[linux-2.6-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/ktime.h>

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

#include "metric.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;

	/* both protected by s_mdsc->cap_dirty_lock */
	struct list_head  s_cap_dirty;	      /* inodes w/ dirty caps */
	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_start_latency;  /* start time to measure latency */
	unsigned long r_end_latency;    /* finish time to measure latency */
	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_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 */

	struct ceph_client_metric metric;

	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 (!IS_ERR_OR_NULL(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>
70c94820	13	#include <linux/ktime.h>
2f2dc053	14
3d14c5d2 YS	15	#include <linux/ceph/types.h>
	16	#include <linux/ceph/messenger.h>
	17	#include <linux/ceph/mdsmap.h>
6c4a1915	18	#include <linux/ceph/auth.h>
2f2dc053	19
f9009efa XL	20	#include "metric.h"
f9009efa XL	21
342ce182	22	/* The first 8 bits are reserved for old ceph releases */
9ba1e224 XL	23	enum ceph_feature_type {
	24	CEPHFS_FEATURE_MIMIC = 8,
	25	CEPHFS_FEATURE_REPLY_ENCODING,
	26	CEPHFS_FEATURE_RECLAIM_CLIENT,
	27	CEPHFS_FEATURE_LAZY_CAP_WANTED,
	28	CEPHFS_FEATURE_MULTI_RECONNECT,
d4846487	29	CEPHFS_FEATURE_DELEG_INO,
d2f8bb27	30
d4846487	31	CEPHFS_FEATURE_MAX = CEPHFS_FEATURE_DELEG_INO,
9ba1e224 XL	32	};
	33
	34	/*
	35	* This will always have the highest feature bit value
	36	* as the last element of the array.
	37	*/
	38	#define CEPHFS_FEATURES_CLIENT_SUPPORTED { \
d2f8bb27 YZ	39	0, 1, 2, 3, 4, 5, 6, 7, \
d2f8bb27 YZ	40	CEPHFS_FEATURE_MIMIC, \
b37fe1f9	41	CEPHFS_FEATURE_REPLY_ENCODING, \
d2f8bb27	42	CEPHFS_FEATURE_LAZY_CAP_WANTED, \
81c5a148	43	CEPHFS_FEATURE_MULTI_RECONNECT, \
d4846487	44	CEPHFS_FEATURE_DELEG_INO, \
9ba1e224 XL	45	\
9ba1e224 XL	46	CEPHFS_FEATURE_MAX, \
342ce182	47	}
342ce182 YZ	48	#define CEPHFS_FEATURES_CLIENT_REQUIRED {}
342ce182 YZ	49
2f2dc053 SW	50	/*
	51	* Some lock dependencies:
	52	*
	53	* session->s_mutex
	54	* mdsc->mutex
	55	*
	56	* mdsc->snap_rwsem
	57	*
be655596	58	* ci->i_ceph_lock
2f2dc053 SW	59	* mdsc->snap_flush_lock
	60	* mdsc->cap_delay_lock
	61	*
	62	*/
	63
3d14c5d2	64	struct ceph_fs_client;
2f2dc053 SW	65	struct ceph_cap;
	66
	67	/*
	68	* parsed info about a single inode. pointers are into the encoded
	69	* on-wire structures within the mds reply message payload.
	70	*/
	71	struct ceph_mds_reply_info_in {
	72	struct ceph_mds_reply_inode *in;
14303d20	73	struct ceph_dir_layout dir_layout;
2f2dc053 SW	74	u32 symlink_len;
	75	char *symlink;
	76	u32 xattr_len;
	77	char *xattr_data;
fb01d1f8 YZ	78	u64 inline_version;
	79	u32 inline_len;
	80	char *inline_data;
5ea5c5e0	81	u32 pool_ns_len;
779fe0fb	82	char *pool_ns_data;
fb18a575 LH	83	u64 max_bytes;
fb18a575 LH	84	u64 max_files;
08796873	85	s32 dir_pin;
245ce991	86	struct ceph_timespec btime;
193e7b37	87	struct ceph_timespec snap_btime;
a35ead31	88	u64 change_attr;
2f2dc053 SW	89	};
2f2dc053 SW	90
2a5beea3 YZ	91	struct ceph_mds_reply_dir_entry {
	92	char *name;
	93	u32 name_len;
	94	struct ceph_mds_reply_lease *lease;
	95	struct ceph_mds_reply_info_in inode;
8974eebd	96	loff_t offset;
2a5beea3 YZ	97	};
2a5beea3 YZ	98
2f2dc053	99	/*
25933abd HS	100	* parsed info about an mds reply, including information about
	101	* either: 1) the target inode and/or its parent directory and dentry,
	102	* and directory contents (for readdir results), or
	103	* 2) the file range lock info (for fcntl F_GETLK results).
