static int rbd_dev_refresh(struct rbd_device *rbd_dev);
static int rbd_dev_v2_header_onetime(struct rbd_device *rbd_dev);
-static int rbd_dev_v2_header_info(struct rbd_device *rbd_dev);
+static int rbd_dev_header_info(struct rbd_device *rbd_dev);
+static int rbd_dev_v2_parent_info(struct rbd_device *rbd_dev);
static const char *rbd_dev_v2_snap_name(struct rbd_device *rbd_dev,
u64 snap_id);
static int _rbd_dev_v2_snap_size(struct rbd_device *rbd_dev, u64 snap_id,
header->snap_names = snap_names;
header->snap_sizes = snap_sizes;
- /* Make sure mapping size is consistent with header info */
-
- if (rbd_dev->spec->snap_id == CEPH_NOSNAP || first_time)
- if (rbd_dev->mapping.size != header->image_size)
- rbd_dev->mapping.size = header->image_size;
-
return 0;
out_2big:
ret = -EIO;
rbd_dev->mapping.features = 0;
}
+static void rbd_segment_name_free(const char *name)
+{
+ /* The explicit cast here is needed to drop the const qualifier */
+
+ kmem_cache_free(rbd_segment_name_cache, (void *)name);
+}
+
static const char *rbd_segment_name(struct rbd_device *rbd_dev, u64 offset)
{
char *name;
if (ret < 0 || ret > CEPH_MAX_OID_NAME_LEN) {
pr_err("error formatting segment name for #%llu (%d)\n",
segment, ret);
- kfree(name);
+ rbd_segment_name_free(name);
name = NULL;
}
return name;
}
-static void rbd_segment_name_free(const char *name)
-{
- /* The explicit cast here is needed to drop the const qualifier */
-
- kmem_cache_free(rbd_segment_name_cache, (void *)name);
-}
-
static u64 rbd_segment_offset(struct rbd_device *rbd_dev, u64 offset)
{
u64 segment_size = (u64) 1 << rbd_dev->header.obj_order;
static int rbd_obj_request_submit(struct ceph_osd_client *osdc,
struct rbd_obj_request *obj_request)
{
- dout("%s: osdc %p obj %p\n", __func__, osdc, obj_request);
-
+ dout("%s %p\n", __func__, obj_request);
return ceph_osdc_start_request(osdc, obj_request->osd_req, false);
}
+static void rbd_obj_request_end(struct rbd_obj_request *obj_request)
+{
+ dout("%s %p\n", __func__, obj_request);
+ ceph_osdc_cancel_request(obj_request->osd_req);
+}
+
+/*
+ * Wait for an object request to complete. If interrupted, cancel the
+ * underlying osd request.
+ */
+static int rbd_obj_request_wait(struct rbd_obj_request *obj_request)
+{
+ int ret;
+
+ dout("%s %p\n", __func__, obj_request);
+
+ ret = wait_for_completion_interruptible(&obj_request->completion);
+ if (ret < 0) {
+ dout("%s %p interrupted\n", __func__, obj_request);
+ rbd_obj_request_end(obj_request);
+ return ret;
+ }
+
+ dout("%s %p done\n", __func__, obj_request);
+ return 0;
+}
+
static void rbd_img_request_complete(struct rbd_img_request *img_request)
{
rbd_img_request_put(img_request);
}
-/* Caller is responsible for rbd_obj_request_destroy(obj_request) */
-
-static int rbd_obj_request_wait(struct rbd_obj_request *obj_request)
-{
- dout("%s: obj %p\n", __func__, obj_request);
-
- return wait_for_completion_interruptible(&obj_request->completion);
-}
-
/*
* The default/initial value for all image request flags is 0. Each
* is conditionally set to 1 at image request initialization time
dout("%s: \"%s\" notify_id %llu opcode %u\n", __func__,
rbd_dev->header_name, (unsigned long long)notify_id,
(unsigned int)opcode);
+
+ /*
+ * Until adequate refresh error handling is in place, there is
+ * not much we can do here, except warn.
+ *
+ * See http://tracker.ceph.com/issues/5040
+ */
ret = rbd_dev_refresh(rbd_dev);
if (ret)
- rbd_warn(rbd_dev, "header refresh error (%d)\n", ret);
+ rbd_warn(rbd_dev, "refresh failed: %d\n", ret);
- rbd_obj_notify_ack_sync(rbd_dev, notify_id);
+ ret = rbd_obj_notify_ack_sync(rbd_dev, notify_id);
+ if (ret)
+ rbd_warn(rbd_dev, "notify_ack ret %d\n", ret);
}
/*
- * Initiate a watch request, synchronously.
