Merge git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client

* git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client:
  rbd: use watch/notify for changes in rbd header
  libceph: add lingering request and watch/notify event framework
  rbd: update email address in Documentation
  ceph: rename dentry_release -> d_release, fix comment
  ceph: add request to the tail of unsafe write list
  ceph: remove request from unsafe list if it is canceled/timed out
  ceph: move readahead default to fs/ceph from libceph
  ceph: add ino32 mount option
  ceph: update common header files
  ceph: remove debugfs debug cruft
  libceph: fix osd request queuing on osdmap updates
  ceph: preserve I_COMPLETE across rename
  libceph: Fix base64-decoding when input ends in newline.

Documentation/ABI/testing/sysfs-bus-rbd

 What:		/sys/bus/rbd/
 Date:		November 2010
-Contact:	Yehuda Sadeh <yehuda@hq.newdream.net>,
+Contact:	Yehuda Sadeh <yehuda@newdream.net>,
 		Sage Weil <sage@newdream.net>
 Description:
 

fs/ceph/debugfs.c

@@ -210,8 +210,6 @@ int ceph_fs_debugfs_init(struct ceph_fs_client *fsc)
 	if (!fsc->debugfs_congestion_kb)
 		goto out;
 
-	dout("a\n");
-
 	snprintf(name, sizeof(name), "../../bdi/%s",
 		 dev_name(fsc->backing_dev_info.dev));
 	fsc->debugfs_bdi =
@@ -221,7 +219,6 @@ int ceph_fs_debugfs_init(struct ceph_fs_client *fsc)
 	if (!fsc->debugfs_bdi)
 		goto out;
 
-	dout("b\n");
 	fsc->debugfs_mdsmap = debugfs_create_file("mdsmap",
 					0600,
 					fsc->client->debugfs_dir,
@@ -230,7 +227,6 @@ int ceph_fs_debugfs_init(struct ceph_fs_client *fsc)
 	if (!fsc->debugfs_mdsmap)
 		goto out;
 
-	dout("ca\n");
 	fsc->debugfs_mdsc = debugfs_create_file("mdsc",
 						0600,
 						fsc->client->debugfs_dir,
@@ -239,7 +235,6 @@ int ceph_fs_debugfs_init(struct ceph_fs_client *fsc)
 	if (!fsc->debugfs_mdsc)
 		goto out;
 
-	dout("da\n");
 	fsc->debugfs_caps = debugfs_create_file("caps",
 						   0400,
 						   fsc->client->debugfs_dir,
@@ -248,7 +243,6 @@ int ceph_fs_debugfs_init(struct ceph_fs_client *fsc)
 	if (!fsc->debugfs_caps)
 		goto out;
 
-	dout("ea\n");
 	fsc->debugfs_dentry_lru = debugfs_create_file("dentry_lru",
 					0600,
 					fsc->client->debugfs_dir,

fs/ceph/dir.c

@@ -161,7 +161,7 @@ static int __dcache_readdir(struct file *filp,
 	filp->f_pos = di->offset;
 	err = filldir(dirent, dentry->d_name.name,
 		      dentry->d_name.len, di->offset,
-		      dentry->d_inode->i_ino,
+		      ceph_translate_ino(dentry->d_sb, dentry->d_inode->i_ino),
 		      dentry->d_inode->i_mode >> 12);
 
 	if (last) {
@@ -245,15 +245,17 @@ static int ceph_readdir(struct file *filp, void *dirent, filldir_t filldir)
 
 		dout("readdir off 0 -> '.'\n");
 		if (filldir(dirent, ".", 1, ceph_make_fpos(0, 0),
-			    inode->i_ino, inode->i_mode >> 12) < 0)
+			    ceph_translate_ino(inode->i_sb, inode->i_ino),
+			    inode->i_mode >> 12) < 0)
 			return 0;
 		filp->f_pos = 1;
 		off = 1;
 	}
 	if (filp->f_pos == 1) {
+		ino_t ino = filp->f_dentry->d_parent->d_inode->i_ino;
 		dout("readdir off 1 -> '..'\n");
 		if (filldir(dirent, "..", 2, ceph_make_fpos(0, 1),
-			    filp->f_dentry->d_parent->d_inode->i_ino,
+			    ceph_translate_ino(inode->i_sb, ino),
 			    inode->i_mode >> 12) < 0)
 			return 0;
 		filp->f_pos = 2;
@@ -377,7 +379,8 @@ static int ceph_readdir(struct file *filp, void *dirent, filldir_t filldir)
 		if (filldir(dirent,
 			    rinfo->dir_dname[off - fi->offset],
 			    rinfo->dir_dname_len[off - fi->offset],
-			    pos, ino, ftype) < 0) {
+			    pos,
+			    ceph_translate_ino(inode->i_sb, ino), ftype) < 0) {
 			dout("filldir stopping us...\n");
 			return 0;
 		}
@@ -1024,14 +1027,13 @@ static int ceph_d_revalidate(struct dentry *dentry, struct nameidata *nd)
 }
 
