static int ioreq_runio_qemu_aio(struct ioreq *ioreq)
{
struct XenBlkDev *blkdev = ioreq->blkdev;
if (ioreq->req.nr_segments && ioreq_map(ioreq) == -1) {
goto err_no_map;
}
ioreq->aio_inflight++;
if (ioreq->presync) {
bdrv_flush(blkdev->bs); /* FIXME: aio_flush() ??? */
}
switch (ioreq->req.operation) {
case BLKIF_OP_READ:
bdrv_acct_start(blkdev->bs, &ioreq->acct, ioreq->v.size, BDRV_ACCT_READ);
ioreq->aio_inflight++;
bdrv_aio_readv(blkdev->bs, ioreq->start / BLOCK_SIZE,
&ioreq->v, ioreq->v.size / BLOCK_SIZE,
qemu_aio_complete, ioreq);
break;
case BLKIF_OP_WRITE:
case BLKIF_OP_WRITE_BARRIER:
if (!ioreq->req.nr_segments) {
break;
}
bdrv_acct_start(blkdev->bs, &ioreq->acct, ioreq->v.size, BDRV_ACCT_WRITE);
ioreq->aio_inflight++;
bdrv_aio_writev(blkdev->bs, ioreq->start / BLOCK_SIZE,
&ioreq->v, ioreq->v.size / BLOCK_SIZE,
qemu_aio_complete, ioreq);
break;
default:
/* unknown operation (shouldn't happen -- parse catches this) */
goto err;
}
if (ioreq->postsync) {
bdrv_flush(blkdev->bs); /* FIXME: aio_flush() ??? */
}
qemu_aio_complete(ioreq, 0);
return 0;
err:
ioreq_unmap(ioreq);
err_no_map:
ioreq_finish(ioreq);
ioreq->status = BLKIF_RSP_ERROR;
return -1;
}
int qcow2_cache_flush(BlockDriverState *bs, Qcow2Cache *c)
{
BDRVQcowState *s = bs->opaque;
int result = 0;
int ret;
int i;
trace_qcow2_cache_flush(qemu_coroutine_self(), c == s->l2_table_cache);
for (i = 0; i < c->size; i++) {
ret = qcow2_cache_entry_flush(bs, c, i);
if (ret < 0 && result != -ENOSPC) {
result = ret;
}
}
if (result == 0) {
ret = bdrv_flush(bs->file);
if (ret < 0) {
result = ret;
}
}
return result;
}
static int qcow2_cache_entry_flush(BlockDriverState *bs, Qcow2Cache *c, int i)
{
BDRVQcowState *s = bs->opaque;
int ret = 0;
if (!c->entries[i].dirty || !c->entries[i].offset) {
return 0;
}
trace_qcow2_cache_entry_flush(qemu_coroutine_self(),
c == s->l2_table_cache, i);
if (c->depends) {
ret = qcow2_cache_flush_dependency(bs, c);
} else if (c->depends_on_flush) {
ret = bdrv_flush(bs->file);
if (ret >= 0) {
c->depends_on_flush = false;
}
}
if (ret < 0) {
return ret;
}
if (c == s->refcount_block_cache) {
ret = qcow2_pre_write_overlap_check(bs,
QCOW2_OL_DEFAULT & ~QCOW2_OL_REFCOUNT_BLOCK,
c->entries[i].offset, s->cluster_size);
} else if (c == s->l2_table_cache) {
ret = qcow2_pre_write_overlap_check(bs,
QCOW2_OL_DEFAULT & ~QCOW2_OL_ACTIVE_L2,
c->entries[i].offset, s->cluster_size);
} else {
ret = qcow2_pre_write_overlap_check(bs, QCOW2_OL_DEFAULT,
c->entries[i].offset, s->cluster_size);
}
if (ret < 0) {
return ret;
}
if (c == s->refcount_block_cache) {
BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_UPDATE_PART);
} else if (c == s->l2_table_cache) {
BLKDBG_EVENT(bs->file, BLKDBG_L2_UPDATE);
}
ret = bdrv_pwrite(bs->file, c->entries[i].offset, c->entries[i].table,
s->cluster_size);
if (ret < 0) {
return ret;
}
c->entries[i].dirty = false;
return 0;
}
static void test_sync_op_flush(BdrvChild *c)
{
int ret;
/* Normal success path */
ret = bdrv_flush(c->bs);
g_assert_cmpint(ret, ==, 0);
/* Early success: Read-only image */
c->bs->read_only = true;
c->bs->open_flags &= ~BDRV_O_RDWR;
ret = bdrv_flush(c->bs);
g_assert_cmpint(ret, ==, 0);
c->bs->read_only = false;
c->bs->open_flags |= BDRV_O_RDWR;
}
static int update_header_sync(BlockDriverState *bs)
{
int ret;
ret = qcow2_update_header(bs);
if (ret < 0) {
return ret;
}
return bdrv_flush(bs);
}
static int ioreq_runio_qemu_aio(struct ioreq *ioreq)
{
struct XenBlkDev *blkdev = ioreq->blkdev;
if (ioreq_map(ioreq) == -1)
goto err;
ioreq->aio_inflight++;
if (ioreq->presync)
bdrv_flush(blkdev->bs); /* FIXME: aio_flush() ??? */
switch (ioreq->req.operation) {
case BLKIF_OP_READ:
ioreq->aio_inflight++;
bdrv_aio_readv(blkdev->bs, ioreq->start / BLOCK_SIZE,
&ioreq->v, ioreq->v.size / BLOCK_SIZE,
qemu_aio_complete, ioreq);
break;
case BLKIF_OP_WRITE:
case BLKIF_OP_WRITE_BARRIER:
ioreq->aio_inflight++;