+ * Send a (un)watch request and wait for the ack. Return a request
+ * with a ref held on success or error.
*/
-static int rbd_dev_header_watch_sync(struct rbd_device *rbd_dev)
+static struct rbd_obj_request *rbd_obj_watch_request_helper(
+ struct rbd_device *rbd_dev,
+ bool watch)
{
struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
struct rbd_obj_request *obj_request;
int ret;
- rbd_assert(!rbd_dev->watch_event);
- rbd_assert(!rbd_dev->watch_request);
-
- ret = ceph_osdc_create_event(osdc, rbd_watch_cb, rbd_dev,
- &rbd_dev->watch_event);
- if (ret < 0)
- return ret;
-
- rbd_assert(rbd_dev->watch_event);
-
obj_request = rbd_obj_request_create(rbd_dev->header_name, 0, 0,
OBJ_REQUEST_NODATA);
- if (!obj_request) {
- ret = -ENOMEM;
- goto out_cancel;
- }
+ if (!obj_request)
+ return ERR_PTR(-ENOMEM);
obj_request->osd_req = rbd_osd_req_create(rbd_dev, true, 1,
obj_request);
if (!obj_request->osd_req) {
ret = -ENOMEM;
- goto out_put;
+ goto out;
}
- ceph_osdc_set_request_linger(osdc, obj_request->osd_req);
-
osd_req_op_watch_init(obj_request->osd_req, 0, CEPH_OSD_OP_WATCH,
- rbd_dev->watch_event->cookie, 0, 1);
+ rbd_dev->watch_event->cookie, 0, watch);
rbd_osd_req_format_write(obj_request);
+ if (watch)
+ ceph_osdc_set_request_linger(osdc, obj_request->osd_req);
+
ret = rbd_obj_request_submit(osdc, obj_request);
if (ret)
- goto out_linger;
+ goto out;
ret = rbd_obj_request_wait(obj_request);
if (ret)
- goto out_linger;
+ goto out;
ret = obj_request->result;
- if (ret)
- goto out_linger;
-
- /*
- * A watch request is set to linger, so the underlying osd
- * request won't go away until we unregister it. We retain
- * a pointer to the object request during that time (in
- * rbd_dev->watch_request), so we'll keep a reference to
- * it. We'll drop that reference (below) after we've
- * unregistered it.
- */
- rbd_dev->watch_request = obj_request;
+ if (ret) {
+ if (watch)
+ rbd_obj_request_end(obj_request);
+ goto out;
+ }
- return 0;
+ return obj_request;
-out_linger:
- ceph_osdc_unregister_linger_request(osdc, obj_request->osd_req);
-out_put:
+out:
rbd_obj_request_put(obj_request);
-out_cancel:
- ceph_osdc_cancel_event(rbd_dev->watch_event);
- rbd_dev->watch_event = NULL;
-
- return ret;
+ return ERR_PTR(ret);
}
/*
- * Tear down a watch request, synchronously.
+ * Initiate a watch request, synchronously.
*/
-static int __rbd_dev_header_unwatch_sync(struct rbd_device *rbd_dev)
+static int rbd_dev_header_watch_sync(struct rbd_device *rbd_dev)
{
struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
struct rbd_obj_request *obj_request;
int ret;
- rbd_assert(rbd_dev->watch_event);
- rbd_assert(rbd_dev->watch_request);
+ rbd_assert(!rbd_dev->watch_event);
+ rbd_assert(!rbd_dev->watch_request);
- obj_request = rbd_obj_request_create(rbd_dev->header_name, 0, 0,
- OBJ_REQUEST_NODATA);
- if (!obj_request) {
- ret = -ENOMEM;
- goto out_cancel;
- }
+ ret = ceph_osdc_create_event(osdc, rbd_watch_cb, rbd_dev,
+ &rbd_dev->watch_event);
+ if (ret < 0)
+ return ret;
- obj_request->osd_req = rbd_osd_req_create(rbd_dev, true, 1,
- obj_request);
- if (!obj_request->osd_req) {
- ret = -ENOMEM;
- goto out_put;
+ obj_request = rbd_obj_watch_request_helper(rbd_dev, true);
+ if (IS_ERR(obj_request)) {
+ ceph_osdc_cancel_event(rbd_dev->watch_event);
+ rbd_dev->watch_event = NULL;
+ return PTR_ERR(obj_request);
}
- osd_req_op_watch_init(obj_request->osd_req, 0, CEPH_OSD_OP_WATCH,
- rbd_dev->watch_event->cookie, 0, 0);
- rbd_osd_req_format_write(obj_request);
-
- ret = rbd_obj_request_submit(osdc, obj_request);
- if (ret)
- goto out_put;
+ /*
+ * A watch request is set to linger, so the underlying osd
+ * request won't go away until we unregister it. We retain
+ * a pointer to the object request during that time (in
+ * rbd_dev->watch_request), so we'll keep a reference to it.
