#define RBD_MINORS_PER_MAJOR 256 /* max minors per blkdev */
#define RBD_MAX_SNAP_NAME_LEN 32
+#define RBD_MAX_SNAP_COUNT 510 /* allows max snapc to fit in 4KB */
#define RBD_MAX_OPT_LEN 1024
#define RBD_SNAP_HEAD_NAME "-"
+#define RBD_IMAGE_ID_LEN_MAX 64
+#define RBD_OBJ_PREFIX_LEN_MAX 64
+
+/* Feature bits */
+
+#define RBD_FEATURE_LAYERING 1
+
+/* Features supported by this (client software) implementation. */
+
+#define RBD_FEATURES_ALL (0)
+
/*
* An RBD device name will be "rbd#", where the "rbd" comes from
* RBD_DRV_NAME above, and # is a unique integer identifier.
struct rbd_image_header {
/* These four fields never change for a given rbd image */
char *object_prefix;
+ u64 features;
__u8 obj_order;
__u8 crypt_type;
__u8 comp_type;
u64 size;
struct list_head node;
u64 id;
+ u64 features;
};
struct rbd_mapping {
char *snap_name;
u64 snap_id;
u64 size;
+ u64 features;
bool snap_exists;
bool read_only;
};
int major; /* blkdev assigned major */
struct gendisk *disk; /* blkdev's gendisk and rq */
+ u32 image_format; /* Either 1 or 2 */
struct rbd_options rbd_opts;
struct rbd_client *rbd_client;
spinlock_t lock; /* queue lock */
struct rbd_image_header header;
+ char *image_id;
+ size_t image_id_len;
char *image_name;
size_t image_name_len;
char *header_name;
static int rbd_dev_snaps_register(struct rbd_device *rbd_dev);
static void rbd_dev_release(struct device *dev);
-static ssize_t rbd_snap_add(struct device *dev,
- struct device_attribute *attr,
- const char *buf,
- size_t count);
static void __rbd_remove_snap_dev(struct rbd_snap *snap);
static ssize_t rbd_add(struct bus_type *bus, const char *buf,
put_device(&rbd_dev->dev);
}
-static int rbd_refresh_header(struct rbd_device *rbd_dev, u64 *hver);
+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);
static int rbd_open(struct block_device *bdev, fmode_t mode)
{
kfree(coll);
}
+static bool rbd_image_format_valid(u32 image_format)
+{
+ return image_format == 1 || image_format == 2;
+}
+
static bool rbd_dev_ondisk_valid(struct rbd_image_header_ondisk *ondisk)
{
size_t size;
header->snap_sizes = NULL;
}
+ header->features = 0; /* No features support in v1 images */
header->obj_order = ondisk->options.order;
header->crypt_type = ondisk->options.crypt_type;
header->comp_type = ondisk->options.comp_type;
if (!strcmp(snap_name, snap->name)) {
rbd_dev->mapping.snap_id = snap->id;
rbd_dev->mapping.size = snap->size;
+ rbd_dev->mapping.features = snap->features;
return 0;
}
sizeof (RBD_SNAP_HEAD_NAME))) {
rbd_dev->mapping.snap_id = CEPH_NOSNAP;
rbd_dev->mapping.size = rbd_dev->header.image_size;
+ rbd_dev->mapping.features = rbd_dev->header.features;
rbd_dev->mapping.snap_exists = false;
rbd_dev->mapping.read_only = rbd_dev->rbd_opts.read_only;
ret = 0;
layout->fl_stripe_count = cpu_to_le32(1);
layout->fl_object_size = cpu_to_le32(1 << RBD_MAX_OBJ_ORDER);
layout->fl_pg_pool = cpu_to_le32(rbd_dev->pool_id);
- ceph_calc_raw_layout(osdc, layout, snapid, ofs, &len, &bno,
- req, ops);
+ ret = ceph_calc_raw_layout(osdc, layout, snapid, ofs, &len, &bno,
+ req, ops);
+ rbd_assert(ret == 0);
ceph_osdc_build_request(req, ofs, &len,
ops,
dout("rbd_watch_cb %s notify_id=%llu opcode=%u\n",
rbd_dev->header_name, (unsigned long long) notify_id,
(unsigned int) opcode);
- rc = rbd_refresh_header(rbd_dev, &hver);
+ rc = rbd_dev_refresh(rbd_dev, &hver);
if (rc)
pr_warning(RBD_DRV_NAME "%d got notification but failed to "
" update snaps: %d\n", rbd_dev->major, rc);
return ret;
}
-struct rbd_notify_info {
- struct rbd_device *rbd_dev;
-};
-
-static void rbd_notify_cb(u64 ver, u64 notify_id, u8 opcode, void *data)
-{
- struct rbd_device *rbd_dev = (struct rbd_device *)data;
- if (!rbd_dev)
- return;
-
- dout("rbd_notify_cb %s notify_id=%llu opcode=%u\n",
- rbd_dev->header_name, (unsigned long long) notify_id,
- (unsigned int) opcode);
-}
-
-/*
- * Request sync osd notify
- */
-static int rbd_req_sync_notify(struct rbd_device *rbd_dev)
-{
- struct ceph_osd_req_op *ops;
- struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
