[linux-block.git] / fs / ceph / mds_client.h

/* SPDX-License-Identifier: GPL-2.0 */
#ifndef _FS_CEPH_MDS_CLIENT_H
#define _FS_CEPH_MDS_CLIENT_H

#include <linux/completion.h>
#include <linux/kref.h>
#include <linux/list.h>
#include <linux/mutex.h>
#include <linux/rbtree.h>
#include <linux/spinlock.h>
#include <linux/refcount.h>
#include <linux/utsname.h>

#include <linux/ceph/types.h>
#include <linux/ceph/messenger.h>
#include <linux/ceph/mdsmap.h>
#include <linux/ceph/auth.h>

/* The first 8 bits are reserved for old ceph releases */
enum ceph_feature_type {
	CEPHFS_FEATURE_MIMIC = 8,
	CEPHFS_FEATURE_REPLY_ENCODING,
	CEPHFS_FEATURE_RECLAIM_CLIENT,
	CEPHFS_FEATURE_LAZY_CAP_WANTED,
	CEPHFS_FEATURE_MULTI_RECONNECT,
	CEPHFS_FEATURE_DELEG_INO,

	CEPHFS_FEATURE_MAX = CEPHFS_FEATURE_DELEG_INO,
};

/*
 * This will always have the highest feature bit value
 * as the last element of the array.
 */
#define CEPHFS_FEATURES_CLIENT_SUPPORTED {	\
	0, 1, 2, 3, 4, 5, 6, 7,			\
	CEPHFS_FEATURE_MIMIC,			\
	CEPHFS_FEATURE_REPLY_ENCODING,		\
	CEPHFS_FEATURE_LAZY_CAP_WANTED,		\
	CEPHFS_FEATURE_MULTI_RECONNECT,		\
	CEPHFS_FEATURE_DELEG_INO,		\
						\
	CEPHFS_FEATURE_MAX,			\
}
#define CEPHFS_FEATURES_CLIENT_REQUIRED {}

/*
 * Some lock dependencies:
 *
 * session->s_mutex
 *         mdsc->mutex
 *
 *         mdsc->snap_rwsem
 *
 *         ci->i_ceph_lock
 *                 mdsc->snap_flush_lock
 *                 mdsc->cap_delay_lock
 *
 */

struct ceph_fs_client;
struct ceph_cap;

/*
 * parsed info about a single inode.  pointers are into the encoded
 * on-wire structures within the mds reply message payload.
 */
struct ceph_mds_reply_info_in {
	struct ceph_mds_reply_inode *in;
	struct ceph_dir_layout dir_layout;
	u32 symlink_len;
	char *symlink;
	u32 xattr_len;
	char *xattr_data;
	u64 inline_version;
	u32 inline_len;
	char *inline_data;
	u32 pool_ns_len;
	char *pool_ns_data;
	u64 max_bytes;
	u64 max_files;
	s32 dir_pin;
	struct ceph_timespec btime;
	struct ceph_timespec snap_btime;
	u64 change_attr;
};

struct ceph_mds_reply_dir_entry {
	char                          *name;
	u32                           name_len;
	struct ceph_mds_reply_lease   *lease;
	struct ceph_mds_reply_info_in inode;
	loff_t			      offset;
};

/*
 * parsed info about an mds reply, including information about
 * either: 1) the target inode and/or its parent directory and dentry,
 * and directory contents (for readdir results), or
 * 2) the file range lock info (for fcntl F_GETLK results).
 */
struct ceph_mds_reply_info_parsed {
	struct ceph_mds_reply_head    *head;

	/* trace */
	struct ceph_mds_reply_info_in diri, targeti;
	struct ceph_mds_reply_dirfrag *dirfrag;
	char                          *dname;
	u32                           dname_len;
	struct ceph_mds_reply_lease   *dlease;

	/* extra */
	union {
		/* for fcntl F_GETLK results */
		struct ceph_filelock *filelock_reply;

		/* for readdir results */
		struct {
			struct ceph_mds_reply_dirfrag *dir_dir;
			size_t			      dir_buf_size;
			int                           dir_nr;
			bool			      dir_end;
			bool			      dir_complete;
			bool			      hash_order;
			bool			      offset_hash;
			struct ceph_mds_reply_dir_entry  *dir_entries;
		};

		/* for create results */
		struct {
			bool has_create_ino;
			u64 ino;
		};
	};

	/* encoded blob describing snapshot contexts for certain
	   operations (e.g., open) */
	void *snapblob;
	int snapblob_len;
};


/*
 * cap releases are batched and sent to the MDS en masse.
 *
 * Account for per-message overhead of mds_cap_release header
 * and __le32 for osd epoch barrier trailing field.
 */
#define CEPH_CAPS_PER_RELEASE ((PAGE_SIZE - sizeof(u32) -		\
				sizeof(struct ceph_mds_cap_release)) /	\
			        sizeof(struct ceph_mds_cap_item))


/*
 * state associated with each MDS<->client session
 */
enum {
	CEPH_MDS_SESSION_NEW = 1,
	CEPH_MDS_SESSION_OPENING = 2,
	CEPH_MDS_SESSION_OPEN = 3,
	CEPH_MDS_SESSION_HUNG = 4,
	CEPH_MDS_SESSION_RESTARTING = 5,
	CEPH_MDS_SESSION_RECONNECTING = 6,
	CEPH_MDS_SESSION_CLOSING = 7,
	CEPH_MDS_SESSION_CLOSED = 8,
	CEPH_MDS_SESSION_REJECTED = 9,
};

struct ceph_mds_session {
	struct ceph_mds_client *s_mdsc;
	int               s_mds;
	int               s_state;
	unsigned long     s_ttl;      /* time until mds kills us */
	unsigned long	  s_features;
	u64               s_seq;      /* incoming msg seq # */
	struct mutex      s_mutex;    /* serialize session messages */

	struct ceph_connection s_con;

	struct ceph_auth_handshake s_auth;

	/* protected by s_gen_ttl_lock */
	spinlock_t        s_gen_ttl_lock;
	u32               s_cap_gen;  /* inc each time we get mds stale msg */
	unsigned long     s_cap_ttl;  /* when session caps expire */