+ * We'll drop that reference after we've unregistered it in
+ * rbd_dev_header_unwatch_sync().
+ */
+ rbd_dev->watch_request = obj_request;
- ret = rbd_obj_request_wait(obj_request);
- if (ret)
- goto out_put;
+ return 0;
+}
- ret = obj_request->result;
- if (ret)
- goto out_put;
+/*
+ * Tear down a watch request, synchronously.
+ */
+static void rbd_dev_header_unwatch_sync(struct rbd_device *rbd_dev)
+{
+ struct rbd_obj_request *obj_request;
- /* We have successfully torn down the watch request */
+ rbd_assert(rbd_dev->watch_event);
+ rbd_assert(rbd_dev->watch_request);
- ceph_osdc_unregister_linger_request(osdc,
- rbd_dev->watch_request->osd_req);
+ rbd_obj_request_end(rbd_dev->watch_request);
rbd_obj_request_put(rbd_dev->watch_request);
rbd_dev->watch_request = NULL;
-out_put:
- rbd_obj_request_put(obj_request);
-out_cancel:
+ obj_request = rbd_obj_watch_request_helper(rbd_dev, false);
+ if (!IS_ERR(obj_request))
+ rbd_obj_request_put(obj_request);
+ else
+ rbd_warn(rbd_dev, "unable to tear down watch request (%ld)",
+ PTR_ERR(obj_request));
+
ceph_osdc_cancel_event(rbd_dev->watch_event);
rbd_dev->watch_event = NULL;
-
- return ret;
-}
-
-static void rbd_dev_header_unwatch_sync(struct rbd_device *rbd_dev)
-{
- int ret;
-
- ret = __rbd_dev_header_unwatch_sync(rbd_dev);
- if (ret) {
- rbd_warn(rbd_dev, "unable to tear down watch request: %d\n",
- ret);
- }
}
/*
u64 mapping_size;
int ret;
- rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
down_write(&rbd_dev->header_rwsem);
mapping_size = rbd_dev->mapping.size;
- if (rbd_dev->image_format == 1)
- ret = rbd_dev_v1_header_info(rbd_dev);
- else
- ret = rbd_dev_v2_header_info(rbd_dev);
- /* If it's a mapped snapshot, validate its EXISTS flag */
+ ret = rbd_dev_header_info(rbd_dev);
+ if (ret)
+ return ret;
+
+ /*
+ * If there is a parent, see if it has disappeared due to the
+ * mapped image getting flattened.
+ */
+ if (rbd_dev->parent) {
+ ret = rbd_dev_v2_parent_info(rbd_dev);
+ if (ret)
+ return ret;
+ }
+
+ if (rbd_dev->spec->snap_id == CEPH_NOSNAP) {
+ if (rbd_dev->mapping.size != rbd_dev->header.image_size)
+ rbd_dev->mapping.size = rbd_dev->header.image_size;
+ } else {
+ /* validate mapped snapshot's EXISTS flag */
+ rbd_exists_validate(rbd_dev);
+ }
- rbd_exists_validate(rbd_dev);
up_write(&rbd_dev->header_rwsem);
- if (mapping_size != rbd_dev->mapping.size) {
+ if (mapping_size != rbd_dev->mapping.size)
rbd_dev_update_size(rbd_dev);
- }
- return ret;
+ return 0;
}
static int rbd_init_disk(struct rbd_device *rbd_dev)
}
/*
- * For an rbd v2 image, shows the pool id, image id, and snapshot id
- * for the parent image. If there is no parent, simply shows
- * "(no parent image)".
+ * For a v2 image, shows the chain of parent images, separated by empty
+ * lines. For v1 images or if there is no parent, shows "(no parent
+ * image)".
