u64 stripe_unit;
u64 stripe_count;
-
- u64 obj_version;
};
/*
static int rbd_dev_snaps_update(struct rbd_device *rbd_dev);
-static void rbd_dev_release(struct device *dev);
+static void rbd_dev_device_release(struct device *dev);
static void rbd_snap_destroy(struct rbd_snap *snap);
static ssize_t rbd_add(struct bus_type *bus, const char *buf,
# define rbd_assert(expr) ((void) 0)
#endif /* !RBD_DEBUG */
-static void rbd_img_parent_read(struct rbd_obj_request *obj_request);
static int rbd_img_obj_request_submit(struct rbd_obj_request *obj_request);
+static void rbd_img_parent_read(struct rbd_obj_request *obj_request);
+static void rbd_dev_remove_parent(struct rbd_device *rbd_dev);
static int rbd_dev_refresh(struct rbd_device *rbd_dev, u64 *hver);
static int rbd_dev_v2_refresh(struct rbd_device *rbd_dev, u64 *hver);
kfree(rbdc);
}
-/* Caller has to fill in snapc->seq and snapc->snaps[0..snap_count-1] */
-
-static struct ceph_snap_context *rbd_snap_context_create(u32 snap_count)
-{
- struct ceph_snap_context *snapc;
- size_t size;
-
- size = sizeof (struct ceph_snap_context);
- size += snap_count * sizeof (snapc->snaps[0]);
- snapc = kzalloc(size, GFP_KERNEL);
- if (!snapc)
- return NULL;
-
- atomic_set(&snapc->nref, 1);
- snapc->num_snaps = snap_count;
-
- return snapc;
-}
-
-static inline void rbd_snap_context_get(struct ceph_snap_context *snapc)
-{
- (void)ceph_get_snap_context(snapc);
-}
-
-static inline void rbd_snap_context_put(struct ceph_snap_context *snapc)
-{
- ceph_put_snap_context(snapc);
-}
-
/*
* Drop reference to ceph client node. If it's not referenced anymore, release
* it.
header->image_size = le64_to_cpu(ondisk->image_size);
- header->snapc = rbd_snap_context_create(snap_count);
+ header->snapc = ceph_create_snap_context(snap_count, GFP_KERNEL);
if (!header->snapc)
goto out_err;
header->snapc->seq = le64_to_cpu(ondisk->snap_seq);
return NULL;
}
-static int rbd_dev_set_mapping(struct rbd_device *rbd_dev)
+static int rbd_dev_mapping_set(struct rbd_device *rbd_dev)
{
if (!memcmp(rbd_dev->spec->snap_name, RBD_SNAP_HEAD_NAME,
sizeof (RBD_SNAP_HEAD_NAME))) {
return 0;
}
-static void rbd_header_free(struct rbd_image_header *header)
+static void rbd_dev_mapping_clear(struct rbd_device *rbd_dev)
{
- kfree(header->object_prefix);
- header->object_prefix = NULL;
- kfree(header->snap_sizes);
- header->snap_sizes = NULL;
- kfree(header->snap_names);
- header->snap_names = NULL;
- rbd_snap_context_put(header->snapc);
- header->snapc = NULL;
+ rbd_dev->mapping.size = 0;
+ rbd_dev->mapping.features = 0;
+ rbd_dev->mapping.read_only = true;
+}
+
+static void rbd_dev_clear_mapping(struct rbd_device *rbd_dev)
+{
+ rbd_dev->mapping.size = 0;
+ rbd_dev->mapping.features = 0;
+ rbd_dev->mapping.read_only = true;
}
static const char *rbd_segment_name(struct rbd_device *rbd_dev, u64 offset)
if (write_request) {
down_read(&rbd_dev->header_rwsem);
- rbd_snap_context_get(rbd_dev->header.snapc);
+ ceph_get_snap_context(rbd_dev->header.snapc);
up_read(&rbd_dev->header_rwsem);
}
