/* commit COW file into the raw image */
int bdrv_commit(BlockDriverState *bs)
{
BlockBackend *src, *backing;
BlockDriverState *backing_file_bs = NULL;
BlockDriverState *commit_top_bs = NULL;
BlockDriver *drv = bs->drv;
int64_t offset, length, backing_length;
int ro;
int64_t n;
int ret = 0;
uint8_t *buf = NULL;
Error *local_err = NULL;
if (!drv)
return -ENOMEDIUM;
if (!bs->backing) {
return -ENOTSUP;
}
if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_COMMIT_SOURCE, NULL) ||
bdrv_op_is_blocked(bs->backing->bs, BLOCK_OP_TYPE_COMMIT_TARGET, NULL)) {
return -EBUSY;
}
ro = bs->backing->bs->read_only;
if (ro) {
if (bdrv_reopen_set_read_only(bs->backing->bs, false, NULL)) {
return -EACCES;
}
}
src = blk_new(BLK_PERM_CONSISTENT_READ, BLK_PERM_ALL);
backing = blk_new(BLK_PERM_WRITE | BLK_PERM_RESIZE, BLK_PERM_ALL);
ret = blk_insert_bs(src, bs, &local_err);
if (ret < 0) {
error_report_err(local_err);
goto ro_cleanup;
}
/* Insert commit_top block node above backing, so we can write to it */
backing_file_bs = backing_bs(bs);
commit_top_bs = bdrv_new_open_driver(&bdrv_commit_top, NULL, BDRV_O_RDWR,
&local_err);
if (commit_top_bs == NULL) {
error_report_err(local_err);
goto ro_cleanup;
}
bdrv_set_aio_context(commit_top_bs, bdrv_get_aio_context(backing_file_bs));
bdrv_set_backing_hd(commit_top_bs, backing_file_bs, &error_abort);
bdrv_set_backing_hd(bs, commit_top_bs, &error_abort);
ret = blk_insert_bs(backing, backing_file_bs, &local_err);
if (ret < 0) {
error_report_err(local_err);
goto ro_cleanup;
}
length = blk_getlength(src);
if (length < 0) {
ret = length;
goto ro_cleanup;
}
backing_length = blk_getlength(backing);
if (backing_length < 0) {
ret = backing_length;
goto ro_cleanup;
}
/* If our top snapshot is larger than the backing file image,
* grow the backing file image if possible. If not possible,
* we must return an error */
if (length > backing_length) {
ret = blk_truncate(backing, length, PREALLOC_MODE_OFF, &local_err);
if (ret < 0) {
error_report_err(local_err);
goto ro_cleanup;
}
}
/* blk_try_blockalign() for src will choose an alignment that works for
* backing as well, so no need to compare the alignment manually. */
buf = blk_try_blockalign(src, COMMIT_BUF_SIZE);
if (buf == NULL) {
ret = -ENOMEM;
goto ro_cleanup;
}
for (offset = 0; offset < length; offset += n) {
ret = bdrv_is_allocated(bs, offset, COMMIT_BUF_SIZE, &n);
if (ret < 0) {
goto ro_cleanup;
}
if (ret) {
ret = blk_pread(src, offset, buf, n);
if (ret < 0) {
goto ro_cleanup;
}
ret = blk_pwrite(backing, offset, buf, n, 0);
if (ret < 0) {
goto ro_cleanup;
}
}
}
if (drv->bdrv_make_empty) {
ret = drv->bdrv_make_empty(bs);
if (ret < 0) {
goto ro_cleanup;
}
blk_flush(src);
}
/*
* Make sure all data we wrote to the backing device is actually
* stable on disk.