*/
static ssize_t rbd_parent_show(struct device *dev,
- struct device_attribute *attr,
- char *buf)
+ struct device_attribute *attr,
+ char *buf)
{
struct rbd_device *rbd_dev = dev_to_rbd_dev(dev);
- struct rbd_spec *spec = rbd_dev->parent_spec;
- int count;
- char *bufp = buf;
+ ssize_t count = 0;
- if (!spec)
+ if (!rbd_dev->parent)
return sprintf(buf, "(no parent image)\n");
- count = sprintf(bufp, "pool_id %llu\npool_name %s\n",
- (unsigned long long) spec->pool_id, spec->pool_name);
- if (count < 0)
- return count;
- bufp += count;
-
- count = sprintf(bufp, "image_id %s\nimage_name %s\n", spec->image_id,
- spec->image_name ? spec->image_name : "(unknown)");
- if (count < 0)
- return count;
- bufp += count;
-
- count = sprintf(bufp, "snap_id %llu\nsnap_name %s\n",
- (unsigned long long) spec->snap_id, spec->snap_name);
- if (count < 0)
- return count;
- bufp += count;
-
- count = sprintf(bufp, "overlap %llu\n", rbd_dev->parent_overlap);
- if (count < 0)
- return count;
- bufp += count;
+ for ( ; rbd_dev->parent; rbd_dev = rbd_dev->parent) {
+ struct rbd_spec *spec = rbd_dev->parent_spec;
+
+ count += sprintf(&buf[count], "%s"
+ "pool_id %llu\npool_name %s\n"
+ "image_id %s\nimage_name %s\n"
+ "snap_id %llu\nsnap_name %s\n"
+ "overlap %llu\n",
+ !count ? "" : "\n", /* first? */
+ spec->pool_id, spec->pool_name,
+ spec->image_id, spec->image_name ?: "(unknown)",
+ spec->snap_id, spec->snap_name,
+ rbd_dev->parent_overlap);
+ }
- return (ssize_t) (bufp - buf);
+ return count;
}
static ssize_t rbd_image_refresh(struct device *dev,
ret = rbd_dev_refresh(rbd_dev);
if (ret)
- rbd_warn(rbd_dev, ": manual header refresh error (%d)\n", ret);
+ return ret;
- return ret < 0 ? ret : size;
+ return size;
}
static DEVICE_ATTR(size, S_IRUGO, rbd_size_show, NULL);
spec = kzalloc(sizeof (*spec), GFP_KERNEL);
if (!spec)
return NULL;
+
+ spec->pool_id = CEPH_NOPOOL;
+ spec->snap_id = CEPH_NOSNAP;
kref_init(&spec->kref);
return spec;
parent_spec->snap_id = snap_id;
rbd_dev->parent_spec = parent_spec;
parent_spec = NULL; /* rbd_dev now owns this */
+ } else {
+ kfree(image_id);
}
/*
}
/*
- * 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.
+ * An image being mapped will have everything but the snap id.
+ */
+static int rbd_spec_fill_snap_id(struct rbd_device *rbd_dev)
+{
+ struct rbd_spec *spec = rbd_dev->spec;
+
+ rbd_assert(spec->pool_id != CEPH_NOPOOL && spec->pool_name);
+ rbd_assert(spec->image_id && spec->image_name);
+ rbd_assert(spec->snap_name);
+
+ if (strcmp(spec->snap_name, RBD_SNAP_HEAD_NAME)) {
+ u64 snap_id;
+
+ snap_id = rbd_snap_id_by_name(rbd_dev, spec->snap_name);
+ if (snap_id == CEPH_NOSNAP)
+ return -ENOENT;
+
+ spec->snap_id = snap_id;
+ } else {
+ spec->snap_id = CEPH_NOSNAP;
+ }
+
+ return 0;
+}
+
+/*
+ * A parent image will have all ids but none of the names.
*
- * 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.
+ * All names in an rbd spec are dynamically allocated. It's OK if we
+ * can't figure out the name for an image id.