2f2dc053 SW	104	*/
	105	struct ceph_mds_reply_info_parsed {
	106	struct ceph_mds_reply_head *head;
	107
25933abd	108	/* trace */
2f2dc053 SW	109	struct ceph_mds_reply_info_in diri, targeti;
	110	struct ceph_mds_reply_dirfrag *dirfrag;
	111	char *dname;
	112	u32 dname_len;
	113	struct ceph_mds_reply_lease *dlease;
	114
25933abd HS	115	/* extra */
	116	union {
	117	/* for fcntl F_GETLK results */
	118	struct ceph_filelock *filelock_reply;
	119
	120	/* for readdir results */
	121	struct {
	122	struct ceph_mds_reply_dirfrag *dir_dir;
54008399	123	size_t dir_buf_size;
25933abd	124	int dir_nr;
f3c4ebe6	125	bool dir_end;
79162547	126	bool dir_complete;
f3c4ebe6	127	bool hash_order;
79162547	128	bool offset_hash;
2a5beea3	129	struct ceph_mds_reply_dir_entry *dir_entries;
25933abd	130	};
6e8575fa SL	131
	132	/* for create results */
	133	struct {
	134	bool has_create_ino;
	135	u64 ino;
	136	};
25933abd	137	};
2f2dc053 SW	138
	139	/* encoded blob describing snapshot contexts for certain
	140	operations (e.g., open) */
	141	void *snapblob;
	142	int snapblob_len;
	143	};
	144
	145
	146	/*
	147	* cap releases are batched and sent to the MDS en masse.
92475f05 JL	148	*
	149	* Account for per-message overhead of mds_cap_release header
	150	* and __le32 for osd epoch barrier trailing field.
2f2dc053	151	*/
92475f05	152	#define CEPH_CAPS_PER_RELEASE ((PAGE_SIZE - sizeof(u32) - \
2f2dc053	153	sizeof(struct ceph_mds_cap_release)) / \
92475f05	154	sizeof(struct ceph_mds_cap_item))
2f2dc053 SW	155
	156
	157	/*
	158	* state associated with each MDS<->client session
	159	*/
	160	enum {
	161	CEPH_MDS_SESSION_NEW = 1,
	162	CEPH_MDS_SESSION_OPENING = 2,
	163	CEPH_MDS_SESSION_OPEN = 3,
	164	CEPH_MDS_SESSION_HUNG = 4,
7e6906c1 YZ	165	CEPH_MDS_SESSION_RESTARTING = 5,
	166	CEPH_MDS_SESSION_RECONNECTING = 6,
	167	CEPH_MDS_SESSION_CLOSING = 7,
4d681c2f XL	168	CEPH_MDS_SESSION_CLOSED = 8,
4d681c2f XL	169	CEPH_MDS_SESSION_REJECTED = 9,
2f2dc053 SW	170	};
	171
	172	struct ceph_mds_session {
	173	struct ceph_mds_client *s_mdsc;
	174	int s_mds;
	175	int s_state;
	176	unsigned long s_ttl; /* time until mds kills us */
84bf3950	177	unsigned long s_features;
2f2dc053 SW	178	u64 s_seq; /* incoming msg seq # */
	179	struct mutex s_mutex; /* serialize session messages */
	180
	181	struct ceph_connection s_con;
	182
6c4a1915	183	struct ceph_auth_handshake s_auth;
4e7a5dcd	184
d8fb02ab AE	185	/* protected by s_gen_ttl_lock */
d8fb02ab AE	186	spinlock_t s_gen_ttl_lock;
2f2dc053 SW	187	u32 s_cap_gen; /* inc each time we get mds stale msg */
2f2dc053 SW	188	unsigned long s_cap_ttl; /* when session caps expire */
d8fb02ab AE	189
	190	/* protected by s_cap_lock */
	191	spinlock_t s_cap_lock;
045100cd	192	refcount_t s_ref;
2f2dc053	193	struct list_head s_caps; /* all caps issued by this session */
e3ec8d68	194	struct ceph_cap *s_cap_iterator;
533a2818	195	int s_nr_caps;
2f2dc053	196	int s_num_cap_releases;
99a9c273	197	int s_cap_reconnect;
03f4fcb0	198	int s_readonly;