 /*
- * When a dentry is released, clear the dir I_COMPLETE if it was part
- * of the current dir gen or if this is in the snapshot namespace.
+ * Release our ceph_dentry_info.
  */
-static void ceph_dentry_release(struct dentry *dentry)
+static void ceph_d_release(struct dentry *dentry)
 {
 	struct ceph_dentry_info *di = ceph_dentry(dentry);
 
-	dout("dentry_release %p\n", dentry);
+	dout("d_release %p\n", dentry);
 	if (di) {
 		ceph_dentry_lru_del(dentry);
 		if (di->lease_session)
@@ -1256,14 +1258,14 @@ const struct inode_operations ceph_dir_iops = {
 
 const struct dentry_operations ceph_dentry_ops = {
 	.d_revalidate = ceph_d_revalidate,
-	.d_release = ceph_dentry_release,
+	.d_release = ceph_d_release,
 };
 
 const struct dentry_operations ceph_snapdir_dentry_ops = {
 	.d_revalidate = ceph_snapdir_d_revalidate,
-	.d_release = ceph_dentry_release,
+	.d_release = ceph_d_release,
 };
 
 const struct dentry_operations ceph_snap_dentry_ops = {
-	.d_release = ceph_dentry_release,
+	.d_release = ceph_d_release,
 };

fs/ceph/file.c

@@ -564,11 +564,19 @@ static ssize_t ceph_sync_write(struct file *file, const char __user *data,
 			 * start_request so that a tid has been assigned.
 			 */
 			spin_lock(&ci->i_unsafe_lock);
-			list_add(&req->r_unsafe_item, &ci->i_unsafe_writes);
+			list_add_tail(&req->r_unsafe_item,
+				      &ci->i_unsafe_writes);
 			spin_unlock(&ci->i_unsafe_lock);
 			ceph_get_cap_refs(ci, CEPH_CAP_FILE_WR);
 		}
+		
 		ret = ceph_osdc_wait_request(&fsc->client->osdc, req);
+		if (ret < 0 && req->r_safe_callback) {
+			spin_lock(&ci->i_unsafe_lock);
+			list_del_init(&req->r_unsafe_item);
+			spin_unlock(&ci->i_unsafe_lock);
+			ceph_put_cap_refs(ci, CEPH_CAP_FILE_WR);
+		}
 	}
 
 	if (file->f_flags & O_DIRECT)

fs/ceph/inode.c

@@ -36,6 +36,13 @@ static void ceph_vmtruncate_work(struct work_struct *work);
 /*
  * find or create an inode, given the ceph ino number
  */
+static int ceph_set_ino_cb(struct inode *inode, void *data)
+{
+	ceph_inode(inode)->i_vino = *(struct ceph_vino *)data;
+	inode->i_ino = ceph_vino_to_ino(*(struct ceph_vino *)data);
+	return 0;
+}
+
 struct inode *ceph_get_inode(struct super_block *sb, struct ceph_vino vino)
 {
 	struct inode *inode;
@@ -1030,9 +1037,6 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req,
 			dout("fill_trace doing d_move %p -> %p\n",
 			     req->r_old_dentry, dn);
 
-			/* d_move screws up d_subdirs order */
-			ceph_i_clear(dir, CEPH_I_COMPLETE);
-
 			d_move(req->r_old_dentry, dn);
 			dout(" src %p '%.*s' dst %p '%.*s'\n",
 			     req->r_old_dentry,
@@ -1044,12 +1048,15 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req,
 			   rehashing bug in vfs_rename_dir */
 			ceph_invalidate_dentry_lease(dn);
 