bdrv_aio_writev(blkdev->bs, ioreq->start / BLOCK_SIZE,
&ioreq->v, ioreq->v.size / BLOCK_SIZE,
qemu_aio_complete, ioreq);
break;
default:
/* unknown operation (shouldn't happen -- parse catches this) */
goto err;
}
if (ioreq->postsync)
bdrv_flush(blkdev->bs); /* FIXME: aio_flush() ??? */
qemu_aio_complete(ioreq, 0);
return 0;
err:
ioreq->status = BLKIF_RSP_ERROR;
return -1;
}
static void bdrv_qed_close(BlockDriverState *bs)
{
BDRVQEDState *s = bs->opaque;
/* Ensure writes reach stable storage */
bdrv_flush(bs->file);
/* Clean shutdown, no check required on next open */
if (s->header.features & QED_F_NEED_CHECK) {
s->header.features &= ~QED_F_NEED_CHECK;
qed_write_header_sync(s);
}
qed_free_l2_cache(&s->l2_cache);
qemu_vfree(s->l1_table);
}
/**
* Mark an image clean once it passes check or has been repaired
*/
static void qed_check_mark_clean(BDRVQEDState *s, BdrvCheckResult *result)
{
/* Skip if there were unfixable corruptions or I/O errors */
if (result->corruptions > 0 || result->check_errors > 0) {
return;
}
/* Skip if image is already marked clean */
if (!(s->header.features & QED_F_NEED_CHECK)) {
return;
}
/* Ensure fixes reach storage before clearing check bit */
bdrv_flush(s->bs);
s->header.features &= ~QED_F_NEED_CHECK;
qed_write_header_sync(s);
}
static void qemu_aio_complete(void *opaque, int ret)
{
struct ioreq *ioreq = opaque;
if (ret != 0) {
xen_be_printf(&ioreq->blkdev->xendev, 0, "%s I/O error\n",
ioreq->req.operation == BLKIF_OP_READ ? "read" : "write");
ioreq->aio_errors++;
}
ioreq->aio_inflight--;
if (ioreq->aio_inflight > 0)
return;
if (ioreq->postsync)
bdrv_flush(ioreq->blkdev->bs);
ioreq->status = ioreq->aio_errors ? BLKIF_RSP_ERROR : BLKIF_RSP_OKAY;
ioreq_unmap(ioreq);
ioreq_finish(ioreq);
qemu_bh_schedule(ioreq->blkdev->bh);
}
int qcow2_cache_flush(BlockDriverState *bs, Qcow2Cache *c)
{
int result = 0;
int ret;
int i;
for (i = 0; i < c->size; i++) {
ret = qcow2_cache_entry_flush(bs, c, i);
if (ret < 0 && result != -ENOSPC) {
result = ret;
}
}
if (result == 0) {
ret = bdrv_flush(bs->file);
if (ret < 0) {
result = ret;
}
}
return result;
}
static int qcow2_cache_entry_flush(BlockDriverState *bs, Qcow2Cache *c, int i)
{
BDRVQcowState *s = bs->opaque;
int ret = 0;
if (!c->entries[i].dirty || !c->entries[i].offset) {
return 0;
}
if (c->depends) {
ret = qcow2_cache_flush_dependency(bs, c);
} else if (c->depends_on_flush) {
ret = bdrv_flush(bs->file);
if (ret >= 0) {
c->depends_on_flush = false;
}
}
if (ret < 0) {
return ret;
}
if (c == s->refcount_block_cache) {
BLKDBG_EVENT(bs->file, BLKDBG_REFBLOCK_UPDATE_PART);
} else if (c == s->l2_table_cache) {
BLKDBG_EVENT(bs->file, BLKDBG_L2_UPDATE);
}
ret = bdrv_pwrite(bs->file, c->entries[i].offset, c->entries[i].table,
s->cluster_size);
if (ret < 0) {
return ret;
}
c->entries[i].dirty = false;
return 0;
}
static void coroutine_fn mirror_run(void *opaque)
{
MirrorBlockJob *s = opaque;
MirrorExitData *data;
BlockDriverState *bs = s->common.bs;
int64_t sector_num, end, sectors_per_chunk, length;
uint64_t last_pause_ns;
BlockDriverInfo bdi;
char backing_filename[2]; /* we only need 2 characters because we are only
checking for a NULL string */
int ret = 0;
int n;
if (block_job_is_cancelled(&s->common)) {
goto immediate_exit;
}
s->bdev_length = bdrv_getlength(bs);
if (s->bdev_length < 0) {
ret = s->bdev_length;
goto immediate_exit;
} else if (s->bdev_length == 0) {
/* Report BLOCK_JOB_READY and wait for complete. */
block_job_event_ready(&s->common);
s->synced = true;
while (!block_job_is_cancelled(&s->common) && !s->should_complete) {
block_job_yield(&s->common);
}
s->common.cancelled = false;
goto immediate_exit;
}
length = DIV_ROUND_UP(s->bdev_length, s->granularity);
s->in_flight_bitmap = bitmap_new(length);
/* If we have no backing file yet in the destination, we cannot let
* the destination do COW. Instead, we copy sectors around the
* dirty data if needed. We need a bitmap to do that.