2f2dc053	199	struct list_head s_cap_releases; /* waiting cap_release messages */
e3ec8d68	200	struct work_struct s_cap_release_work;
2f2dc053	201
1cf03a68 JL	202	/* both protected by s_mdsc->cap_dirty_lock */
1cf03a68 JL	203	struct list_head s_cap_dirty; /* inodes w/ dirty caps */
2f2dc053	204	struct list_head s_cap_flushing; /* inodes w/ flushing caps */
1cf03a68	205
2f2dc053 SW	206	unsigned long s_renew_requested; /* last time we sent a renew req */
	207	u64 s_renew_seq;
	208
2f2dc053 SW	209	struct list_head s_waiting; /* waiting requests */
2f2dc053 SW	210	struct list_head s_unsafe; /* unsafe requests */
d4846487	211	struct xarray s_delegated_inos;
2f2dc053 SW	212	};
	213
	214	/*
	215	* modes of choosing which MDS to send a request to
	216	*/
	217	enum {
	218	USE_ANY_MDS,
	219	USE_RANDOM_MDS,
	220	USE_AUTH_MDS, /* prefer authoritative mds for this metadata item */
	221	};
	222
	223	struct ceph_mds_request;
	224	struct ceph_mds_client;
	225
	226	/*
	227	* request completion callback
	228	*/
	229	typedef void (ceph_mds_request_callback_t) (struct ceph_mds_client mdsc,
	230	struct ceph_mds_request *req);
9280be24 YZ	231	/*
	232	* wait for request completion callback
	233	*/
	234	typedef int (ceph_mds_request_wait_callback_t) (struct ceph_mds_client mdsc,
	235	struct ceph_mds_request *req);
2f2dc053 SW	236
	237	/*
	238	* an in-flight mds request
	239	*/
	240	struct ceph_mds_request {
	241	u64 r_tid; /* transaction id */
44ca18f2	242	struct rb_node r_node;
37151668	243	struct ceph_mds_client *r_mdsc;
2f2dc053	244
045100cd	245	struct kref r_kref;
2f2dc053	246	int r_op; /* mds op code */
2f2dc053 SW	247
	248	/* operation on what? */
	249	struct inode r_inode; / arg1 */
	250	struct dentry r_dentry; / arg1 */
	251	struct dentry r_old_dentry; / arg2: rename from or link from */
41b02e1f	252	struct inode r_old_dentry_dir; / arg2: old dentry's parent dir */
2f2dc053 SW	253	char r_path1, r_path2;
	254	struct ceph_vino r_ino1, r_ino2;
	255
3dd69aab	256	struct inode r_parent; / parent dir inode */
2f2dc053 SW	257	struct inode r_target_inode; / resulting inode */
2f2dc053 SW	258
bc2de10d JL	259	#define CEPH_MDS_R_DIRECT_IS_HASH (1) /* r_direct_hash is valid */
	260	#define CEPH_MDS_R_ABORTED (2) /* call was aborted */
	261	#define CEPH_MDS_R_GOT_UNSAFE (3) /* got an unsafe reply */
	262	#define CEPH_MDS_R_GOT_SAFE (4) /* got a safe reply */
	263	#define CEPH_MDS_R_GOT_RESULT (5) /* got a result */
	264	#define CEPH_MDS_R_DID_PREPOPULATE (6) /* prepopulated readdir */
3dd69aab	265	#define CEPH_MDS_R_PARENT_LOCKED (7) /* is r_parent->i_rwsem wlocked? */
3bb48b41	266	#define CEPH_MDS_R_ASYNC (8) /* async request */
bc2de10d JL	267	unsigned long r_req_flags;
bc2de10d JL	268
b4556396 SW	269	struct mutex r_fill_mutex;
b4556396 SW	270
2f2dc053 SW	271	union ceph_mds_request_args r_args;
2f2dc053 SW	272	int r_fmode; /* file mode, if expecting cap */
ff3d0046 EB	273	kuid_t r_uid;
ff3d0046 EB	274	kgid_t r_gid;
c36d6414	275	int r_request_release_offset;
0ed1e90a	276	struct timespec64 r_stamp;
2f2dc053 SW	277
