/* Features supported by this (client software) implementation. */
-#define RBD_FEATURES_SUPPORTED (0)
+#define RBD_FEATURES_SUPPORTED (RBD_FEATURES_ALL)
/*
* An RBD device name will be "rbd#", where the "rbd" comes from
char *snap_names;
u64 *snap_sizes;
+ u64 stripe_unit;
+ u64 stripe_count;
+
u64 obj_version;
};
*/
struct rbd_spec {
u64 pool_id;
- char *pool_name;
+ const char *pool_name;
- char *image_id;
- char *image_name;
+ const char *image_id;
+ const char *image_name;
u64 snap_id;
- char *snap_name;
+ const char *snap_name;
struct kref kref;
};
enum obj_req_flags {
OBJ_REQ_DONE, /* completion flag: not done = 0, done = 1 */
OBJ_REQ_IMG_DATA, /* object usage: standalone = 0, image = 1 */
+ OBJ_REQ_KNOWN, /* EXISTS flag valid: no = 0, yes = 1 */
+ OBJ_REQ_EXISTS, /* target exists: no = 0, yes = 1 */
};
struct rbd_obj_request {
u64 length; /* bytes from offset */
unsigned long flags;
- struct rbd_img_request *img_request;
- u64 img_offset; /* image relative offset */
- struct list_head links; /* img_request->obj_requests */
+ /*
+ * An object request associated with an image will have its
+ * img_data flag set; a standalone object request will not.
+ *
+ * A standalone object request will have which == BAD_WHICH
+ * and a null obj_request pointer.
+ *
+ * An object request initiated in support of a layered image
+ * object (to check for its existence before a write) will
+ * have which == BAD_WHICH and a non-null obj_request pointer.
+ *
+ * Finally, an object request for rbd image data will have
+ * which != BAD_WHICH, and will have a non-null img_request
+ * pointer. The value of which will be in the range
+ * 0..(img_request->obj_request_count-1).
+ */
+ union {
+ struct rbd_obj_request *obj_request; /* STAT op */
+ struct {
+ struct rbd_img_request *img_request;
+ u64 img_offset;
+ /* links for img_request->obj_requests list */
+ struct list_head links;
+ };
+ };
u32 which; /* posn image request list */
enum obj_request_type type;
u32 page_count;
};
};
+ struct page **copyup_pages;
struct ceph_osd_request *osd_req;
struct request *rq; /* block request */
struct rbd_obj_request *obj_request; /* obj req initiator */
};
+ struct page **copyup_pages;
spinlock_t completion_lock;/* protects next_completion */
u32 next_completion;
rbd_img_callback_t callback;
list_for_each_entry_safe_reverse(oreq, n, &(ireq)->obj_requests, links)
struct rbd_snap {
- struct device dev;
const char *name;
u64 size;
struct list_head node;
static LIST_HEAD(rbd_client_list); /* clients */
static DEFINE_SPINLOCK(rbd_client_list_lock);
+static int rbd_img_request_submit(struct rbd_img_request *img_request);
+
static int rbd_dev_snaps_update(struct rbd_device *rbd_dev);
-static int rbd_dev_snaps_register(struct rbd_device *rbd_dev);
static void rbd_dev_release(struct device *dev);
-static void rbd_remove_snap_dev(struct rbd_snap *snap);
+static void rbd_snap_destroy(struct rbd_snap *snap);
static ssize_t rbd_add(struct bus_type *bus, const char *buf,
size_t count);
static ssize_t rbd_remove(struct bus_type *bus, const char *buf,
size_t count);
-static int rbd_dev_probe(struct rbd_device *rbd_dev);
+static int rbd_dev_image_probe(struct rbd_device *rbd_dev);
static struct bus_attribute rbd_bus_attrs[] = {
__ATTR(add, S_IWUSR, NULL, rbd_add),
# define rbd_assert(expr) ((void) 0)
#endif /* !RBD_DEBUG */
+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->snap_sizes[i] =
le64_to_cpu(ondisk->snaps[i].image_size);
} else {
- WARN_ON(ondisk->snap_names_len);
header->snap_names = NULL;
header->snap_sizes = NULL;
}
/* Allocate and fill in the snapshot context */
header->image_size = le64_to_cpu(ondisk->image_size);
- size = sizeof (struct ceph_snap_context);
- size += snap_count * sizeof (header->snapc->snaps[0]);
- header->snapc = kzalloc(size, GFP_KERNEL);
+
+ header->snapc = rbd_snap_context_create(snap_count);
if (!header->snapc)
goto out_err;
-
- atomic_set(&header->snapc->nref, 1);
header->snapc->seq = le64_to_cpu(ondisk->snap_seq);
- header->snapc->num_snaps = snap_count;
for (i = 0; i < snap_count; i++)
- header->snapc->snaps[i] =
- le64_to_cpu(ondisk->snaps[i].id);
+ header->snapc->snaps[i] = le64_to_cpu(ondisk->snaps[i].id);
return 0;
return NULL;
}
-static int snap_by_name(struct rbd_device *rbd_dev, const char *snap_name)
+static struct rbd_snap *snap_by_name(struct rbd_device *rbd_dev,
+ const char *snap_name)
{
-
struct rbd_snap *snap;
- list_for_each_entry(snap, &rbd_dev->snaps, node) {
- if (!strcmp(snap_name, snap->name)) {
- rbd_dev->spec->snap_id = snap->id;
- rbd_dev->mapping.size = snap->size;
- rbd_dev->mapping.features = snap->features;
-
- return 0;
- }
- }
+ list_for_each_entry(snap, &rbd_dev->snaps, node)
+ if (!strcmp(snap_name, snap->name))
+ return snap;
- return -ENOENT;
+ return NULL;
}
-static int rbd_dev_set_mapping(struct rbd_device *rbd_dev)
+static int rbd_dev_mapping_set(struct rbd_device *rbd_dev)
{
- int ret;
-
if (!memcmp(rbd_dev->spec->snap_name, RBD_SNAP_HEAD_NAME,
sizeof (RBD_SNAP_HEAD_NAME))) {
- rbd_dev->spec->snap_id = CEPH_NOSNAP;
rbd_dev->mapping.size = rbd_dev->header.image_size;
rbd_dev->mapping.features = rbd_dev->header.features;
- ret = 0;
} else {
- ret = snap_by_name(rbd_dev, rbd_dev->spec->snap_name);
- if (ret < 0)
- goto done;
+ struct rbd_snap *snap;
+
+ snap = snap_by_name(rbd_dev, rbd_dev->spec->snap_name);
+ if (!snap)
+ return -ENOENT;
+ rbd_dev->mapping.size = snap->size;
+ rbd_dev->mapping.features = snap->features;
rbd_dev->mapping.read_only = true;
}
- set_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags);
-done:
- return ret;
+ return 0;
+}
+
+static void rbd_dev_mapping_clear(struct rbd_device *rbd_dev)
+{
+ rbd_dev->mapping.size = 0;
+ rbd_dev->mapping.features = 0;
+ rbd_dev->mapping.read_only = true;
}
static void rbd_header_free(struct rbd_image_header *header)
header->snap_sizes = NULL;
kfree(header->snap_names);
header->snap_names = NULL;
- ceph_put_snap_context(header->snapc);
+ rbd_snap_context_put(header->snapc);
header->snapc = NULL;
}
}
}
+/*
+ * similar to zero_bio_chain(), zeros data defined by a page array,
+ * starting at the given byte offset from the start of the array and
+ * continuing up to the given end offset. The pages array is
+ * assumed to be big enough to hold all bytes up to the end.
+ */
+static void zero_pages(struct page **pages, u64 offset, u64 end)
+{
+ struct page **page = &pages[offset >> PAGE_SHIFT];
+
+ rbd_assert(end > offset);
+ rbd_assert(end - offset <= (u64)SIZE_MAX);
+ while (offset < end) {
+ size_t page_offset;
+ size_t length;
+ unsigned long flags;
+ void *kaddr;
+
+ page_offset = (size_t)(offset & ~PAGE_MASK);
+ length = min(PAGE_SIZE - page_offset, (size_t)(end - offset));
+ local_irq_save(flags);
+ kaddr = kmap_atomic(*page);
+ memset(kaddr + page_offset, 0, length);
+ kunmap_atomic(kaddr);
+ local_irq_restore(flags);
+
+ offset += length;
+ page++;
+ }
+}
+
/*
* Clone a portion of a bio, starting at the given byte offset
* and continuing for the number of bytes indicated.
* each flag, once its value is set to 1 it is never reset to 0
* again.
*/
+static void obj_request_img_data_set(struct rbd_obj_request *obj_request)
+{
+ if (test_and_set_bit(OBJ_REQ_IMG_DATA, &obj_request->flags)) {
+ struct rbd_device *rbd_dev;
+
+ rbd_dev = obj_request->img_request->rbd_dev;
+ rbd_warn(rbd_dev, "obj_request %p already marked img_data\n",
+ obj_request);
+ }
+}
+
+static bool obj_request_img_data_test(struct rbd_obj_request *obj_request)
+{
+ smp_mb();
+ return test_bit(OBJ_REQ_IMG_DATA, &obj_request->flags) != 0;
+}
+
static void obj_request_done_set(struct rbd_obj_request *obj_request)
{
if (test_and_set_bit(OBJ_REQ_DONE, &obj_request->flags)) {
- struct rbd_img_request *img_request = obj_request->img_request;
- struct rbd_device *rbd_dev;
+ struct rbd_device *rbd_dev = NULL;
- rbd_dev = img_request ? img_request->rbd_dev : NULL;
+ if (obj_request_img_data_test(obj_request))
+ rbd_dev = obj_request->img_request->rbd_dev;
rbd_warn(rbd_dev, "obj_request %p already marked done\n",
obj_request);
}
return test_bit(OBJ_REQ_DONE, &obj_request->flags) != 0;
}
-static void obj_request_img_data_set(struct rbd_obj_request *obj_request)
+/*
+ * This sets the KNOWN flag after (possibly) setting the EXISTS
+ * flag. The latter is set based on the "exists" value provided.
+ *
+ * Note that for our purposes once an object exists it never goes
+ * away again. It's possible that the response from two existence
+ * checks are separated by the creation of the target object, and
+ * the first ("doesn't exist") response arrives *after* the second
+ * ("does exist"). In that case we ignore the second one.