- struct ceph_osd_event *event;
- struct rbd_notify_info info;
- int payload_len = sizeof(u32) + sizeof(u32);
- int ret;
-
- ops = rbd_create_rw_ops(1, CEPH_OSD_OP_NOTIFY, payload_len);
- if (!ops)
- return -ENOMEM;
-
- info.rbd_dev = rbd_dev;
-
- ret = ceph_osdc_create_event(osdc, rbd_notify_cb, 1,
- (void *)&info, &event);
- if (ret < 0)
- goto fail;
-
- ops[0].watch.ver = 1;
- ops[0].watch.flag = 1;
- ops[0].watch.cookie = event->cookie;
- ops[0].watch.prot_ver = RADOS_NOTIFY_VER;
- ops[0].watch.timeout = 12;
-
- ret = rbd_req_sync_op(rbd_dev, NULL,
- CEPH_NOSNAP,
- CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK,
- ops,
- rbd_dev->header_name,
- 0, 0, NULL, NULL, NULL);
- if (ret < 0)
- goto fail_event;
-
- ret = ceph_osdc_wait_event(event, CEPH_OSD_TIMEOUT_DEFAULT);
- dout("ceph_osdc_wait_event returned %d\n", ret);
- rbd_destroy_ops(ops);
- return 0;
-
-fail_event:
- ceph_osdc_cancel_event(event);
-fail:
- rbd_destroy_ops(ops);
- return ret;
-}
-
/*
* Synchronous osd object method call
*/
return ret;
}
-/*
- * create a snapshot
- */
-static int rbd_header_add_snap(struct rbd_device *rbd_dev,
- const char *snap_name,
- gfp_t gfp_flags)
-{
- int name_len = strlen(snap_name);
- u64 new_snapid;
- int ret;
- void *data, *p, *e;
- struct ceph_mon_client *monc;
-
- /* we should create a snapshot only if we're pointing at the head */
- if (rbd_dev->mapping.snap_id != CEPH_NOSNAP)
- return -EINVAL;
-
- monc = &rbd_dev->rbd_client->client->monc;
- ret = ceph_monc_create_snapid(monc, rbd_dev->pool_id, &new_snapid);
- dout("created snapid=%llu\n", (unsigned long long) new_snapid);
- if (ret < 0)
- return ret;
-
- data = kmalloc(name_len + 16, gfp_flags);
- if (!data)
- return -ENOMEM;
-
- p = data;
- e = data + name_len + 16;
-
- ceph_encode_string_safe(&p, e, snap_name, name_len, bad);
- ceph_encode_64_safe(&p, e, new_snapid, bad);
-
- ret = rbd_req_sync_exec(rbd_dev, rbd_dev->header_name,
- "rbd", "snap_add",
- data, (size_t) (p - data), NULL, 0,
- CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK,
- NULL);
-
- kfree(data);
-
- return ret < 0 ? ret : 0;
-bad:
- return -ERANGE;
-}
-
static void __rbd_remove_all_snaps(struct rbd_device *rbd_dev)
{
struct rbd_snap *snap;
__rbd_remove_snap_dev(snap);
}
+static void rbd_update_mapping_size(struct rbd_device *rbd_dev)
+{
+ sector_t size;
+
+ if (rbd_dev->mapping.snap_id != CEPH_NOSNAP)
+ return;
+
+ size = (sector_t) rbd_dev->header.image_size / SECTOR_SIZE;
+ dout("setting size to %llu sectors", (unsigned long long) size);
+ rbd_dev->mapping.size = (u64) size;
+ set_capacity(rbd_dev->disk, size);
+}
+
/*
* only read the first part of the ondisk header, without the snaps info
*/
-static int __rbd_refresh_header(struct rbd_device *rbd_dev, u64 *hver)
+static int rbd_dev_v1_refresh(struct rbd_device *rbd_dev, u64 *hver)
{
int ret;
struct rbd_image_header h;
down_write(&rbd_dev->header_rwsem);
- /* resized? */
- if (rbd_dev->mapping.snap_id == CEPH_NOSNAP) {
- sector_t size = (sector_t) h.image_size / SECTOR_SIZE;
-
- if (size != (sector_t) rbd_dev->mapping.size) {
- dout("setting size to %llu sectors",
- (unsigned long long) size);
- rbd_dev->mapping.size = (u64) size;
- set_capacity(rbd_dev->disk, size);
- }
- }
+ /* Update image size, and check for resize of mapped image */
+ rbd_dev->header.image_size = h.image_size;
+ rbd_update_mapping_size(rbd_dev);
/* rbd_dev->header.object_prefix shouldn't change */
kfree(rbd_dev->header.snap_sizes);
return ret;
}
-static int rbd_refresh_header(struct rbd_device *rbd_dev, u64 *hver)
+static int rbd_dev_refresh(struct rbd_device *rbd_dev, u64 *hver)
{
int ret;
+ rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
- ret = __rbd_refresh_header(rbd_dev, hver);
+ 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);
return ret;
return sprintf(buf, "%llu\n", (unsigned long long) size * SECTOR_SIZE);
}
+/*
+ * Note this shows the features for whatever's mapped, which is not
+ * necessarily the base image.