	278	/* for choosing which mds to send this request to */
	279	int r_direct_mode;
	280	u32 r_direct_hash; /* choose dir frag based on this dentry hash */
2f2dc053 SW	281
2f2dc053 SW	282	/* data payload is used for xattr ops */
25e6bae3	283	struct ceph_pagelist *r_pagelist;
2f2dc053 SW	284
	285	/* what caps shall we drop? */
	286	int r_inode_drop, r_inode_unless;
	287	int r_dentry_drop, r_dentry_unless;
	288	int r_old_dentry_drop, r_old_dentry_unless;
	289	struct inode *r_old_inode;
	290	int r_old_inode_drop, r_old_inode_unless;
	291
	292	struct ceph_msg r_request; / original request */
	293	struct ceph_msg *r_reply;
	294	struct ceph_mds_reply_info_parsed r_reply_info;
	295	int r_err;
a25949b9 JL	296
	297
	298	struct page *r_locked_page;
	299	int r_dir_caps;
c36d6414 JL	300	int r_num_caps;
c36d6414 JL	301	u32 r_readdir_offset;
2f2dc053	302
a319bf56	303	unsigned long r_timeout; /* optional. jiffies, 0 is "wait forever" */
2f2dc053	304	unsigned long r_started; /* start time to measure timeout against */
70c94820 XL	305	unsigned long r_start_latency; /* start time to measure latency */
70c94820 XL	306	unsigned long r_end_latency; /* finish time to measure latency */
2f2dc053 SW	307	unsigned long r_request_started; /* start time for mds request only,
	308	used to measure lease durations */
	309
	310	/* link unsafe requests to parent directory, for fsync */
	311	struct inode *r_unsafe_dir;
	312	struct list_head r_unsafe_dir_item;
	313
68cd5b4b YZ	314	/* unsafe requests that modify the target inode */
	315	struct list_head r_unsafe_target_item;
	316
2f2dc053 SW	317	struct ceph_mds_session *r_session;
	318
	319	int r_attempts; /* resend attempts */
	320	int r_num_fwd; /* number of forward attempts */
2f2dc053	321	int r_resend_mds; /* mds to resend to next, if any*/
e55b71f8	322	u32 r_sent_on_mseq; /* cap mseq request was sent at*/
6deb8008	323	u64 r_deleg_ino;
2f2dc053	324
2f2dc053 SW	325	struct list_head r_wait;
	326	struct completion r_completion;
	327	struct completion r_safe_completion;
	328	ceph_mds_request_callback_t r_callback;
9280be24	329	ceph_mds_request_wait_callback_t r_wait_for_completion;
2f2dc053	330	struct list_head r_unsafe_item; /* per-session unsafe list item */
2f2dc053	331
fdd4e158 YZ	332	long long r_dir_release_cnt;
	333	long long r_dir_ordered_cnt;
	334	int r_readdir_cache_idx;
2f2dc053 SW	335
2f2dc053 SW	336	struct ceph_cap_reservation r_caps_reservation;
2f2dc053 SW	337	};
2f2dc053 SW	338
10183a69 YZ	339	struct ceph_pool_perm {
10183a69 YZ	340	struct rb_node node;
10183a69	341	int perm;
7627151e	342	s64 pool;
779fe0fb YZ	343	size_t pool_ns_len;
779fe0fb YZ	344	char pool_ns[];
10183a69 YZ	345	};
10183a69 YZ	346
75c9627e YZ	347	struct ceph_snapid_map {
	348	struct rb_node node;
	349	struct list_head lru;
	350	atomic_t ref;
	351	u64 snap;
	352	dev_t dev;
	353	unsigned long last_used;
	354	};
	355
0c44a8e0 LH	356	/*
	357	* node for list of quotarealm inodes that are not visible from the filesystem
	358	* mountpoint, but required to handle, e.g. quotas.