*/
bdrv_get_backing_filename(s->target, backing_filename,
sizeof(backing_filename));
if (backing_filename[0] && !s->target->backing_hd) {
ret = bdrv_get_info(s->target, &bdi);
if (ret < 0) {
goto immediate_exit;
}
if (s->granularity < bdi.cluster_size) {
s->buf_size = MAX(s->buf_size, bdi.cluster_size);
s->cow_bitmap = bitmap_new(length);
}
}
end = s->bdev_length / BDRV_SECTOR_SIZE;
s->buf = qemu_try_blockalign(bs, s->buf_size);
if (s->buf == NULL) {
ret = -ENOMEM;
goto immediate_exit;
}
sectors_per_chunk = s->granularity >> BDRV_SECTOR_BITS;
mirror_free_init(s);
if (!s->is_none_mode) {
/* First part, loop on the sectors and initialize the dirty bitmap. */
BlockDriverState *base = s->base;
for (sector_num = 0; sector_num < end; ) {
int64_t next = (sector_num | (sectors_per_chunk - 1)) + 1;
ret = bdrv_is_allocated_above(bs, base,
sector_num, next - sector_num, &n);
if (ret < 0) {
goto immediate_exit;
}
assert(n > 0);
if (ret == 1) {
bdrv_set_dirty_bitmap(s->dirty_bitmap, sector_num, n);
sector_num = next;
} else {
sector_num += n;
}
}
}
bdrv_dirty_iter_init(s->dirty_bitmap, &s->hbi);
last_pause_ns = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
for (;;) {
uint64_t delay_ns = 0;
int64_t cnt;
bool should_complete;
if (s->ret < 0) {
ret = s->ret;
goto immediate_exit;
}
cnt = bdrv_get_dirty_count(s->dirty_bitmap);
/* s->common.offset contains the number of bytes already processed so
* far, cnt is the number of dirty sectors remaining and
* s->sectors_in_flight is the number of sectors currently being
* processed; together those are the current total operation length */
s->common.len = s->common.offset +
(cnt + s->sectors_in_flight) * BDRV_SECTOR_SIZE;
/* Note that even when no rate limit is applied we need to yield
* periodically with no pending I/O so that bdrv_drain_all() returns.
* We do so every SLICE_TIME nanoseconds, or when there is an error,
* or when the source is clean, whichever comes first.
*/
if (qemu_clock_get_ns(QEMU_CLOCK_REALTIME) - last_pause_ns < SLICE_TIME &&
s->common.iostatus == BLOCK_DEVICE_IO_STATUS_OK) {
if (s->in_flight == MAX_IN_FLIGHT || s->buf_free_count == 0 ||
(cnt == 0 && s->in_flight > 0)) {
trace_mirror_yield(s, s->in_flight, s->buf_free_count, cnt);
qemu_coroutine_yield();
continue;
} else if (cnt != 0) {
delay_ns = mirror_iteration(s);
}
}
should_complete = false;
if (s->in_flight == 0 && cnt == 0) {
trace_mirror_before_flush(s);
ret = bdrv_flush(s->target);
if (ret < 0) {
if (mirror_error_action(s, false, -ret) ==
BLOCK_ERROR_ACTION_REPORT) {
goto immediate_exit;
}
} else {
/* We're out of the streaming phase. From now on, if the job
* is cancelled we will actually complete all pending I/O and
* report completion. This way, block-job-cancel will leave
* the target in a consistent state.
*/
if (!s->synced) {
block_job_event_ready(&s->common);
s->synced = true;
}
should_complete = s->should_complete ||
block_job_is_cancelled(&s->common);
cnt = bdrv_get_dirty_count(s->dirty_bitmap);
}
}
if (cnt == 0 && should_complete) {
/* The dirty bitmap is not updated while operations are pending.
* If we're about to exit, wait for pending operations before
* calling bdrv_get_dirty_count(bs), or we may exit while the
* source has dirty data to copy!
*
* Note that I/O can be submitted by the guest while
* mirror_populate runs.
*/
trace_mirror_before_drain(s, cnt);
bdrv_drain(bs);
cnt = bdrv_get_dirty_count(s->dirty_bitmap);
}
ret = 0;
trace_mirror_before_sleep(s, cnt, s->synced, delay_ns);
if (!s->synced) {
block_job_sleep_ns(&s->common, QEMU_CLOCK_REALTIME, delay_ns);
if (block_job_is_cancelled(&s->common)) {
break;
}
} else if (!should_complete) {
delay_ns = (s->in_flight == 0 && cnt == 0 ? SLICE_TIME : 0);
block_job_sleep_ns(&s->common, QEMU_CLOCK_REALTIME, delay_ns);
} else if (cnt == 0) {
/* The two disks are in sync. Exit and report successful
* completion.
*/
assert(QLIST_EMPTY(&bs->tracked_requests));
s->common.cancelled = false;
break;
}
last_pause_ns = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
}
immediate_exit:
if (s->in_flight > 0) {
/* We get here only if something went wrong. Either the job failed,
* or it was cancelled prematurely so that we do not guarantee that
* the target is a copy of the source.