+ */
+static void obj_request_existence_set(struct rbd_obj_request *obj_request,
+ bool exists)
{
- if (test_and_set_bit(OBJ_REQ_IMG_DATA, &obj_request->flags)) {
- struct rbd_img_request *img_request = obj_request->img_request;
- struct rbd_device *rbd_dev;
+ if (exists)
+ set_bit(OBJ_REQ_EXISTS, &obj_request->flags);
+ set_bit(OBJ_REQ_KNOWN, &obj_request->flags);
+ smp_mb();
+}
- rbd_dev = img_request ? img_request->rbd_dev : NULL;
- rbd_warn(rbd_dev, "obj_request %p already marked img_data\n",
- obj_request);
- }
+static bool obj_request_known_test(struct rbd_obj_request *obj_request)
+{
+ smp_mb();
+ return test_bit(OBJ_REQ_KNOWN, &obj_request->flags) != 0;
}
-static bool obj_request_img_data_test(struct rbd_obj_request *obj_request)
+static bool obj_request_exists_test(struct rbd_obj_request *obj_request)
{
smp_mb();
- return test_bit(OBJ_REQ_IMG_DATA, &obj_request->flags) != 0;
+ return test_bit(OBJ_REQ_EXISTS, &obj_request->flags) != 0;
}
static void rbd_obj_request_get(struct rbd_obj_request *obj_request)
static void
rbd_img_obj_request_read_callback(struct rbd_obj_request *obj_request)
{
+ u64 xferred = obj_request->xferred;
+ u64 length = obj_request->length;
+
dout("%s: obj %p img %p result %d %llu/%llu\n", __func__,
obj_request, obj_request->img_request, obj_request->result,
- obj_request->xferred, obj_request->length);
+ xferred, length);
/*
* ENOENT means a hole in the image. We zero-fill the
* entire length of the request. A short read also implies
* update the xferred count to indicate the whole request
* was satisfied.
*/
- BUG_ON(obj_request->type != OBJ_REQUEST_BIO);
+ rbd_assert(obj_request->type != OBJ_REQUEST_NODATA);
if (obj_request->result == -ENOENT) {
- zero_bio_chain(obj_request->bio_list, 0);
+ if (obj_request->type == OBJ_REQUEST_BIO)
+ zero_bio_chain(obj_request->bio_list, 0);
+ else
+ zero_pages(obj_request->pages, 0, length);
obj_request->result = 0;
- obj_request->xferred = obj_request->length;
- } else if (obj_request->xferred < obj_request->length &&
- !obj_request->result) {
- zero_bio_chain(obj_request->bio_list, obj_request->xferred);
- obj_request->xferred = obj_request->length;
+ obj_request->xferred = length;
+ } else if (xferred < length && !obj_request->result) {
+ if (obj_request->type == OBJ_REQUEST_BIO)
+ zero_bio_chain(obj_request->bio_list, xferred);
+ else
+ zero_pages(obj_request->pages, xferred, length);
+ obj_request->xferred = length;
}
obj_request_done_set(obj_request);
}
static void rbd_osd_read_callback(struct rbd_obj_request *obj_request)
{
- struct rbd_img_request *img_request = obj_request->img_request;
- bool layered = img_request && img_request_layered_test(img_request);
+ struct rbd_img_request *img_request = NULL;
+ struct rbd_device *rbd_dev = NULL;
+ bool layered = false;
+
+ if (obj_request_img_data_test(obj_request)) {
+ img_request = obj_request->img_request;
+ layered = img_request && img_request_layered_test(img_request);
+ rbd_dev = img_request->rbd_dev;
+ }
dout("%s: obj %p img %p result %d %llu/%llu\n", __func__,
obj_request, img_request, obj_request->result,
obj_request->xferred, obj_request->length);
- if (layered && obj_request->result == -ENOENT)
+ if (layered && obj_request->result == -ENOENT &&
+ obj_request->img_offset < rbd_dev->parent_overlap)
rbd_img_parent_read(obj_request);
else if (img_request)
rbd_img_obj_request_read_callback(obj_request);
dout("%s: osd_req %p msg %p\n", __func__, osd_req, msg);
rbd_assert(osd_req == obj_request->osd_req);
- rbd_assert(obj_request_img_data_test(obj_request) ^
- !obj_request->img_request);
- rbd_assert(obj_request_img_data_test(obj_request) ^
- (obj_request->which == BAD_WHICH));
+ if (obj_request_img_data_test(obj_request)) {
+ rbd_assert(obj_request->img_request);
+ rbd_assert(obj_request->which != BAD_WHICH);
+ } else {
+ rbd_assert(obj_request->which == BAD_WHICH);
+ }
if (osd_req->r_result < 0)
obj_request->result = osd_req->r_result;
obj_request->version = le64_to_cpu(osd_req->r_reassert_version.version);
- WARN_ON(osd_req->r_num_ops != 1); /* For now */
+ BUG_ON(osd_req->r_num_ops > 2);
/*
* We support a 64-bit length, but ultimately it has to be
rbd_obj_request_complete(obj_request);
}
-static void rbd_osd_req_format(struct rbd_obj_request *obj_request,
- bool write_request)
+static void rbd_osd_req_format_read(struct rbd_obj_request *obj_request)
{
struct rbd_img_request *img_request = obj_request->img_request;
struct ceph_osd_request *osd_req = obj_request->osd_req;
- struct ceph_snap_context *snapc = NULL;
- u64 snap_id = CEPH_NOSNAP;
- struct timespec *mtime = NULL;
- struct timespec now;
+ u64 snap_id;
rbd_assert(osd_req != NULL);
- if (write_request) {
- now = CURRENT_TIME;
- mtime = &now;
- if (img_request)
- snapc = img_request->snapc;
- } else if (img_request) {
- snap_id = img_request->snap_id;
- }
+ snap_id = img_request ? img_request->snap_id : CEPH_NOSNAP;
+ ceph_osdc_build_request(osd_req, obj_request->offset,
+ NULL, snap_id, NULL);
+}
+
+static void rbd_osd_req_format_write(struct rbd_obj_request *obj_request)
+{
+ struct rbd_img_request *img_request = obj_request->img_request;
+ struct ceph_osd_request *osd_req = obj_request->osd_req;
+ struct ceph_snap_context *snapc;
+ struct timespec mtime = CURRENT_TIME;
+
+ rbd_assert(osd_req != NULL);
+
+ snapc = img_request ? img_request->snapc : NULL;
ceph_osdc_build_request(osd_req, obj_request->offset,
- snapc, snap_id, mtime);
+ snapc, CEPH_NOSNAP, &mtime);
}
static struct ceph_osd_request *rbd_osd_req_create(
return osd_req;
}
+/*
+ * Create a copyup osd request based on the information in the
+ * object request supplied. A copyup request has two osd ops,
+ * a copyup method call, and a "normal" write request.
+ */
+static struct ceph_osd_request *
+rbd_osd_req_create_copyup(struct rbd_obj_request *obj_request)
+{
+ struct rbd_img_request *img_request;
+ struct ceph_snap_context *snapc;
+ struct rbd_device *rbd_dev;
+ struct ceph_osd_client *osdc;
+ struct ceph_osd_request *osd_req;
+
+ rbd_assert(obj_request_img_data_test(obj_request));
+ img_request = obj_request->img_request;
+ rbd_assert(img_request);
+ rbd_assert(img_request_write_test(img_request));
+
+ /* Allocate and initialize the request, for the two ops */
+
+ snapc = img_request->snapc;
+ rbd_dev = img_request->rbd_dev;
+ osdc = &rbd_dev->rbd_client->client->osdc;
+ osd_req = ceph_osdc_alloc_request(osdc, snapc, 2, false, GFP_ATOMIC);
+ if (!osd_req)
+ return NULL; /* ENOMEM */
+
+ osd_req->r_flags = CEPH_OSD_FLAG_WRITE | CEPH_OSD_FLAG_ONDISK;
+ osd_req->r_callback = rbd_osd_req_callback;
+ osd_req->r_priv = obj_request;
+
+ osd_req->r_oid_len = strlen(obj_request->object_name);
+ rbd_assert(osd_req->r_oid_len < sizeof (osd_req->r_oid));
+ memcpy(osd_req->r_oid, obj_request->object_name, osd_req->r_oid_len);
+
+ osd_req->r_file_layout = rbd_dev->layout; /* struct */
+
+ return osd_req;
+}
+
+
static void rbd_osd_req_destroy(struct ceph_osd_request *osd_req)
{
ceph_osdc_put_request(osd_req);
bool child_request)
{
struct rbd_img_request *img_request;
- struct ceph_snap_context *snapc = NULL;
img_request = kmalloc(sizeof (*img_request), GFP_ATOMIC);
if (!img_request)
if (write_request) {
down_read(&rbd_dev->header_rwsem);
- snapc = ceph_get_snap_context(rbd_dev->header.snapc);
+ rbd_snap_context_get(rbd_dev->header.snapc);
up_read(&rbd_dev->header_rwsem);
- if (WARN_ON(!snapc)) {
- kfree(img_request);
- return NULL; /* Shouldn't happen */
- }
-
}
img_request->rq = NULL;
img_request->flags = 0;
if (write_request) {
img_request_write_set(img_request);
- img_request->snapc = snapc;
+ img_request->snapc = rbd_dev->header.snapc;
} else {
img_request->snap_id = rbd_dev->spec->snap_id;
}
rbd_assert(img_request->obj_request_count == 0);
if (img_request_write_test(img_request))
- ceph_put_snap_context(img_request->snapc);
+ rbd_snap_context_put(img_request->snapc);
if (img_request_child_test(img_request))
rbd_obj_request_put(img_request->obj_request);
img_request->result = result;
}
+ /* Image object requests don't own their page array */
+
+ if (obj_request->type == OBJ_REQUEST_PAGES) {
+ obj_request->pages = NULL;
+ obj_request->page_count = 0;
+ }
+
if (img_request_child_test(img_request)) {
rbd_assert(img_request->obj_request != NULL);
more = obj_request->which < img_request->obj_request_count - 1;
rbd_img_request_complete(img_request);
}
-static int rbd_img_request_fill_bio(struct rbd_img_request *img_request,
- struct bio *bio_list)
+/*
+ * Split up an image request into one or more object requests, each
+ * to a different object. The "type" parameter indicates whether
+ * "data_desc" is the pointer to the head of a list of bio
+ * structures, or the base of a page array. In either case this
+ * function assumes data_desc describes memory sufficient to hold
+ * all data described by the image request.