	/* protected by s_cap_lock */
	spinlock_t        s_cap_lock;
	refcount_t        s_ref;
	struct list_head  s_caps;     /* all caps issued by this session */
	struct ceph_cap  *s_cap_iterator;
	int               s_nr_caps;
	int               s_num_cap_releases;
	int		  s_cap_reconnect;
	int		  s_readonly;
	struct list_head  s_cap_releases; /* waiting cap_release messages */
	struct work_struct s_cap_release_work;

	/* protected by mutex */
	struct list_head  s_cap_flushing;     /* inodes w/ flushing caps */
	unsigned long     s_renew_requested; /* last time we sent a renew req */
	u64               s_renew_seq;

	struct list_head  s_waiting;  /* waiting requests */
	struct list_head  s_unsafe;   /* unsafe requests */
	struct xarray	  s_delegated_inos;
};

/*
 * modes of choosing which MDS to send a request to
 */
enum {
	USE_ANY_MDS,
	USE_RANDOM_MDS,
	USE_AUTH_MDS,   /* prefer authoritative mds for this metadata item */
};

struct ceph_mds_request;
struct ceph_mds_client;

/*
 * request completion callback
 */
typedef void (*ceph_mds_request_callback_t) (struct ceph_mds_client *mdsc,
					     struct ceph_mds_request *req);
/*
 * wait for request completion callback
 */
typedef int (*ceph_mds_request_wait_callback_t) (struct ceph_mds_client *mdsc,
						 struct ceph_mds_request *req);

/*
 * an in-flight mds request
 */
struct ceph_mds_request {
	u64 r_tid;                   /* transaction id */
	struct rb_node r_node;
	struct ceph_mds_client *r_mdsc;

	struct kref       r_kref;
	int r_op;                    /* mds op code */

	/* operation on what? */
	struct inode *r_inode;              /* arg1 */
	struct dentry *r_dentry;            /* arg1 */
	struct dentry *r_old_dentry;        /* arg2: rename from or link from */
	struct inode *r_old_dentry_dir;     /* arg2: old dentry's parent dir */
	char *r_path1, *r_path2;
	struct ceph_vino r_ino1, r_ino2;

	struct inode *r_parent;		    /* parent dir inode */
	struct inode *r_target_inode;       /* resulting inode */

#define CEPH_MDS_R_DIRECT_IS_HASH	(1) /* r_direct_hash is valid */
#define CEPH_MDS_R_ABORTED		(2) /* call was aborted */
#define CEPH_MDS_R_GOT_UNSAFE		(3) /* got an unsafe reply */
#define CEPH_MDS_R_GOT_SAFE		(4) /* got a safe reply */
#define CEPH_MDS_R_GOT_RESULT		(5) /* got a result */
#define CEPH_MDS_R_DID_PREPOPULATE	(6) /* prepopulated readdir */
#define CEPH_MDS_R_PARENT_LOCKED	(7) /* is r_parent->i_rwsem wlocked? */
#define CEPH_MDS_R_ASYNC		(8) /* async request */
	unsigned long	r_req_flags;

	struct mutex r_fill_mutex;

	union ceph_mds_request_args r_args;
	int r_fmode;        /* file mode, if expecting cap */
	kuid_t r_uid;
	kgid_t r_gid;
	int r_request_release_offset;
	struct timespec64 r_stamp;

	/* for choosing which mds to send this request to */
	int r_direct_mode;
	u32 r_direct_hash;      /* choose dir frag based on this dentry hash */

	/* data payload is used for xattr ops */
	struct ceph_pagelist *r_pagelist;

	/* what caps shall we drop? */
	int r_inode_drop, r_inode_unless;
	int r_dentry_drop, r_dentry_unless;
	int r_old_dentry_drop, r_old_dentry_unless;
	struct inode *r_old_inode;
	int r_old_inode_drop, r_old_inode_unless;

	struct ceph_msg  *r_request;  /* original request */
	struct ceph_msg  *r_reply;
	struct ceph_mds_reply_info_parsed r_reply_info;
	int r_err;


	struct page *r_locked_page;
	int r_dir_caps;
	int r_num_caps;
	u32               r_readdir_offset;

	unsigned long r_timeout;  /* optional.  jiffies, 0 is "wait forever" */
	unsigned long r_started;  /* start time to measure timeout against */
	unsigned long r_request_started; /* start time for mds request only,
					    used to measure lease durations */

	/* link unsafe requests to parent directory, for fsync */
	struct inode	*r_unsafe_dir;
	struct list_head r_unsafe_dir_item;

	/* unsafe requests that modify the target inode */
	struct list_head r_unsafe_target_item;

	struct ceph_mds_session *r_session;

	int               r_attempts;   /* resend attempts */
	int               r_num_fwd;    /* number of forward attempts */
	int               r_resend_mds; /* mds to resend to next, if any*/
	u32               r_sent_on_mseq; /* cap mseq request was sent at*/
	u64		  r_deleg_ino;

	struct list_head  r_wait;
	struct completion r_completion;
	struct completion r_safe_completion;
	ceph_mds_request_callback_t r_callback;
	ceph_mds_request_wait_callback_t r_wait_for_completion;
	struct list_head  r_unsafe_item;  /* per-session unsafe list item */

	long long	  r_dir_release_cnt;
	long long	  r_dir_ordered_cnt;
	int		  r_readdir_cache_idx;

	struct ceph_cap_reservation r_caps_reservation;
};

struct ceph_pool_perm {
	struct rb_node node;
	int perm;
	s64 pool;
	size_t pool_ns_len;
	char pool_ns[];
};

struct ceph_snapid_map {
	struct rb_node node;
	struct list_head lru;
	atomic_t ref;
	u64 snap;
	dev_t dev;
	unsigned long last_used;
};