-			/* take overwritten dentry's readdir offset */
-			dout("dn %p gets %p offset %lld (old offset %lld)\n",
-			     req->r_old_dentry, dn, ceph_dentry(dn)->offset,
+			/*
+			 * d_move() puts the renamed dentry at the end of
+			 * d_subdirs.  We need to assign it an appropriate
+			 * directory offset so we can behave when holding
+			 * I_COMPLETE.
+			 */
+			ceph_set_dentry_offset(req->r_old_dentry);
+			dout("dn %p gets new offset %lld\n", req->r_old_dentry, 
 			     ceph_dentry(req->r_old_dentry)->offset);
-			ceph_dentry(req->r_old_dentry)->offset =
-				ceph_dentry(dn)->offset;
 
 			dn = req->r_old_dentry;  /* use old_dentry */
 			in = dn->d_inode;
@@ -1809,7 +1816,7 @@ int ceph_getattr(struct vfsmount *mnt, struct dentry *dentry,
 	err = ceph_do_getattr(inode, CEPH_STAT_CAP_INODE_ALL);
 	if (!err) {
 		generic_fillattr(inode, stat);
-		stat->ino = inode->i_ino;
+		stat->ino = ceph_translate_ino(inode->i_sb, inode->i_ino);
 		if (ceph_snap(inode) != CEPH_NOSNAP)
 			stat->dev = ceph_snap(inode);
 		else

fs/ceph/super.c

@@ -131,6 +131,7 @@ enum {
 	Opt_rbytes,
 	Opt_norbytes,
 	Opt_noasyncreaddir,
+	Opt_ino32,
 };
 
 static match_table_t fsopt_tokens = {
@@ -150,6 +151,7 @@ static match_table_t fsopt_tokens = {
 	{Opt_rbytes, "rbytes"},
 	{Opt_norbytes, "norbytes"},
 	{Opt_noasyncreaddir, "noasyncreaddir"},
+	{Opt_ino32, "ino32"},
 	{-1, NULL}
 };
 
@@ -225,6 +227,9 @@ static int parse_fsopt_token(char *c, void *private)
 	case Opt_noasyncreaddir:
 		fsopt->flags |= CEPH_MOUNT_OPT_NOASYNCREADDIR;
 		break;
+	case Opt_ino32:
+		fsopt->flags |= CEPH_MOUNT_OPT_INO32;
+		break;
 	default:
 		BUG_ON(token);
 	}
@@ -288,7 +293,7 @@ static int parse_mount_options(struct ceph_mount_options **pfsopt,
         fsopt->sb_flags = flags;
         fsopt->flags = CEPH_MOUNT_OPT_DEFAULT;
 
-        fsopt->rsize = CEPH_MOUNT_RSIZE_DEFAULT;
+        fsopt->rsize = CEPH_RSIZE_DEFAULT;
         fsopt->snapdir_name = kstrdup(CEPH_SNAPDIRNAME_DEFAULT, GFP_KERNEL);
 	fsopt->caps_wanted_delay_min = CEPH_CAPS_WANTED_DELAY_MIN_DEFAULT;
 	fsopt->caps_wanted_delay_max = CEPH_CAPS_WANTED_DELAY_MAX_DEFAULT;
@@ -370,7 +375,7 @@ static int ceph_show_options(struct seq_file *m, struct vfsmount *mnt)
 
 	if (fsopt->wsize)
 		seq_printf(m, ",wsize=%d", fsopt->wsize);
-	if (fsopt->rsize != CEPH_MOUNT_RSIZE_DEFAULT)
+	if (fsopt->rsize != CEPH_RSIZE_DEFAULT)
 		seq_printf(m, ",rsize=%d", fsopt->rsize);
 	if (fsopt->congestion_kb != default_congestion_kb())
 		seq_printf(m, ",write_congestion_kb=%d", fsopt->congestion_kb);

fs/ceph/super.h

@@ -27,6 +27,7 @@
 #define CEPH_MOUNT_OPT_DIRSTAT         (1<<4) /* `cat dirname` for stats */
 #define CEPH_MOUNT_OPT_RBYTES          (1<<5) /* dir st_bytes = rbytes */
 #define CEPH_MOUNT_OPT_NOASYNCREADDIR  (1<<7) /* no dcache readdir */
+#define CEPH_MOUNT_OPT_INO32           (1<<8) /* 32 bit inos */
 