rbd_assert(img_request->obj_request_count == 0);
if (img_request_write_test(img_request))
- rbd_snap_context_put(img_request->snapc);
+ ceph_put_snap_context(img_request->snapc);
if (img_request_child_test(img_request))
rbd_obj_request_put(img_request->obj_request);
rbd_dev->watch_request->osd_req);
osd_req_op_watch_init(obj_request->osd_req, 0, CEPH_OSD_OP_WATCH,
- rbd_dev->watch_event->cookie,
- rbd_dev->header.obj_version, start);
+ rbd_dev->watch_event->cookie, 0, start);
rbd_osd_req_format_write(obj_request);
ret = rbd_obj_request_submit(osdc, obj_request);
if (IS_ERR(ondisk))
return PTR_ERR(ondisk);
ret = rbd_header_from_disk(header, ondisk);
- if (ret >= 0)
- header->obj_version = ver;
kfree(ondisk);
return ret;
kfree(rbd_dev->header.snap_sizes);
kfree(rbd_dev->header.snap_names);
/* osd requests may still refer to snapc */
- rbd_snap_context_put(rbd_dev->header.snapc);
+ ceph_put_snap_context(rbd_dev->header.snapc);
- if (hver)
- *hver = h.obj_version;
- rbd_dev->header.obj_version = h.obj_version;
rbd_dev->header.image_size = h.image_size;
rbd_dev->header.snapc = h.snapc;
rbd_dev->header.snap_names = h.snap_names;
static int rbd_dev_refresh(struct rbd_device *rbd_dev, u64 *hver)
{
+ u64 image_size;
int ret;
rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
+ image_size = rbd_dev->header.image_size;
mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
if (rbd_dev->image_format == 1)
ret = rbd_dev_v1_refresh(rbd_dev, hver);
else
ret = rbd_dev_v2_refresh(rbd_dev, hver);
mutex_unlock(&ctl_mutex);
- revalidate_disk(rbd_dev->disk);
if (ret)
rbd_warn(rbd_dev, "got notification but failed to "
" update snaps: %d\n", ret);
+ if (image_size != rbd_dev->header.image_size)
+ revalidate_disk(rbd_dev->disk);
return ret;
}
rbd_dev->disk = disk;
- set_capacity(rbd_dev->disk, rbd_dev->mapping.size / SECTOR_SIZE);
-
return 0;
out_disk:
put_disk(disk);
* Returns a dynamically-allocated snapshot name if successful, or a
* pointer-coded error otherwise.
*/
-static char *rbd_dev_v1_snap_info(struct rbd_device *rbd_dev, u32 which,
+static const char *rbd_dev_v1_snap_info(struct rbd_device *rbd_dev, u32 which,
u64 *snap_size, u64 *snap_features)
{
- char *snap_name;
+ const char *snap_name;
int i;
rbd_assert(which < rbd_dev->header.snapc->num_snaps);
goto out;
ret = 0;
- snapc = rbd_snap_context_create(snap_count);
+ snapc = ceph_create_snap_context(snap_count, GFP_KERNEL);
if (!snapc) {
ret = -ENOMEM;
goto out;
return ret;
}
-static char *rbd_dev_v2_snap_name(struct rbd_device *rbd_dev, u32 which)
+static const char *rbd_dev_v2_snap_name(struct rbd_device *rbd_dev, u32 which)
{
size_t size;
void *reply_buf;
return snap_name;
}
-static char *rbd_dev_v2_snap_info(struct rbd_device *rbd_dev, u32 which,
+static const char *rbd_dev_v2_snap_info(struct rbd_device *rbd_dev, u32 which,
u64 *snap_size, u64 *snap_features)
{
u64 snap_id;
u64 size;
u64 features;
- char *snap_name;
+ const char *snap_name;
int ret;
rbd_assert(which < rbd_dev->header.snapc->num_snaps);
return ERR_PTR(ret);
}