/*
 * node for list of quotarealm inodes that are not visible from the filesystem
 * mountpoint, but required to handle, e.g. quotas.
 */
struct ceph_quotarealm_inode {
	struct rb_node node;
	u64 ino;
	unsigned long timeout; /* last time a lookup failed for this inode */
	struct mutex mutex;
	struct inode *inode;
};

struct cap_wait {
	struct list_head	list;
	unsigned long		ino;
	pid_t			tgid;
	int			need;
	int			want;
};

/*
 * mds client state
 */
struct ceph_mds_client {
	struct ceph_fs_client  *fsc;
	struct mutex            mutex;         /* all nested structures */

	struct ceph_mdsmap      *mdsmap;
	struct completion       safe_umount_waiters;
	wait_queue_head_t       session_close_wq;
	struct list_head        waiting_for_map;
	int 			mdsmap_err;

	struct ceph_mds_session **sessions;    /* NULL for mds if no session */
	atomic_t		num_sessions;
	int                     max_sessions;  /* len of s_mds_sessions */
	int                     stopping;      /* true if shutting down */

	atomic64_t		quotarealms_count; /* # realms with quota */
	/*
	 * We keep a list of inodes we don't see in the mountpoint but that we
	 * need to track quota realms.
	 */
	struct rb_root		quotarealms_inodes;
	struct mutex		quotarealms_inodes_mutex;

	/*
	 * snap_rwsem will cover cap linkage into snaprealms, and
	 * realm snap contexts.  (later, we can do per-realm snap
	 * contexts locks..)  the empty list contains realms with no
	 * references (implying they contain no inodes with caps) that
	 * should be destroyed.
	 */
	u64			last_snap_seq;
	struct rw_semaphore     snap_rwsem;
	struct rb_root          snap_realms;
	struct list_head        snap_empty;
	int			num_snap_realms;
	spinlock_t              snap_empty_lock;  /* protect snap_empty */

	u64                    last_tid;      /* most recent mds request */
	u64                    oldest_tid;    /* oldest incomplete mds request,
						 excluding setfilelock requests */
	struct rb_root         request_tree;  /* pending mds requests */
	struct delayed_work    delayed_work;  /* delayed work */
	unsigned long    last_renew_caps;  /* last time we renewed our caps */
	struct list_head cap_delay_list;   /* caps with delayed release */
	spinlock_t       cap_delay_lock;   /* protects cap_delay_list */
	struct list_head snap_flush_list;  /* cap_snaps ready to flush */
	spinlock_t       snap_flush_lock;

	u64               last_cap_flush_tid;
	struct list_head  cap_flush_list;
	struct list_head  cap_dirty;        /* inodes with dirty caps */
	struct list_head  cap_dirty_migrating; /* ...that are migration... */
	int               num_cap_flushing; /* # caps we are flushing */
	spinlock_t        cap_dirty_lock;   /* protects above items */
	wait_queue_head_t cap_flushing_wq;

	struct work_struct cap_reclaim_work;
	atomic_t	   cap_reclaim_pending;

	/*
	 * Cap reservations
	 *
	 * Maintain a global pool of preallocated struct ceph_caps, referenced
	 * by struct ceph_caps_reservations.  This ensures that we preallocate
	 * memory needed to successfully process an MDS response.  (If an MDS
	 * sends us cap information and we fail to process it, we will have
	 * problems due to the client and MDS being out of sync.)
	 *
	 * Reservations are 'owned' by a ceph_cap_reservation context.
	 */
	spinlock_t	caps_list_lock;
	struct		list_head caps_list; /* unused (reserved or
						unreserved) */
	struct		list_head cap_wait_list;
	int		caps_total_count;    /* total caps allocated */
	int		caps_use_count;      /* in use */
	int		caps_use_max;	     /* max used caps */
	int		caps_reserve_count;  /* unused, reserved */
	int		caps_avail_count;    /* unused, unreserved */
	int		caps_min_count;      /* keep at least this many
						(unreserved) */
	spinlock_t	  dentry_list_lock;
	struct list_head  dentry_leases;     /* fifo list */
	struct list_head  dentry_dir_leases; /* lru list */

	spinlock_t		snapid_map_lock;
	struct rb_root		snapid_map_tree;
	struct list_head	snapid_map_lru;

	struct rw_semaphore     pool_perm_rwsem;
	struct rb_root		pool_perm_tree;

	char nodename[__NEW_UTS_LEN + 1];
};

extern const char *ceph_mds_op_name(int op);

extern struct ceph_mds_session *
__ceph_lookup_mds_session(struct ceph_mds_client *, int mds);

extern const char *ceph_session_state_name(int s);

extern struct ceph_mds_session *
ceph_get_mds_session(struct ceph_mds_session *s);
extern void ceph_put_mds_session(struct ceph_mds_session *s);

extern int ceph_send_msg_mds(struct ceph_mds_client *mdsc,
			     struct ceph_msg *msg, int mds);

extern int ceph_mdsc_init(struct ceph_fs_client *fsc);
extern void ceph_mdsc_close_sessions(struct ceph_mds_client *mdsc);
extern void ceph_mdsc_force_umount(struct ceph_mds_client *mdsc);
extern void ceph_mdsc_destroy(struct ceph_fs_client *fsc);

extern void ceph_mdsc_sync(struct ceph_mds_client *mdsc);

extern void ceph_invalidate_dir_request(struct ceph_mds_request *req);
extern int ceph_alloc_readdir_reply_buffer(struct ceph_mds_request *req,
					   struct inode *dir);
extern struct ceph_mds_request *
ceph_mdsc_create_request(struct ceph_mds_client *mdsc, int op, int mode);
extern int ceph_mdsc_submit_request(struct ceph_mds_client *mdsc,
				    struct inode *dir,
				    struct ceph_mds_request *req);
extern int ceph_mdsc_do_request(struct ceph_mds_client *mdsc,
				struct inode *dir,
				struct ceph_mds_request *req);
extern void ceph_mdsc_release_dir_caps(struct ceph_mds_request *req);
static inline void ceph_mdsc_get_request(struct ceph_mds_request *req)
{
	kref_get(&req->r_kref);
}
extern void ceph_mdsc_release_request(struct kref *kref);
static inline void ceph_mdsc_put_request(struct ceph_mds_request *req)
{
	kref_put(&req->r_kref, ceph_mdsc_release_request);
}