	359	*/
	360	struct ceph_quotarealm_inode {
	361	struct rb_node node;
	362	u64 ino;
	363	unsigned long timeout; /* last time a lookup failed for this inode */
	364	struct mutex mutex;
	365	struct inode *inode;
	366	};
	367
3a3430af JL	368	struct cap_wait {
	369	struct list_head list;
	370	unsigned long ino;
	371	pid_t tgid;
	372	int need;
	373	int want;
	374	};
	375
2f2dc053 SW	376	/*
	377	* mds client state
	378	*/
	379	struct ceph_mds_client {
3d14c5d2	380	struct ceph_fs_client *fsc;
2f2dc053 SW	381	struct mutex mutex; /* all nested structures */
	382
	383	struct ceph_mdsmap *mdsmap;
f3c60c59 SW	384	struct completion safe_umount_waiters;
f3c60c59 SW	385	wait_queue_head_t session_close_wq;
2f2dc053	386	struct list_head waiting_for_map;
430afbad	387	int mdsmap_err;
2f2dc053 SW	388
2f2dc053 SW	389	struct ceph_mds_session *sessions; / NULL for mds if no session */
86d8f67b	390	atomic_t num_sessions;
2f2dc053 SW	391	int max_sessions; /* len of s_mds_sessions */
	392	int stopping; /* true if shutting down */
	393
d557c48d	394	atomic64_t quotarealms_count; /* # realms with quota */
0c44a8e0 LH	395	/*
	396	* We keep a list of inodes we don't see in the mountpoint but that we
	397	* need to track quota realms.
	398	*/
	399	struct rb_root quotarealms_inodes;
	400	struct mutex quotarealms_inodes_mutex;
d557c48d	401
2f2dc053 SW	402	/*
	403	* snap_rwsem will cover cap linkage into snaprealms, and
	404	* realm snap contexts. (later, we can do per-realm snap
	405	* contexts locks..) the empty list contains realms with no
	406	* references (implying they contain no inodes with caps) that
	407	* should be destroyed.
	408	*/
affbc19a	409	u64 last_snap_seq;
2f2dc053	410	struct rw_semaphore snap_rwsem;
a105f00c	411	struct rb_root snap_realms;
2f2dc053	412	struct list_head snap_empty;
81c5a148	413	int num_snap_realms;
2f2dc053 SW	414	spinlock_t snap_empty_lock; /* protect snap_empty */
	415
	416	u64 last_tid; /* most recent mds request */
e8a7b8b1 YZ	417	u64 oldest_tid; /* oldest incomplete mds request,
e8a7b8b1 YZ	418	excluding setfilelock requests */
44ca18f2	419	struct rb_root request_tree; /* pending mds requests */
2f2dc053 SW	420	struct delayed_work delayed_work; /* delayed work */
	421	unsigned long last_renew_caps; /* last time we renewed our caps */
	422	struct list_head cap_delay_list; /* caps with delayed release */
	423	spinlock_t cap_delay_lock; /* protects cap_delay_list */
	424	struct list_head snap_flush_list; /* cap_snaps ready to flush */
	425	spinlock_t snap_flush_lock;
	426
553adfd9	427	u64 last_cap_flush_tid;
e4500b5e	428	struct list_head cap_flush_list;
db354052	429	struct list_head cap_dirty_migrating; /* ...that are migration... */
2f2dc053 SW	430	int num_cap_flushing; /* # caps we are flushing */
	431	spinlock_t cap_dirty_lock; /* protects above items */
	432	wait_queue_head_t cap_flushing_wq;
	433
37c4efc1	434	struct work_struct cap_reclaim_work;
fe33032d	435	atomic_t cap_reclaim_pending;
37c4efc1	436
37151668 YS	437	/*
	438	* Cap reservations
	439	*
	440	* Maintain a global pool of preallocated struct ceph_caps, referenced
	441	* by struct ceph_caps_reservations. This ensures that we preallocate
	442	* memory needed to successfully process an MDS response. (If an MDS
	443	* sends us cap information and we fail to process it, we will have
	444	* problems due to the client and MDS being out of sync.)