*/
-static int rbd_dev_spec_update(struct rbd_device *rbd_dev)
+static int rbd_spec_fill_names(struct rbd_device *rbd_dev)
{
struct ceph_osd_client *osdc = &rbd_dev->rbd_client->client->osdc;
struct rbd_spec *spec = rbd_dev->spec;
const char *snap_name;
int ret;
- /*
- * 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)) {
- u64 snap_id;
-
- snap_id = rbd_snap_id_by_name(rbd_dev, spec->snap_name);
- if (snap_id == CEPH_NOSNAP)
- return -ENOENT;
- spec->snap_id = snap_id;
- } else {
- spec->snap_id = CEPH_NOSNAP;
- }
-
- return 0;
- }
+ rbd_assert(spec->pool_id != CEPH_NOPOOL);
+ rbd_assert(spec->image_id);
+ rbd_assert(spec->snap_id != CEPH_NOSNAP);
/* Get the pool name; we have to make our own copy of this */
if (!image_name)
rbd_warn(rbd_dev, "unable to get image name");
- /* Look up the snapshot name, and make a copy */
+ /* Fetch the snapshot name */
snap_name = rbd_snap_name(rbd_dev, spec->snap_id);
if (IS_ERR(snap_name)) {
spec->snap_name = snap_name;
return 0;
+
out_err:
kfree(image_name);
kfree(pool_name);
-
return ret;
}
return ret;
}
- /*
- * If the image supports layering, get the parent info. We
- * need to probe the first time regardless. Thereafter we
- * only need to if there's a parent, to see if it has
- * disappeared due to the mapped image getting flattened.
- */
- if (rbd_dev->header.features & RBD_FEATURE_LAYERING &&
- (first_time || rbd_dev->parent_spec)) {
- bool warn;
-
- ret = rbd_dev_v2_parent_info(rbd_dev);
- if (ret)
- return ret;
-
- /*
- * Print a warning if this is the initial probe and
- * the image has a parent. Don't print it if the
- * image now being probed is itself a parent. We
- * can tell at this point because we won't know its
- * pool name yet (just its pool id).
- */
- warn = rbd_dev->parent_spec && rbd_dev->spec->pool_name;
- if (first_time && warn)
- rbd_warn(rbd_dev, "WARNING: kernel layering "
- "is EXPERIMENTAL!");
- }
-
- if (rbd_dev->spec->snap_id == CEPH_NOSNAP)
- if (rbd_dev->mapping.size != rbd_dev->header.image_size)
- rbd_dev->mapping.size = rbd_dev->header.image_size;
-
ret = rbd_dev_v2_snap_context(rbd_dev);
dout("rbd_dev_v2_snap_context returned %d\n", ret);
return ret;
}
+static int rbd_dev_header_info(struct rbd_device *rbd_dev)
+{
+ rbd_assert(rbd_image_format_valid(rbd_dev->image_format));
+
+ if (rbd_dev->image_format == 1)
+ return rbd_dev_v1_header_info(rbd_dev);
+
+ return rbd_dev_v2_header_info(rbd_dev);
+}
+
static int rbd_bus_add_dev(struct rbd_device *rbd_dev)
{
struct device *dev;
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));
ret = rbd_dev_header_name(rbd_dev);
if (ret)
goto out_header_name;
}
- if (rbd_dev->image_format == 1)
- ret = rbd_dev_v1_header_info(rbd_dev);
- else
- ret = rbd_dev_v2_header_info(rbd_dev);
+ ret = rbd_dev_header_info(rbd_dev);
if (ret)
goto err_out_watch;
- ret = rbd_dev_spec_update(rbd_dev);
+ /*
+ * If this image is the one being mapped, we have pool name and
+ * id, image name and id, and snap name - need to fill snap id.
+ * Otherwise this is a parent image, identified by pool, image
+ * and snap ids - need to fill in names for those ids.
+ */
+ if (mapping)
+ ret = rbd_spec_fill_snap_id(rbd_dev);
+ else
+ ret = rbd_spec_fill_names(rbd_dev);
if (ret)
goto err_out_probe;
+ if (rbd_dev->header.features & RBD_FEATURE_LAYERING) {
+ ret = rbd_dev_v2_parent_info(rbd_dev);
+ if (ret)
+ goto err_out_probe;
+
+ /*
+ * Need to warn users if this image is the one being
+ * mapped and has a parent.
+ */
+ if (mapping && rbd_dev->parent_spec)
+ rbd_warn(rbd_dev,
+ "WARNING: kernel layering is EXPERIMENTAL!");
+ }
+
ret = rbd_dev_probe_parent(rbd_dev);
if (ret)
goto err_out_probe;
dout("discovered format %u image, header name is %s\n",
rbd_dev->image_format, rbd_dev->header_name);
-
return 0;
+
err_out_probe:
rbd_dev_unprobe(rbd_dev);
err_out_watch:
rbd_dev->image_format = 0;
kfree(rbd_dev->spec->image_id);
rbd_dev->spec->image_id = NULL;
-
- dout("probe failed, returning %d\n", ret);
-
return ret;
}
projects / firefly-linux-kernel-4.4.55.git / blobdiff