+ */
+static ssize_t rbd_features_show(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
+
+ return sprintf(buf, "0x%016llx\n",
+ (unsigned long long) rbd_dev->mapping.features);
+}
+
static ssize_t rbd_major_show(struct device *dev,
struct device_attribute *attr, char *buf)
{
return sprintf(buf, "%s\n", rbd_dev->image_name);
}
+static ssize_t rbd_image_id_show(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
+
+ return sprintf(buf, "%s\n", rbd_dev->image_id);
+}
+
+/*
+ * Shows the name of the currently-mapped snapshot (or
+ * RBD_SNAP_HEAD_NAME for the base image).
+ */
static ssize_t rbd_snap_show(struct device *dev,
struct device_attribute *attr,
char *buf)
struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
int ret;
- ret = rbd_refresh_header(rbd_dev, NULL);
+ ret = rbd_dev_refresh(rbd_dev, NULL);
return ret < 0 ? ret : size;
}
static DEVICE_ATTR(size, S_IRUGO, rbd_size_show, NULL);
+static DEVICE_ATTR(features, S_IRUGO, rbd_features_show, NULL);
static DEVICE_ATTR(major, S_IRUGO, rbd_major_show, NULL);
static DEVICE_ATTR(client_id, S_IRUGO, rbd_client_id_show, NULL);
static DEVICE_ATTR(pool, S_IRUGO, rbd_pool_show, NULL);
static DEVICE_ATTR(pool_id, S_IRUGO, rbd_pool_id_show, NULL);
static DEVICE_ATTR(name, S_IRUGO, rbd_name_show, NULL);
+static DEVICE_ATTR(image_id, S_IRUGO, rbd_image_id_show, NULL);
static DEVICE_ATTR(refresh, S_IWUSR, NULL, rbd_image_refresh);
static DEVICE_ATTR(current_snap, S_IRUGO, rbd_snap_show, NULL);
-static DEVICE_ATTR(create_snap, S_IWUSR, NULL, rbd_snap_add);
static struct attribute *rbd_attrs[] = {
&dev_attr_size.attr,
+ &dev_attr_features.attr,
&dev_attr_major.attr,
&dev_attr_client_id.attr,
&dev_attr_pool.attr,
&dev_attr_pool_id.attr,
&dev_attr_name.attr,
+ &dev_attr_image_id.attr,
&dev_attr_current_snap.attr,
&dev_attr_refresh.attr,
- &dev_attr_create_snap.attr,
NULL
};
return sprintf(buf, "%llu\n", (unsigned long long)snap->id);
}
+static ssize_t rbd_snap_features_show(struct device *dev,
+ struct device_attribute *attr,
+ char *buf)
+{
+ struct rbd_snap *snap = container_of(dev, struct rbd_snap, dev);
+
+ return sprintf(buf, "0x%016llx\n",
+ (unsigned long long) snap->features);
+}
+
static DEVICE_ATTR(snap_size, S_IRUGO, rbd_snap_size_show, NULL);
static DEVICE_ATTR(snap_id, S_IRUGO, rbd_snap_id_show, NULL);
+static DEVICE_ATTR(snap_features, S_IRUGO, rbd_snap_features_show, NULL);
static struct attribute *rbd_snap_attrs[] = {
&dev_attr_snap_size.attr,
&dev_attr_snap_id.attr,
+ &dev_attr_snap_features.attr,
NULL,
};
}
static struct rbd_snap *__rbd_add_snap_dev(struct rbd_device *rbd_dev,
- int i, const char *name)
+ const char *snap_name,
+ u64 snap_id, u64 snap_size,
+ u64 snap_features)
{
struct rbd_snap *snap;
int ret;
return ERR_PTR(-ENOMEM);
ret = -ENOMEM;
- snap->name = kstrdup(name, GFP_KERNEL);
+ snap->name = kstrdup(snap_name, GFP_KERNEL);
if (!snap->name)
goto err;
- snap->size = rbd_dev->header.snap_sizes[i];
- snap->id = rbd_dev->header.snapc->snaps[i];
+ snap->id = snap_id;
+ snap->size = snap_size;
+ snap->features = snap_features;
return snap;
return ERR_PTR(ret);
}
+static char *rbd_dev_v1_snap_info(struct rbd_device *rbd_dev, u32 which,
+ u64 *snap_size, u64 *snap_features)
+{