-static char *rbd_dev_snap_info(struct rbd_device *rbd_dev, u32 which,
+static const char *rbd_dev_snap_info(struct rbd_device *rbd_dev, u32 which,
u64 *snap_size, u64 *snap_features)
{
if (rbd_dev->image_format == 1)
static int rbd_dev_v2_refresh(struct rbd_device *rbd_dev, u64 *hver)
{
int ret;
- __u8 obj_order;
down_write(&rbd_dev->header_rwsem);
- /* Grab old order first, to see if it changes */
-
- obj_order = rbd_dev->header.obj_order,
ret = rbd_dev_v2_image_size(rbd_dev);
if (ret)
goto out;
- if (rbd_dev->header.obj_order != obj_order) {
- ret = -EIO;
- goto out;
- }
rbd_update_mapping_size(rbd_dev);
ret = rbd_dev_v2_snap_context(rbd_dev, hver);
while (index < snap_count || links != head) {
u64 snap_id;
struct rbd_snap *snap;
- char *snap_name;
+ const char *snap_name;
u64 snap_size = 0;
u64 snap_features = 0;
dev->bus = &rbd_bus_type;
dev->type = &rbd_device_type;
dev->parent = &rbd_root_dev;
- dev->release = rbd_dev_release;
+ dev->release = rbd_dev_device_release;
dev_set_name(dev, "%d", rbd_dev->dev_id);
ret = device_register(dev);
return ret;
}
-static int rbd_dev_v1_probe(struct rbd_device *rbd_dev)
+/* Undo whatever state changes are made by v1 or v2 image probe */
+
+static void rbd_dev_unprobe(struct rbd_device *rbd_dev)
{
- int ret;
- size_t size;
+ struct rbd_image_header *header;
- /* Record the header object name for this rbd image. */
+ rbd_dev_remove_parent(rbd_dev);
+ rbd_spec_put(rbd_dev->parent_spec);
+ rbd_dev->parent_spec = NULL;
+ rbd_dev->parent_overlap = 0;
- size = strlen(rbd_dev->spec->image_name) + sizeof (RBD_SUFFIX);
- rbd_dev->header_name = kmalloc(size, GFP_KERNEL);
- if (!rbd_dev->header_name) {
- ret = -ENOMEM;
- goto out_err;
- }
- sprintf(rbd_dev->header_name, "%s%s",
- rbd_dev->spec->image_name, RBD_SUFFIX);
+ /* Free dynamic fields from the header, then zero it out */
+
+ header = &rbd_dev->header;
+ ceph_put_snap_context(header->snapc);
+ kfree(header->snap_sizes);
+ kfree(header->snap_names);
+ kfree(header->object_prefix);
+ memset(header, 0, sizeof (*header));
+}
+
+static int rbd_dev_v1_probe(struct rbd_device *rbd_dev)
+{
+ int ret;
/* Populate rbd image metadata */
static int rbd_dev_v2_probe(struct rbd_device *rbd_dev)
{
- size_t size;
int ret;
u64 ver = 0;
- /*
- * Image id was filled in by the caller. Record the header
- * object name for this rbd image.
- */
- size = sizeof (RBD_HEADER_PREFIX) + strlen(rbd_dev->spec->image_id);
- rbd_dev->header_name = kmalloc(size, GFP_KERNEL);
- if (!rbd_dev->header_name)
- return -ENOMEM;
- sprintf(rbd_dev->header_name, "%s%s",
- RBD_HEADER_PREFIX, rbd_dev->spec->image_id);
-
- /* Get the size and object order for the image */
ret = rbd_dev_v2_image_size(rbd_dev);
if (ret)
goto out_err;
ret = rbd_dev_v2_parent_info(rbd_dev);
if (ret)
goto out_err;
- rbd_warn(rbd_dev, "WARNING: kernel support for "
- "layered rbd images is EXPERIMENTAL!");
+
+ /*
+ * Don't print a warning for parent images. We can
+ * tell this point because we won't know its pool
+ * name yet (just its pool id).
