_qcow2, tmp_filename, options); - free_option_parameters(options); + ret = bdrv_create(bdrv_qcow2, tmp_filename, create_options); + free_option_parameters(create_options); if (ret < 0) { - return ret; + goto fail; } filename = tmp_filename; @@ -852,9 +1034,12 @@ int bdrv_open(BlockDriverState *bs, const char *filename, int flags, flags |= BDRV_O_ALLOW_RDWR; } - ret = bdrv_file_open(&file, filename, bdrv_open_flags(bs, flags)); + extract_subqdict(options, &file_options, "file."); + + ret = bdrv_file_open(&file, filename, file_options, + bdrv_open_flags(bs, flags)); if (ret < 0) { - return ret; + goto fail; } /* Find the right image format driver */ @@ -867,7 +1052,7 @@ int bdrv_open(BlockDriverState *bs, const char *filename, int flags, } /* Open the image */ - ret = bdrv_open_common(bs, file, filename, flags, drv); + ret = bdrv_open_common(bs, file, options, flags, drv); if (ret < 0) { goto unlink_and_fail; } @@ -879,13 +1064,27 @@ int bdrv_open(BlockDriverState *bs, const char *filename, int flags, /* If there is a backing file, use it */ if ((flags & BDRV_O_NO_BACKING) == 0) { - ret = bdrv_open_backing_file(bs); + QDict *backing_options; + + extract_subqdict(options, &backing_options, "backing."); + ret = bdrv_open_backing_file(bs, backing_options); if (ret < 0) { - bdrv_close(bs); - return ret; + goto close_and_fail; } } + /* Check if any unknown options were used */ + if (qdict_size(options) != 0) { + const QDictEntry *entry = qdict_first(options); + qerror_report(ERROR_CLASS_GENERIC_ERROR, "Block format '%s' used by " + "device '%s' doesn't support the option '%s'", + drv->format_name, bs->device_name, entry->key); + + ret = -EINVAL; + goto close_and_fail; + } + QDECREF(options); + if (!bdrv_key_required(bs)) { bdrv_dev_change_media_cb(bs, true); } @@ -904,6 +1103,15 @@ unlink_and_fail: if (bs->is_temporary) { unlink(filename); } +fail: + QDECREF(bs->options); + QDECREF(options); + bs->options = NULL; + return ret; + +close_and_fail: + bdrv_close(bs); + QDECREF(options); return ret; } @@ -1173,6 +1381,8 @@ void bdrv_close(BlockDriverState *bs) bs->valid_key = 0; bs->sg = 0; bs->growable = 0; + QDECREF(bs->options); + bs->options = NULL; if (bs->file != NULL) { bdrv_delete(bs->file); @@ -1268,11 +1478,10 @@ static void bdrv_move_feature_fields(BlockDriverState *bs_dest, bs_dest->enable_write_cache = bs_src->enable_write_cache; /* i/o timing parameters */ - bs_dest->slice_time = bs_src->slice_time; bs_dest->slice_start = bs_src->slice_start; bs_dest->slice_end = bs_src->slice_end; + bs_dest->slice_submitted = bs_src->slice_submitted; bs_dest->io_limits = bs_src->io_limits; - bs_dest->io_base = bs_src->io_base; bs_dest->throttled_reqs = bs_src->throttled_reqs; bs_dest->block_timer = bs_src->block_timer; bs_dest->io_limits_enabled = bs_src->io_limits_enabled; @@ -1286,7 +1495,6 @@ static void bdrv_move_feature_fields(BlockDriverState *bs_dest, bs_dest->iostatus = bs_src->iostatus; /* dirty bitmap */ - bs_dest->dirty_count = bs_src->dirty_count; bs_dest->dirty_bitmap = bs_src->dirty_bitmap; /* job */ @@ -1621,9 +1829,11 @@ int bdrv_commit_all(void) BlockDriverState *bs; QTAILQ_FOREACH(bs, &bdrv_states, list) { - int ret = bdrv_commit(bs); - if (ret < 0) { - return ret; + if (bs->drv && bs->backing_hd) { + int ret = bdrv_commit(bs); + if (ret < 0) { + return ret; + } } } return 0; @@ -1674,10 +1884,10 @@ static void tracked_request_begin(BdrvTrackedRequest *req, /** * Round a region to cluster boundaries */ -static void round_to_clusters(BlockDriverState *bs, - int64_t sector_num, int nb_sectors, - int64_t *cluster_sector_num, - int *cluster_nb_sectors) +void bdrv_round_to_clusters(BlockDriverState *bs, + int64_t sector_num, int nb_sectors, + int64_t *cluster_sector_num, + int *cluster_nb_sectors) { BlockDriverInfo bdi; @@ -1719,8 +1929,8 @@ static void coroutine_fn wait_for_overlapping_requests(BlockDriverState *bs, * CoR read and write operations are atomic and guest writes cannot * interleave between them. */ - round_to_clusters(bs, sector_num, nb_sectors, - &cluster_sector_num, &cluster_nb_sectors); + bdrv_round_to_clusters(bs, sector_num, nb_sectors, + &cluster_sector_num, &cluster_nb_sectors); do { retry = false; @@ -1968,27 +2178,21 @@ static void coroutine_fn bdrv_rw_co_entry(void *opaque) } /* - * Process a synchronous request using coroutines + * Process a vectored synchronous request using coroutines */ -static int bdrv_rw_co(BlockDriverState *bs, int64_t sector_num, uint8_t *buf, - int nb_sectors, bool is_write) +static int bdrv_rwv_co(BlockDriverState *bs, int64_t sector_num, + QEMUIOVector *qiov, bool is_write) { - QEMUIOVector qiov; - struct iovec iov = { - .iov_base = (void *)buf, - .iov_len = nb_sectors * BDRV_SECTOR_SIZE, - }; Coroutine *co; RwCo rwco = { .bs = bs, .sector_num = sector_num, - .nb_sectors = nb_sectors, - .qiov = &qiov, + .nb_sectors = qiov->size >> BDRV_SECTOR_BITS, + .qiov = qiov, .is_write = is_write, .ret = NOT_DONE, }; - - qemu_iovec_init_external(&qiov, &iov, 1); + assert((qiov->size & (BDRV_SECTOR_SIZE - 1)) == 0); /** * In sync call context, when the vcpu is blocked, this throttling timer @@ -2014,6 +2218,22 @@ static int bdrv_rw_co(BlockDriverState *bs, int64_t sector_num, uint8_t *buf, return rwco.ret; } +/* + * Process a synchronous request using coroutines + */ +static int bdrv_rw_co(BlockDriverState *bs, int64_t sector_num, uint8_t *buf, + int nb_sectors, bool is_write) +{ + QEMUIOVector qiov; + struct iovec iov = { + .iov_base = (void *)buf, + .iov_len = nb_sectors * BDRV_SECTOR_SIZE, + }; + + qemu_iovec_init_external(&qiov, &iov, 1); + return bdrv_rwv_co(bs, sector_num, &qiov, is_write); +} + /* return < 0 if error. See bdrv_write() for the return codes */ int bdrv_read(BlockDriverState *bs, int64_t sector_num, uint8_t *buf, int nb_sectors) @@ -2035,36 +2255,6 @@ int bdrv_read_unthrottled(BlockDriverState *bs, int64_t sector_num, return ret; } -#define BITS_PER_LONG (sizeof(unsigned long) * 8) - -static void set_dirty_bitmap(BlockDriverState *bs, int64_t sector_num, - int nb_sectors, int dirty) -{ - int64_t start, end; - unsigned long val, idx, bit; - - start = sector_num / BDRV_SECTORS_PER_DIRTY_CHUNK; - end = (sector_num + nb_sectors - 1) / BDRV_SECTORS_PER_DIRTY_CHUNK; - - for (; start <= end; start++) { - idx = start / BITS_PER_LONG; - bit = start % BITS_PER_LONG; - val = bs->dirty_bitmap[idx]; - if (dirty) { - if (!