extern void __ceph_queue_cap_release(struct ceph_mds_session *session,
				    struct ceph_cap *cap);
extern void ceph_flush_cap_releases(struct ceph_mds_client *mdsc,
				    struct ceph_mds_session *session);
extern void ceph_queue_cap_reclaim_work(struct ceph_mds_client *mdsc);
extern void ceph_reclaim_caps_nr(struct ceph_mds_client *mdsc, int nr);
extern int ceph_iterate_session_caps(struct ceph_mds_session *session,
				     int (*cb)(struct inode *,
					       struct ceph_cap *, void *),
				     void *arg);
extern void ceph_mdsc_pre_umount(struct ceph_mds_client *mdsc);

static inline void ceph_mdsc_free_path(char *path, int len)
{
	if (path)
		__putname(path - (PATH_MAX - 1 - len));
}

extern char *ceph_mdsc_build_path(struct dentry *dentry, int *plen, u64 *base,
				  int stop_on_nosnap);

extern void __ceph_mdsc_drop_dentry_lease(struct dentry *dentry);
extern void ceph_mdsc_lease_send_msg(struct ceph_mds_session *session,
				     struct dentry *dentry, char action,
				     u32 seq);

extern void ceph_mdsc_handle_mdsmap(struct ceph_mds_client *mdsc,
				    struct ceph_msg *msg);
extern void ceph_mdsc_handle_fsmap(struct ceph_mds_client *mdsc,
				   struct ceph_msg *msg);

extern struct ceph_mds_session *
ceph_mdsc_open_export_target_session(struct ceph_mds_client *mdsc, int target);
extern void ceph_mdsc_open_export_target_sessions(struct ceph_mds_client *mdsc,
					  struct ceph_mds_session *session);

extern int ceph_trim_caps(struct ceph_mds_client *mdsc,
			  struct ceph_mds_session *session,
			  int max_caps);

static inline int ceph_wait_on_async_create(struct inode *inode)
{
	struct ceph_inode_info *ci = ceph_inode(inode);

	return wait_on_bit(&ci->i_ceph_flags, CEPH_ASYNC_CREATE_BIT,
			   TASK_INTERRUPTIBLE);
}