 #define CEPH_MOUNT_OPT_DEFAULT    (CEPH_MOUNT_OPT_RBYTES)
 
@@ -35,6 +36,7 @@
 #define ceph_test_mount_opt(fsc, opt) \
 	(!!((fsc)->mount_options->flags & CEPH_MOUNT_OPT_##opt))
 
+#define CEPH_RSIZE_DEFAULT             (512*1024) /* readahead */
 #define CEPH_MAX_READDIR_DEFAULT        1024
 #define CEPH_MAX_READDIR_BYTES_DEFAULT  (512*1024)
 #define CEPH_SNAPDIRNAME_DEFAULT        ".snap"
@@ -319,6 +321,16 @@ static inline struct ceph_inode_info *ceph_inode(struct inode *inode)
 	return container_of(inode, struct ceph_inode_info, vfs_inode);
 }
 
+static inline struct ceph_fs_client *ceph_inode_to_client(struct inode *inode)
+{
+	return (struct ceph_fs_client *)inode->i_sb->s_fs_info;
+}
+
+static inline struct ceph_fs_client *ceph_sb_to_client(struct super_block *sb)
+{
+	return (struct ceph_fs_client *)sb->s_fs_info;
+}
+
 static inline struct ceph_vino ceph_vino(struct inode *inode)
 {
 	return ceph_inode(inode)->i_vino;
@@ -327,19 +339,49 @@ static inline struct ceph_vino ceph_vino(struct inode *inode)
 /*
  * ino_t is <64 bits on many architectures, blech.
  *
- * don't include snap in ino hash, at least for now.
+ *               i_ino (kernel inode)   st_ino (userspace)
+ * i386          32                     32
+ * x86_64+ino32  64                     32
+ * x86_64        64                     64
+ */
+static inline u32 ceph_ino_to_ino32(ino_t ino)
+{
+	ino ^= ino >> (sizeof(ino) * 8 - 32);
+	if (!ino)
+		ino = 1;
+	return ino;
+}
+
+/*
+ * kernel i_ino value
  */
 static inline ino_t ceph_vino_to_ino(struct ceph_vino vino)
 {
 	ino_t ino = (ino_t)vino.ino;  /* ^ (vino.snap << 20); */
 #if BITS_PER_LONG == 32
-	ino ^= vino.ino >> (sizeof(u64)-sizeof(ino_t)) * 8;
-	if (!ino)
-		ino = 1;
+	ino = ceph_ino_to_ino32(ino);
 #endif
 	return ino;
 }
 
+/*
+ * user-visible ino (stat, filldir)
+ */
+#if BITS_PER_LONG == 32
+static inline ino_t ceph_translate_ino(struct super_block *sb, ino_t ino)
+{
+	return ino;
+}
+#else
+static inline ino_t ceph_translate_ino(struct super_block *sb, ino_t ino)
+{
+	if (ceph_test_mount_opt(ceph_sb_to_client(sb), INO32))
+		ino = ceph_ino_to_ino32(ino);
+	return ino;
+}
+#endif
+
+
 /* for printf-style formatting */
 #define ceph_vinop(i) ceph_inode(i)->i_vino.ino, ceph_inode(i)->i_vino.snap
 
@@ -428,13 +470,6 @@ static inline loff_t ceph_make_fpos(unsigned frag, unsigned off)
 	return ((loff_t)frag << 32) | (loff_t)off;
 }
 
-static inline int ceph_set_ino_cb(struct inode *inode, void *data)
-{
-	ceph_inode(inode)->i_vino = *(struct ceph_vino *)data;
-	inode->i_ino = ceph_vino_to_ino(*(struct ceph_vino *)data);
-	return 0;
-}
-
 /*
  * caps helpers
  */
@@ -503,15 +538,6 @@ extern void ceph_reservation_status(struct ceph_fs_client *client,
 				    int *total, int *avail, int *used,
 				    int *reserved, int *min);
 