+ */
+static int rbd_img_request_fill(struct rbd_img_request *img_request,
+ enum obj_request_type type,
+ void *data_desc)
{
struct rbd_device *rbd_dev = img_request->rbd_dev;
struct rbd_obj_request *obj_request = NULL;
struct rbd_obj_request *next_obj_request;
bool write_request = img_request_write_test(img_request);
- unsigned int bio_offset;
+ struct bio *bio_list;
+ unsigned int bio_offset = 0;
+ struct page **pages;
u64 img_offset;
u64 resid;
u16 opcode;
- dout("%s: img %p bio %p\n", __func__, img_request, bio_list);
+ dout("%s: img %p type %d data_desc %p\n", __func__, img_request,
+ (int)type, data_desc);
opcode = write_request ? CEPH_OSD_OP_WRITE : CEPH_OSD_OP_READ;
- bio_offset = 0;
img_offset = img_request->offset;
- rbd_assert(img_offset == bio_list->bi_sector << SECTOR_SHIFT);
resid = img_request->length;
rbd_assert(resid > 0);
+
+ if (type == OBJ_REQUEST_BIO) {
+ bio_list = data_desc;
+ rbd_assert(img_offset == bio_list->bi_sector << SECTOR_SHIFT);
+ } else {
+ rbd_assert(type == OBJ_REQUEST_PAGES);
+ pages = data_desc;
+ }
+
while (resid) {
struct ceph_osd_request *osd_req;
const char *object_name;
- unsigned int clone_size;
u64 offset;
u64 length;
offset = rbd_segment_offset(rbd_dev, img_offset);
length = rbd_segment_length(rbd_dev, img_offset, resid);
obj_request = rbd_obj_request_create(object_name,
- offset, length,
- OBJ_REQUEST_BIO);
+ offset, length, type);
kfree(object_name); /* object request has its own copy */
if (!obj_request)
goto out_unwind;
- rbd_assert(length <= (u64) UINT_MAX);
- clone_size = (unsigned int) length;
- obj_request->bio_list = bio_chain_clone_range(&bio_list,
- &bio_offset, clone_size,
- GFP_ATOMIC);
- if (!obj_request->bio_list)
- goto out_partial;
+ if (type == OBJ_REQUEST_BIO) {
+ unsigned int clone_size;
+
+ rbd_assert(length <= (u64)UINT_MAX);
+ clone_size = (unsigned int)length;
+ obj_request->bio_list =
+ bio_chain_clone_range(&bio_list,
+ &bio_offset,
+ clone_size,
+ GFP_ATOMIC);
+ if (!obj_request->bio_list)
+ goto out_partial;
+ } else {
+ unsigned int page_count;
+
+ obj_request->pages = pages;
+ page_count = (u32)calc_pages_for(offset, length);
+ obj_request->page_count = page_count;
+ if ((offset + length) & ~PAGE_MASK)
+ page_count--; /* more on last page */
+ pages += page_count;
+ }
osd_req = rbd_osd_req_create(rbd_dev, write_request,
obj_request);
osd_req_op_extent_init(osd_req, 0, opcode, offset, length,
0, 0);
- osd_req_op_extent_osd_data_bio(osd_req, 0,
- obj_request->bio_list, obj_request->length);
- rbd_osd_req_format(obj_request, write_request);
+ if (type == OBJ_REQUEST_BIO)
+ osd_req_op_extent_osd_data_bio(osd_req, 0,
+ obj_request->bio_list, length);
+ else
+ osd_req_op_extent_osd_data_pages(osd_req, 0,
+ obj_request->pages, length,
+ offset & ~PAGE_MASK, false, false);
+
+ if (write_request)
+ rbd_osd_req_format_write(obj_request);
+ else
+ rbd_osd_req_format_read(obj_request);
obj_request->img_offset = img_offset;
rbd_img_obj_request_add(img_request, obj_request);
return -ENOMEM;
}
-static int rbd_img_request_submit(struct rbd_img_request *img_request)
+static void
+rbd_img_obj_copyup_callback(struct rbd_obj_request *obj_request)
{
- struct rbd_device *rbd_dev = img_request->rbd_dev;
- struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
- struct rbd_obj_request *obj_request;
- struct rbd_obj_request *next_obj_request;
-
- dout("%s: img %p\n", __func__, img_request);
- for_each_obj_request_safe(img_request, obj_request, next_obj_request) {
- int ret;
+ struct rbd_img_request *img_request;
+ struct rbd_device *rbd_dev;
+ u64 length;
+ u32 page_count;
- ret = rbd_obj_request_submit(osdc, obj_request);
- if (ret)
- return ret;
- }
+ rbd_assert(obj_request->type == OBJ_REQUEST_BIO);
+ rbd_assert(obj_request_img_data_test(obj_request));
+ img_request = obj_request->img_request;
+ rbd_assert(img_request);
- return 0;
-}
+ rbd_dev = img_request->rbd_dev;
+ rbd_assert(rbd_dev);
+ length = (u64)1 << rbd_dev->header.obj_order;
+ page_count = (u32)calc_pages_for(0, length);
-static void rbd_img_parent_read_callback(struct rbd_img_request *img_request)
-{
- struct rbd_obj_request *obj_request;
+ rbd_assert(obj_request->copyup_pages);
+ ceph_release_page_vector(obj_request->copyup_pages, page_count);
+ obj_request->copyup_pages = NULL;
- rbd_assert(img_request_child_test(img_request));
+ /*
+ * We want the transfer count to reflect the size of the
+ * original write request. There is no such thing as a
+ * successful short write, so if the request was successful
+ * we can just set it to the originally-requested length.
+ */
+ if (!obj_request->result)
+ obj_request->xferred = obj_request->length;
- obj_request = img_request->obj_request;
- rbd_assert(obj_request != NULL);
- obj_request->result = img_request->result;
- obj_request->xferred = img_request->xferred;
+ /* Finish up with the normal image object callback */
- rbd_img_obj_request_read_callback(obj_request);
- rbd_obj_request_complete(obj_request);
+ rbd_img_obj_callback(obj_request);
}
-static void rbd_img_parent_read(struct rbd_obj_request *obj_request)
+static void
+rbd_img_obj_parent_read_full_callback(struct rbd_img_request *img_request)
{
+ struct rbd_obj_request *orig_request;
+ struct ceph_osd_request *osd_req;
+ struct ceph_osd_client *osdc;
struct rbd_device *rbd_dev;
- struct rbd_img_request *img_request;
+ struct page **pages;
int result;
+ u64 obj_size;
+ u64 xferred;
- rbd_assert(obj_request_img_data_test(obj_request));
- rbd_assert(obj_request->img_request != NULL);
- rbd_assert(obj_request->result == (s32) -ENOENT);
- rbd_assert(obj_request->type == OBJ_REQUEST_BIO);
+ rbd_assert(img_request_child_test(img_request));
- rbd_dev = obj_request->img_request->rbd_dev;
- rbd_assert(rbd_dev->parent != NULL);
- /* rbd_read_finish(obj_request, obj_request->length); */
- img_request = rbd_img_request_create(rbd_dev->parent,
- obj_request->img_offset,
- obj_request->length,
- false, true);
- result = -ENOMEM;
- if (!img_request)
- goto out_err;
+ /* First get what we need from the image request */
- rbd_obj_request_get(obj_request);
- img_request->obj_request = obj_request;
+ pages = img_request->copyup_pages;
+ rbd_assert(pages != NULL);
+ img_request->copyup_pages = NULL;
+
+ orig_request = img_request->obj_request;
+ rbd_assert(orig_request != NULL);
+ rbd_assert(orig_request->type == OBJ_REQUEST_BIO);
+ result = img_request->result;
+ obj_size = img_request->length;
+ xferred = img_request->xferred;
+
+ rbd_dev = img_request->rbd_dev;
+ rbd_assert(rbd_dev);
+ rbd_assert(obj_size == (u64)1 << rbd_dev->header.obj_order);
+
+ rbd_img_request_put(img_request);
- result = rbd_img_request_fill_bio(img_request, obj_request->bio_list);
if (result)
goto out_err;
- img_request->callback = rbd_img_parent_read_callback;
- result = rbd_img_request_submit(img_request);
- if (result)
+ /* Allocate the new copyup osd request for the original request */
+
+ result = -ENOMEM;
+ rbd_assert(!orig_request->osd_req);
+ osd_req = rbd_osd_req_create_copyup(orig_request);
+ if (!osd_req)
goto out_err;
+ orig_request->osd_req = osd_req;
+ orig_request->copyup_pages = pages;
- return;
-out_err:
- if (img_request)
- rbd_img_request_put(img_request);
- obj_request->result = result;
- obj_request->xferred = 0;
- obj_request_done_set(obj_request);
-}
+ /* Initialize the copyup op */
-static int rbd_obj_notify_ack(struct rbd_device *rbd_dev,
- u64 ver, u64 notify_id)
-{
- struct rbd_obj_request *obj_request;
- struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
- int ret;
+ osd_req_op_cls_init(osd_req, 0, CEPH_OSD_OP_CALL, "rbd", "copyup");
+ osd_req_op_cls_request_data_pages(osd_req, 0, pages, obj_size, 0,
+ false, false);
- obj_request = rbd_obj_request_create(rbd_dev->header_name, 0, 0,
- OBJ_REQUEST_NODATA);
- if (!obj_request)
- return -ENOMEM;
+ /* Then the original write request op */
- ret = -ENOMEM;
- obj_request->osd_req = rbd_osd_req_create(rbd_dev, false, obj_request);
- if (!obj_request->osd_req)
- goto out;
- obj_request->callback = rbd_obj_request_put;
+ osd_req_op_extent_init(osd_req, 1, CEPH_OSD_OP_WRITE,
+ orig_request->offset,
+ orig_request->length, 0, 0);
+ osd_req_op_extent_osd_data_bio(osd_req, 1, orig_request->bio_list,
+ orig_request->length);
- osd_req_op_watch_init(obj_request->osd_req, 0, CEPH_OSD_OP_NOTIFY_ACK,
- notify_id, ver, 0);
- rbd_osd_req_format(obj_request, false);
+ rbd_osd_req_format_write(orig_request);
- ret = rbd_obj_request_submit(osdc, obj_request);
-out:
- if (ret)
+ /* All set, send it off. */
+
+ orig_request->callback = rbd_img_obj_copyup_callback;
+ osdc = &rbd_dev->rbd_client->client->osdc;
+ result = rbd_obj_request_submit(osdc, orig_request);
+ if (!result)
+ return;
+out_err:
+ /* Record the error code and complete the request */
+
+ orig_request->result = result;
+ orig_request->xferred = 0;
+ obj_request_done_set(orig_request);
+ rbd_obj_request_complete(orig_request);
+}
+
+/*
+ * Read from the parent image the range of data that covers the
+ * entire target of the given object request. This is used for
+ * satisfying a layered image write request when the target of an
+ * object request from the image request does not exist.