+ char *snap_name;
+
+ rbd_assert(which < rbd_dev->header.snapc->num_snaps);
+
+ *snap_size = rbd_dev->header.snap_sizes[which];
+ *snap_features = 0; /* No features for v1 */
+
+ /* Skip over names until we find the one we are looking for */
+
+ snap_name = rbd_dev->header.snap_names;
+ while (which--)
+ snap_name += strlen(snap_name) + 1;
+
+ return snap_name;
+}
+
+/*
+ * Get the size and object order for an image snapshot, or if
+ * snap_id is CEPH_NOSNAP, gets this information for the base
+ * image.
+ */
+static int _rbd_dev_v2_snap_size(struct rbd_device *rbd_dev, u64 snap_id,
+ u8 *order, u64 *snap_size)
+{
+ __le64 snapid = cpu_to_le64(snap_id);
+ int ret;
+ struct {
+ u8 order;
+ __le64 size;
+ } __attribute__ ((packed)) size_buf = { 0 };
+
+ ret = rbd_req_sync_exec(rbd_dev, rbd_dev->header_name,
+ "rbd", "get_size",
+ (char *) &snapid, sizeof (snapid),
+ (char *) &size_buf, sizeof (size_buf),
+ CEPH_OSD_FLAG_READ, NULL);
+ dout("%s: rbd_req_sync_exec returned %d\n", __func__, ret);
+ if (ret < 0)
+ return ret;
+
+ *order = size_buf.order;
+ *snap_size = le64_to_cpu(size_buf.size);
+
+ dout(" snap_id 0x%016llx order = %u, snap_size = %llu\n",
+ (unsigned long long) snap_id, (unsigned int) *order,
+ (unsigned long long) *snap_size);
+
+ return 0;
+}
+
+static int rbd_dev_v2_image_size(struct rbd_device *rbd_dev)
+{
+ return _rbd_dev_v2_snap_size(rbd_dev, CEPH_NOSNAP,
+ &rbd_dev->header.obj_order,
+ &rbd_dev->header.image_size);
+}
+
+static int rbd_dev_v2_object_prefix(struct rbd_device *rbd_dev)
+{
+ void *reply_buf;
+ int ret;
+ void *p;
+
+ reply_buf = kzalloc(RBD_OBJ_PREFIX_LEN_MAX, GFP_KERNEL);
+ if (!reply_buf)
+ return -ENOMEM;
+
+ ret = rbd_req_sync_exec(rbd_dev, rbd_dev->header_name,
+ "rbd", "get_object_prefix",
+ NULL, 0,
+ reply_buf, RBD_OBJ_PREFIX_LEN_MAX,
+ CEPH_OSD_FLAG_READ, NULL);
+ dout("%s: rbd_req_sync_exec returned %d\n", __func__, ret);
+ if (ret < 0)
+ goto out;
+
+ p = reply_buf;
+ rbd_dev->header.object_prefix = ceph_extract_encoded_string(&p,
+ p + RBD_OBJ_PREFIX_LEN_MAX,
+ NULL, GFP_NOIO);
+
+ if (IS_ERR(rbd_dev->header.object_prefix)) {
+ ret = PTR_ERR(rbd_dev->header.object_prefix);
+ rbd_dev->header.object_prefix = NULL;
+ } else {
+ dout(" object_prefix = %s\n", rbd_dev->header.object_prefix);
+ }
+
+out:
+ kfree(reply_buf);
+
+ return ret;
+}
+
+static int _rbd_dev_v2_snap_features(struct rbd_device *rbd_dev, u64 snap_id,
+ u64 *snap_features)
+{
+ __le64 snapid = cpu_to_le64(snap_id);
+ struct {
+ __le64 features;
+ __le64 incompat;
+ } features_buf = { 0 };
+ u64 incompat;
+ int ret;
+
+ ret = rbd_req_sync_exec(rbd_dev, rbd_dev->header_name,
+ "rbd", "get_features",
+ (char *) &snapid, sizeof (snapid),
+ (char *) &features_buf, sizeof (features_buf),
+ CEPH_OSD_FLAG_READ, NULL);
+ dout("%s: rbd_req_sync_exec returned %d\n", __func__, ret);
+ if (ret < 0)
+ return ret;
+
+ incompat = le64_to_cpu(features_buf.incompat);
+ if (incompat & ~RBD_FEATURES_ALL)
+ return -ENOTSUPP;
+
+ *snap_features = le64_to_cpu(features_buf.features);
+
+ dout(" snap_id 0x%016llx features = 0x%016llx incompat = 0x%016llx\n",
+ (unsigned long long) snap_id,
+ (unsigned long long) *snap_features,
+ (unsigned long long) le64_to_cpu(features_buf.incompat));