-static inline struct ceph_fs_client *ceph_inode_to_client(struct inode *inode)
-{
-	return (struct ceph_fs_client *)inode->i_sb->s_fs_info;
-}
-
-static inline struct ceph_fs_client *ceph_sb_to_client(struct super_block *sb)
-{
-	return (struct ceph_fs_client *)sb->s_fs_info;
-}
 
 
 /*

include/linux/ceph/ceph_fs.h

@@ -136,9 +136,18 @@ struct ceph_dir_layout {
 
 
 /* osd */
-#define CEPH_MSG_OSD_MAP          41
-#define CEPH_MSG_OSD_OP           42
-#define CEPH_MSG_OSD_OPREPLY      43
+#define CEPH_MSG_OSD_MAP                41
+#define CEPH_MSG_OSD_OP                 42
+#define CEPH_MSG_OSD_OPREPLY            43
+#define CEPH_MSG_WATCH_NOTIFY           44
+
+
+/* watch-notify operations */
+enum {
+  WATCH_NOTIFY				= 1, /* notifying watcher */
+  WATCH_NOTIFY_COMPLETE			= 2, /* notifier notified when done */
+};
+
 
 /* pool operations */
 enum {
@@ -213,8 +222,10 @@ struct ceph_client_mount {
 	struct ceph_mon_request_header monhdr;
 } __attribute__ ((packed));
 
+#define CEPH_SUBSCRIBE_ONETIME    1  /* i want only 1 update after have */
+
 struct ceph_mon_subscribe_item {
-	__le64 have_version;	__le64 have;
+	__le64 have_version;    __le64 have;
 	__u8 onetime;
 } __attribute__ ((packed));
 

include/linux/ceph/libceph.h

@@ -71,7 +71,6 @@ struct ceph_options {
 #define CEPH_OSD_TIMEOUT_DEFAULT    60  /* seconds */
 #define CEPH_OSD_KEEPALIVE_DEFAULT  5
 #define CEPH_OSD_IDLE_TTL_DEFAULT    60
-#define CEPH_MOUNT_RSIZE_DEFAULT    (512*1024) /* readahead */
 
 #define CEPH_MSG_MAX_FRONT_LEN	(16*1024*1024)
 #define CEPH_MSG_MAX_DATA_LEN	(16*1024*1024)

include/linux/ceph/osd_client.h

@@ -32,6 +32,7 @@ struct ceph_osd {
 	struct rb_node o_node;
 	struct ceph_connection o_con;
 	struct list_head o_requests;
+	struct list_head o_linger_requests;
 	struct list_head o_osd_lru;
 	struct ceph_authorizer *o_authorizer;
 	void *o_authorizer_buf, *o_authorizer_reply_buf;
@@ -47,6 +48,8 @@ struct ceph_osd_request {
 	struct rb_node  r_node;
 	struct list_head r_req_lru_item;
 	struct list_head r_osd_item;
+	struct list_head r_linger_item;
+	struct list_head r_linger_osd;
 	struct ceph_osd *r_osd;
 	struct ceph_pg   r_pgid;
 	int              r_pg_osds[CEPH_PG_MAX_SIZE];
@@ -59,6 +62,7 @@ struct ceph_osd_request {
 	int               r_flags;     /* any additional flags for the osd */
 	u32               r_sent;      /* >0 if r_request is sending/sent */
 	int               r_got_reply;
+	int		  r_linger;
 
 	struct ceph_osd_client *r_osdc;
 	struct kref       r_kref;
@@ -74,7 +78,6 @@ struct ceph_osd_request {
 	char              r_oid[40];          /* object name */
 	int               r_oid_len;
 	unsigned long     r_stamp;            /* send OR check time */
-	bool              r_resend;           /* msg send failed, needs retry */
 
 	struct ceph_file_layout r_file_layout;
 	struct ceph_snap_context *r_snapc;    /* snap context for writes */
@@ -90,6 +93,26 @@ struct ceph_osd_request {
 	struct ceph_pagelist *r_trail;	      /* trailing part of the data */
 };
 
+struct ceph_osd_event {
+	u64 cookie;
+	int one_shot;
+	struct ceph_osd_client *osdc;
+	void (*cb)(u64, u64, u8, void *);
+	void *data;
+	struct rb_node node;
+	struct list_head osd_node;
+	struct kref kref;
+	struct completion completion;
+};
+
+struct ceph_osd_event_work {
+	struct work_struct work;
+	struct ceph_osd_event *event;
+        u64 ver;
+        u64 notify_id;
+        u8 opcode;
+};
+
 struct ceph_osd_client {
 	struct ceph_client     *client;
 