extern u64 ceph_get_deleg_ino(struct ceph_mds_session *session);
extern int ceph_restore_deleg_ino(struct ceph_mds_session *session, u64 ino);
#endif
Commit	Line	Data
b2441318	1	/* SPDX-License-Identifier: GPL-2.0 */
2f2dc053 SW	2	#ifndef _FS_CEPH_MDS_CLIENT_H
	3	#define _FS_CEPH_MDS_CLIENT_H
	4
	5	#include <linux/completion.h>
153c8e6b	6	#include <linux/kref.h>
2f2dc053 SW	7	#include <linux/list.h>
2f2dc053 SW	8	#include <linux/mutex.h>
44ca18f2	9	#include <linux/rbtree.h>
2f2dc053	10	#include <linux/spinlock.h>
3997c01d	11	#include <linux/refcount.h>
717e6f28	12	#include <linux/utsname.h>
2f2dc053	13
3d14c5d2 YS	14	#include <linux/ceph/types.h>
	15	#include <linux/ceph/messenger.h>
	16	#include <linux/ceph/mdsmap.h>
6c4a1915	17	#include <linux/ceph/auth.h>
2f2dc053	18
342ce182	19	/* The first 8 bits are reserved for old ceph releases */
9ba1e224 XL	20	enum ceph_feature_type {
	21	CEPHFS_FEATURE_MIMIC = 8,
	22	CEPHFS_FEATURE_REPLY_ENCODING,
	23	CEPHFS_FEATURE_RECLAIM_CLIENT,
	24	CEPHFS_FEATURE_LAZY_CAP_WANTED,
	25	CEPHFS_FEATURE_MULTI_RECONNECT,
d4846487	26	CEPHFS_FEATURE_DELEG_INO,
d2f8bb27	27
d4846487	28	CEPHFS_FEATURE_MAX = CEPHFS_FEATURE_DELEG_INO,
9ba1e224 XL	29	};
	30
	31	/*
	32	* This will always have the highest feature bit value
	33	* as the last element of the array.
	34	*/
	35	#define CEPHFS_FEATURES_CLIENT_SUPPORTED { \
d2f8bb27 YZ	36	0, 1, 2, 3, 4, 5, 6, 7, \
d2f8bb27 YZ	37	CEPHFS_FEATURE_MIMIC, \
b37fe1f9	38	CEPHFS_FEATURE_REPLY_ENCODING, \
d2f8bb27	39	CEPHFS_FEATURE_LAZY_CAP_WANTED, \
81c5a148	40	CEPHFS_FEATURE_MULTI_RECONNECT, \
d4846487	41	CEPHFS_FEATURE_DELEG_INO, \
9ba1e224 XL	42	\
9ba1e224 XL	43	CEPHFS_FEATURE_MAX, \
342ce182	44	}
342ce182 YZ	45	#define CEPHFS_FEATURES_CLIENT_REQUIRED {}
342ce182 YZ	46
2f2dc053 SW	47	/*
	48	* Some lock dependencies:
	49	*
	50	* session->s_mutex
	51	* mdsc->mutex
	52	*
	53	* mdsc->snap_rwsem
	54	*
be655596	55	* ci->i_ceph_lock
2f2dc053 SW	56	* mdsc->snap_flush_lock
	57	* mdsc->cap_delay_lock
	58	*
	59	*/
	60
3d14c5d2	61	struct ceph_fs_client;
2f2dc053 SW	62	struct ceph_cap;
	63
	64	/*
	65	* parsed info about a single inode. pointers are into the encoded
	66	* on-wire structures within the mds reply message payload.
	67	*/
	68	struct ceph_mds_reply_info_in {
	69	struct ceph_mds_reply_inode *in;
14303d20	70	struct ceph_dir_layout dir_layout;
2f2dc053 SW	71	u32 symlink_len;
	72	char *symlink;
	73	u32 xattr_len;
	74	char *xattr_data;
fb01d1f8 YZ	75	u64 inline_version;
	76	u32 inline_len;
	77	char *inline_data;
5ea5c5e0	78	u32 pool_ns_len;
779fe0fb	79	char *pool_ns_data;
fb18a575 LH	80	u64 max_bytes;
fb18a575 LH	81	u64 max_files;
08796873	82	s32 dir_pin;
245ce991	83	struct ceph_timespec btime;
193e7b37	84	struct ceph_timespec snap_btime;
a35ead31	85	u64 change_attr;
2f2dc053 SW	86	};
2f2dc053 SW	87
2a5beea3 YZ	88	struct ceph_mds_reply_dir_entry {
	89	char *name;
	90	u32 name_len;
	91	struct ceph_mds_reply_lease *lease;
	92	struct ceph_mds_reply_info_in inode;
8974eebd	93	loff_t offset;
2a5beea3 YZ	94	};
2a5beea3 YZ	95
2f2dc053	96	/*
25933abd HS	97	* parsed info about an mds reply, including information about
	98	* either: 1) the target inode and/or its parent directory and dentry,
	99	* and directory contents (for readdir results), or
	100	* 2) the file range lock info (for fcntl F_GETLK results).
2f2dc053 SW	101	*/
	102	struct ceph_mds_reply_info_parsed {
	103	struct ceph_mds_reply_head *head;
	104
25933abd	105	/* trace */
2f2dc053 SW	106	struct ceph_mds_reply_info_in diri, targeti;
	107	struct ceph_mds_reply_dirfrag *dirfrag;
	108	char *dname;
	109	u32 dname_len;
	110	struct ceph_mds_reply_lease *dlease;
	111
25933abd HS	112	/* extra */
	113	union {
	114	/* for fcntl F_GETLK results */
	115	struct ceph_filelock *filelock_reply;
	116
	117	/* for readdir results */
	118	struct {
	119	struct ceph_mds_reply_dirfrag *dir_dir;
54008399	120	size_t dir_buf_size;
25933abd	121	int dir_nr;
f3c4ebe6	122	bool dir_end;
79162547	123	bool dir_complete;
f3c4ebe6	124	bool hash_order;
79162547	125	bool offset_hash;
2a5beea3	126	struct ceph_mds_reply_dir_entry *dir_entries;
25933abd	127	};
6e8575fa SL	128
	129	/* for create results */
	130	struct {
	131	bool has_create_ino;
	132	u64 ino;
	133	};
25933abd	134	};
2f2dc053 SW	135
	136	/* encoded blob describing snapshot contexts for certain
	137	operations (e.g., open) */
	138	void *snapblob;
	139	int snapblob_len;
	140	};
	141
	142
	143	/*
	144	* cap releases are batched and sent to the MDS en masse.
92475f05 JL	145	*
	146	* Account for per-message overhead of mds_cap_release header
	147	* and __le32 for osd epoch barrier trailing field.
2f2dc053	148	*/
92475f05	149	#define CEPH_CAPS_PER_RELEASE ((PAGE_SIZE - sizeof(u32) - \
2f2dc053	150	sizeof(struct ceph_mds_cap_release)) / \
92475f05	151	sizeof(struct ceph_mds_cap_item))
2f2dc053 SW	152
	153
	154	/*
	155	* state associated with each MDS<->client session
	156	*/
	157	enum {
	158	CEPH_MDS_SESSION_NEW = 1,
	159	CEPH_MDS_SESSION_OPENING = 2,
	160	CEPH_MDS_SESSION_OPEN = 3,
	161	CEPH_MDS_SESSION_HUNG = 4,
7e6906c1 YZ	162	CEPH_MDS_SESSION_RESTARTING = 5,
	163	CEPH_MDS_SESSION_RECONNECTING = 6,
	164	CEPH_MDS_SESSION_CLOSING = 7,
4d681c2f XL	165	CEPH_MDS_SESSION_CLOSED = 8,