+ *
+ * A page array big enough to hold the returned data is allocated
+ * and supplied to rbd_img_request_fill() as the "data descriptor."
+ * When the read completes, this page array will be transferred to
+ * the original object request for the copyup operation.
+ *
+ * If an error occurs, record it as the result of the original
+ * object request and mark it done so it gets completed.
+ */
+static int rbd_img_obj_parent_read_full(struct rbd_obj_request *obj_request)
+{
+ struct rbd_img_request *img_request = NULL;
+ struct rbd_img_request *parent_request = NULL;
+ struct rbd_device *rbd_dev;
+ u64 img_offset;
+ u64 length;
+ struct page **pages = NULL;
+ u32 page_count;
+ int result;
+
+ rbd_assert(obj_request_img_data_test(obj_request));
+ rbd_assert(obj_request->type == OBJ_REQUEST_BIO);
+
+ img_request = obj_request->img_request;
+ rbd_assert(img_request != NULL);
+ rbd_dev = img_request->rbd_dev;
+ rbd_assert(rbd_dev->parent != NULL);
+
+ /*
+ * First things first. The original osd request is of no
+ * use to use any more, we'll need a new one that can hold
+ * the two ops in a copyup request. We'll get that later,
+ * but for now we can release the old one.
+ */
+ rbd_osd_req_destroy(obj_request->osd_req);
+ obj_request->osd_req = NULL;
+
+ /*
+ * Determine the byte range covered by the object in the
+ * child image to which the original request was to be sent.
+ */
+ img_offset = obj_request->img_offset - obj_request->offset;
+ length = (u64)1 << rbd_dev->header.obj_order;
+
+ /*
+ * There is no defined parent data beyond the parent
+ * overlap, so limit what we read at that boundary if
+ * necessary.
+ */
+ if (img_offset + length > rbd_dev->parent_overlap) {
+ rbd_assert(img_offset < rbd_dev->parent_overlap);
+ length = rbd_dev->parent_overlap - img_offset;
+ }
+
+ /*
+ * Allocate a page array big enough to receive the data read
+ * from the parent.
+ */
+ page_count = (u32)calc_pages_for(0, length);
+ pages = ceph_alloc_page_vector(page_count, GFP_KERNEL);
+ if (IS_ERR(pages)) {
+ result = PTR_ERR(pages);
+ pages = NULL;
+ goto out_err;
+ }
+
+ result = -ENOMEM;
+ parent_request = rbd_img_request_create(rbd_dev->parent,
+ img_offset, length,
+ false, true);
+ if (!parent_request)
+ goto out_err;
+ rbd_obj_request_get(obj_request);
+ parent_request->obj_request = obj_request;
+
+ result = rbd_img_request_fill(parent_request, OBJ_REQUEST_PAGES, pages);
+ if (result)
+ goto out_err;
+ parent_request->copyup_pages = pages;
+
+ parent_request->callback = rbd_img_obj_parent_read_full_callback;
+ result = rbd_img_request_submit(parent_request);
+ if (!result)
+ return 0;
+
+ parent_request->copyup_pages = NULL;
+ parent_request->obj_request = NULL;
+ rbd_obj_request_put(obj_request);
+out_err:
+ if (pages)
+ ceph_release_page_vector(pages, page_count);
+ if (parent_request)
+ rbd_img_request_put(parent_request);
+ obj_request->result = result;
+ obj_request->xferred = 0;
+ obj_request_done_set(obj_request);
+
+ return result;
+}
+
+static void rbd_img_obj_exists_callback(struct rbd_obj_request *obj_request)
+{
+ struct rbd_obj_request *orig_request;
+ int result;
+
+ rbd_assert(!obj_request_img_data_test(obj_request));
+
+ /*
+ * All we need from the object request is the original
+ * request and the result of the STAT op. Grab those, then
+ * we're done with the request.
+ */
+ orig_request = obj_request->obj_request;
+ obj_request->obj_request = NULL;
+ rbd_assert(orig_request);
+ rbd_assert(orig_request->img_request);
+
+ result = obj_request->result;
+ obj_request->result = 0;
+
+ dout("%s: obj %p for obj %p result %d %llu/%llu\n", __func__,
+ obj_request, orig_request, result,
+ obj_request->xferred, obj_request->length);
+ rbd_obj_request_put(obj_request);
+
+ rbd_assert(orig_request);
+ rbd_assert(orig_request->img_request);
+
+ /*
+ * Our only purpose here is to determine whether the object
+ * exists, and we don't want to treat the non-existence as
+ * an error. If something else comes back, transfer the
+ * error to the original request and complete it now.
+ */
+ if (!result) {
+ obj_request_existence_set(orig_request, true);
+ } else if (result == -ENOENT) {
+ obj_request_existence_set(orig_request, false);
+ } else if (result) {
+ orig_request->result = result;
+ goto out;
+ }
+
+ /*
+ * Resubmit the original request now that we have recorded
+ * whether the target object exists.
+ */
+ orig_request->result = rbd_img_obj_request_submit(orig_request);
+out:
+ if (orig_request->result)
+ rbd_obj_request_complete(orig_request);
+ rbd_obj_request_put(orig_request);
+}
+
+static int rbd_img_obj_exists_submit(struct rbd_obj_request *obj_request)
+{
+ struct rbd_obj_request *stat_request;
+ struct rbd_device *rbd_dev;
+ struct ceph_osd_client *osdc;
+ struct page **pages = NULL;
+ u32 page_count;
+ size_t size;
+ int ret;
+
+ /*
+ * The response data for a STAT call consists of:
+ * le64 length;
+ * struct {
+ * le32 tv_sec;
+ * le32 tv_nsec;
+ * } mtime;
+ */
+ size = sizeof (__le64) + sizeof (__le32) + sizeof (__le32);
+ page_count = (u32)calc_pages_for(0, size);
+ pages = ceph_alloc_page_vector(page_count, GFP_KERNEL);
+ if (IS_ERR(pages))
+ return PTR_ERR(pages);
+
+ ret = -ENOMEM;
+ stat_request = rbd_obj_request_create(obj_request->object_name, 0, 0,
+ OBJ_REQUEST_PAGES);
+ if (!stat_request)
+ goto out;
+
+ rbd_obj_request_get(obj_request);
+ stat_request->obj_request = obj_request;
+ stat_request->pages = pages;
+ stat_request->page_count = page_count;
+
+ rbd_assert(obj_request->img_request);
+ rbd_dev = obj_request->img_request->rbd_dev;
+ stat_request->osd_req = rbd_osd_req_create(rbd_dev, false,
+ stat_request);
+ if (!stat_request->osd_req)
+ goto out;
+ stat_request->callback = rbd_img_obj_exists_callback;
+
+ osd_req_op_init(stat_request->osd_req, 0, CEPH_OSD_OP_STAT);
+ osd_req_op_raw_data_in_pages(stat_request->osd_req, 0, pages, size, 0,
+ false, false);
+ rbd_osd_req_format_read(stat_request);
+
+ osdc = &rbd_dev->rbd_client->client->osdc;
+ ret = rbd_obj_request_submit(osdc, stat_request);
+out:
+ if (ret)
+ rbd_obj_request_put(obj_request);
+
+ return ret;
+}
+
+static int rbd_img_obj_request_submit(struct rbd_obj_request *obj_request)
+{
+ struct rbd_img_request *img_request;
+ struct rbd_device *rbd_dev;
+ bool known;
+
+ rbd_assert(obj_request_img_data_test(obj_request));
+
+ img_request = obj_request->img_request;
+ rbd_assert(img_request);
+ rbd_dev = img_request->rbd_dev;
+
+ /*
+ * Only writes to layered images need special handling.
+ * Reads and non-layered writes are simple object requests.
+ * Layered writes that start beyond the end of the overlap
+ * with the parent have no parent data, so they too are
+ * simple object requests. Finally, if the target object is
+ * known to already exist, its parent data has already been
+ * copied, so a write to the object can also be handled as a
+ * simple object request.
+ */
+ if (!img_request_write_test(img_request) ||
+ !img_request_layered_test(img_request) ||
+ rbd_dev->parent_overlap <= obj_request->img_offset ||
+ ((known = obj_request_known_test(obj_request)) &&
+ obj_request_exists_test(obj_request))) {
+
+ struct rbd_device *rbd_dev;
+ struct ceph_osd_client *osdc;
+
+ rbd_dev = obj_request->img_request->rbd_dev;
+ osdc = &rbd_dev->rbd_client->client->osdc;
+
+ return rbd_obj_request_submit(osdc, obj_request);
+ }
+
+ /*
+ * It's a layered write. The target object might exist but
+ * we may not know that yet. If we know it doesn't exist,
+ * start by reading the data for the full target object from
+ * the parent so we can use it for a copyup to the target.
+ */
+ if (known)
+ return rbd_img_obj_parent_read_full(obj_request);
+
+ /* We don't know whether the target exists. Go find out. */
+
+ return rbd_img_obj_exists_submit(obj_request);
+}
+
+static int rbd_img_request_submit(struct rbd_img_request *img_request)
+{
+ struct rbd_obj_request *obj_request;
+ struct rbd_obj_request *next_obj_request;
+
+ dout("%s: img %p\n", __func__, img_request);
+ for_each_obj_request_safe(img_request, obj_request, next_obj_request) {
+ int ret;
+
+ ret = rbd_img_obj_request_submit(obj_request);
+ if (ret)
+ return ret;
+ }
+
+ return 0;
+}
+
+static void rbd_img_parent_read_callback(struct rbd_img_request *img_request)
+{
+ struct rbd_obj_request *obj_request;
+ struct rbd_device *rbd_dev;
+ u64 obj_end;
+
+ rbd_assert(img_request_child_test(img_request));
+
+ obj_request = img_request->obj_request;
+ rbd_assert(obj_request);
+ rbd_assert(obj_request->img_request);
+
+ obj_request->result = img_request->result;
+ if (obj_request->result)
+ goto out;
+
+ /*
+ * We need to zero anything beyond the parent overlap
+ * boundary. Since rbd_img_obj_request_read_callback()
+ * will zero anything beyond the end of a short read, an
+ * easy way to do this is to pretend the data from the
+ * parent came up short--ending at the overlap boundary.