	445	*
	446	* Reservations are 'owned' by a ceph_cap_reservation context.
	447	*/
	448	spinlock_t caps_list_lock;
	449	struct list_head caps_list; /* unused (reserved or
	450	unreserved) */
3a3430af	451	struct list_head cap_wait_list;
37151668 YS	452	int caps_total_count; /* total caps allocated */
37151668 YS	453	int caps_use_count; /* in use */
fe33032d	454	int caps_use_max; /* max used caps */
37151668 YS	455	int caps_reserve_count; /* unused, reserved */
	456	int caps_avail_count; /* unused, unreserved */
	457	int caps_min_count; /* keep at least this many
	458	(unreserved) */
37c4efc1 YZ	459	spinlock_t dentry_list_lock;
	460	struct list_head dentry_leases; /* fifo list */
	461	struct list_head dentry_dir_leases; /* lru list */
10183a69	462
f9009efa XL	463	struct ceph_client_metric metric;
f9009efa XL	464
75c9627e YZ	465	spinlock_t snapid_map_lock;
	466	struct rb_root snapid_map_tree;
	467	struct list_head snapid_map_lru;
	468
10183a69 YZ	469	struct rw_semaphore pool_perm_rwsem;
10183a69 YZ	470	struct rb_root pool_perm_tree;
717e6f28 YZ	471
717e6f28 YZ	472	char nodename[__NEW_UTS_LEN + 1];
2f2dc053 SW	473	};
	474
	475	extern const char *ceph_mds_op_name(int op);
	476
	477	extern struct ceph_mds_session *
	478	__ceph_lookup_mds_session(struct ceph_mds_client *, int mds);
	479
a687ecaf JS	480	extern const char *ceph_session_state_name(int s);
a687ecaf JS	481
5b3248c6 XL	482	extern struct ceph_mds_session *
5b3248c6 XL	483	ceph_get_mds_session(struct ceph_mds_session *s);
2f2dc053 SW	484	extern void ceph_put_mds_session(struct ceph_mds_session *s);
	485
	486	extern int ceph_send_msg_mds(struct ceph_mds_client *mdsc,
	487	struct ceph_msg *msg, int mds);
	488
3d14c5d2	489	extern int ceph_mdsc_init(struct ceph_fs_client *fsc);
2f2dc053	490	extern void ceph_mdsc_close_sessions(struct ceph_mds_client *mdsc);
48fec5d0	491	extern void ceph_mdsc_force_umount(struct ceph_mds_client *mdsc);
3d14c5d2	492	extern void ceph_mdsc_destroy(struct ceph_fs_client *fsc);
2f2dc053 SW	493
	494	extern void ceph_mdsc_sync(struct ceph_mds_client *mdsc);
	495
167c9e35	496	extern void ceph_invalidate_dir_request(struct ceph_mds_request *req);
54008399 YZ	497	extern int ceph_alloc_readdir_reply_buffer(struct ceph_mds_request *req,
54008399 YZ	498	struct inode *dir);
2f2dc053 SW	499	extern struct ceph_mds_request *
2f2dc053 SW	500	ceph_mdsc_create_request(struct ceph_mds_client *mdsc, int op, int mode);
86bda539 JL	501	extern int ceph_mdsc_submit_request(struct ceph_mds_client *mdsc,
	502	struct inode *dir,
	503	struct ceph_mds_request *req);
2f2dc053 SW	504	extern int ceph_mdsc_do_request(struct ceph_mds_client *mdsc,
	505	struct inode *dir,
	506	struct ceph_mds_request *req);