+
+ return 0;
+}
+
+static int rbd_dev_v2_features(struct rbd_device *rbd_dev)
+{
+ return _rbd_dev_v2_snap_features(rbd_dev, CEPH_NOSNAP,
+ &rbd_dev->header.features);
+}
+
+static int rbd_dev_v2_snap_context(struct rbd_device *rbd_dev, u64 *ver)
+{
+ size_t size;
+ int ret;
+ void *reply_buf;
+ void *p;
+ void *end;
+ u64 seq;
+ u32 snap_count;
+ struct ceph_snap_context *snapc;
+ u32 i;
+
+ /*
+ * We'll need room for the seq value (maximum snapshot id),
+ * snapshot count, and array of that many snapshot ids.
+ * For now we have a fixed upper limit on the number we're
+ * prepared to receive.
+ */
+ size = sizeof (__le64) + sizeof (__le32) +
+ RBD_MAX_SNAP_COUNT * sizeof (__le64);
+ reply_buf = kzalloc(size, GFP_KERNEL);
+ if (!reply_buf)
+ return -ENOMEM;
+
+ ret = rbd_req_sync_exec(rbd_dev, rbd_dev->header_name,
+ "rbd", "get_snapcontext",
+ NULL, 0,
+ reply_buf, size,
+ CEPH_OSD_FLAG_READ, ver);
+ dout("%s: rbd_req_sync_exec returned %d\n", __func__, ret);
+ if (ret < 0)
+ goto out;
+
+ ret = -ERANGE;
+ p = reply_buf;
+ end = (char *) reply_buf + size;
+ ceph_decode_64_safe(&p, end, seq, out);
+ ceph_decode_32_safe(&p, end, snap_count, out);
+
+ /*
+ * Make sure the reported number of snapshot ids wouldn't go
+ * beyond the end of our buffer. But before checking that,
+ * make sure the computed size of the snapshot context we
+ * allocate is representable in a size_t.
+ */
+ if (snap_count > (SIZE_MAX - sizeof (struct ceph_snap_context))
+ / sizeof (u64)) {
+ ret = -EINVAL;
+ goto out;
+ }
+ if (!ceph_has_room(&p, end, snap_count * sizeof (__le64)))
+ goto out;
+
+ size = sizeof (struct ceph_snap_context) +
+ snap_count * sizeof (snapc->snaps[0]);
+ snapc = kmalloc(size, GFP_KERNEL);
+ if (!snapc) {
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ atomic_set(&snapc->nref, 1);
+ snapc->seq = seq;
+ snapc->num_snaps = snap_count;
+ for (i = 0; i < snap_count; i++)
+ snapc->snaps[i] = ceph_decode_64(&p);
+
+ rbd_dev->header.snapc = snapc;
+
+ dout(" snap context seq = %llu, snap_count = %u\n",
+ (unsigned long long) seq, (unsigned int) snap_count);
+
+out:
+ kfree(reply_buf);
+
+ return 0;
+}
+
+static char *rbd_dev_v2_snap_name(struct rbd_device *rbd_dev, u32 which)
+{
+ size_t size;
+ void *reply_buf;
+ __le64 snap_id;
+ int ret;
+ void *p;
+ void *end;
+ size_t snap_name_len;
+ char *snap_name;
+
+ size = sizeof (__le32) + RBD_MAX_SNAP_NAME_LEN;
+ reply_buf = kmalloc(size, GFP_KERNEL);
+ if (!reply_buf)
+ return ERR_PTR(-ENOMEM);
+
+ snap_id = cpu_to_le64(rbd_dev->header.snapc->snaps[which]);
+ ret = rbd_req_sync_exec(rbd_dev, rbd_dev->header_name,
+ "rbd", "get_snapshot_name",
+ (char *) &snap_id, sizeof (snap_id),
+ reply_buf, size,
+ CEPH_OSD_FLAG_READ, NULL);
+ dout("%s: rbd_req_sync_exec returned %d\n", __func__, ret);
+ if (ret < 0)
+ goto out;
+
+ p = reply_buf;
+ end = (char *) reply_buf + size;
+ snap_name_len = 0;
+ snap_name = ceph_extract_encoded_string(&p, end, &snap_name_len,
+ GFP_KERNEL);
+ if (IS_ERR(snap_name)) {
+ ret = PTR_ERR(snap_name);
+ goto out;
+ } else {
+ dout(" snap_id 0x%016llx snap_name = %s\n",
+ (unsigned long long) le64_to_cpu(snap_id), snap_name);
+ }
+ kfree(reply_buf);
+
+ return snap_name;
+out:
+ kfree(reply_buf);
+
+ return ERR_PTR(ret);
+}
+
+static char *rbd_dev_v2_snap_info(struct rbd_device *rbd_dev, u32 which,
+ u64 *snap_size, u64 *snap_features)
+{
+ __le64 snap_id;
+ u8 order;
+ int ret;
+
+ snap_id = rbd_dev->header.snapc->snaps[which];
+ ret = _rbd_dev_v2_snap_size(rbd_dev, snap_id, &order, snap_size);
+ if (ret)
+ return ERR_PTR(ret);
+ ret = _rbd_dev_v2_snap_features(rbd_dev, snap_id, snap_features);
+ if (ret)
+ return ERR_PTR(ret);
+
+ return rbd_dev_v2_snap_name(rbd_dev, which);
+}
+
+static char *rbd_dev_snap_info(struct rbd_device *rbd_dev, u32 which,
+ u64 *snap_size, u64 *snap_features)
+{
+ if (rbd_dev->image_format == 1)
+ return rbd_dev_v1_snap_info(rbd_dev, which,
+ snap_size, snap_features);
+ if (rbd_dev->image_format == 2)
+ return rbd_dev_v2_snap_info(rbd_dev, which,
+ snap_size, snap_features);
+ return ERR_PTR(-EINVAL);
+}
+
+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);
+ dout("rbd_dev_v2_snap_context returned %d\n", ret);
+ if (ret)
+ goto out;
+ ret = rbd_dev_snaps_update(rbd_dev);
+ dout("rbd_dev_snaps_update returned %d\n", ret);
+ if (ret)
+ goto out;
+ ret = rbd_dev_snaps_register(rbd_dev);
+ dout("rbd_dev_snaps_register returned %d\n", ret);
+out:
+ up_write(&rbd_dev->header_rwsem);
+
+ return ret;
+}
+
/*
* Scan the rbd device's current snapshot list and compare it to the
* newly-received snapshot context. Remove any existing snapshots
{
struct ceph_snap_context *snapc = rbd_dev->header.snapc;
const u32 snap_count = snapc->num_snaps;
- char *snap_name = rbd_dev->header.snap_names;
struct list_head *head = &rbd_dev->snaps;
struct list_head *links = head->next;
u32 index = 0;
while (index < snap_count || links != head) {
u64 snap_id;
struct rbd_snap *snap;
+ char *snap_name;
+ u64 snap_size = 0;
+ u64 snap_features = 0;
snap_id = index < snap_count ? snapc->snaps[index]
: CEPH_NOSNAP;
continue;
}
+ snap_name = rbd_dev_snap_info(rbd_dev, index,
+ &snap_size, &snap_features);
+ if (IS_ERR(snap_name))
+ return PTR_ERR(snap_name);
+
dout("entry %u: snap_id = %llu\n", (unsigned int) snap_count,
(unsigned long long) snap_id);
if (!snap || (snap_id != CEPH_NOSNAP && snap->id < snap_id)) {
/* We haven't seen this snapshot before */
- new_snap = __rbd_add_snap_dev(rbd_dev, index,
- snap_name);
+ new_snap = __rbd_add_snap_dev(rbd_dev, snap_name,
+ snap_id, snap_size, snap_features);
if (IS_ERR(new_snap)) {
int err = PTR_ERR(new_snap);
dout(" already present\n");
- rbd_assert(snap->size ==
- rbd_dev->header.snap_sizes[index]);
+ rbd_assert(snap->size == snap_size);
rbd_assert(!strcmp(snap->name, snap_name));
+ rbd_assert(snap->features == snap_features);
/* Done with this list entry; advance */
/* Advance to the next entry in the snapshot context */
index++;
- snap_name += strlen(snap_name) + 1;
}
dout("%s: done\n", __func__);
do {
ret = rbd_req_sync_watch(rbd_dev);
if (ret == -ERANGE) {
- rc = rbd_refresh_header(rbd_dev, NULL);
+ rc = rbd_dev_refresh(rbd_dev, NULL);
if (rc < 0)
return rc;
}
return err_ptr;
}
+/*
+ * An rbd format 2 image has a unique identifier, distinct from the
+ * name given to it by the user. Internally, that identifier is
+ * what's used to specify the names of objects related to the image.