+ */
+ rbd_assert(obj_request->img_offset < U64_MAX - obj_request->length);
+ obj_end = obj_request->img_offset + obj_request->length;
+ rbd_dev = obj_request->img_request->rbd_dev;
+ if (obj_end > rbd_dev->parent_overlap) {
+ u64 xferred = 0;
+
+ if (obj_request->img_offset < rbd_dev->parent_overlap)
+ xferred = rbd_dev->parent_overlap -
+ obj_request->img_offset;
+
+ obj_request->xferred = min(img_request->xferred, xferred);
+ } else {
+ obj_request->xferred = img_request->xferred;
+ }
+out:
+ rbd_img_obj_request_read_callback(obj_request);
+ rbd_obj_request_complete(obj_request);
+}
+
+static void rbd_img_parent_read(struct rbd_obj_request *obj_request)
+{
+ struct rbd_device *rbd_dev;
+ struct rbd_img_request *img_request;
+ int result;
+
+ rbd_assert(obj_request_img_data_test(obj_request));
+ rbd_assert(obj_request->img_request != NULL);
+ rbd_assert(obj_request->result == (s32) -ENOENT);
+ rbd_assert(obj_request->type == OBJ_REQUEST_BIO);
+
+ rbd_dev = obj_request->img_request->rbd_dev;
+ rbd_assert(rbd_dev->parent != NULL);
+ /* rbd_read_finish(obj_request, obj_request->length); */
+ img_request = rbd_img_request_create(rbd_dev->parent,
+ obj_request->img_offset,
+ obj_request->length,
+ false, true);
+ result = -ENOMEM;
+ if (!img_request)
+ goto out_err;
+
+ rbd_obj_request_get(obj_request);
+ img_request->obj_request = obj_request;
+
+ result = rbd_img_request_fill(img_request, OBJ_REQUEST_BIO,
+ obj_request->bio_list);
+ if (result)
+ goto out_err;
+
+ img_request->callback = rbd_img_parent_read_callback;
+ result = rbd_img_request_submit(img_request);
+ if (result)
+ goto out_err;
+
+ return;
+out_err:
+ if (img_request)
+ rbd_img_request_put(img_request);
+ obj_request->result = result;
+ obj_request->xferred = 0;
+ obj_request_done_set(obj_request);
+}
+
+static int rbd_obj_notify_ack(struct rbd_device *rbd_dev,
+ u64 ver, u64 notify_id)
+{
+ struct rbd_obj_request *obj_request;
+ struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
+ int ret;
+
+ obj_request = rbd_obj_request_create(rbd_dev->header_name, 0, 0,
+ OBJ_REQUEST_NODATA);
+ if (!obj_request)
+ return -ENOMEM;
+
+ ret = -ENOMEM;
+ obj_request->osd_req = rbd_osd_req_create(rbd_dev, false, obj_request);
+ if (!obj_request->osd_req)
+ goto out;
+ obj_request->callback = rbd_obj_request_put;
+
+ osd_req_op_watch_init(obj_request->osd_req, 0, CEPH_OSD_OP_NOTIFY_ACK,
+ notify_id, ver, 0);
+ rbd_osd_req_format_read(obj_request);
+
+ ret = rbd_obj_request_submit(osdc, obj_request);
+out:
+ if (ret)
rbd_obj_request_put(obj_request);
return ret;
{
struct rbd_device *rbd_dev = (struct rbd_device *)data;
u64 hver;
- int rc;
if (!rbd_dev)
return;
dout("%s: \"%s\" notify_id %llu opcode %u\n", __func__,
rbd_dev->header_name, (unsigned long long) notify_id,
(unsigned int) opcode);
- rc = rbd_dev_refresh(rbd_dev, &hver);
- if (rc)
- rbd_warn(rbd_dev, "got notification but failed to "
- " update snaps: %d\n", rc);
+ (void)rbd_dev_refresh(rbd_dev, &hver);
rbd_obj_notify_ack(rbd_dev, hver, notify_id);
}
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_osd_req_format(obj_request, true);
+ rbd_osd_req_format_write(obj_request);
ret = rbd_obj_request_submit(osdc, obj_request);
if (ret)
}
/*
- * Synchronous osd object method call
+ * Synchronous osd object method call. Returns the number of bytes
+ * returned in the outbound buffer, or a negative error code.
*/
static int rbd_obj_method_sync(struct rbd_device *rbd_dev,
const char *object_name,
const char *class_name,
const char *method_name,
- const char *outbound,
+ const void *outbound,
size_t outbound_size,
- char *inbound,
+ void *inbound,
size_t inbound_size,
u64 *version)
{
* method. Currently if this is present it will be a
* snapshot id.
*/
- page_count = (u32) calc_pages_for(0, inbound_size);
+ page_count = (u32)calc_pages_for(0, inbound_size);
pages = ceph_alloc_page_vector(page_count, GFP_KERNEL);
if (IS_ERR(pages))
return PTR_ERR(pages);
osd_req_op_cls_response_data_pages(obj_request->osd_req, 0,
obj_request->pages, inbound_size,
0, false, false);
- rbd_osd_req_format(obj_request, false);
+ rbd_osd_req_format_read(obj_request);
ret = rbd_obj_request_submit(osdc, obj_request);
if (ret)
ret = obj_request->result;
if (ret < 0)
goto out;
- ret = 0;
+
+ rbd_assert(obj_request->xferred < (u64)INT_MAX);
+ ret = (int)obj_request->xferred;
ceph_copy_from_page_vector(pages, inbound, 0, obj_request->xferred);
if (version)
*version = obj_request->version;
}
result = -EINVAL;
- if (WARN_ON(offset && length > U64_MAX - offset + 1))
+ if (offset && length > U64_MAX - offset + 1) {
+ rbd_warn(rbd_dev, "bad request range (%llu~%llu)\n",
+ offset, length);
goto end_request; /* Shouldn't happen */
+ }
result = -ENOMEM;
img_request = rbd_img_request_create(rbd_dev, offset, length,
img_request->rq = rq;
- result = rbd_img_request_fill_bio(img_request, rq->bio);
+ result = rbd_img_request_fill(img_request, OBJ_REQUEST_BIO,
+ rq->bio);
if (!result)
result = rbd_img_request_submit(img_request);
if (result)
if (!disk)
return;
- if (disk->flags & GENHD_FL_UP)
+ rbd_dev->disk = NULL;
+ if (disk->flags & GENHD_FL_UP) {
del_gendisk(disk);
- if (disk->queue)
- blk_cleanup_queue(disk->queue);
+ if (disk->queue)
+ blk_cleanup_queue(disk->queue);
+ }
put_disk(disk);
}
static int rbd_obj_read_sync(struct rbd_device *rbd_dev,
const char *object_name,
u64 offset, u64 length,
- char *buf, u64 *version)
+ void *buf, u64 *version)
{
struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
obj_request->length,
obj_request->offset & ~PAGE_MASK,
false, false);
- rbd_osd_req_format(obj_request, false);
+ rbd_osd_req_format_read(obj_request);
ret = rbd_obj_request_submit(osdc, obj_request);
if (ret)
return ERR_PTR(-ENOMEM);
ret = rbd_obj_read_sync(rbd_dev, rbd_dev->header_name,
- 0, size,
- (char *) ondisk, version);
+ 0, size, ondisk, version);
if (ret < 0)
goto out_err;
- if (WARN_ON((size_t) ret < size)) {
+ if ((size_t)ret < size) {
ret = -ENXIO;
rbd_warn(rbd_dev, "short header read (want %zd got %d)",
size, ret);
struct rbd_snap *snap;
struct rbd_snap *next;
- list_for_each_entry_safe(snap, next, &rbd_dev->snaps, node)
- rbd_remove_snap_dev(snap);
+ list_for_each_entry_safe(snap, next, &rbd_dev->snaps, node) {
+ list_del(&snap->node);
+ rbd_snap_destroy(snap);
+ }
}
static void rbd_update_mapping_size(struct rbd_device *rbd_dev)
{
- sector_t size;
-
if (rbd_dev->spec->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);
+ if (rbd_dev->mapping.size != rbd_dev->header.image_size) {
+ sector_t size;
+
+ rbd_dev->mapping.size = rbd_dev->header.image_size;
+ size = (sector_t)rbd_dev->mapping.size / SECTOR_SIZE;
+ dout("setting size to %llu sectors", (unsigned long long)size);
+ set_capacity(rbd_dev->disk, size);
+ }
}
/*
kfree(rbd_dev->header.snap_sizes);
kfree(rbd_dev->header.snap_names);
/* osd requests may still refer to snapc */
- ceph_put_snap_context(rbd_dev->header.snapc);
+ rbd_snap_context_put(rbd_dev->header.snapc);
if (hver)
*hver = h.obj_version;
rbd_dev->header.snap_names = h.snap_names;
rbd_dev->header.snap_sizes = h.snap_sizes;
/* Free the extra copy of the object prefix */
- WARN_ON(strcmp(rbd_dev->header.object_prefix, h.object_prefix));
+ if (strcmp(rbd_dev->header.object_prefix, h.object_prefix))
+ rbd_warn(rbd_dev, "object prefix changed (ignoring)");
kfree(h.object_prefix);
ret = rbd_dev_snaps_update(rbd_dev);
- if (!ret)
- ret = rbd_dev_snaps_register(rbd_dev);
up_write(&rbd_dev->header_rwsem);
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);
return ret;
}
rbd_dev->disk = disk;
- set_capacity(rbd_dev->disk, rbd_dev->mapping.size / SECTOR_SIZE);
-
return 0;
out_disk:
put_disk(disk);
struct device_attribute *attr, char *buf)
{
struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
- sector_t size;
-
- down_read(&rbd_dev->header_rwsem);
- size = get_capacity(rbd_dev->disk);
- up_read(&rbd_dev->header_rwsem);
- return sprintf(buf, "%llu\n", (unsigned long long) size * SECTOR_SIZE);
+ return sprintf(buf, "%llu\n",
+ (unsigned long long)rbd_dev->mapping.size);
}
/*
struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
return sprintf(buf, "0x%016llx\n",
- (unsigned long long) rbd_dev->mapping.features);
+ (unsigned long long)rbd_dev->mapping.features);
}
static ssize_t rbd_major_show(struct device *dev,
{
struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
- return sprintf(buf, "%d\n", rbd_dev->major);
+ if (rbd_dev->major)
+ return sprintf(buf, "%d\n", rbd_dev->major);
+
+ return sprintf(buf, "(none)\n");
+
}
static ssize_t rbd_client_id_show(struct device *dev,
struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
return sprintf(buf, "%llu\n",
- (unsigned long long) rbd_dev->spec->pool_id);
+ (unsigned long long) rbd_dev->spec->pool_id);
}