a25949b9	507	extern void ceph_mdsc_release_dir_caps(struct ceph_mds_request *req);
2f2dc053 SW	508	static inline void ceph_mdsc_get_request(struct ceph_mds_request *req)
2f2dc053 SW	509	{
153c8e6b SW	510	kref_get(&req->r_kref);
	511	}
	512	extern void ceph_mdsc_release_request(struct kref *kref);
	513	static inline void ceph_mdsc_put_request(struct ceph_mds_request *req)
	514	{
	515	kref_put(&req->r_kref, ceph_mdsc_release_request);
2f2dc053	516	}
2f2dc053	517
e3ec8d68 YZ	518	extern void __ceph_queue_cap_release(struct ceph_mds_session *session,
	519	struct ceph_cap *cap);
	520	extern void ceph_flush_cap_releases(struct ceph_mds_client *mdsc,
	521	struct ceph_mds_session *session);
37c4efc1	522	extern void ceph_queue_cap_reclaim_work(struct ceph_mds_client *mdsc);
fe33032d	523	extern void ceph_reclaim_caps_nr(struct ceph_mds_client *mdsc, int nr);
f5d77269 JL	524	extern int ceph_iterate_session_caps(struct ceph_mds_session *session,
	525	int (cb)(struct inode ,
	526	struct ceph_cap , void ),
	527	void *arg);
2f2dc053 SW	528	extern void ceph_mdsc_pre_umount(struct ceph_mds_client *mdsc);
2f2dc053 SW	529
f77f21bb JL	530	static inline void ceph_mdsc_free_path(char *path, int len)
f77f21bb JL	531	{
2a575f13	532	if (!IS_ERR_OR_NULL(path))
f77f21bb JL	533	__putname(path - (PATH_MAX - 1 - len));
	534	}
	535
2f2dc053 SW	536	extern char ceph_mdsc_build_path(struct dentry dentry, int plen, u64 base,
	537	int stop_on_nosnap);
	538
	539	extern void __ceph_mdsc_drop_dentry_lease(struct dentry *dentry);
	540	extern void ceph_mdsc_lease_send_msg(struct ceph_mds_session *session,
2f2dc053 SW	541	struct dentry *dentry, char action,
	542	u32 seq);
	543
430afbad YZ	544	extern void ceph_mdsc_handle_mdsmap(struct ceph_mds_client *mdsc,
	545	struct ceph_msg *msg);
	546	extern void ceph_mdsc_handle_fsmap(struct ceph_mds_client *mdsc,
	547	struct ceph_msg *msg);
2f2dc053	548
5d72d13c YZ	549	extern struct ceph_mds_session *
5d72d13c YZ	550	ceph_mdsc_open_export_target_session(struct ceph_mds_client *mdsc, int target);
154f42c2 SW	551	extern void ceph_mdsc_open_export_target_sessions(struct ceph_mds_client *mdsc,
	552	struct ceph_mds_session *session);
	553
e30ee581 ZZ	554	extern int ceph_trim_caps(struct ceph_mds_client *mdsc,
	555	struct ceph_mds_session *session,
	556	int max_caps);
d4846487	557
891f3f5a JL	558	static inline int ceph_wait_on_async_create(struct inode *inode)
	559	{
	560	struct ceph_inode_info *ci = ceph_inode(inode);
	561
	562	return wait_on_bit(&ci->i_ceph_flags, CEPH_ASYNC_CREATE_BIT,
	563	TASK_INTERRUPTIBLE);
	564	}
d4846487 JL	565
	566	extern u64 ceph_get_deleg_ino(struct ceph_mds_session *session);
	567	extern int ceph_restore_deleg_ino(struct ceph_mds_session *session, u64 ino);
2f2dc053	568	#endif