rbd: re-submit write request for flattened clone
authorAlex Elder <elder@inktank.com>
Mon, 6 May 2013 22:40:33 +0000 (17:40 -0500)
committerAlex Elder <elder@inktank.com>
Mon, 13 May 2013 20:06:45 +0000 (15:06 -0500)
Add code to rbd_img_parent_read_full_callback() to detect when a
clone's parent image has disappeared, and re-submit the original
write request in that case.  (See the previous commit for more
reasoning about why this is appropriate.)

Rename some variables in rbd_img_obj_parent_read_full_callback()
to match the convention used in the previous patch.

Signed-off-by: Alex Elder <elder@inktank.com>
Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
drivers/block/rbd.c

index 4edcb6d85f01debc7678753b5cf49a6481350e6d..5c2731859e8acade5214c6224a07219029a0e488 100644 (file)
@@ -2319,7 +2319,7 @@ rbd_img_obj_parent_read_full_callback(struct rbd_img_request *img_request)
        struct rbd_device *rbd_dev;
        struct page **pages;
        u32 page_count;
-       int result;
+       int img_result;
        u64 parent_length;
        u64 offset;
        u64 length;
@@ -2338,7 +2338,7 @@ rbd_img_obj_parent_read_full_callback(struct rbd_img_request *img_request)
        orig_request = img_request->obj_request;
        rbd_assert(orig_request != NULL);
        rbd_assert(obj_request_type_valid(orig_request->type));
-       result = img_request->result;
+       img_result = img_request->result;
        parent_length = img_request->length;
        rbd_assert(parent_length == img_request->xferred);
        rbd_img_request_put(img_request);
@@ -2347,7 +2347,22 @@ rbd_img_obj_parent_read_full_callback(struct rbd_img_request *img_request)
        rbd_dev = orig_request->img_request->rbd_dev;
        rbd_assert(rbd_dev);
 
-       if (result)
+       /*
+        * If the overlap has become 0 (most likely because the
+        * image has been flattened) we need to free the pages
+        * and re-submit the original write request.
+        */
+       if (!rbd_dev->parent_overlap) {
+               struct ceph_osd_client *osdc;
+
+               ceph_release_page_vector(pages, page_count);
+               osdc = &rbd_dev->rbd_client->client->osdc;
+               img_result = rbd_obj_request_submit(osdc, orig_request);
+               if (!img_result)
+                       return;
+       }
+
+       if (img_result)
                goto out_err;
 
        /*
@@ -2356,7 +2371,7 @@ rbd_img_obj_parent_read_full_callback(struct rbd_img_request *img_request)
         * request.  Allocate the new copyup osd request for the
         * original request, and release the old one.
         */
-       result = -ENOMEM;
+       img_result = -ENOMEM;
        osd_req = rbd_osd_req_create_copyup(orig_request);
        if (!osd_req)
                goto out_err;
@@ -2391,13 +2406,13 @@ rbd_img_obj_parent_read_full_callback(struct rbd_img_request *img_request)
 
        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)
+       img_result = rbd_obj_request_submit(osdc, orig_request);
+       if (!img_result)
                return;
 out_err:
        /* Record the error code and complete the request */
 
-       orig_request->result = result;
+       orig_request->result = img_result;
        orig_request->xferred = 0;
        obj_request_done_set(orig_request);
        rbd_obj_request_complete(orig_request);