static ssize_t rbd_name_show(struct device *dev,
.release = rbd_sysfs_dev_release,
};
-
-/*
- sysfs - snapshots
-*/
-
-static ssize_t rbd_snap_size_show(struct device *dev,
- struct device_attribute *attr,
- char *buf)
-{
- struct rbd_snap *snap = container_of(dev, struct rbd_snap, dev);
-
- return sprintf(buf, "%llu\n", (unsigned long long)snap->size);
-}
-
-static ssize_t rbd_snap_id_show(struct device *dev,
- struct device_attribute *attr,
- char *buf)
-{
- struct rbd_snap *snap = container_of(dev, struct rbd_snap, dev);
-
- 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 attribute_group rbd_snap_attr_group = {
- .attrs = rbd_snap_attrs,
-};
-
-static void rbd_snap_dev_release(struct device *dev)
-{
- struct rbd_snap *snap = container_of(dev, struct rbd_snap, dev);
- kfree(snap->name);
- kfree(snap);
-}
-
-static const struct attribute_group *rbd_snap_attr_groups[] = {
- &rbd_snap_attr_group,
- NULL
-};
-
-static struct device_type rbd_snap_device_type = {
- .groups = rbd_snap_attr_groups,
- .release = rbd_snap_dev_release,
-};
-
static struct rbd_spec *rbd_spec_get(struct rbd_spec *spec)
{
kref_get(&spec->kref);
static void rbd_dev_destroy(struct rbd_device *rbd_dev)
{
- rbd_spec_put(rbd_dev->parent_spec);
- kfree(rbd_dev->header_name);
rbd_put_client(rbd_dev->rbd_client);
rbd_spec_put(rbd_dev->spec);
kfree(rbd_dev);
}
-static bool rbd_snap_registered(struct rbd_snap *snap)
+static void rbd_snap_destroy(struct rbd_snap *snap)
{
- bool ret = snap->dev.type == &rbd_snap_device_type;
- bool reg = device_is_registered(&snap->dev);
-
- rbd_assert(!ret ^ reg);
-
- return ret;
-}
-
-static void rbd_remove_snap_dev(struct rbd_snap *snap)
-{
- list_del(&snap->node);
- if (device_is_registered(&snap->dev))
- device_unregister(&snap->dev);
-}
-
-static int rbd_register_snap_dev(struct rbd_snap *snap,
- struct device *parent)
-{
- struct device *dev = &snap->dev;
- int ret;
-
- dev->type = &rbd_snap_device_type;
- dev->parent = parent;
- dev->release = rbd_snap_dev_release;
- dev_set_name(dev, "%s%s", RBD_SNAP_DEV_NAME_PREFIX, snap->name);
- dout("%s: registering device for snapshot %s\n", __func__, snap->name);
-
- ret = device_register(dev);
-
- return ret;
+ kfree(snap->name);
+ kfree(snap);
}
-static struct rbd_snap *__rbd_add_snap_dev(struct rbd_device *rbd_dev,
+static struct rbd_snap *rbd_snap_create(struct rbd_device *rbd_dev,
const char *snap_name,
u64 snap_id, u64 snap_size,
u64 snap_features)
{
struct rbd_snap *snap;
- int ret;
snap = kzalloc(sizeof (*snap), GFP_KERNEL);
if (!snap)
return ERR_PTR(-ENOMEM);
- ret = -ENOMEM;
- snap->name = kstrdup(snap_name, GFP_KERNEL);
- if (!snap->name)
- goto err;
-
+ snap->name = snap_name;
snap->id = snap_id;
snap->size = snap_size;
snap->features = snap_features;
return snap;
-
-err:
- kfree(snap->name);
- kfree(snap);
-
- return ERR_PTR(ret);
}
+/*
+ * 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,
u64 *snap_size, u64 *snap_features)
{
char *snap_name;
+ int i;
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--)
+ for (i = 0; i < which; i++)
snap_name += strlen(snap_name) + 1;
+ snap_name = kstrdup(snap_name, GFP_KERNEL);
+ if (!snap_name)
+ return ERR_PTR(-ENOMEM);
+
+ *snap_size = rbd_dev->header.snap_sizes[which];
+ *snap_features = 0; /* No features for v1 */
+
return snap_name;
}
ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
"rbd", "get_size",
- (char *) &snapid, sizeof (snapid),
- (char *) &size_buf, sizeof (size_buf), NULL);
+ &snapid, sizeof (snapid),
+ &size_buf, sizeof (size_buf), NULL);
dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
if (ret < 0)
return ret;
+ if (ret < sizeof (size_buf))
+ return -ERANGE;
- *order = size_buf.order;
+ if (order)
+ *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);
+ (unsigned long long)snap_id, (unsigned int)*order,
+ (unsigned long long)*snap_size);
return 0;
}
return -ENOMEM;
ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
- "rbd", "get_object_prefix",
- NULL, 0,
+ "rbd", "get_object_prefix", NULL, 0,
reply_buf, RBD_OBJ_PREFIX_LEN_MAX, NULL);
dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
if (ret < 0)
p = reply_buf;
rbd_dev->header.object_prefix = ceph_extract_encoded_string(&p,
- p + RBD_OBJ_PREFIX_LEN_MAX,
- NULL, GFP_NOIO);
+ p + ret, NULL, GFP_NOIO);
+ ret = 0;
if (IS_ERR(rbd_dev->header.object_prefix)) {
ret = PTR_ERR(rbd_dev->header.object_prefix);
} else {
dout(" object_prefix = %s\n", rbd_dev->header.object_prefix);
}
-
out:
kfree(reply_buf);
struct {
__le64 features;
__le64 incompat;
- } features_buf = { 0 };
+ } __attribute__ ((packed)) features_buf = { 0 };
u64 incompat;
int ret;
ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
"rbd", "get_features",
- (char *) &snapid, sizeof (snapid),
- (char *) &features_buf, sizeof (features_buf),
- NULL);
+ &snapid, sizeof (snapid),
+ &features_buf, sizeof (features_buf), NULL);
dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
if (ret < 0)
return ret;
+ if (ret < sizeof (features_buf))
+ return -ERANGE;
incompat = le64_to_cpu(features_buf.incompat);
if (incompat & ~RBD_FEATURES_SUPPORTED)
*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));
+ (unsigned long long)snap_id,
+ (unsigned long long)*snap_features,
+ (unsigned long long)le64_to_cpu(features_buf.incompat));
return 0;
}
snapid = cpu_to_le64(CEPH_NOSNAP);
ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
"rbd", "get_parent",
- (char *) &snapid, sizeof (snapid),
- (char *) reply_buf, size, NULL);
+ &snapid, sizeof (snapid),
+ reply_buf, size, NULL);
dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
if (ret < 0)
goto out_err;
- ret = -ERANGE;
p = reply_buf;
- end = (char *) reply_buf + size;
+ end = reply_buf + ret;
+ ret = -ERANGE;
ceph_decode_64_safe(&p, end, parent_spec->pool_id, out_err);
if (parent_spec->pool_id == CEPH_NOPOOL)
goto out; /* No parent? No problem. */
/* The ceph file layout needs to fit pool id in 32 bits */
ret = -EIO;
- if (WARN_ON(parent_spec->pool_id > (u64) U32_MAX))
- goto out;
+ if (parent_spec->pool_id > (u64)U32_MAX) {
+ rbd_warn(NULL, "parent pool id too large (%llu > %u)\n",
+ (unsigned long long)parent_spec->pool_id, U32_MAX);
+ goto out_err;
+ }
image_id = ceph_extract_encoded_string(&p, end, NULL, GFP_KERNEL);
if (IS_ERR(image_id)) {
return ret;
}
+static int rbd_dev_v2_striping_info(struct rbd_device *rbd_dev)
+{
+ struct {
+ __le64 stripe_unit;
+ __le64 stripe_count;
+ } __attribute__ ((packed)) striping_info_buf = { 0 };
+ size_t size = sizeof (striping_info_buf);
+ void *p;
+ u64 obj_size;
+ u64 stripe_unit;
+ u64 stripe_count;
+ int ret;
+
+ ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
+ "rbd", "get_stripe_unit_count", NULL, 0,
+ (char *)&striping_info_buf, size, NULL);
+ dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
+ if (ret < 0)
+ return ret;
+ if (ret < size)
+ return -ERANGE;
+
+ /*
+ * We don't actually support the "fancy striping" feature
+ * (STRIPINGV2) yet, but if the striping sizes are the
+ * defaults the behavior is the same as before. So find
+ * out, and only fail if the image has non-default values.
+ */
+ ret = -EINVAL;
+ obj_size = (u64)1 << rbd_dev->header.obj_order;
+ p = &striping_info_buf;
+ stripe_unit = ceph_decode_64(&p);
+ if (stripe_unit != obj_size) {
+ rbd_warn(rbd_dev, "unsupported stripe unit "
+ "(got %llu want %llu)",
+ stripe_unit, obj_size);
+ return -EINVAL;
+ }
+ stripe_count = ceph_decode_64(&p);
+ if (stripe_count != 1) {
+ rbd_warn(rbd_dev, "unsupported stripe count "
+ "(got %llu want 1)", stripe_count);
+ return -EINVAL;
+ }
+ rbd_dev->header.stripe_unit = stripe_unit;
+ rbd_dev->header.stripe_count = stripe_count;
+
+ return 0;
+}
+
static char *rbd_dev_image_name(struct rbd_device *rbd_dev)
{
size_t image_id_size;
return NULL;
p = image_id;
- end = (char *) image_id + image_id_size;
- ceph_encode_string(&p, end, rbd_dev->spec->image_id, (u32) len);
+ end = image_id + image_id_size;
+ ceph_encode_string(&p, end, rbd_dev->spec->image_id, (u32)len);
size = sizeof (__le32) + RBD_IMAGE_NAME_LEN_MAX;
reply_buf = kmalloc(size, GFP_KERNEL);
ret = rbd_obj_method_sync(rbd_dev, RBD_DIRECTORY,
"rbd", "dir_get_name",
image_id, image_id_size,
- (char *) reply_buf, size, NULL);
+ reply_buf, size, NULL);
if (ret < 0)
goto out;
p = reply_buf;
- end = (char *) reply_buf + size;
+ end = reply_buf + ret;
+
image_name = ceph_extract_encoded_string(&p, end, &len, GFP_KERNEL);
if (IS_ERR(image_name))
image_name = NULL;
}
/*
- * When a parent image gets probed, we only have the pool, image,
- * and snapshot ids but not the names of any of them. This call
- * is made later to fill in those names. It has to be done after
- * rbd_dev_snaps_update() has completed because some of the
- * information (in particular, snapshot name) is not available
- * until then.