4d681c2f XL	166	CEPH_MDS_SESSION_REJECTED = 9,
2f2dc053 SW	167	};
	168
	169	struct ceph_mds_session {
	170	struct ceph_mds_client *s_mdsc;
	171	int s_mds;
	172	int s_state;
	173	unsigned long s_ttl; /* time until mds kills us */
84bf3950	174	unsigned long s_features;
2f2dc053 SW	175	u64 s_seq; /* incoming msg seq # */
	176	struct mutex s_mutex; /* serialize session messages */
	177
	178	struct ceph_connection s_con;
	179
6c4a1915	180	struct ceph_auth_handshake s_auth;
4e7a5dcd	181
d8fb02ab AE	182	/* protected by s_gen_ttl_lock */
d8fb02ab AE	183	spinlock_t s_gen_ttl_lock;
2f2dc053 SW	184	u32 s_cap_gen; /* inc each time we get mds stale msg */
2f2dc053 SW	185	unsigned long s_cap_ttl; /* when session caps expire */
d8fb02ab AE	186
	187	/* protected by s_cap_lock */
	188	spinlock_t s_cap_lock;
045100cd	189	refcount_t s_ref;
2f2dc053	190	struct list_head s_caps; /* all caps issued by this session */
e3ec8d68	191	struct ceph_cap *s_cap_iterator;
533a2818	192	int s_nr_caps;
2f2dc053	193	int s_num_cap_releases;
99a9c273	194	int s_cap_reconnect;
03f4fcb0	195	int s_readonly;
2f2dc053	196	struct list_head s_cap_releases; /* waiting cap_release messages */
e3ec8d68	197	struct work_struct s_cap_release_work;
2f2dc053 SW	198
	199	/* protected by mutex */
	200	struct list_head s_cap_flushing; /* inodes w/ flushing caps */
2f2dc053 SW	201	unsigned long s_renew_requested; /* last time we sent a renew req */
	202	u64 s_renew_seq;
	203
2f2dc053 SW	204	struct list_head s_waiting; /* waiting requests */
2f2dc053 SW	205	struct list_head s_unsafe; /* unsafe requests */
d4846487	206	struct xarray s_delegated_inos;
2f2dc053 SW	207	};
	208
	209	/*
	210	* modes of choosing which MDS to send a request to
	211	*/
	212	enum {
	213	USE_ANY_MDS,
	214	USE_RANDOM_MDS,
	215	USE_AUTH_MDS, /* prefer authoritative mds for this metadata item */
	216	};
	217
	218	struct ceph_mds_request;
	219	struct ceph_mds_client;
	220
	221	/*
	222	* request completion callback
	223	*/
	224	typedef void (ceph_mds_request_callback_t) (struct ceph_mds_client mdsc,
	225	struct ceph_mds_request *req);
9280be24 YZ	226	/*
	227	* wait for request completion callback
	228	*/
	229	typedef int (ceph_mds_request_wait_callback_t) (struct ceph_mds_client mdsc,
	230	struct ceph_mds_request *req);
2f2dc053 SW	231
	232	/*
	233	* an in-flight mds request
	234	*/
	235	struct ceph_mds_request {
	236	u64 r_tid; /* transaction id */
44ca18f2	237	struct rb_node r_node;
37151668	238	struct ceph_mds_client *r_mdsc;
2f2dc053	239
045100cd	240	struct kref r_kref;
2f2dc053	241	int r_op; /* mds op code */
2f2dc053 SW	242
	243	/* operation on what? */
	244	struct inode r_inode; / arg1 */
	245	struct dentry r_dentry; / arg1 */
	246	struct dentry r_old_dentry; / arg2: rename from or link from */
41b02e1f	247	struct inode r_old_dentry_dir; / arg2: old dentry's parent dir */
2f2dc053 SW	248	char r_path1, r_path2;
	249	struct ceph_vino r_ino1, r_ino2;
	250
3dd69aab	251	struct inode r_parent; / parent dir inode */
2f2dc053 SW	252	struct inode r_target_inode; / resulting inode */
2f2dc053 SW	253
bc2de10d JL	254	#define CEPH_MDS_R_DIRECT_IS_HASH (1) /* r_direct_hash is valid */
	255	#define CEPH_MDS_R_ABORTED (2) /* call was aborted */
	256	#define CEPH_MDS_R_GOT_UNSAFE (3) /* got an unsafe reply */
	257	#define CEPH_MDS_R_GOT_SAFE (4) /* got a safe reply */
	258	#define CEPH_MDS_R_GOT_RESULT (5) /* got a result */
	259	#define CEPH_MDS_R_DID_PREPOPULATE (6) /* prepopulated readdir */
3dd69aab	260	#define CEPH_MDS_R_PARENT_LOCKED (7) /* is r_parent->i_rwsem wlocked? */
3bb48b41	261	#define CEPH_MDS_R_ASYNC (8) /* async request */
bc2de10d JL	262	unsigned long r_req_flags;
bc2de10d JL	263
b4556396 SW	264	struct mutex r_fill_mutex;
b4556396 SW	265
2f2dc053 SW	266	union ceph_mds_request_args r_args;
2f2dc053 SW	267	int r_fmode; /* file mode, if expecting cap */
ff3d0046 EB	268	kuid_t r_uid;
ff3d0046 EB	269	kgid_t r_gid;
c36d6414	270	int r_request_release_offset;
0ed1e90a	271	struct timespec64 r_stamp;
2f2dc053 SW	272
	273	/* for choosing which mds to send this request to */
	274	int r_direct_mode;
	275	u32 r_direct_hash; /* choose dir frag based on this dentry hash */
2f2dc053 SW	276
2f2dc053 SW	277	/* data payload is used for xattr ops */
25e6bae3	278	struct ceph_pagelist *r_pagelist;
2f2dc053 SW	279
	280	/* what caps shall we drop? */
	281	int r_inode_drop, r_inode_unless;
	282	int r_dentry_drop, r_dentry_unless;
	283	int r_old_dentry_drop, r_old_dentry_unless;
	284	struct inode *r_old_inode;
	285	int r_old_inode_drop, r_old_inode_unless;
	286
	287	struct ceph_msg r_request; / original request */
	288	struct ceph_msg *r_reply;
	289	struct ceph_mds_reply_info_parsed r_reply_info;
	290	int r_err;
a25949b9 JL	291
	292
	293	struct page *r_locked_page;
	294	int r_dir_caps;
c36d6414 JL	295	int r_num_caps;
c36d6414 JL	296	u32 r_readdir_offset;
2f2dc053	297
a319bf56	298	unsigned long r_timeout; /* optional. jiffies, 0 is "wait forever" */
2f2dc053 SW	299	unsigned long r_started; /* start time to measure timeout against */
	300	unsigned long r_request_started; /* start time for mds request only,
	301	used to measure lease durations */
	302
	303	/* link unsafe requests to parent directory, for fsync */