*/
assert(ret < 0 || (!s->synced && block_job_is_cancelled(&s->common)));
mirror_drain(s);
}
assert(s->in_flight == 0);
qemu_vfree(s->buf);
g_free(s->cow_bitmap);
g_free(s->in_flight_bitmap);
bdrv_release_dirty_bitmap(bs, s->dirty_bitmap);
bdrv_iostatus_disable(s->target);
data = g_malloc(sizeof(*data));
data->ret = ret;
block_job_defer_to_main_loop(&s->common, mirror_exit, data);
}
static int ioreq_runio_qemu_sync(struct ioreq *ioreq)
{
struct XenBlkDev *blkdev = ioreq->blkdev;
int i, rc, len = 0;
off_t pos;
if (ioreq->req.nr_segments && ioreq_map(ioreq) == -1)
goto err;
if (ioreq->presync)
bdrv_flush(blkdev->bs);
switch (ioreq->req.operation) {
case BLKIF_OP_READ:
pos = ioreq->start;
for (i = 0; i < ioreq->v.niov; i++) {
rc = bdrv_read(blkdev->bs, pos / BLOCK_SIZE,
ioreq->v.iov[i].iov_base,
ioreq->v.iov[i].iov_len / BLOCK_SIZE);
if (rc != 0) {
xen_be_printf(&blkdev->xendev, 0, "rd I/O error (%p, len %zd)\n",
ioreq->v.iov[i].iov_base,
ioreq->v.iov[i].iov_len);
goto err;
}
len += ioreq->v.iov[i].iov_len;
pos += ioreq->v.iov[i].iov_len;
}
break;
case BLKIF_OP_WRITE:
case BLKIF_OP_WRITE_BARRIER:
if (!ioreq->req.nr_segments)
break;
pos = ioreq->start;
for (i = 0; i < ioreq->v.niov; i++) {
rc = bdrv_write(blkdev->bs, pos / BLOCK_SIZE,
ioreq->v.iov[i].iov_base,
ioreq->v.iov[i].iov_len / BLOCK_SIZE);
if (rc != 0) {
xen_be_printf(&blkdev->xendev, 0, "wr I/O error (%p, len %zd)\n",
ioreq->v.iov[i].iov_base,
ioreq->v.iov[i].iov_len);
goto err;
}
len += ioreq->v.iov[i].iov_len;
pos += ioreq->v.iov[i].iov_len;
}
break;
default:
/* unknown operation (shouldn't happen -- parse catches this) */
goto err;
}
if (ioreq->postsync)
bdrv_flush(blkdev->bs);
ioreq->status = BLKIF_RSP_OKAY;
ioreq_unmap(ioreq);
ioreq_finish(ioreq);
return 0;
err:
ioreq->status = BLKIF_RSP_ERROR;
return -1;
}
static int bdrv_qed_open(BlockDriverState *bs, QDict *options, int flags,
Error **errp)
{
BDRVQEDState *s = bs->opaque;
QEDHeader le_header;
int64_t file_size;
int ret;
s->bs = bs;
QSIMPLEQ_INIT(&s->allocating_write_reqs);
ret = bdrv_pread(bs->file, 0, &le_header, sizeof(le_header));
if (ret < 0) {
return ret;
}
qed_header_le_to_cpu(&le_header, &s->header);
if (s->header.magic != QED_MAGIC) {
error_setg(errp, "Image not in QED format");
return -EINVAL;
}
if (s->header.features & ~QED_FEATURE_MASK) {
/* image uses unsupported feature bits */
char buf[64];
snprintf(buf, sizeof(buf), "%" PRIx64,
s->header.features & ~QED_FEATURE_MASK);
error_set(errp, QERR_UNKNOWN_BLOCK_FORMAT_FEATURE,
bs->device_name, "QED", buf);
return -ENOTSUP;
}
if (!qed_is_cluster_size_valid(s->header.cluster_size)) {
return -EINVAL;
}
/* Round down file size to the last cluster */
file_size = bdrv_getlength(bs->file);
if (file_size < 0) {
return file_size;
}
s->file_size = qed_start_of_cluster(s, file_size);
if (!qed_is_table_size_valid(s->header.table_size)) {
return -EINVAL;
}
if (!qed_is_image_size_valid(s->header.image_size,
s->header.cluster_size,
s->header.table_size)) {
return -EINVAL;
}
if (!qed_check_table_offset(s, s->header.l1_table_offset)) {
return -EINVAL;
}
s->table_nelems = (s->header.cluster_size * s->header.table_size) /
sizeof(uint64_t);
s->l2_shift = ffs(s->header.cluster_size) - 1;
s->l2_mask = s->table_nelems - 1;
s->l1_shift = s->l2_shift + ffs(s->table_nelems) - 1;
if ((s->header.features & QED_F_BACKING_FILE)) {
if ((uint64_t)s->header.backing_filename_offset +
s->header.backing_filename_size >
s->header.cluster_size * s->header.header_size) {
return -EINVAL;
}
ret = qed_read_string(bs->file, s->header.backing_filename_offset,
s->header.backing_filename_size, bs->backing_file,
sizeof(bs->backing_file));
if (ret < 0) {
return ret;
}
if (s->header.features & QED_F_BACKING_FORMAT_NO_PROBE) {
pstrcpy(bs->backing_format, sizeof(bs->backing_format), "raw");
}
}
/* Reset unknown autoclear feature bits. This is a backwards
* compatibility mechanism that allows images to be opened by older
* programs, which "knock out" unknown feature bits. When an image is
* opened by a newer program again it can detect that the autoclear
* feature is no longer valid.