+ * When an rbd image has a parent image, it is identified by the
+ * pool, image, and snapshot ids (not names). This function fills
+ * in the names for those ids. (It's OK if we can't figure out the
+ * name for an image id, but the pool and snapshot ids should always
+ * exist and have names.) All names in an rbd spec are dynamically
+ * allocated.
+ *
+ * When an image being mapped (not a parent) is probed, we have the
+ * pool name and pool id, image name and image id, and the snapshot
+ * name. The only thing we're missing is the snapshot id.
+ *
+ * The set of snapshots for an image is not known until they have
+ * been read by rbd_dev_snaps_update(), so we can't completely fill
+ * in this information until after that has been called.
*/
-static int rbd_dev_probe_update_spec(struct rbd_device *rbd_dev)
+static int rbd_dev_spec_update(struct rbd_device *rbd_dev)
{
- struct ceph_osd_client *osdc;
- const char *name;
- void *reply_buf = NULL;
+ struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
+ struct rbd_spec *spec = rbd_dev->spec;
+ const char *pool_name;
+ const char *image_name;
+ const char *snap_name;
int ret;
- if (rbd_dev->spec->pool_name)
- return 0; /* Already have the names */
+ /*
+ * An image being mapped will have the pool name (etc.), but
+ * we need to look up the snapshot id.
+ */
+ if (spec->pool_name) {
+ if (strcmp(spec->snap_name, RBD_SNAP_HEAD_NAME)) {
+ struct rbd_snap *snap;
+
+ snap = snap_by_name(rbd_dev, spec->snap_name);
+ if (!snap)
+ return -ENOENT;
+ spec->snap_id = snap->id;
+ } else {
+ spec->snap_id = CEPH_NOSNAP;
+ }
+
+ return 0;
+ }
- /* Look up the pool name */
+ /* Get the pool name; we have to make our own copy of this */
- osdc = &rbd_dev->rbd_client->client->osdc;
- name = ceph_pg_pool_name_by_id(osdc->osdmap, rbd_dev->spec->pool_id);
- if (!name) {
- rbd_warn(rbd_dev, "there is no pool with id %llu",
- rbd_dev->spec->pool_id); /* Really a BUG() */
+ pool_name = ceph_pg_pool_name_by_id(osdc->osdmap, spec->pool_id);
+ if (!pool_name) {
+ rbd_warn(rbd_dev, "no pool with id %llu", spec->pool_id);
return -EIO;
}
-
- rbd_dev->spec->pool_name = kstrdup(name, GFP_KERNEL);
- if (!rbd_dev->spec->pool_name)
+ pool_name = kstrdup(pool_name, GFP_KERNEL);
+ if (!pool_name)
return -ENOMEM;
/* Fetch the image name; tolerate failure here */
- name = rbd_dev_image_name(rbd_dev);
- if (name)
- rbd_dev->spec->image_name = (char *) name;
- else
+ image_name = rbd_dev_image_name(rbd_dev);
+ if (!image_name)
rbd_warn(rbd_dev, "unable to get image name");
- /* Look up the snapshot name. */
+ /* Look up the snapshot name, and make a copy */
- name = rbd_snap_name(rbd_dev, rbd_dev->spec->snap_id);
- if (!name) {
- rbd_warn(rbd_dev, "no snapshot with id %llu",
- rbd_dev->spec->snap_id); /* Really a BUG() */
+ snap_name = rbd_snap_name(rbd_dev, spec->snap_id);
+ if (!snap_name) {
+ rbd_warn(rbd_dev, "no snapshot with id %llu", spec->snap_id);
ret = -EIO;
goto out_err;
}
- rbd_dev->spec->snap_name = kstrdup(name, GFP_KERNEL);
- if(!rbd_dev->spec->snap_name)
+ snap_name = kstrdup(snap_name, GFP_KERNEL);
+ if (!snap_name) {
+ ret = -ENOMEM;
goto out_err;
+ }
+
+ spec->pool_name = pool_name;
+ spec->image_name = image_name;
+ spec->snap_name = snap_name;
return 0;
out_err:
- kfree(reply_buf);
- kfree(rbd_dev->spec->pool_name);
- rbd_dev->spec->pool_name = NULL;
+ kfree(image_name);
+ kfree(pool_name);
return ret;
}
return -ENOMEM;
ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
- "rbd", "get_snapcontext",
- NULL, 0,
+ "rbd", "get_snapcontext", NULL, 0,
reply_buf, size, ver);
dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
if (ret < 0)
goto out;
- ret = -ERANGE;
p = reply_buf;
- end = (char *) reply_buf + size;
+ end = reply_buf + ret;
+ ret = -ERANGE;
ceph_decode_64_safe(&p, end, seq, out);
ceph_decode_32_safe(&p, end, snap_count, out);
}
if (!ceph_has_room(&p, end, snap_count * sizeof (__le64)))
goto out;
+ ret = 0;
- size = sizeof (struct ceph_snap_context) +
- snap_count * sizeof (snapc->snaps[0]);
- snapc = kmalloc(size, GFP_KERNEL);
+ snapc = rbd_snap_context_create(snap_count);
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);
-
+ (unsigned long long)seq, (unsigned int)snap_count);
out:
kfree(reply_buf);
- return 0;
+ return ret;
}
static char *rbd_dev_v2_snap_name(struct rbd_device *rbd_dev, u32 which)
if (!reply_buf)
return ERR_PTR(-ENOMEM);
+ rbd_assert(which < rbd_dev->header.snapc->num_snaps);
snap_id = cpu_to_le64(rbd_dev->header.snapc->snaps[which]);
ret = rbd_obj_method_sync(rbd_dev, rbd_dev->header_name,
"rbd", "get_snapshot_name",
- (char *) &snap_id, sizeof (snap_id),
+ &snap_id, sizeof (snap_id),
reply_buf, size, NULL);
dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
- if (ret < 0)
+ if (ret < 0) {
+ snap_name = ERR_PTR(ret);
goto out;
+ }
p = reply_buf;
- end = (char *) reply_buf + size;
+ end = reply_buf + ret;
snap_name = ceph_extract_encoded_string(&p, end, NULL, GFP_KERNEL);
- if (IS_ERR(snap_name)) {
- ret = PTR_ERR(snap_name);
+ if (IS_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;
+ dout(" snap_id 0x%016llx snap_name = %s\n",
+ (unsigned long long)le64_to_cpu(snap_id), snap_name);
out:
kfree(reply_buf);
- return ERR_PTR(ret);
+ return snap_name;
}
static char *rbd_dev_v2_snap_info(struct rbd_device *rbd_dev, u32 which,
u64 *snap_size, u64 *snap_features)
{
u64 snap_id;
- u8 order;
+ u64 size;
+ u64 features;
+ char *snap_name;
int ret;
+ rbd_assert(which < rbd_dev->header.snapc->num_snaps);
snap_id = rbd_dev->header.snapc->snaps[which];
- ret = _rbd_dev_v2_snap_size(rbd_dev, snap_id, &order, snap_size);
+ ret = _rbd_dev_v2_snap_size(rbd_dev, snap_id, NULL, &size);
if (ret)
- return ERR_PTR(ret);
- ret = _rbd_dev_v2_snap_features(rbd_dev, snap_id, snap_features);
+ goto out_err;
+
+ ret = _rbd_dev_v2_snap_features(rbd_dev, snap_id, &features);
if (ret)
- return ERR_PTR(ret);
+ goto out_err;
- return rbd_dev_v2_snap_name(rbd_dev, which);
+ snap_name = rbd_dev_v2_snap_name(rbd_dev, which);
+ if (!IS_ERR(snap_name)) {
+ *snap_size = size;
+ *snap_features = features;
+ }
+
+ return snap_name;
+out_err:
+ return ERR_PTR(ret);
}
static char *rbd_dev_snap_info(struct rbd_device *rbd_dev, u32 which,
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);
* Assumes the snapshots in the snapshot context are sorted by
* snapshot id, highest id first. (Snapshots in the rbd_dev's list
* are also maintained in that order.)
+ *
+ * Note that any error occurs while updating the snapshot list
+ * aborts the update, and the entire list is cleared. The snapshot
+ * list becomes inconsistent at that point anyway, so it might as
+ * well be empty.
*/
static int rbd_dev_snaps_update(struct rbd_device *rbd_dev)
{
struct list_head *head = &rbd_dev->snaps;
struct list_head *links = head->next;
u32 index = 0;
+ int ret = 0;
- dout("%s: snap count is %u\n", __func__, (unsigned int) snap_count);
+ dout("%s: snap count is %u\n", __func__, (unsigned int)snap_count);
while (index < snap_count || links != head) {
u64 snap_id;
struct rbd_snap *snap;
* A previously-existing snapshot is not in
* the new snap context.
*
- * If the now missing snapshot is the one the
- * image is mapped to, clear its exists flag
- * so we can avoid sending any more requests
- * to it.
+ * If the now-missing snapshot is the one
+ * the image represents, clear its existence
+ * flag so we can avoid sending any more
+ * requests to it.
*/
if (rbd_dev->spec->snap_id == snap->id)
clear_bit(RBD_DEV_FLAG_EXISTS, &rbd_dev->flags);
- rbd_remove_snap_dev(snap);
- dout("%ssnap id %llu has been removed\n",
+ dout("removing %ssnap id %llu\n",
rbd_dev->spec->snap_id == snap->id ?