@@ -104,7 +127,10 @@ struct ceph_osd_client {
 	u64                    timeout_tid;   /* tid of timeout triggering rq */
 	u64                    last_tid;      /* tid of last request */
 	struct rb_root         requests;      /* pending requests */
-	struct list_head       req_lru;	      /* pending requests lru */
+	struct list_head       req_lru;	      /* in-flight lru */
+	struct list_head       req_unsent;    /* unsent/need-resend queue */
+	struct list_head       req_notarget;  /* map to no osd */
+	struct list_head       req_linger;    /* lingering requests */
 	int                    num_requests;
 	struct delayed_work    timeout_work;
 	struct delayed_work    osds_timeout_work;
@@ -116,6 +142,12 @@ struct ceph_osd_client {
 
 	struct ceph_msgpool	msgpool_op;
 	struct ceph_msgpool	msgpool_op_reply;
+
+	spinlock_t		event_lock;
+	struct rb_root		event_tree;
+	u64			event_count;
+
+	struct workqueue_struct	*notify_wq;
 };
 
 struct ceph_osd_req_op {
@@ -150,6 +182,13 @@ struct ceph_osd_req_op {
 	        struct {
 		        u64 snapid;
 	        } snap;
+		struct {
+			u64 cookie;
+			u64 ver;
+			__u8 flag;
+			u32 prot_ver;
+			u32 timeout;
+		} watch;
 	};
 	u32 payload_len;
 };
@@ -198,6 +237,11 @@ extern struct ceph_osd_request *ceph_osdc_new_request(struct ceph_osd_client *,
 				      bool use_mempool, int num_reply,
 				      int page_align);
 
+extern void ceph_osdc_set_request_linger(struct ceph_osd_client *osdc,
+					 struct ceph_osd_request *req);
+extern void ceph_osdc_unregister_linger_request(struct ceph_osd_client *osdc,
+						struct ceph_osd_request *req);
+
 static inline void ceph_osdc_get_request(struct ceph_osd_request *req)
 {
 	kref_get(&req->r_kref);
@@ -233,5 +277,14 @@ extern int ceph_osdc_writepages(struct ceph_osd_client *osdc,
 				struct page **pages, int nr_pages,
 				int flags, int do_sync, bool nofail);
 
+/* watch/notify events */
+extern int ceph_osdc_create_event(struct ceph_osd_client *osdc,
+				  void (*event_cb)(u64, u64, u8, void *),
+				  int one_shot, void *data,
+				  struct ceph_osd_event **pevent);
+extern void ceph_osdc_cancel_event(struct ceph_osd_event *event);
+extern int ceph_osdc_wait_event(struct ceph_osd_event *event,
+				unsigned long timeout);
+extern void ceph_osdc_put_event(struct ceph_osd_event *event);
 #endif
 

include/linux/ceph/rados.h

@@ -12,9 +12,9 @@
  * osdmap encoding versions
  */
 #define CEPH_OSDMAP_INC_VERSION     5
-#define CEPH_OSDMAP_INC_VERSION_EXT 5
+#define CEPH_OSDMAP_INC_VERSION_EXT 6
 #define CEPH_OSDMAP_VERSION         5
-#define CEPH_OSDMAP_VERSION_EXT     5
+#define CEPH_OSDMAP_VERSION_EXT     6
 
 /*
  * fs id
@@ -181,9 +181,17 @@ enum {
 	/* read */
 	CEPH_OSD_OP_READ      = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 1,
 	CEPH_OSD_OP_STAT      = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 2,
+	CEPH_OSD_OP_MAPEXT    = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 3,
 
 	/* fancy read */
-	CEPH_OSD_OP_MASKTRUNC = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 4,
+	CEPH_OSD_OP_MASKTRUNC   = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 4,
+	CEPH_OSD_OP_SPARSE_READ = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 5,
+
+	CEPH_OSD_OP_NOTIFY    = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 6,
+	CEPH_OSD_OP_NOTIFY_ACK = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 7,
+
+	/* versioning */
+	CEPH_OSD_OP_ASSERT_VER = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_DATA | 8,
 