(val & (1UL << bit))) { - bs->dirty_count++; - val |= 1UL << bit; - } - } else { - if (val & (1UL << bit)) { - bs->dirty_count--; - val &= ~(1UL << bit); - } - } - bs->dirty_bitmap[idx] = val; - } -} - /* Return < 0 if error. Important errors are: -EIO generic I/O error (may happen for all errors) -ENOMEDIUM No media inserted. @@ -2077,6 +2267,11 @@ int bdrv_write(BlockDriverState *bs, int64_t sector_num, return bdrv_rw_co(bs, sector_num, (uint8_t *)buf, nb_sectors, true); } +int bdrv_writev(BlockDriverState *bs, int64_t sector_num, QEMUIOVector *qiov) +{ + return bdrv_rwv_co(bs, sector_num, qiov, true); +} + int bdrv_pread(BlockDriverState *bs, int64_t offset, void *buf, int count1) { @@ -2122,15 +2317,15 @@ int bdrv_pread(BlockDriverState *bs, int64_t offset, return count1; } -int bdrv_pwrite(BlockDriverState *bs, int64_t offset, - const void *buf, int count1) +int bdrv_pwritev(BlockDriverState *bs, int64_t offset, QEMUIOVector *qiov) { uint8_t tmp_buf[BDRV_SECTOR_SIZE]; int len, nb_sectors, count; int64_t sector_num; int ret; - count = count1; + count = qiov->size; + /* first write to align to sector start */ len = (BDRV_SECTOR_SIZE - offset) & (BDRV_SECTOR_SIZE - 1); if (len > count) @@ -2139,24 +2334,32 @@ int bdrv_pwrite(BlockDriverState *bs, int64_t offset, if (len > 0) { if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0) return ret; - memcpy(tmp_buf + (offset & (BDRV_SECTOR_SIZE - 1)), buf, len); + qemu_iovec_to_buf(qiov, 0, tmp_buf + (offset & (BDRV_SECTOR_SIZE - 1)), + len); if ((ret = bdrv_write(bs, sector_num, tmp_buf, 1)) < 0) return ret; count -= len; if (count == 0) - return count1; + return qiov->size; sector_num++; - buf += len; } /* write the sectors "in place" */ nb_sectors = count >> BDRV_SECTOR_BITS; if (nb_sectors > 0) { - if ((ret = bdrv_write(bs, sector_num, buf, nb_sectors)) < 0) + QEMUIOVector qiov_inplace; + + qemu_iovec_init(&qiov_inplace, qiov->niov); + qemu_iovec_concat(&qiov_inplace, qiov, len, + nb_sectors << BDRV_SECTOR_BITS); + ret = bdrv_writev(bs, sector_num, &qiov_inplace); + qemu_iovec_destroy(&qiov_inplace); + if (ret < 0) { return ret; + } + sector_num += nb_sectors; len = nb_sectors << BDRV_SECTOR_BITS; - buf += len; count -= len; } @@ -2164,11 +2367,24 @@ int bdrv_pwrite(BlockDriverState *bs, int64_t offset, if (count > 0) { if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0) return ret; - memcpy(tmp_buf, buf, count); + qemu_iovec_to_buf(qiov, qiov->size - count, tmp_buf, count); if ((ret = bdrv_write(bs, sector_num, tmp_buf, 1)) < 0) return ret; } - return count1; + return qiov->size; +} + +int bdrv_pwrite(BlockDriverState *bs, int64_t offset, + const void *buf, int count1) +{ + QEMUIOVector qiov; + struct iovec iov = { + .iov_base = (void *) buf, + .iov_len = count1, + }; + + qemu_iovec_init_external(&qiov, &iov, 1); + return bdrv_pwritev(bs, offset, &qiov); } /* @@ -2216,8 +2432,8 @@ static int coroutine_fn bdrv_co_do_copy_on_readv(BlockDriverState *bs, /* Cover entire cluster so no additional backing file I/O is required when * allocating cluster in the image file. */ - round_to_clusters(bs, sector_num, nb_sectors, - &cluster_sector_num, &cluster_nb_sectors); + bdrv_round_to_clusters(bs, sector_num, nb_sectors, + &cluster_sector_num, &cluster_nb_sectors); trace_bdrv_co_do_copy_on_readv(bs, sector_num, nb_sectors, cluster_sector_num, cluster_nb_sectors); @@ -2712,6 +2928,7 @@ int bdrv_has_zero_init(BlockDriverState *bs) typedef struct BdrvCoIsAllocatedData { BlockDriverState *bs; + BlockDriverState *base; int64_t sector_num; int nb_sectors; int *pnum; @@ -2831,7 +3048,9 @@ int coroutine_fn bdrv_co_is_allocated_above(BlockDriverState *top, * * [sector_num+x, nr_sectors] allocated. */ - if (n > pnum_inter) { + if (n > pnum_inter && + (intermediate == top || + sector_num + pnum_inter < intermediate->total_sectors)) { n = pnum_inter; } @@ -2842,6 +3061,44 @@ int coroutine_fn bdrv_co_is_allocated_above(BlockDriverState *top, return 0; } +/* Coroutine wrapper for bdrv_is_allocated_above() */ +static void coroutine_fn bdrv_is_allocated_above_co_entry(void *opaque) +{ + BdrvCoIsAllocatedData *data = opaque; + BlockDriverState *top = data->bs; + BlockDriverState *base = data->base; + + data->ret = bdrv_co_is_allocated_above(top, base, data->sector_num, + data->nb_sectors, data->pnum); + data->done = true; +} + +/* + * Synchronous wrapper around bdrv_co_is_allocated_above(). + * + * See bdrv_co_is_allocated_above() for details. + */ +int bdrv_is_allocated_above(BlockDriverState *top, BlockDriverState *base, + int64_t sector_num, int nb_sectors, int *pnum) +{ + Coroutine *co; + BdrvCoIsAllocatedData data = { + .bs = top, + .base = base, + .sector_num = sector_num, + .nb_sectors = nb_sectors, + .pnum = pnum, + .done = false, + }; + + co = qemu_coroutine_create(bdrv_is_allocated_above_co_entry); + qemu_coroutine_enter(co, &data); + while (!data.done) { + qemu_aio_wait(); + } + return data.ret; +} + BlockInfo *bdrv_query_info(BlockDriverState *bs) { BlockInfo *info = g_malloc0(sizeof(*info)); @@ -2863,8 +3120,9 @@ BlockInfo *bdrv_query_info(BlockDriverState *bs) if (bs->dirty_bitmap) { info->has_dirty = true; info->dirty = g_malloc0(sizeof(*info->dirty)); - info->dirty->count = bdrv_get_dirty_count(bs) * - BDRV_SECTORS_PER_DIRTY_CHUNK * BDRV_SECTOR_SIZE; + info->dirty->count = bdrv_get_dirty_count(bs) * BDRV_SECTOR_SIZE; + info->dirty->granularity = + ((int64_t) BDRV_SECTOR_SIZE << hbitmap_granularity(bs->dirty_bitmap)); } if (bs->drv) { @@ -3008,14 +3266,29 @@ int bdrv_get_info(BlockDriverState *bs, BlockDriverInfo *bdi) int bdrv_save_vmstate(BlockDriverState *bs, const uint8_t *buf, int64_t pos, int size) +{ + QEMUIOVector qiov; + struct iovec iov = { + .iov_base = (void *) buf, + .iov_len = size, + }; + + qemu_iovec_init_external(&qiov, &iov, 1); + return bdrv_writev_vmstate(bs, &qiov, pos); +} + +int bdrv_writev_vmstate(BlockDriverState *bs, QEMUIOVector *qiov, int64_t pos) { BlockDriver *drv = bs->drv; - if (!drv) + + if (!drv) { return -ENOMEDIUM; - if (drv->bdrv_save_vmstate) - return drv->bdrv_save_vmstate(bs, buf, pos, size); - if (bs->file) - return bdrv_save_vmstate(bs->file, buf, pos, size); + } else if (drv->bdrv_save_vmstate) { + return drv->bdrv_save_vmstate(bs, qiov, pos); + } else if (bs->file) { + return bdrv_writev_vmstate(bs->file, qiov, pos); + } + return -ENOTSUP; } @@ -3153,7 +3426,7 @@ int bdrv_snapshot_goto(BlockDriverState *bs, if (bs->file) { drv->bdrv_close(bs); ret = bdrv_snapshot_goto(bs->file, snapshot_id); - open_ret = drv->bdrv_open(bs, bs->open_flags); + open_ret = drv->bdrv_open(bs, NULL, bs->open_flags); if (open_ret < 0) { bdrv_delete(bs->file); bs->drv = NULL; @@ -3572,6 +3845,7 @@ static bool bdrv_exceed_bps_limits(BlockDriverState *bs, int nb_sectors, bool is_write, double elapsed_time, uint64_t *wait) { uint64_t bps_limit = 0; + uint64_t extension; double bytes_limit, bytes_base, bytes_res; double slice_time, wait_time; @@ -3590,9 +3864,9 @@ static bool bdrv_exceed_bps_limits(BlockDriverState *bs, int nb_sectors, slice_time = bs->slice_end - bs->slice_start; slice_time /= (NANOSECONDS_PER_SECOND); bytes_limit = bps_limit * slice_time; - bytes_base = bs->nr_bytes[is_write] - bs->io_base.bytes[is_write]; + bytes_base = bs->slice_submitted.bytes[is_write]; if (bs->io_limits.bps[BLOCK_IO_LIMIT_TOTAL]) { - bytes_base += bs->nr_bytes[!is_write] - bs->io_base.bytes[!is_write]; + bytes_base += bs->slice_submitted.bytes[!is_write]; } /* bytes_base: the bytes of data which have been read/written; and @@ -3619,10 +3893,12 @@ static bool bdrv_exceed_bps_limits(BlockDriverState *bs, int nb_sectors, * info can be kept until the timer fire, so it is increased and tuned * based on the result of experiment. */ - bs->slice_time = wait_time * BLOCK_IO_SLICE_TIME * 10; - bs->slice_end += bs->slice_time - 3 * BLOCK_IO_SLICE_TIME; + extension = wait_time * NANOSECONDS_PER_SECOND; + extension = DIV_ROUND_UP(extension, BLOCK_IO_SLICE_TIME) * + BLOCK_IO_SLICE_TIME; + bs->slice_end += extension; if (wait) { - *wait = wait_time * BLOCK_IO_SLICE_TIME * 10; + *wait = wait_time * NANOSECONDS_PER_SECOND; } return true; @@ -3650,9 +3926,9 @@ static bool bdrv_exceed_iops_limits(BlockDriverState *bs, bool is_write, slice_time = bs->slice_end - bs->slice_start; slice_time /= (NANOSECONDS_PER_SECOND); ios_limit = iops_limit * slice_time; - ios_base = bs->nr_ops[is_write] - bs->io_base.ios[is_write]; + ios_base = bs->slice_submitted.ios[is_write]; if (bs->io_limits.iops[BLOCK_IO_LIMIT_TOTAL]) { - ios_base += bs->nr_ops[!is_write] - bs->io_base.ios[!is_write]; + ios_base += bs->slice_submitted.ios[!is_write]; } if (ios_base + 1 <= ios_limit) { @@ -3663,7 +3939,7 @@ static bool bdrv_exceed_iops_limits(BlockDriverState *bs, bool is_write, return false; } - /* Calc approx time to dispatch */ + /* Calc approx time to dispatch, in seconds */ wait_time = (ios_base + 1) / iops_limit; if (wait_time > elapsed_time) { wait_time = wait_time - elapsed_time; @@ -3671,10 +3947,10 @@ static bool bdrv_exceed_iops_limits(BlockDriverState *bs, bool is_write, wait_time = 0; } - bs->slice_time = wait_time * BLOCK_IO_SLICE_TIME * 10; - bs->slice_end += bs->slice_time - 3 * BLOCK_IO_SLICE_TIME; + /* Exceeded current slice, extend it by another slice time */ + bs->slice_end += BLOCK_IO_SLICE_TIME; if (wait) { - *wait = wait_time * BLOCK_IO_SLICE_TIME * 10; + *wait = wait_time * NANOSECONDS_PER_SECOND; } return true; @@ -3689,19 +3965,10 @@ static bool bdrv_exceed_io_limits(BlockDriverState *bs, int nb_sectors, int bps_ret, iops_ret; now = qemu_get_clock_ns(vm_clock); - if ((bs->slice_start < now) - && (bs->slice_end > now)) { - bs->slice_end = now + bs->slice_time; - } else { - bs->slice_time = 5 * BLOCK_IO_SLICE_TIME; + if (now > bs->slice_end) { bs->slice_start = now; - bs->slice_end = now + bs->slice_time; - - bs->io_base.bytes[is_write] = bs->nr_bytes[is_write]; - bs->io_base.bytes[!is_write] = bs->nr_bytes[!is_write]; - - bs->io_base.ios[is_write] = bs->nr_ops[is_write]; - bs->io_base.ios[!is_write] = bs->nr_ops[!is_write]; + bs->slice_end = now + BLOCK_IO_SLICE_TIME; + memset(&bs->slice_submitted, 0, sizeof(bs->slice_submitted)); } elapsed_time = now - bs->slice_start; @@ -3729,6 +3996,10 @@ static bool bdrv_exceed_io_limits(BlockDriverState *bs, int nb_sectors, *wait = 0; } + bs->slice_submitted.bytes[is_write] += (int64_t)nb_sectors * + BDRV_SECTOR_SIZE; + bs->slice_submitted.ios[is_write]++; + return false; } @@ -4170,7 +4441,18 @@ int coroutine_fn bdrv_co_discard(BlockDriverState *bs, int64_t sector_num, return -EIO; } else if (bs->read_only) { return -EROFS; - } else if (bs->drv->bdrv_co_discard) { + } + + if (bs->dirty_bitmap) { + bdrv_reset_dirty(bs, sector_num, nb_sectors); + } + + /* Do nothing if disabled. */ + if (!(bs->open_flags & BDRV_O_UNMAP)) { + return 0; + } + + if (bs->drv->bdrv_co_discard) { return bs->drv->bdrv_co_discard(bs, sector_num, nb_sectors); } else if (bs->drv->bdrv_aio_discard) { BlockDriverAIOCB *acb; @@ -4325,22 +4607,20 @@ bool bdrv_qiov_is_aligned(BlockDriverState *bs, QEMUIOVector *qiov) return true; } -void bdrv_set_dirty_tracking(BlockDriverState *bs, int enable) +void bdrv_set_dirty_tracking(BlockDriverState *bs, int granularity) { int64_t bitmap_size; - bs->dirty_count = 0; - if (enable) { - if (!bs->dirty_bitmap) { - bitmap_size = (bdrv_getlength(bs) >> BDRV_SECTOR_BITS) + - BDRV_SECTORS_PER_DIRTY_CHUNK * BITS_PER_LONG - 1; - bitmap_size /= BDRV_SECTORS_PER_DIRTY_CHUNK * BITS_PER_LONG; + assert((granularity & (granularity - 1)) == 0); - bs->dirty_bitmap = g_new0(unsigned long, bitmap_size); - } + if (granularity) { + granularity >>= BDRV_SECTOR_BITS; + assert(!bs->dirty_bitmap); + bitmap_size = (bdrv_getlength(bs) >> BDRV_SECTOR_BITS); + bs->dirty_bitmap = hbitmap_alloc(bitmap_size, ffs(granularity) - 1); } else { if (bs->dirty_bitmap) { - g_free(bs->dirty_bitmap); + hbitmap_free(bs->dirty_bitmap); bs->dirty_bitmap = NULL; } } @@ -4348,67 +4628,37 @@ void bdrv_set_dirty_tracking(BlockDriverState *bs, int enable) int bdrv_get_dirty(BlockDriverState *bs, int64_t sector) { - int64_t chunk = sector / (int64_t)BDRV_SECTORS_PER_DIRTY_CHUNK; - - if (bs->dirty_bitmap && - (sector << BDRV_SECTOR_BITS) < bdrv_getlength(bs)) { - return !!(bs->dirty_bitmap[chunk / BITS_PER_LONG] & - (1UL << (chunk % BITS_PER_LONG))); + if (bs->dirty_bitmap) { + return hbitmap_get(bs->dirty_bitmap, sector); } else { return 0; } } -int64_t bdrv_get_next_dirty(BlockDriverState *bs, int64_t sector) +void bdrv_dirty_iter_init(BlockDriverState *bs, HBitmapIter *hbi) { - int64_t chunk; - int bit, elem; - - /* Avoid an infinite loop. */ - assert(bs->dirty_count > 0); - - sector = (sector | (BDRV_SECTORS_PER_DIRTY_CHUNK - 1)) + 1; - chunk = sector / (int64_t)BDRV_SECTORS_PER_DIRTY_CHUNK; - - QEMU_BUILD_BUG_ON(sizeof(bs->dirty_bitmap[0]) * 8 != BITS_PER_LONG); - elem = chunk / BITS_PER_LONG; - bit = chunk % BITS_PER_LONG; - for (;;) { - if (sector >= bs->total_sectors) { - sector = 0; - bit = elem = 0; - } - if (bit == 0 && bs->dirty_bitmap[elem] == 0) { - sector += BDRV_SECTORS_PER_DIRTY_CHUNK * BITS_PER_LONG; - elem++; - } else { - if (bs->dirty_bitmap[elem] & (1UL << bit)) { - return sector; - } - sector += BDRV_SECTORS_PER_DIRTY_CHUNK; - if (++bit == BITS_PER_LONG) { - bit = 0; - elem++; - } - } - } + hbitmap_iter_init(hbi, bs->dirty_bitmap, 0); } void bdrv_set_dirty(BlockDriverState *bs, int64_t cur_sector, int nr_sectors) { - set_dirty_bitmap(bs, cur_sector, nr_sectors, 1); + hbitmap_set(bs->dirty_bitmap, cur_sector, nr_sectors); } void bdrv_reset_dirty(BlockDriverState *bs, int64_t cur_sector, int nr_sectors) { - set_dirty_bitmap(bs, cur_sector, nr_sectors, 0); + hbitmap_reset(bs->dirty_bitmap, cur_sector, nr_sectors); } int64_t bdrv_get_dirty_count(BlockDriverState *bs) { - return bs->dirty_count; + if (bs->dirty_bitmap) { + return hbitmap_count(bs->dirty_bitmap); + } else { + return 0; + } } void bdrv_set_in_use(BlockDriverState *bs, int in_use) @@ -4485,7 +4735,8 @@ bdrv_acct_done(BlockDriverState *bs, BlockAcctCookie *cookie) void bdrv_img_create(const char *filename, const char *fmt, const char *base_filename, const char *base_fmt, - char *options, uint64_t img_size, int flags, Error **errp) + char *options, uint64_t img_size, int flags, + Error **errp, bool quiet) { QEMUOptionParameter *param = NULL, *create_options = NULL; QEMUOptionParameter *backing_fmt, *backing_file, *size; @@ -4577,7 +4828,8 @@ void bdrv_img_create(const char *filename, const char *fmt, bs = bdrv_new(""); - ret = bdrv_open(bs, backing_file->value.s, back_flags, backing_drv); + ret = bdrv_open(bs, backing_file->value.s, NULL, back_flags, + backing_drv); if (ret < 0) { error_setg_errno(errp, -ret, "Could not open '%s'", backing_file->value.s); @@ -4594,18 +4846,23 @@ void bdrv_img_create(const char *filename, const char *fmt, } } - printf("Formatting '%s', fmt=%s ", filename, fmt); - print_option_parameters(param); - puts(""); - + if (!quiet) { + printf("Formatting '%s', fmt=%s ", filename, fmt); + print_option_parameters(param); + puts(""); + } ret = bdrv_create(drv, filename, param); if (ret < 0) { if (ret == -ENOTSUP) { error_setg(errp,"Formatting or formatting option not supported for " "file format '%s'", fmt); } else if (ret == -EFBIG) { - error_setg(errp, "The image size is too large for file format '%s'", - fmt); + const char *cluster_size_hint = ""; + if (get_option_parameter(create_options, BLOCK_OPT_CLUSTER_SIZE)) { + cluster_size_hint = " (try using a larger cluster size)"; + } + error_setg(errp, "The image size is too large for file format '%s'%s", + fmt, cluster_size_hint); } else { error_setg(errp, "%s: error while creating %s: %s", filename, fmt, strerror(-ret)); @@ -4620,3 +4877,9 @@ out: bdrv_delete(bs); } } + +AioContext *bdrv_get_aio_context(BlockDriverState *bs) +{ + /* Currently BlockDriverState always uses the main loop AioContext */ + return qemu_get_aio_context(); +}