*/
if ((s->header.autoclear_features & ~QED_AUTOCLEAR_FEATURE_MASK) != 0 &&
!bdrv_is_read_only(bs->file) && !(flags & BDRV_O_INCOMING)) {
s->header.autoclear_features &= QED_AUTOCLEAR_FEATURE_MASK;
ret = qed_write_header_sync(s);
if (ret) {
return ret;
}
/* From here on only known autoclear feature bits are valid */
bdrv_flush(bs->file);
}
s->l1_table = qed_alloc_table(s);
qed_init_l2_cache(&s->l2_cache);
ret = qed_read_l1_table_sync(s);
if (ret) {
goto out;
}
/* If image was not closed cleanly, check consistency */
if (!(flags & BDRV_O_CHECK) && (s->header.features & QED_F_NEED_CHECK)) {
/* Read-only images cannot be fixed. There is no risk of corruption
* since write operations are not possible. Therefore, allow
* potentially inconsistent images to be opened read-only. This can
* aid data recovery from an otherwise inconsistent image.
*/
if (!bdrv_is_read_only(bs->file) &&
!(flags & BDRV_O_INCOMING)) {
BdrvCheckResult result = {0};
ret = qed_check(s, &result, true);
if (ret) {
goto out;
}
}
}
bdrv_qed_attach_aio_context(bs, bdrv_get_aio_context(bs));
out:
if (ret) {
qed_free_l2_cache(&s->l2_cache);
qemu_vfree(s->l1_table);
}
return ret;
}
static int raw_flush(BlockDriverState *bs)
{
return bdrv_flush(bs->file);
}
/* add at the end of the file a new list of snapshots */
static int qcow2_write_snapshots(BlockDriverState *bs)
{
BDRVQcowState *s = bs->opaque;
QCowSnapshot *sn;
QCowSnapshotHeader h;
int i, name_size, id_str_size, snapshots_size;
uint64_t data64;
uint32_t data32;
int64_t offset, snapshots_offset;
/* compute the size of the snapshots */
offset = 0;
for(i = 0; i < s->nb_snapshots; i++) {
sn = s->snapshots + i;
offset = align_offset(offset, 8);
offset += sizeof(h);
offset += strlen(sn->id_str);
offset += strlen(sn->name);
}
snapshots_size = offset;
snapshots_offset = qcow2_alloc_clusters(bs, snapshots_size);
bdrv_flush(bs->file);
offset = snapshots_offset;
if (offset < 0) {
return offset;
}
for(i = 0; i < s->nb_snapshots; i++) {
sn = s->snapshots + i;
memset(&h, 0, sizeof(h));
h.l1_table_offset = cpu_to_be64(sn->l1_table_offset);
h.l1_size = cpu_to_be32(sn->l1_size);
h.vm_state_size = cpu_to_be32(sn->vm_state_size);
h.date_sec = cpu_to_be32(sn->date_sec);
h.date_nsec = cpu_to_be32(sn->date_nsec);
h.vm_clock_nsec = cpu_to_be64(sn->vm_clock_nsec);
id_str_size = strlen(sn->id_str);
name_size = strlen(sn->name);
h.id_str_size = cpu_to_be16(id_str_size);
h.name_size = cpu_to_be16(name_size);
offset = align_offset(offset, 8);
if (bdrv_pwrite_sync(bs->file, offset, &h, sizeof(h)) < 0)
goto fail;
offset += sizeof(h);
if (bdrv_pwrite_sync(bs->file, offset, sn->id_str, id_str_size) < 0)
goto fail;
offset += id_str_size;
if (bdrv_pwrite_sync(bs->file, offset, sn->name, name_size) < 0)
goto fail;
offset += name_size;
}
/* update the various header fields */
data64 = cpu_to_be64(snapshots_offset);
if (bdrv_pwrite_sync(bs->file, offsetof(QCowHeader, snapshots_offset),
&data64, sizeof(data64)) < 0)
goto fail;
data32 = cpu_to_be32(s->nb_snapshots);
if (bdrv_pwrite_sync(bs->file, offsetof(QCowHeader, nb_snapshots),
&data32, sizeof(data32)) < 0)
goto fail;
/* free the old snapshot table */
qcow2_free_clusters(bs, s->snapshots_offset, s->snapshots_size);
s->snapshots_offset = snapshots_offset;
s->snapshots_size = snapshots_size;
return 0;
fail:
return -1;
}
static int bdrv_fclose(void *opaque)
{
return bdrv_flush(opaque);
}
/* if no id is provided, a new one is constructed */
int qcow2_snapshot_create(BlockDriverState *bs, QEMUSnapshotInfo *sn_info)
{
BDRVQcowState *s = bs->opaque;
QCowSnapshot *snapshots1, sn1, *sn = &sn1;
int i, ret;
uint64_t *l1_table = NULL;
int64_t l1_table_offset;
memset(sn, 0, sizeof(*sn));
if (sn_info->id_str[0] == '\0') {
/* compute a new id */
find_new_snapshot_id(bs, sn_info->id_str, sizeof(sn_info->id_str));
}
/* check that the ID is unique */
if (find_snapshot_by_id(bs, sn_info->id_str) >= 0)
return -ENOENT;
sn->id_str = g_strdup(sn_info->id_str);
if (!sn->id_str)
goto fail;
sn->name = g_strdup(sn_info->name);
if (!sn->name)
goto fail;
sn->vm_state_size = sn_info->vm_state_size;
sn->date_sec = sn_info->date_sec;
sn->date_nsec = sn_info->date_nsec;
sn->vm_clock_nsec = sn_info->vm_clock_nsec;