+ *
+ * A special "rbd id" object is used to map an rbd image name to its
+ * id. If that object doesn't exist, then there is no v2 rbd image
+ * with the supplied name.
+ *
+ * This function will record the given rbd_dev's image_id field if
+ * it can be determined, and in that case will return 0. If any
+ * errors occur a negative errno will be returned and the rbd_dev's
+ * image_id field will be unchanged (and should be NULL).
+ */
+static int rbd_dev_image_id(struct rbd_device *rbd_dev)
+{
+ int ret;
+ size_t size;
+ char *object_name;
+ void *response;
+ void *p;
+
+ /*
+ * First, see if the format 2 image id file exists, and if
+ * so, get the image's persistent id from it.
+ */
+ size = sizeof (RBD_ID_PREFIX) + rbd_dev->image_name_len;
+ object_name = kmalloc(size, GFP_NOIO);
+ if (!object_name)
+ return -ENOMEM;
+ sprintf(object_name, "%s%s", RBD_ID_PREFIX, rbd_dev->image_name);
+ dout("rbd id object name is %s\n", object_name);
+
+ /* Response will be an encoded string, which includes a length */
+
+ size = sizeof (__le32) + RBD_IMAGE_ID_LEN_MAX;
+ response = kzalloc(size, GFP_NOIO);
+ if (!response) {
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ ret = rbd_req_sync_exec(rbd_dev, object_name,
+ "rbd", "get_id",
+ NULL, 0,
+ response, RBD_IMAGE_ID_LEN_MAX,
+ CEPH_OSD_FLAG_READ, NULL);
+ dout("%s: rbd_req_sync_exec returned %d\n", __func__, ret);
+ if (ret < 0)
+ goto out;
+
+ p = response;
+ rbd_dev->image_id = ceph_extract_encoded_string(&p,
+ p + RBD_IMAGE_ID_LEN_MAX,
+ &rbd_dev->image_id_len,
+ GFP_NOIO);
+ if (IS_ERR(rbd_dev->image_id)) {
+ ret = PTR_ERR(rbd_dev->image_id);
+ rbd_dev->image_id = NULL;
+ } else {
+ dout("image_id is %s\n", rbd_dev->image_id);
+ }
+out:
+ kfree(response);
+ kfree(object_name);
+
+ return ret;
+}
+
+static int rbd_dev_v1_probe(struct rbd_device *rbd_dev)
+{
+ int ret;
+ size_t size;
+
+ /* Version 1 images have no id; empty string is used */
+
+ rbd_dev->image_id = kstrdup("", GFP_KERNEL);
+ if (!rbd_dev->image_id)
+ return -ENOMEM;
+ rbd_dev->image_id_len = 0;
+
+ /* Record the header object name for this rbd image. */
+
+ size = rbd_dev->image_name_len + 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->image_name, RBD_SUFFIX);
+
+ /* Populate rbd image metadata */
+
+ ret = rbd_read_header(rbd_dev, &rbd_dev->header);
+ if (ret < 0)
+ goto out_err;
+ rbd_dev->image_format = 1;
+
+ dout("discovered version 1 image, header name is %s\n",
+ rbd_dev->header_name);
+
+ return 0;
+
+out_err:
+ kfree(rbd_dev->header_name);
+ rbd_dev->header_name = NULL;
+ kfree(rbd_dev->image_id);
+ rbd_dev->image_id = NULL;
+
+ return ret;
+}
+
+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) + rbd_dev->image_id_len;
+ 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->image_id);
+
+ /* Get the size and object order for the image */
+
+ ret = rbd_dev_v2_image_size(rbd_dev);
+ if (ret < 0)
+ goto out_err;
+
+ /* Get the object prefix (a.k.a. block_name) for the image */
+
+ ret = rbd_dev_v2_object_prefix(rbd_dev);
+ if (ret < 0)
+ goto out_err;
+
+ /* Get the and check features for the image */
+
+ ret = rbd_dev_v2_features(rbd_dev);
+ if (ret < 0)
+ goto out_err;
+
+ /* crypto and compression type aren't (yet) supported for v2 images */
+
+ rbd_dev->header.crypt_type = 0;
+ rbd_dev->header.comp_type = 0;
+
+ /* Get the snapshot context, plus the header version */
+
+ ret = rbd_dev_v2_snap_context(rbd_dev, &ver);
+ if (ret)
+ goto out_err;
+ rbd_dev->header.obj_version = ver;
+
+ rbd_dev->image_format = 2;
+
+ dout("discovered version 2 image, header name is %s\n",
+ rbd_dev->header_name);
+
+ return -ENOTSUPP;
+out_err:
+ kfree(rbd_dev->header_name);
+ rbd_dev->header_name = NULL;
+ kfree(rbd_dev->header.object_prefix);
+ rbd_dev->header.object_prefix = NULL;
+
+ return ret;
+}
+
+/*
+ * Probe for the existence of the header object for the given rbd
+ * device. For format 2 images this includes determining the image
+ * id.