*/
blk_flush(backing);
ret = 0;
ro_cleanup:
qemu_vfree(buf);
blk_unref(backing);
if (backing_file_bs) {
bdrv_set_backing_hd(bs, backing_file_bs, &error_abort);
}
bdrv_unref(commit_top_bs);
blk_unref(src);
if (ro) {
/* ignoring error return here */
bdrv_reopen_set_read_only(bs->backing->bs, true, NULL);
}
return ret;
}
static void handle_notify(EventNotifier *e)
{
VirtIOBlockDataPlane *s = container_of(e, VirtIOBlockDataPlane,
host_notifier);
VirtIOBlock *vblk = VIRTIO_BLK(s->vdev);
event_notifier_test_and_clear(&s->host_notifier);
bdrv_io_plug(s->blk->conf.bs);
for (;;) {
MultiReqBuffer mrb = {
.num_writes = 0,
};
int ret;
/* Disable guest->host notifies to avoid unnecessary vmexits */
vring_disable_notification(s->vdev, &s->vring);
for (;;) {
VirtIOBlockReq *req = virtio_blk_alloc_request(vblk);
ret = vring_pop(s->vdev, &s->vring, &req->elem);
if (ret < 0) {
virtio_blk_free_request(req);
break; /* no more requests */
}
trace_virtio_blk_data_plane_process_request(s, req->elem.out_num,
req->elem.in_num,
req->elem.index);
virtio_blk_handle_request(req, &mrb);
}
virtio_submit_multiwrite(s->blk->conf.bs, &mrb);
if (likely(ret == -EAGAIN)) { /* vring emptied */
/* Re-enable guest->host notifies and stop processing the vring.
* But if the guest has snuck in more descriptors, keep processing.
*/
if (vring_enable_notification(s->vdev, &s->vring)) {
break;
}
} else { /* fatal error */
break;
}
}
bdrv_io_unplug(s->blk->conf.bs);
}
/* Context: QEMU global mutex held */
void virtio_blk_data_plane_create(VirtIODevice *vdev, VirtIOBlkConf *blk,
VirtIOBlockDataPlane **dataplane,
Error **errp)
{
VirtIOBlockDataPlane *s;
Error *local_err = NULL;
BusState *qbus = BUS(qdev_get_parent_bus(DEVICE(vdev)));
VirtioBusClass *k = VIRTIO_BUS_GET_CLASS(qbus);
*dataplane = NULL;
if (!blk->data_plane && !blk->iothread) {
return;
}
/* Don't try if transport does not support notifiers. */
if (!k->set_guest_notifiers || !k->set_host_notifier) {
error_setg(errp,
"device is incompatible with x-data-plane "
"(transport does not support notifiers)");
return;
}
/* If dataplane is (re-)enabled while the guest is running there could be
* block jobs that can conflict.
*/
if (bdrv_op_is_blocked(blk->conf.bs, BLOCK_OP_TYPE_DATAPLANE, &local_err)) {
error_setg(errp, "cannot start dataplane thread: %s",
error_get_pretty(local_err));
error_free(local_err);
return;
}
s = g_new0(VirtIOBlockDataPlane, 1);
s->vdev = vdev;
s->blk = blk;
if (blk->iothread) {
s->iothread = blk->iothread;
object_ref(OBJECT(s->iothread));
} else {
/* Create per-device IOThread if none specified. This is for
* x-data-plane option compatibility. If x-data-plane is removed we
* can drop this.
*/
object_initialize(&s->internal_iothread_obj,
sizeof(s->internal_iothread_obj),
TYPE_IOTHREAD);
user_creatable_complete(OBJECT(&s->internal_iothread_obj), &error_abort);
s->iothread = &s->internal_iothread_obj;
}
s->ctx = iothread_get_aio_context(s->iothread);
s->bh = aio_bh_new(s->ctx, notify_guest_bh, s);
error_setg(&s->blocker, "block device is in use by data plane");
bdrv_op_block_all(blk->conf.bs, s->blocker);
bdrv_op_unblock(blk->conf.bs, BLOCK_OP_TYPE_RESIZE, s->blocker);
bdrv_op_unblock(blk->conf.bs, BLOCK_OP_TYPE_DRIVE_DEL, s->blocker);
*dataplane = s;
}