ret = qcow2_update_snapshot_refcount(bs, s->l1_table_offset, s->l1_size, 1);
if (ret < 0)
goto fail;
/* create the L1 table of the snapshot */
l1_table_offset = qcow2_alloc_clusters(bs, s->l1_size * sizeof(uint64_t));
if (l1_table_offset < 0) {
goto fail;
}
bdrv_flush(bs->file);
sn->l1_table_offset = l1_table_offset;
sn->l1_size = s->l1_size;
if (s->l1_size != 0) {
l1_table = g_malloc(s->l1_size * sizeof(uint64_t));
} else {
l1_table = NULL;
}
for(i = 0; i < s->l1_size; i++) {
l1_table[i] = cpu_to_be64(s->l1_table[i]);
}
if (bdrv_pwrite_sync(bs->file, sn->l1_table_offset,
l1_table, s->l1_size * sizeof(uint64_t)) < 0)
goto fail;
g_free(l1_table);
l1_table = NULL;
snapshots1 = g_malloc((s->nb_snapshots + 1) * sizeof(QCowSnapshot));
if (s->snapshots) {
memcpy(snapshots1, s->snapshots, s->nb_snapshots * sizeof(QCowSnapshot));
g_free(s->snapshots);
}
s->snapshots = snapshots1;
s->snapshots[s->nb_snapshots++] = *sn;
if (qcow2_write_snapshots(bs) < 0)
goto fail;
#ifdef DEBUG_ALLOC
qcow2_check_refcounts(bs);
#endif
return 0;
fail:
g_free(sn->name);
g_free(l1_table);
return -1;
}
static void coroutine_fn mirror_run(void *opaque)
{
MirrorBlockJob *s = opaque;
BlockDriverState *bs = s->common.bs;
int64_t sector_num, end, sectors_per_chunk, length;
uint64_t last_pause_ns;
BlockDriverInfo bdi;
char backing_filename[1024];
int ret = 0;
int n;
if (block_job_is_cancelled(&s->common)) {
goto immediate_exit;
}
s->common.len = bdrv_getlength(bs);
if (s->common.len <= 0) {
block_job_completed(&s->common, s->common.len);
return;
}
length = (bdrv_getlength(bs) + s->granularity - 1) / s->granularity;
s->in_flight_bitmap = bitmap_new(length);
/* If we have no backing file yet in the destination, we cannot let
* the destination do COW. Instead, we copy sectors around the
* dirty data if needed. We need a bitmap to do that.
*/
bdrv_get_backing_filename(s->target, backing_filename,
sizeof(backing_filename));
if (backing_filename[0] && !s->target->backing_hd) {
bdrv_get_info(s->target, &bdi);
if (s->granularity < bdi.cluster_size) {
s->buf_size = MAX(s->buf_size, bdi.cluster_size);
s->cow_bitmap = bitmap_new(length);
}
}
end = s->common.len >> BDRV_SECTOR_BITS;
s->buf = qemu_blockalign(bs, s->buf_size);
sectors_per_chunk = s->granularity >> BDRV_SECTOR_BITS;
mirror_free_init(s);
if (s->mode != MIRROR_SYNC_MODE_NONE) {
/* First part, loop on the sectors and initialize the dirty bitmap. */
BlockDriverState *base;
base = s->mode == MIRROR_SYNC_MODE_FULL ? NULL : bs->backing_hd;
for (sector_num = 0; sector_num < end; ) {
int64_t next = (sector_num | (sectors_per_chunk - 1)) + 1;
ret = bdrv_is_allocated_above(bs, base,
sector_num, next - sector_num, &n);
if (ret < 0) {
goto immediate_exit;
}
assert(n > 0);
if (ret == 1) {
bdrv_set_dirty(bs, sector_num, n);
sector_num = next;
} else {
sector_num += n;
}
}
}
bdrv_dirty_iter_init(bs, &s->hbi);
last_pause_ns = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
for (;;) {
uint64_t delay_ns;
int64_t cnt;
bool should_complete;
if (s->ret < 0) {
ret = s->ret;
goto immediate_exit;
}
cnt = bdrv_get_dirty_count(bs);
/* Note that even when no rate limit is applied we need to yield
* periodically with no pending I/O so that qemu_aio_flush() returns.
* We do so every SLICE_TIME nanoseconds, or when there is an error,
* or when the source is clean, whichever comes first.
*/
if (qemu_clock_get_ns(QEMU_CLOCK_REALTIME) - last_pause_ns < SLICE_TIME &&
s->common.iostatus == BLOCK_DEVICE_IO_STATUS_OK) {
if (s->in_flight == MAX_IN_FLIGHT || s->buf_free_count == 0 ||
(cnt == 0 && s->in_flight > 0)) {
trace_mirror_yield(s, s->in_flight, s->buf_free_count, cnt);
qemu_coroutine_yield();
continue;
} else if (cnt != 0) {
mirror_iteration(s);
continue;
}
}
should_complete = false;
if (s->in_flight == 0 && cnt == 0) {
trace_mirror_before_flush(s);
ret = bdrv_flush(s->target);
if (ret < 0) {
if (mirror_error_action(s, false, -ret) == BDRV_ACTION_REPORT) {
goto immediate_exit;
}
} else {
/* We're out of the streaming phase. From now on, if the job
* is cancelled we will actually complete all pending I/O and
* report completion. This way, block-job-cancel will leave
* the target in a consistent state.