 	/* write */
 	CEPH_OSD_OP_WRITE     = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 1,
@@ -205,6 +213,8 @@ enum {
 	CEPH_OSD_OP_CREATE  = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 13,
 	CEPH_OSD_OP_ROLLBACK= CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 14,
 
+	CEPH_OSD_OP_WATCH   = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_DATA | 15,
+
 	/** attrs **/
 	/* read */
 	CEPH_OSD_OP_GETXATTR  = CEPH_OSD_OP_MODE_RD | CEPH_OSD_OP_TYPE_ATTR | 1,
@@ -218,11 +228,14 @@ enum {
 	CEPH_OSD_OP_RMXATTR   = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_ATTR | 4,
 
 	/** subop **/
-	CEPH_OSD_OP_PULL           = CEPH_OSD_OP_MODE_SUB | 1,
-	CEPH_OSD_OP_PUSH           = CEPH_OSD_OP_MODE_SUB | 2,
-	CEPH_OSD_OP_BALANCEREADS   = CEPH_OSD_OP_MODE_SUB | 3,
-	CEPH_OSD_OP_UNBALANCEREADS = CEPH_OSD_OP_MODE_SUB | 4,
-	CEPH_OSD_OP_SCRUB          = CEPH_OSD_OP_MODE_SUB | 5,
+	CEPH_OSD_OP_PULL            = CEPH_OSD_OP_MODE_SUB | 1,
+	CEPH_OSD_OP_PUSH            = CEPH_OSD_OP_MODE_SUB | 2,
+	CEPH_OSD_OP_BALANCEREADS    = CEPH_OSD_OP_MODE_SUB | 3,
+	CEPH_OSD_OP_UNBALANCEREADS  = CEPH_OSD_OP_MODE_SUB | 4,
+	CEPH_OSD_OP_SCRUB           = CEPH_OSD_OP_MODE_SUB | 5,
+	CEPH_OSD_OP_SCRUB_RESERVE   = CEPH_OSD_OP_MODE_SUB | 6,
+	CEPH_OSD_OP_SCRUB_UNRESERVE = CEPH_OSD_OP_MODE_SUB | 7,
+	CEPH_OSD_OP_SCRUB_STOP      = CEPH_OSD_OP_MODE_SUB | 8,
 
 	/** lock **/
 	CEPH_OSD_OP_WRLOCK    = CEPH_OSD_OP_MODE_WR | CEPH_OSD_OP_TYPE_LOCK | 1,
@@ -328,6 +341,8 @@ enum {
 	CEPH_OSD_CMPXATTR_MODE_U64    = 2
 };
 
+#define RADOS_NOTIFY_VER	1
+
 /*
  * an individual object operation.  each may be accompanied by some data
  * payload
@@ -359,7 +374,12 @@ struct ceph_osd_op {
 	        struct {
 		        __le64 snapid;
 	        } __attribute__ ((packed)) snap;
-	};
+		struct {
+			__le64 cookie;
+			__le64 ver;
+			__u8 flag;	/* 0 = unwatch, 1 = watch */
+		} __attribute__ ((packed)) watch;
+};
 	__le32 payload_len;
 } __attribute__ ((packed));
 
@@ -402,4 +422,5 @@ struct ceph_osd_reply_head {
 } __attribute__ ((packed));
 
 
+
 #endif

net/ceph/armor.c

@@ -78,8 +78,10 @@ int ceph_unarmor(char *dst, const char *src, const char *end)
 	while (src < end) {
 		int a, b, c, d;
 
-		if (src < end && src[0] == '\n')
+		if (src[0] == '\n') {
 			src++;
+			continue;
+		}
 		if (src + 4 > end)
 			return -EINVAL;
 		a = decode_bits(src[0]);

net/ceph/ceph_common.c

@@ -62,6 +62,7 @@ const char *ceph_msg_type_name(int type)
 	case CEPH_MSG_OSD_MAP: return "osd_map";
 	case CEPH_MSG_OSD_OP: return "osd_op";
 	case CEPH_MSG_OSD_OPREPLY: return "osd_opreply";
+	case CEPH_MSG_WATCH_NOTIFY: return "watch_notify";
 	default: return "unknown";
 	}
 }