"mapped " : "",
- (unsigned long long) snap->id);
+ (unsigned long long)snap->id);
+
+ list_del(&snap->node);
+ rbd_snap_destroy(snap);
/* Done with this list entry; advance */
snap_name = rbd_dev_snap_info(rbd_dev, index,
&snap_size, &snap_features);
- if (IS_ERR(snap_name))
- return PTR_ERR(snap_name);
+ if (IS_ERR(snap_name)) {
+ ret = PTR_ERR(snap_name);
+ dout("failed to get snap info, error %d\n", ret);
+ goto out_err;
+ }
- dout("entry %u: snap_id = %llu\n", (unsigned int) snap_count,
- (unsigned long long) snap_id);
+ 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)) {
struct rbd_snap *new_snap;
/* We haven't seen this snapshot before */
- new_snap = __rbd_add_snap_dev(rbd_dev, snap_name,
+ new_snap = rbd_snap_create(rbd_dev, snap_name,
snap_id, snap_size, snap_features);
if (IS_ERR(new_snap)) {
- int err = PTR_ERR(new_snap);
-
- dout(" failed to add dev, error %d\n", err);
-
- return err;
+ ret = PTR_ERR(new_snap);
+ dout(" failed to add dev, error %d\n", ret);
+ goto out_err;
}
/* New goes before existing, or at end of list */
dout("%s: done\n", __func__);
return 0;
-}
-
-/*
- * Scan the list of snapshots and register the devices for any that
- * have not already been registered.
- */
-static int rbd_dev_snaps_register(struct rbd_device *rbd_dev)
-{
- struct rbd_snap *snap;
- int ret = 0;
-
- dout("%s:\n", __func__);
- if (WARN_ON(!device_is_registered(&rbd_dev->dev)))
- return -EIO;
-
- list_for_each_entry(snap, &rbd_dev->snaps, node) {
- if (!rbd_snap_registered(snap)) {
- ret = rbd_register_snap_dev(snap, &rbd_dev->dev);
- if (ret < 0)
- break;
- }
- }
- dout("%s: returning %d\n", __func__, ret);
+out_err:
+ rbd_remove_all_snaps(rbd_dev);
return ret;
}
size_t len;
char *options;
const char *mon_addrs;
+ char *snap_name;
size_t mon_addrs_size;
struct rbd_spec *spec = NULL;
struct rbd_options *rbd_opts = NULL;
ret = -ENAMETOOLONG;
goto out_err;
}
- spec->snap_name = kmemdup(buf, len + 1, GFP_KERNEL);
- if (!spec->snap_name)
+ snap_name = kmemdup(buf, len + 1, GFP_KERNEL);
+ if (!snap_name)
goto out_mem;
- *(spec->snap_name + len) = '\0';
+ *(snap_name + len) = '\0';
+ spec->snap_name = snap_name;
/* Initialize all rbd options to the defaults */
size_t size;
char *object_name;
void *response;
- void *p;
-
- /* If we already have it we don't need to look it up */
-
- if (rbd_dev->spec->image_id)
- return 0;
+ char *image_id;
/*
* When probing a parent image, the image id is already
* known (and the image name likely is not). There's no
- * need to fetch the image id again in this case.
+ * need to fetch the image id again in this case. We
+ * do still need to set the image format though.
*/
- if (rbd_dev->spec->image_id)
+ if (rbd_dev->spec->image_id) {
+ rbd_dev->image_format = *rbd_dev->spec->image_id ? 2 : 1;
+
return 0;
+ }
/*
* First, see if the format 2 image id file exists, and if
goto out;
}
+ /* If it doesn't exist we'll assume it's a format 1 image */
+
ret = rbd_obj_method_sync(rbd_dev, object_name,
- "rbd", "get_id",
- NULL, 0,
+ "rbd", "get_id", NULL, 0,
response, RBD_IMAGE_ID_LEN_MAX, NULL);
dout("%s: rbd_obj_method_sync returned %d\n", __func__, ret);
- if (ret < 0)
- goto out;
-
- p = response;
- rbd_dev->spec->image_id = ceph_extract_encoded_string(&p,
- p + RBD_IMAGE_ID_LEN_MAX,
+ if (ret == -ENOENT) {
+ image_id = kstrdup("", GFP_KERNEL);
+ ret = image_id ? 0 : -ENOMEM;
+ if (!ret)
+ rbd_dev->image_format = 1;
+ } else if (ret > sizeof (__le32)) {
+ void *p = response;
+
+ image_id = ceph_extract_encoded_string(&p, p + ret,
NULL, GFP_NOIO);
- if (IS_ERR(rbd_dev->spec->image_id)) {
- ret = PTR_ERR(rbd_dev->spec->image_id);
- rbd_dev->spec->image_id = NULL;
+ ret = IS_ERR(image_id) ? PTR_ERR(image_id) : 0;
+ if (!ret)
+ rbd_dev->image_format = 2;
} else {
- dout("image_id is %s\n", rbd_dev->spec->image_id);
+ ret = -EINVAL;
+ }
+
+ if (!ret) {
+ rbd_dev->spec->image_id = image_id;
+ dout("image_id is %s\n", image_id);
}
out:
kfree(response);
int ret;
size_t size;
- /* Version 1 images have no id; empty string is used */
-
- rbd_dev->spec->image_id = kstrdup("", GFP_KERNEL);
- if (!rbd_dev->spec->image_id)
- return -ENOMEM;
-
/* Record the header object name for this rbd image. */
size = strlen(rbd_dev->spec->image_name) + sizeof (RBD_SUFFIX);
rbd_dev->parent_spec = NULL;
rbd_dev->parent_overlap = 0;
- rbd_dev->image_format = 1;
-
dout("discovered version 1 image, header name is %s\n",
rbd_dev->header_name);
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 < 0)
+ if (ret)
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)
+ if (ret)
goto out_err;
/* Get the and check features for the image */
ret = rbd_dev_v2_features(rbd_dev);
- if (ret < 0)
+ if (ret)
goto out_err;
/* If the image supports layering, get the parent info */
if (rbd_dev->header.features & RBD_FEATURE_LAYERING) {
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!");
+ }
+
+ /* If the image supports fancy striping, get its parameters */
+
+ if (rbd_dev->header.features & RBD_FEATURE_STRIPINGV2) {
+ ret = rbd_dev_v2_striping_info(rbd_dev);
if (ret < 0)
goto out_err;
}
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 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 = -ENOMEM;
+ parent = rbd_dev_create(rbdc, parent_spec);
+ if (!parent)
+ goto out_err;
+
+ 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_probe_finish(struct rbd_device *rbd_dev)
+{
+ int ret;
- ret = rbd_dev_probe_update_spec(rbd_dev);
+ ret = rbd_dev_header_watch_sync(rbd_dev, 1);
if (ret)
- goto err_out_snaps;
+ return ret;
- ret = rbd_dev_set_mapping(rbd_dev);
+ 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_probe(parent);
- if (ret < 0)
- goto err_out_parent;
- rbd_dev->parent = parent;
- }
-
- down_write(&rbd_dev->header_rwsem);
- ret = rbd_dev_snaps_register(rbd_dev);
- up_write(&rbd_dev->header_rwsem);
- if (ret)
- goto err_out_bus;
-
- 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);
pr_info("%s: added with size 0x%llx\n", rbd_dev->disk->disk_name,
return ret;
-err_out_parent:
- 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;
}
* device. For format 2 images this includes determining the image
* id.
*/
-static int rbd_dev_probe(struct rbd_device *rbd_dev)
+static int rbd_dev_image_probe(struct rbd_device *rbd_dev)
{
int ret;
*/
ret = rbd_dev_image_id(rbd_dev);
if (ret)
+ return ret;
+ rbd_assert(rbd_dev->spec->image_id);
+ rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
+
+ if (rbd_dev->image_format == 1)
ret = rbd_dev_v1_probe(rbd_dev);
else
ret = rbd_dev_v2_probe(rbd_dev);
- if (ret) {
- dout("probe failed, returning %d\n", ret);
+ if (ret)
+ goto out_err;
- return ret;
- }
+ ret = rbd_dev_snaps_update(rbd_dev);
+ if (ret)
+ goto out_err;
+
+ ret = rbd_dev_spec_update(rbd_dev);
+ if (ret)
+ goto err_out_snaps;
+
+ ret = rbd_dev_probe_parent(rbd_dev);
+ if (ret)
+ goto err_out_snaps;
ret = rbd_dev_probe_finish(rbd_dev);
if (ret)
- rbd_header_free(&rbd_dev->header);
+ goto err_out_parent;
+
+ return ret;
+err_out_parent:
+ rbd_dev_remove_parent(rbd_dev);
+ rbd_header_free(&rbd_dev->header);
+err_out_snaps:
+ rbd_remove_all_snaps(rbd_dev);
+out_err:
+ kfree(rbd_dev->spec->image_id);
+ rbd_dev->spec->image_id = NULL;
+
+ dout("probe failed, returning %d\n", ret);
return ret;
}
rc = ceph_pg_poolid_by_name(osdc->osdmap, spec->pool_name);
if (rc < 0)
goto err_out_client;
- spec->pool_id = (u64) rc;
+ spec->pool_id = (u64)rc;
/* The ceph file layout needs to fit pool id in 32 bits */
- if (WARN_ON(spec->pool_id > (u64) U32_MAX)) {
+ if (spec->pool_id > (u64)U32_MAX) {
+ rbd_warn(NULL, "pool id too large (%llu > %u)\n",
+ (unsigned long long)spec->pool_id, U32_MAX);
rc = -EIO;
goto err_out_client;
}
kfree(rbd_opts);
rbd_opts = NULL; /* done with this */
- rc = rbd_dev_probe(rbd_dev);
+ rc = rbd_dev_image_probe(rbd_dev);
if (rc < 0)
goto err_out_rbd_dev;
return count;
err_out_rbd_dev:
+ kfree(rbd_dev->header_name);
rbd_dev_destroy(rbd_dev);
err_out_client:
rbd_put_client(rbdc);
dout("Error adding device %s\n", buf);
- return (ssize_t) rc;
+ return (ssize_t)rc;
}
static struct rbd_device *__rbd_get_dev(unsigned long dev_id)
/* done with the id, and with the rbd_dev */
rbd_dev_id_put(rbd_dev);
+ rbd_dev_mapping_clear(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);
}
-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_remove_all_snaps(second);
+ rbd_bus_del_dev(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,
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);
+ rbd_dev_remove_parent(rbd_dev);
+ rbd_remove_all_snaps(rbd_dev);
+ rbd_bus_del_dev(rbd_dev);
done:
mutex_unlock(&ctl_mutex);