Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client

Pull Ceph fixes from Sage Weil:
 "There are two critical regression fixes for CephFS from Zheng, and an
  RBD completion fix for layered images from Ilya"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client:
  rbd: fix copyup completion race
  ceph: always re-send cap flushes when MDS recovers
  ceph: fix ceph_encode_locks_to_buffer()

4 files changed, 23 insertions, 24 deletions

diff --git a/drivers/block/rbd.c b/drivers/block/rbd.c
index d94529d5c8e9..bc67a93aa4f4 100644
--- a/drivers/block/rbd.c
+++ b/drivers/block/rbd.c
@@ -523,6 +523,7 @@ void rbd_warn(struct rbd_device *rbd_dev, const char *fmt, ...)
 #  define rbd_assert(expr)	((void) 0)
 #endif /* !RBD_DEBUG */
 
+static void rbd_osd_copyup_callback(struct rbd_obj_request *obj_request);
 static int rbd_img_obj_request_submit(struct rbd_obj_request *obj_request);
 static void rbd_img_parent_read(struct rbd_obj_request *obj_request);
 static void rbd_dev_remove_parent(struct rbd_device *rbd_dev);
@@ -1818,6 +1819,16 @@ static void rbd_osd_stat_callback(struct rbd_obj_request *obj_request)
 	obj_request_done_set(obj_request);
 }
 
+static void rbd_osd_call_callback(struct rbd_obj_request *obj_request)
+{
+	dout("%s: obj %p\n", __func__, obj_request);
+
+	if (obj_request_img_data_test(obj_request))
+		rbd_osd_copyup_callback(obj_request);
+	else
+		obj_request_done_set(obj_request);
+}
+
 static void rbd_osd_req_callback(struct ceph_osd_request *osd_req,
 				struct ceph_msg *msg)
 {
@@ -1866,6 +1877,8 @@ static void rbd_osd_req_callback(struct ceph_osd_request *osd_req,
 		rbd_osd_discard_callback(obj_request);
 		break;
 	case CEPH_OSD_OP_CALL:
+		rbd_osd_call_callback(obj_request);
+		break;
 	case CEPH_OSD_OP_NOTIFY_ACK:
 	case CEPH_OSD_OP_WATCH:
 		rbd_osd_trivial_callback(obj_request);
@@ -2530,13 +2543,15 @@ out_unwind:
 }
 
 static void
-rbd_img_obj_copyup_callback(struct rbd_obj_request *obj_request)
+rbd_osd_copyup_callback(struct rbd_obj_request *obj_request)
 {
 	struct rbd_img_request *img_request;
 	struct rbd_device *rbd_dev;
 	struct page **pages;
 	u32 page_count;
 
+	dout("%s: obj %p\n", __func__, obj_request);
+
 	rbd_assert(obj_request->type == OBJ_REQUEST_BIO ||
 		obj_request->type == OBJ_REQUEST_NODATA);
 	rbd_assert(obj_request_img_data_test(obj_request));
@@ -2563,9 +2578,7 @@ rbd_img_obj_copyup_callback(struct rbd_obj_request *obj_request)
 	if (!obj_request->result)
 		obj_request->xferred = obj_request->length;
 
-	/* Finish up with the normal image object callback */
-
-	rbd_img_obj_callback(obj_request);
+	obj_request_done_set(obj_request);
 }
 
 static void
@@ -2650,7 +2663,6 @@ rbd_img_obj_parent_read_full_callback(struct rbd_img_request *img_request)
 
 	/* All set, send it off. */
 
-	orig_request->callback = rbd_img_obj_copyup_callback;
 	osdc = &rbd_dev->rbd_client->client->osdc;
 	img_result = rbd_obj_request_submit(osdc, orig_request);
 	if (!img_result)
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index dc10c9dd36c1..ddd5e9471290 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -1506,7 +1506,6 @@ static int __mark_caps_flushing(struct inode *inode,
 
 	swap(cf, ci->i_prealloc_cap_flush);
 	cf->caps = flushing;
-	cf->kick = false;
 
 	spin_lock(&mdsc->cap_dirty_lock);
 	list_del_init(&ci->i_dirty_item);
@@ -2123,8 +2122,7 @@ static void kick_flushing_capsnaps(struct ceph_mds_client *mdsc,
 