+ */
+ if (rbd_dev->spec->pool_name)
+ rbd_warn(rbd_dev, "WARNING: kernel layering "
+ "is EXPERIMENTAL!");
}
/* If the image supports fancy striping, get its parameters */
ret = rbd_dev_v2_snap_context(rbd_dev, &ver);
if (ret)
goto out_err;
- rbd_dev->header.obj_version = ver;
dout("discovered version 2 image, header name is %s\n",
rbd_dev->header_name);
return ret;
}
-static int rbd_dev_probe_finish(struct rbd_device *rbd_dev)
+static int rbd_dev_probe_parent(struct rbd_device *rbd_dev)
{
struct rbd_device *parent = NULL;
- struct rbd_spec *parent_spec = NULL;
- struct rbd_client *rbdc = NULL;
+ struct rbd_spec *parent_spec;
+ struct rbd_client *rbdc;
int ret;
- /* no need to lock here, as rbd_dev is not registered yet */
- ret = rbd_dev_snaps_update(rbd_dev);
- if (ret)
- return ret;
+ if (!rbd_dev->parent_spec)
+ return 0;
+ /*
+ * We need to pass a reference to the client and the parent
+ * spec when creating the parent rbd_dev. Images related by
+ * parent/child relationships always share both.
+ */
+ parent_spec = rbd_spec_get(rbd_dev->parent_spec);
+ rbdc = __rbd_get_client(rbd_dev->rbd_client);
- ret = rbd_dev_spec_update(rbd_dev);
- if (ret)
- goto err_out_snaps;
+ ret = -ENOMEM;
+ parent = rbd_dev_create(rbdc, parent_spec);
+ if (!parent)
+ goto out_err;
- ret = rbd_dev_set_mapping(rbd_dev);
+ ret = rbd_dev_image_probe(parent);
+ if (ret < 0)
+ goto out_err;
+ rbd_dev->parent = parent;
+
+ return 0;
+out_err:
+ if (parent) {
+ rbd_spec_put(rbd_dev->parent_spec);
+ kfree(rbd_dev->header_name);
+ rbd_dev_destroy(parent);
+ } else {
+ rbd_put_client(rbdc);
+ rbd_spec_put(parent_spec);
+ }
+
+ return ret;
+}
+
+static int rbd_dev_device_setup(struct rbd_device *rbd_dev)
+{
+ int ret;
+
+ ret = rbd_dev_mapping_set(rbd_dev);
if (ret)
- goto err_out_snaps;
+ return ret;
/* generate unique id: find highest unique id, add one */
rbd_dev_id_get(rbd_dev);
if (ret)
goto err_out_disk;
- /*
- * At this point cleanup in the event of an error is the job
- * of the sysfs code (initiated by rbd_bus_del_dev()).
- */
- /* Probe the parent if there is one */
-
- if (rbd_dev->parent_spec) {
- /*
- * We need to pass a reference to the client and the
- * parent spec when creating the parent rbd_dev.
- * Images related by parent/child relationships
- * always share both.