+ */
+static int rbd_dev_probe(struct rbd_device *rbd_dev)
+{
+ int ret;
+
+ /*
+ * Get the id from the image id object. If it's not a
+ * format 2 image, we'll get ENOENT back, and we'll assume
+ * it's a format 1 image.
+ */
+ ret = rbd_dev_image_id(rbd_dev);
+ if (ret)
+ ret = rbd_dev_v1_probe(rbd_dev);
+ else
+ ret = rbd_dev_v2_probe(rbd_dev);
+ if (ret)
+ dout("probe failed, returning %d\n", ret);
+
+ return ret;
+}
+
static ssize_t rbd_add(struct bus_type *bus,
const char *buf,
size_t count)
goto err_out_client;
rbd_dev->pool_id = rc;
- /* Create the name of the header object */
-
- rbd_dev->header_name = kmalloc(rbd_dev->image_name_len
- + sizeof (RBD_SUFFIX),
- GFP_KERNEL);
- if (!rbd_dev->header_name)
- goto err_out_client;
- sprintf(rbd_dev->header_name, "%s%s", rbd_dev->image_name, RBD_SUFFIX);
-
- /* Get information about the image being mapped */
-
- rc = rbd_read_header(rbd_dev, &rbd_dev->header);
- if (rc)
+ rc = rbd_dev_probe(rbd_dev);
+ if (rc < 0)
goto err_out_client;
/* no need to lock here, as rbd_dev is not registered yet */
err_out_client:
kfree(rbd_dev->header_name);
rbd_put_client(rbd_dev);
+ kfree(rbd_dev->image_id);
err_out_args:
kfree(rbd_dev->mapping.snap_name);
kfree(rbd_dev->image_name);
/* done with the id, and with the rbd_dev */
kfree(rbd_dev->mapping.snap_name);
+ kfree(rbd_dev->image_id);
kfree(rbd_dev->header_name);
kfree(rbd_dev->pool_name);
kfree(rbd_dev->image_name);
return ret;
}
-static ssize_t rbd_snap_add(struct device *dev,
- struct device_attribute *attr,
- const char *buf,
- size_t count)
-{
- struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
- int ret;
- char *name = kmalloc(count + 1, GFP_KERNEL);
- if (!name)
- return -ENOMEM;
-
- snprintf(name, count, "%s", buf);
-
- mutex_lock_nested(&ctl_mutex, SINGLE_DEPTH_NESTING);
-
- ret = rbd_header_add_snap(rbd_dev,
- name, GFP_KERNEL);
- if (ret < 0)
- goto err_unlock;
-
- ret = __rbd_refresh_header(rbd_dev, NULL);
- if (ret < 0)
- goto err_unlock;
-
- /* shouldn't hold ctl_mutex when notifying.. notify might
- trigger a watch callback that would need to get that mutex */
- mutex_unlock(&ctl_mutex);
-
- /* make a best effort, don't error if failed */
- rbd_req_sync_notify(rbd_dev);
-
- ret = count;
- kfree(name);
- return ret;
-
-err_unlock:
- mutex_unlock(&ctl_mutex);
- kfree(name);
- return ret;
-}
-
/*
* create control files in sysfs
* /sys/bus/rbd/...
projects / mirror_ubuntu-artful-kernel.git / blobdiff