*/
s->common.offset = end * BDRV_SECTOR_SIZE;
if (!s->synced) {
block_job_ready(&s->common);
s->synced = true;
}
should_complete = s->should_complete ||
block_job_is_cancelled(&s->common);
cnt = bdrv_get_dirty_count(bs);
}
}
if (cnt == 0 && should_complete) {
/* The dirty bitmap is not updated while operations are pending.
* If we're about to exit, wait for pending operations before
* calling bdrv_get_dirty_count(bs), or we may exit while the
* source has dirty data to copy!
*
* Note that I/O can be submitted by the guest while
* mirror_populate runs.
*/
trace_mirror_before_drain(s, cnt);
bdrv_drain_all();
cnt = bdrv_get_dirty_count(bs);
}
ret = 0;
trace_mirror_before_sleep(s, cnt, s->synced);
if (!s->synced) {
/* Publish progress */
s->common.offset = (end - cnt) * BDRV_SECTOR_SIZE;
if (s->common.speed) {
delay_ns = ratelimit_calculate_delay(&s->limit, sectors_per_chunk);
} else {
delay_ns = 0;
}
block_job_sleep_ns(&s->common, QEMU_CLOCK_REALTIME, delay_ns);
if (block_job_is_cancelled(&s->common)) {
break;
}
} else if (!should_complete) {
delay_ns = (s->in_flight == 0 && cnt == 0 ? SLICE_TIME : 0);
block_job_sleep_ns(&s->common, QEMU_CLOCK_REALTIME, delay_ns);
} else if (cnt == 0) {
/* The two disks are in sync. Exit and report successful
* completion.
*/
assert(QLIST_EMPTY(&bs->tracked_requests));
s->common.cancelled = false;
break;
}
last_pause_ns = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
}
immediate_exit:
if (s->in_flight > 0) {
/* We get here only if something went wrong. Either the job failed,
* or it was cancelled prematurely so that we do not guarantee that
* the target is a copy of the source.
*/
assert(ret < 0 || (!s->synced && block_job_is_cancelled(&s->common)));
mirror_drain(s);
}
assert(s->in_flight == 0);
qemu_vfree(s->buf);
g_free(s->cow_bitmap);
g_free(s->in_flight_bitmap);
bdrv_set_dirty_tracking(bs, 0);
bdrv_iostatus_disable(s->target);
if (s->should_complete && ret == 0) {
if (bdrv_get_flags(s->target) != bdrv_get_flags(s->common.bs)) {
bdrv_reopen(s->target, bdrv_get_flags(s->common.bs), NULL);
}
bdrv_swap(s->target, s->common.bs);
}
bdrv_close(s->target);
bdrv_unref(s->target);
block_job_completed(&s->common, ret);
}
/* add at the end of the file a new list of snapshots */
static int qcow2_write_snapshots(BlockDriverState *bs)
{
BDRVQcowState *s = bs->opaque;
QCowSnapshot *sn;
QCowSnapshotHeader h;
QCowSnapshotExtraData extra;
int i, name_size, id_str_size, snapshots_size;
struct {
uint32_t nb_snapshots;
uint64_t snapshots_offset;
} QEMU_PACKED header_data;
int64_t offset, snapshots_offset;
int ret;
/* compute the size of the snapshots */
offset = 0;
for(i = 0; i < s->nb_snapshots; i++) {
sn = s->snapshots + i;
offset = align_offset(offset, 8);
offset += sizeof(h);
offset += sizeof(extra);
offset += strlen(sn->id_str);
offset += strlen(sn->name);
}
snapshots_size = offset;
/* Allocate space for the new snapshot list */
snapshots_offset = qcow2_alloc_clusters(bs, snapshots_size);
offset = snapshots_offset;
if (offset < 0) {
return offset;
}
ret = bdrv_flush(bs);
if (ret < 0) {
return ret;
}
/* The snapshot list position has not yet been updated, so these clusters
* must indeed be completely free */
ret = qcow2_pre_write_overlap_check(bs, QCOW2_OL_DEFAULT, offset,
s->snapshots_size);
if (ret < 0) {
return ret;
}
/* Write all snapshots to the new list */
for(i = 0; i < s->nb_snapshots; i++) {
sn = s->snapshots + i;
memset(&h, 0, sizeof(h));
h.l1_table_offset = cpu_to_be64(sn->l1_table_offset);
h.l1_size = cpu_to_be32(sn->l1_size);
/* If it doesn't fit in 32 bit, older implementations should treat it
* as a disk-only snapshot rather than truncate the VM state */
if (sn->vm_state_size <= 0xffffffff) {
h.vm_state_size = cpu_to_be32(sn->vm_state_size);
}
h.date_sec = cpu_to_be32(sn->date_sec);
h.date_nsec = cpu_to_be32(sn->date_nsec);
h.vm_clock_nsec = cpu_to_be64(sn->vm_clock_nsec);
h.extra_data_size = cpu_to_be32(sizeof(extra));
memset(&extra, 0, sizeof(extra));
extra.vm_state_size_large = cpu_to_be64(sn->vm_state_size);
extra.disk_size = cpu_to_be64(sn->disk_size);
id_str_size = strlen(sn->id_str);
name_size = strlen(sn->name);
h.id_str_size = cpu_to_be16(id_str_size);
h.name_size = cpu_to_be16(name_size);
offset = align_offset(offset, 8);
ret = bdrv_pwrite(bs->file, offset, &h, sizeof(h));
if (ret < 0) {
goto fail;
}
offset += sizeof(h);
ret = bdrv_pwrite(bs->file, offset, &extra, sizeof(extra));
if (ret < 0) {
goto fail;
}
offset += sizeof(extra);
ret = bdrv_pwrite(bs->file, offset, sn->id_str, id_str_size);
if (ret < 0) {
goto fail;
}
offset += id_str_size;
ret = bdrv_pwrite(bs->file, offset, sn->name, name_size);
if (ret < 0) {
goto fail;
}
offset += name_size;
}
/*
* Update the header to point to the new snapshot table. This requires the
* new table and its refcounts to be stable on disk.