- */
- parent_spec = rbd_spec_get(rbd_dev->parent_spec);
- rbdc = __rbd_get_client(rbd_dev->rbd_client);
-
- parent = rbd_dev_create(rbdc, parent_spec);
- if (!parent) {
- ret = -ENOMEM;
- goto err_out_spec;
- }
- rbdc = NULL; /* parent now owns reference */
- parent_spec = NULL; /* parent now owns reference */
- ret = rbd_dev_image_probe(parent);
- if (ret < 0)
- goto err_out_parent;
- rbd_dev->parent = parent;
- }
-
- ret = rbd_dev_header_watch_sync(rbd_dev, 1);
- if (ret)
- goto err_out_bus;
-
/* Everything's ready. Announce the disk to the world. */
+ set_capacity(rbd_dev->disk, rbd_dev->mapping.size / SECTOR_SIZE);
set_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags);
add_disk(rbd_dev->disk);
return ret;
-err_out_parent:
- rbd_spec_put(rbd_dev->parent_spec);
- kfree(rbd_dev->header_name);
- rbd_dev_destroy(parent);
-err_out_spec:
- rbd_spec_put(parent_spec);
- rbd_put_client(rbdc);
-err_out_bus:
- /* this will also clean up rest of rbd_dev stuff */
-
- rbd_bus_del_dev(rbd_dev);
-
- return ret;
err_out_disk:
rbd_free_disk(rbd_dev);
err_out_blkdev:
unregister_blkdev(rbd_dev->major, rbd_dev->name);
err_out_id:
rbd_dev_id_put(rbd_dev);
-err_out_snaps:
- rbd_remove_all_snaps(rbd_dev);
+ rbd_dev_mapping_clear(rbd_dev);
return ret;
}
+static int rbd_dev_header_name(struct rbd_device *rbd_dev)
+{
+ struct rbd_spec *spec = rbd_dev->spec;
+ size_t size;
+
+ /* Record the header object name for this rbd image. */
+
+ rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
+
+ if (rbd_dev->image_format == 1)
+ size = strlen(spec->image_name) + sizeof (RBD_SUFFIX);
+ else
+ size = sizeof (RBD_HEADER_PREFIX) + strlen(spec->image_id);
+
+ rbd_dev->header_name = kmalloc(size, GFP_KERNEL);
+ if (!rbd_dev->header_name)
+ return -ENOMEM;
+
+ if (rbd_dev->image_format == 1)
+ sprintf(rbd_dev->header_name, "%s%s",
+ spec->image_name, RBD_SUFFIX);
+ else
+ sprintf(rbd_dev->header_name, "%s%s",
+ RBD_HEADER_PREFIX, spec->image_id);
+ return 0;
+}
+
+static void rbd_dev_image_release(struct rbd_device *rbd_dev)
+{
+ int ret;
+
+ rbd_remove_all_snaps(rbd_dev);
+ rbd_dev_unprobe(rbd_dev);
+ ret = rbd_dev_header_watch_sync(rbd_dev, 0);
+ if (ret)
+ rbd_warn(rbd_dev, "failed to cancel watch event (%d)\n", ret);
+ kfree(rbd_dev->header_name);
+ rbd_dev->header_name = NULL;
+ rbd_dev->image_format = 0;
+ kfree(rbd_dev->spec->image_id);
+ rbd_dev->spec->image_id = NULL;
+
+ rbd_dev_destroy(rbd_dev);
+}
+
/*
* Probe for the existence of the header object for the given rbd
* device. For format 2 images this includes determining the image
static int rbd_dev_image_probe(struct rbd_device *rbd_dev)
{
int ret;
+ int tmp;
/*
* Get the id from the image id object. If it's not a
rbd_assert(rbd_dev->spec->image_id);
rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
+ ret = rbd_dev_header_name(rbd_dev);
+ if (ret)
+ goto err_out_format;
+
+ ret = rbd_dev_header_watch_sync(rbd_dev, 1);
+ if (ret)
+ goto out_header_name;
+
if (rbd_dev->image_format == 1)
ret = rbd_dev_v1_probe(rbd_dev);
else
ret = rbd_dev_v2_probe(rbd_dev);
if (ret)
- goto out_err;
+ goto err_out_watch;
- ret = rbd_dev_probe_finish(rbd_dev);
+ ret = rbd_dev_snaps_update(rbd_dev);
if (ret)
- rbd_header_free(&rbd_dev->header);
+ goto err_out_probe;
- return ret;
-out_err:
+ ret = rbd_dev_spec_update(rbd_dev);
+ if (ret)
+ goto err_out_snaps;
+
+ ret = rbd_dev_probe_parent(rbd_dev);
+ if (!ret)