	304	struct inode *r_unsafe_dir;
	305	struct list_head r_unsafe_dir_item;
	306
68cd5b4b YZ	307	/* unsafe requests that modify the target inode */
	308	struct list_head r_unsafe_target_item;
	309
2f2dc053 SW	310	struct ceph_mds_session *r_session;
	311
	312	int r_attempts; /* resend attempts */
	313	int r_num_fwd; /* number of forward attempts */
2f2dc053	314	int r_resend_mds; /* mds to resend to next, if any*/
e55b71f8	315	u32 r_sent_on_mseq; /* cap mseq request was sent at*/
6deb8008	316	u64 r_deleg_ino;
2f2dc053	317
2f2dc053 SW	318	struct list_head r_wait;
	319	struct completion r_completion;
	320	struct completion r_safe_completion;
	321	ceph_mds_request_callback_t r_callback;
9280be24	322	ceph_mds_request_wait_callback_t r_wait_for_completion;
2f2dc053	323	struct list_head r_unsafe_item; /* per-session unsafe list item */
2f2dc053	324
fdd4e158 YZ	325	long long r_dir_release_cnt;
	326	long long r_dir_ordered_cnt;
	327	int r_readdir_cache_idx;
2f2dc053 SW	328
2f2dc053 SW	329	struct ceph_cap_reservation r_caps_reservation;
2f2dc053 SW	330	};
2f2dc053 SW	331
10183a69 YZ	332	struct ceph_pool_perm {
10183a69 YZ	333	struct rb_node node;
10183a69	334	int perm;
7627151e	335	s64 pool;
779fe0fb YZ	336	size_t pool_ns_len;
779fe0fb YZ	337	char pool_ns[];
10183a69 YZ	338	};
10183a69 YZ	339
75c9627e YZ	340	struct ceph_snapid_map {
	341	struct rb_node node;
	342	struct list_head lru;
	343	atomic_t ref;
	344	u64 snap;
	345	dev_t dev;
	346	unsigned long last_used;
	347	};
	348
0c44a8e0 LH	349	/*
	350	* node for list of quotarealm inodes that are not visible from the filesystem
	351	* mountpoint, but required to handle, e.g. quotas.
	352	*/
	353	struct ceph_quotarealm_inode {
	354	struct rb_node node;
	355	u64 ino;
	356	unsigned long timeout; /* last time a lookup failed for this inode */
	357	struct mutex mutex;
	358	struct inode *inode;
	359	};
	360
3a3430af JL	361	struct cap_wait {
	362	struct list_head list;
	363	unsigned long ino;
	364	pid_t tgid;
	365	int need;
	366	int want;
	367	};
	368
2f2dc053 SW	369	/*
	370	* mds client state
	371	*/
	372	struct ceph_mds_client {
3d14c5d2	373	struct ceph_fs_client *fsc;
2f2dc053 SW	374	struct mutex mutex; /* all nested structures */
	375
	376	struct ceph_mdsmap *mdsmap;
f3c60c59 SW	377	struct completion safe_umount_waiters;
f3c60c59 SW	378	wait_queue_head_t session_close_wq;
2f2dc053	379	struct list_head waiting_for_map;
430afbad	380	int mdsmap_err;
2f2dc053 SW	381
2f2dc053 SW	382	struct ceph_mds_session *sessions; / NULL for mds if no session */
86d8f67b	383	atomic_t num_sessions;
2f2dc053 SW	384	int max_sessions; /* len of s_mds_sessions */
	385	int stopping; /* true if shutting down */
	386
d557c48d	387	atomic64_t quotarealms_count; /* # realms with quota */
0c44a8e0 LH	388	/*
	389	* We keep a list of inodes we don't see in the mountpoint but that we
	390	* need to track quota realms.
	391	*/
	392	struct rb_root quotarealms_inodes;
	393	struct mutex quotarealms_inodes_mutex;
d557c48d	394
2f2dc053 SW	395	/*
	396	* snap_rwsem will cover cap linkage into snaprealms, and
	397	* realm snap contexts. (later, we can do per-realm snap
	398	* contexts locks..) the empty list contains realms with no
	399	* references (implying they contain no inodes with caps) that
	400	* should be destroyed.
	401	*/
affbc19a	402	u64 last_snap_seq;
2f2dc053	403	struct rw_semaphore snap_rwsem;
a105f00c	404	struct rb_root snap_realms;
2f2dc053	405	struct list_head snap_empty;
81c5a148	406	int num_snap_realms;
2f2dc053 SW	407	spinlock_t snap_empty_lock; /* protect snap_empty */
	408
	409	u64 last_tid; /* most recent mds request */
e8a7b8b1 YZ	410	u64 oldest_tid; /* oldest incomplete mds request,
e8a7b8b1 YZ	411	excluding setfilelock requests */
44ca18f2	412	struct rb_root request_tree; /* pending mds requests */
2f2dc053 SW	413	struct delayed_work delayed_work; /* delayed work */
	414	unsigned long last_renew_caps; /* last time we renewed our caps */
	415	struct list_head cap_delay_list; /* caps with delayed release */
	416	spinlock_t cap_delay_lock; /* protects cap_delay_list */
	417	struct list_head snap_flush_list; /* cap_snaps ready to flush */
	418	spinlock_t snap_flush_lock;
	419
553adfd9	420	u64 last_cap_flush_tid;
e4500b5e	421	struct list_head cap_flush_list;
2f2dc053	422	struct list_head cap_dirty; /* inodes with dirty caps */
db354052	423	struct list_head cap_dirty_migrating; /* ...that are migration... */
2f2dc053 SW	424	int num_cap_flushing; /* # caps we are flushing */
	425	spinlock_t cap_dirty_lock; /* protects above items */
	426	wait_queue_head_t cap_flushing_wq;
	427
37c4efc1	428	struct work_struct cap_reclaim_work;
fe33032d	429	atomic_t cap_reclaim_pending;
37c4efc1	430
37151668 YS	431	/*
	432	* Cap reservations
	433	*
	434	* Maintain a global pool of preallocated struct ceph_caps, referenced
	435	* by struct ceph_caps_reservations. This ensures that we preallocate
	436	* memory needed to successfully process an MDS response. (If an MDS
	437	* sends us cap information and we fail to process it, we will have
	438	* problems due to the client and MDS being out of sync.)
	439	*
	440	* Reservations are 'owned' by a ceph_cap_reservation context.
	441	*/
	442	spinlock_t caps_list_lock;