 static int __kick_flushing_caps(struct ceph_mds_client *mdsc,
 				struct ceph_mds_session *session,
-				struct ceph_inode_info *ci,
-				bool kick_all)
+				struct ceph_inode_info *ci)
 {
 	struct inode *inode = &ci->vfs_inode;
 	struct ceph_cap *cap;
@@ -2150,9 +2148,7 @@ static int __kick_flushing_caps(struct ceph_mds_client *mdsc,
 
 		for (n = rb_first(&ci->i_cap_flush_tree); n; n = rb_next(n)) {
 			cf = rb_entry(n, struct ceph_cap_flush, i_node);
-			if (cf->tid < first_tid)
-				continue;
-			if (kick_all || cf->kick)
+			if (cf->tid >= first_tid)
 				break;
 		}
 		if (!n) {
@@ -2161,7 +2157,6 @@ static int __kick_flushing_caps(struct ceph_mds_client *mdsc,
 		}
 
 		cf = rb_entry(n, struct ceph_cap_flush, i_node);
-		cf->kick = false;
 
 		first_tid = cf->tid + 1;
 
@@ -2181,8 +2176,6 @@ void ceph_early_kick_flushing_caps(struct ceph_mds_client *mdsc,
 {
 	struct ceph_inode_info *ci;
 	struct ceph_cap *cap;
-	struct ceph_cap_flush *cf;
-	struct rb_node *n;
 
 	dout("early_kick_flushing_caps mds%d\n", session->s_mds);
 	list_for_each_entry(ci, &session->s_cap_flushing, i_flushing_item) {
@@ -2205,16 +2198,11 @@ void ceph_early_kick_flushing_caps(struct ceph_mds_client *mdsc,
 		if ((cap->issued & ci->i_flushing_caps) !=
 		    ci->i_flushing_caps) {
 			spin_unlock(&ci->i_ceph_lock);
-			if (!__kick_flushing_caps(mdsc, session, ci, true))
+			if (!__kick_flushing_caps(mdsc, session, ci))
 				continue;
 			spin_lock(&ci->i_ceph_lock);
 		}
 
-		for (n = rb_first(&ci->i_cap_flush_tree); n; n = rb_next(n)) {
-			cf = rb_entry(n, struct ceph_cap_flush, i_node);
-			cf->kick = true;
-		}
-
 		spin_unlock(&ci->i_ceph_lock);
 	}
 }
@@ -2228,7 +2216,7 @@ void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc,
 
 	dout("kick_flushing_caps mds%d\n", session->s_mds);
 	list_for_each_entry(ci, &session->s_cap_flushing, i_flushing_item) {
-		int delayed = __kick_flushing_caps(mdsc, session, ci, false);
+		int delayed = __kick_flushing_caps(mdsc, session, ci);
 		if (delayed) {
 			spin_lock(&ci->i_ceph_lock);
 			__cap_delay_requeue(mdsc, ci);
@@ -2261,7 +2249,7 @@ static void kick_flushing_inode_caps(struct ceph_mds_client *mdsc,
 
 		spin_unlock(&ci->i_ceph_lock);
 
-		delayed = __kick_flushing_caps(mdsc, session, ci, true);
+		delayed = __kick_flushing_caps(mdsc, session, ci);
 		if (delayed) {
 			spin_lock(&ci->i_ceph_lock);
 			__cap_delay_requeue(mdsc, ci);
diff --git a/fs/ceph/locks.c b/fs/ceph/locks.c
index 4347039ecc18..6706bde9ad1b 100644
--- a/fs/ceph/locks.c
+++ b/fs/ceph/locks.c
@@ -287,7 +287,7 @@ int ceph_encode_locks_to_buffer(struct inode *inode,
 		return 0;
 
 	spin_lock(&ctx->flc_lock);
-	list_for_each_entry(lock, &ctx->flc_flock, fl_list) {
+	list_for_each_entry(lock, &ctx->flc_posix, fl_list) {
 		++seen_fcntl;
 		if (seen_fcntl > num_fcntl_locks) {
 			err = -ENOSPC;
diff --git a/fs/ceph/super.h b/fs/ceph/super.h
index 860cc016e70d..2f2460d23a06 100644
--- a/fs/ceph/super.h
+++ b/fs/ceph/super.h
@@ -189,7 +189,6 @@ static inline void ceph_put_cap_snap(struct ceph_cap_snap *capsnap)
 struct ceph_cap_flush {
 	u64 tid;
 	int caps;
-	bool kick;
 	struct rb_node g_node; // global
 	union {
 		struct rb_node i_node; // inode