*/
ret = bdrv_flush(bs);
if (ret < 0) {
goto fail;
}
QEMU_BUILD_BUG_ON(offsetof(QCowHeader, snapshots_offset) !=
offsetof(QCowHeader, nb_snapshots) + sizeof(header_data.nb_snapshots));
header_data.nb_snapshots = cpu_to_be32(s->nb_snapshots);
header_data.snapshots_offset = cpu_to_be64(snapshots_offset);
ret = bdrv_pwrite_sync(bs->file, offsetof(QCowHeader, nb_snapshots),
&header_data, sizeof(header_data));
if (ret < 0) {
goto fail;
}
/* free the old snapshot table */
qcow2_free_clusters(bs, s->snapshots_offset, s->snapshots_size,
QCOW2_DISCARD_SNAPSHOT);
s->snapshots_offset = snapshots_offset;
s->snapshots_size = snapshots_size;
return 0;
fail:
return ret;
}
static int flush_f(int argc, char **argv)
{
bdrv_flush(bs);
return 0;
}
/* if no id is provided, a new one is constructed */
int qcow2_snapshot_create(BlockDriverState *bs, QEMUSnapshotInfo *sn_info)
{
BDRVQcowState *s = bs->opaque;
QCowSnapshot *new_snapshot_list = NULL;
QCowSnapshot *old_snapshot_list = NULL;
QCowSnapshot sn1, *sn = &sn1;
int i, ret;
uint64_t *l1_table = NULL;
int64_t l1_table_offset;
memset(sn, 0, sizeof(*sn));
/* Generate an ID if it wasn't passed */
if (sn_info->id_str[0] == '\0') {
find_new_snapshot_id(bs, sn_info->id_str, sizeof(sn_info->id_str));
}
/* Check that the ID is unique */
if (find_snapshot_by_id(bs, sn_info->id_str) >= 0) {
return -EEXIST;
}
/* Populate sn with passed data */
sn->id_str = g_strdup(sn_info->id_str);
sn->name = g_strdup(sn_info->name);
sn->disk_size = bs->total_sectors * BDRV_SECTOR_SIZE;
sn->vm_state_size = sn_info->vm_state_size;
sn->date_sec = sn_info->date_sec;
sn->date_nsec = sn_info->date_nsec;
sn->vm_clock_nsec = sn_info->vm_clock_nsec;
/* Allocate the L1 table of the snapshot and copy the current one there. */
l1_table_offset = qcow2_alloc_clusters(bs, s->l1_size * sizeof(uint64_t));
if (l1_table_offset < 0) {
ret = l1_table_offset;
goto fail;
}
sn->l1_table_offset = l1_table_offset;
sn->l1_size = s->l1_size;
l1_table = g_malloc(s->l1_size * sizeof(uint64_t));
for(i = 0; i < s->l1_size; i++) {
l1_table[i] = cpu_to_be64(s->l1_table[i]);
}
ret = bdrv_pwrite(bs->file, sn->l1_table_offset, l1_table,
s->l1_size * sizeof(uint64_t));
if (ret < 0) {
goto fail;
}
g_free(l1_table);
l1_table = NULL;
/*
* Increase the refcounts of all clusters and make sure everything is
* stable on disk before updating the snapshot table to contain a pointer
* to the new L1 table.
*/
ret = qcow2_update_snapshot_refcount(bs, s->l1_table_offset, s->l1_size, 1);
if (ret < 0) {
goto fail;
}
ret = bdrv_flush(bs);
if (ret < 0) {
goto fail;
}
/* Append the new snapshot to the snapshot list */
new_snapshot_list = g_malloc((s->nb_snapshots + 1) * sizeof(QCowSnapshot));
if (s->snapshots) {
memcpy(new_snapshot_list, s->snapshots,
s->nb_snapshots * sizeof(QCowSnapshot));
old_snapshot_list = s->snapshots;
}
s->snapshots = new_snapshot_list;
s->snapshots[s->nb_snapshots++] = *sn;
ret = qcow2_write_snapshots(bs);
if (ret < 0) {
g_free(s->snapshots);
s->snapshots = old_snapshot_list;
goto fail;
}
g_free(old_snapshot_list);
#ifdef DEBUG_ALLOC
{
BdrvCheckResult result = {0};
qcow2_check_refcounts(bs, &result);
}
#endif
return 0;
fail:
g_free(sn->id_str);
g_free(sn->name);
g_free(l1_table);
return ret;
}