	443	struct list_head caps_list; /* unused (reserved or
	444	unreserved) */
3a3430af	445	struct list_head cap_wait_list;
37151668 YS	446	int caps_total_count; /* total caps allocated */
37151668 YS	447	int caps_use_count; /* in use */
fe33032d	448	int caps_use_max; /* max used caps */
37151668 YS	449	int caps_reserve_count; /* unused, reserved */
	450	int caps_avail_count; /* unused, unreserved */
	451	int caps_min_count; /* keep at least this many
	452	(unreserved) */
37c4efc1 YZ	453	spinlock_t dentry_list_lock;
	454	struct list_head dentry_leases; /* fifo list */
	455	struct list_head dentry_dir_leases; /* lru list */
10183a69	456
75c9627e YZ	457	spinlock_t snapid_map_lock;
	458	struct rb_root snapid_map_tree;
	459	struct list_head snapid_map_lru;
	460
10183a69 YZ	461	struct rw_semaphore pool_perm_rwsem;
10183a69 YZ	462	struct rb_root pool_perm_tree;
717e6f28 YZ	463
717e6f28 YZ	464	char nodename[__NEW_UTS_LEN + 1];
2f2dc053 SW	465	};
	466
	467	extern const char *ceph_mds_op_name(int op);
	468
	469	extern struct ceph_mds_session *
	470	__ceph_lookup_mds_session(struct ceph_mds_client *, int mds);
	471
a687ecaf JS	472	extern const char *ceph_session_state_name(int s);
a687ecaf JS	473
5b3248c6 XL	474	extern struct ceph_mds_session *
5b3248c6 XL	475	ceph_get_mds_session(struct ceph_mds_session *s);
2f2dc053 SW	476	extern void ceph_put_mds_session(struct ceph_mds_session *s);
	477
	478	extern int ceph_send_msg_mds(struct ceph_mds_client *mdsc,
	479	struct ceph_msg *msg, int mds);
	480
3d14c5d2	481	extern int ceph_mdsc_init(struct ceph_fs_client *fsc);
2f2dc053	482	extern void ceph_mdsc_close_sessions(struct ceph_mds_client *mdsc);
48fec5d0	483	extern void ceph_mdsc_force_umount(struct ceph_mds_client *mdsc);
3d14c5d2	484	extern void ceph_mdsc_destroy(struct ceph_fs_client *fsc);
2f2dc053 SW	485
	486	extern void ceph_mdsc_sync(struct ceph_mds_client *mdsc);
	487
167c9e35	488	extern void ceph_invalidate_dir_request(struct ceph_mds_request *req);
54008399 YZ	489	extern int ceph_alloc_readdir_reply_buffer(struct ceph_mds_request *req,
54008399 YZ	490	struct inode *dir);
2f2dc053 SW	491	extern struct ceph_mds_request *
2f2dc053 SW	492	ceph_mdsc_create_request(struct ceph_mds_client *mdsc, int op, int mode);
86bda539 JL	493	extern int ceph_mdsc_submit_request(struct ceph_mds_client *mdsc,
	494	struct inode *dir,
	495	struct ceph_mds_request *req);
2f2dc053 SW	496	extern int ceph_mdsc_do_request(struct ceph_mds_client *mdsc,
	497	struct inode *dir,
	498	struct ceph_mds_request *req);
a25949b9	499	extern void ceph_mdsc_release_dir_caps(struct ceph_mds_request *req);
2f2dc053 SW	500	static inline void ceph_mdsc_get_request(struct ceph_mds_request *req)
2f2dc053 SW	501	{
153c8e6b SW	502	kref_get(&req->r_kref);
	503	}
	504	extern void ceph_mdsc_release_request(struct kref *kref);
	505	static inline void ceph_mdsc_put_request(struct ceph_mds_request *req)
	506	{
	507	kref_put(&req->r_kref, ceph_mdsc_release_request);
2f2dc053	508	}
2f2dc053	509
e3ec8d68 YZ	510	extern void __ceph_queue_cap_release(struct ceph_mds_session *session,
	511	struct ceph_cap *cap);
	512	extern void ceph_flush_cap_releases(struct ceph_mds_client *mdsc,
	513	struct ceph_mds_session *session);
37c4efc1	514	extern void ceph_queue_cap_reclaim_work(struct ceph_mds_client *mdsc);
fe33032d	515	extern void ceph_reclaim_caps_nr(struct ceph_mds_client *mdsc, int nr);
f5d77269 JL	516	extern int ceph_iterate_session_caps(struct ceph_mds_session *session,
	517	int (cb)(struct inode ,
	518	struct ceph_cap , void ),
	519	void *arg);
2f2dc053 SW	520	extern void ceph_mdsc_pre_umount(struct ceph_mds_client *mdsc);
2f2dc053 SW	521
f77f21bb JL	522	static inline void ceph_mdsc_free_path(char *path, int len)
	523	{
	524	if (path)
	525	__putname(path - (PATH_MAX - 1 - len));
	526	}
	527
2f2dc053 SW	528	extern char ceph_mdsc_build_path(struct dentry dentry, int plen, u64 base,
	529	int stop_on_nosnap);
	530
	531	extern void __ceph_mdsc_drop_dentry_lease(struct dentry *dentry);
	532	extern void ceph_mdsc_lease_send_msg(struct ceph_mds_session *session,
2f2dc053 SW	533	struct dentry *dentry, char action,
	534	u32 seq);
	535
430afbad YZ	536	extern void ceph_mdsc_handle_mdsmap(struct ceph_mds_client *mdsc,
	537	struct ceph_msg *msg);
	538	extern void ceph_mdsc_handle_fsmap(struct ceph_mds_client *mdsc,
	539	struct ceph_msg *msg);
2f2dc053	540
5d72d13c YZ	541	extern struct ceph_mds_session *
5d72d13c YZ	542	ceph_mdsc_open_export_target_session(struct ceph_mds_client *mdsc, int target);
154f42c2 SW	543	extern void ceph_mdsc_open_export_target_sessions(struct ceph_mds_client *mdsc,
	544	struct ceph_mds_session *session);
	545
e30ee581 ZZ	546	extern int ceph_trim_caps(struct ceph_mds_client *mdsc,
	547	struct ceph_mds_session *session,
	548	int max_caps);
d4846487	549
891f3f5a JL	550	static inline int ceph_wait_on_async_create(struct inode *inode)
	551	{
	552	struct ceph_inode_info *ci = ceph_inode(inode);
	553
	554	return wait_on_bit(&ci->i_ceph_flags, CEPH_ASYNC_CREATE_BIT,
	555	TASK_INTERRUPTIBLE);
	556	}
d4846487 JL	557
	558	extern u64 ceph_get_deleg_ino(struct ceph_mds_session *session);
	559	extern int ceph_restore_deleg_ino(struct ceph_mds_session *session, u64 ino);
2f2dc053	560	#endif