+ return 0;
+
+err_out_snaps:
+ rbd_remove_all_snaps(rbd_dev);
+err_out_probe:
+ rbd_dev_unprobe(rbd_dev);
+err_out_watch:
+ tmp = rbd_dev_header_watch_sync(rbd_dev, 0);
+ if (tmp)
+ rbd_warn(rbd_dev, "unable to tear down watch request\n");
+out_header_name:
+ kfree(rbd_dev->header_name);
+ rbd_dev->header_name = NULL;
+err_out_format:
+ rbd_dev->image_format = 0;
kfree(rbd_dev->spec->image_id);
rbd_dev->spec->image_id = NULL;
if (rc < 0)
goto err_out_rbd_dev;
- return count;
+ rc = rbd_dev_device_setup(rbd_dev);
+ if (!rc)
+ return count;
+
+ rbd_dev_image_release(rbd_dev);
err_out_rbd_dev:
- rbd_spec_put(rbd_dev->parent_spec);
- kfree(rbd_dev->header_name);
rbd_dev_destroy(rbd_dev);
err_out_client:
rbd_put_client(rbdc);
return NULL;
}
-static void rbd_dev_release(struct device *dev)
+static void rbd_dev_device_release(struct device *dev)
{
struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
- if (rbd_dev->watch_event)
- rbd_dev_header_watch_sync(rbd_dev, 0);
-
- /* clean up and free blkdev */
rbd_free_disk(rbd_dev);
+ clear_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags);
+ rbd_dev_clear_mapping(rbd_dev);
unregister_blkdev(rbd_dev->major, rbd_dev->name);
-
- /* release allocated disk header fields */
- rbd_header_free(&rbd_dev->header);
-
- /* done with the id, and with the rbd_dev */
+ rbd_dev->major = 0;
rbd_dev_id_put(rbd_dev);
- rbd_assert(rbd_dev->rbd_client != NULL);
- rbd_spec_put(rbd_dev->parent_spec);
- kfree(rbd_dev->header_name);
- rbd_dev_destroy(rbd_dev);
-
- /* release module ref */
- module_put(THIS_MODULE);
+ rbd_dev_mapping_clear(rbd_dev);
}
-static void __rbd_remove(struct rbd_device *rbd_dev)
+static void rbd_dev_remove_parent(struct rbd_device *rbd_dev)
{
- rbd_remove_all_snaps(rbd_dev);
- rbd_bus_del_dev(rbd_dev);
+ while (rbd_dev->parent) {
+ struct rbd_device *first = rbd_dev;
+ struct rbd_device *second = first->parent;
+ struct rbd_device *third;
+
+ /*
+ * Follow to the parent with no grandparent and
+ * remove it.
+ */
+ while (second && (third = second->parent)) {
+ first = second;
+ second = third;
+ }
+ rbd_assert(second);
+ rbd_dev_image_release(second);
+ first->parent = NULL;
+ first->parent_overlap = 0;
+
+ rbd_assert(first->parent_spec);
+ rbd_spec_put(first->parent_spec);
+ first->parent_spec = NULL;
+ }
}
static ssize_t rbd_remove(struct bus_type *bus,
size_t count)
{
struct rbd_device *rbd_dev = NULL;
- int target_id, rc;
+ int target_id;
unsigned long ul;
- int ret = count;
+ int ret;
- rc = strict_strtoul(buf, 10, &ul);
- if (rc)
- return rc;
+ ret = strict_strtoul(buf, 10, &ul);
+ if (ret)
+ return ret;
/* convert to int; abort if we lost anything in the conversion */
target_id = (int) ul;
spin_unlock_irq(&rbd_dev->lock);
if (ret < 0)
goto done;
-
- while (rbd_dev->parent_spec) {
- struct rbd_device *first = rbd_dev;
- struct rbd_device *second = first->parent;
- struct rbd_device *third;
-
- /*
- * Follow to the parent with no grandparent and
- * remove it.
- */
- while (second && (third = second->parent)) {
- first = second;
- second = third;
- }
- __rbd_remove(second);
- rbd_spec_put(first->parent_spec);
- first->parent_spec = NULL;
- first->parent_overlap = 0;
- first->parent = NULL;
- }
- __rbd_remove(rbd_dev);
-
+ ret = count;
+ rbd_bus_del_dev(rbd_dev);
+ rbd_dev_image_release(rbd_dev);
+ module_put(THIS_MODULE);
done:
mutex_unlock(&ctl_mutex);
projects / ~andy / linux / blobdiff