X-Git-Url: https://git.whamcloud.com/?a=blobdiff_plain;f=lustre%2Finclude%2Flu_object.h;h=f431e0b4a5d2c3bc985f35c29c4e759def937c94;hb=e4872648bfaea1795133f10695c40e0fd3069831;hp=1b00b02c429f61502bc214eb6c710f3e2c3a2230;hpb=90d8e7fd28746a572c8de488222f5312fe927fc3;p=fs%2Flustre-release.git diff --git a/lustre/include/lu_object.h b/lustre/include/lu_object.h index 1b00b02..f431e0b 100644 --- a/lustre/include/lu_object.h +++ b/lustre/include/lu_object.h @@ -26,10 +26,13 @@ * GPL HEADER END */ /* - * Copyright 2008 Sun Microsystems, Inc. All rights reserved + * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved. * Use is subject to license terms. */ /* + * Copyright (c) 2011 Whamcloud, Inc. + */ +/* * This file is part of Lustre, http://www.lustre.org/ * Lustre is a trademark of Sun Microsystems, Inc. */ @@ -42,10 +45,10 @@ /* * struct lu_fid */ -#include - #include +#include + #include struct seq_file; @@ -169,11 +172,23 @@ struct lu_device_operations { }; /** + * For lu_object_conf flags + */ +typedef enum { + /* Currently, only used for client-side object initialization. */ + LOC_F_NEW = 0x1, +} loc_flags_t; + +/** * Object configuration, describing particulars of object being created. On * server this is not used, as server objects are full identified by fid. On * client configuration contains struct lustre_md. */ struct lu_object_conf { + /** + * Some hints for obj find and alloc. + */ + loc_flags_t loc_flags; }; /** @@ -257,7 +272,7 @@ struct lu_device { * * \todo XXX which means that atomic_t is probably too small. */ - atomic_t ld_ref; + cfs_atomic_t ld_ref; /** * Pointer to device type. Never modified once set. */ @@ -278,6 +293,10 @@ struct lu_device { * A list of references to this object, for debugging. */ struct lu_ref ld_reference; + /** + * Link the device to the site. + **/ + cfs_list_t ld_linkage; }; struct lu_device_type_operations; @@ -328,7 +347,7 @@ struct lu_device_type { * * \see lu_device_types. */ - struct list_head ldt_linkage; + cfs_list_t ldt_linkage; }; /** @@ -399,11 +418,11 @@ struct lu_attr { /** size in bytes */ __u64 la_size; /** modification time in seconds since Epoch */ - __u64 la_mtime; + obd_time la_mtime; /** access time in seconds since Epoch */ - __u64 la_atime; + obd_time la_atime; /** change time in seconds since Epoch */ - __u64 la_ctime; + obd_time la_ctime; /** 512-byte blocks allocated to object */ __u64 la_blocks; /** permission bits and file type */ @@ -445,6 +464,8 @@ enum la_valid { LA_NLINK = 1 << 10, LA_RDEV = 1 << 11, LA_BLKSIZE = 1 << 12, + LA_KILL_SUID = 1 << 13, + LA_KILL_SGID = 1 << 14, }; /** @@ -466,7 +487,7 @@ struct lu_object { /** * Linkage into list of all layers. */ - struct list_head lo_linkage; + cfs_list_t lo_linkage; /** * Depth. Top level layer depth is 0. */ @@ -496,8 +517,8 @@ enum lu_object_header_attr { /** * UNIX file type is stored in S_IFMT bits. */ - LOHA_FT_START = 1 << 12, /**< S_IFIFO */ - LOHA_FT_END = 1 << 15, /**< S_IFREG */ + LOHA_FT_START = 001 << 12, /**< S_IFIFO */ + LOHA_FT_END = 017 << 12, /**< S_IFMT */ }; /** @@ -516,41 +537,76 @@ struct lu_object_header { * Object flags from enum lu_object_header_flags. Set and checked * atomically. */ - unsigned long loh_flags; + unsigned long loh_flags; /** * Object reference count. Protected by lu_site::ls_guard. */ - atomic_t loh_ref; + cfs_atomic_t loh_ref; /** * Fid, uniquely identifying this object. */ - struct lu_fid loh_fid; + struct lu_fid loh_fid; /** * Common object attributes, cached for efficiency. From enum * lu_object_header_attr. */ - __u32 loh_attr; + __u32 loh_attr; /** * Linkage into per-site hash table. Protected by lu_site::ls_guard. */ - struct hlist_node loh_hash; + cfs_hlist_node_t loh_hash; /** * Linkage into per-site LRU list. Protected by lu_site::ls_guard. */ - struct list_head loh_lru; + cfs_list_t loh_lru; /** * Linkage into list of layers. Never modified once set (except lately * during object destruction). No locking is necessary. */ - struct list_head loh_layers; + cfs_list_t loh_layers; /** * A list of references to this object, for debugging. */ - struct lu_ref loh_reference; + struct lu_ref loh_reference; }; struct fld; +struct lu_site_bkt_data { + /** + * number of busy object on this bucket + */ + long lsb_busy; + /** + * LRU list, updated on each access to object. Protected by + * bucket lock of lu_site::ls_obj_hash. + * + * "Cold" end of LRU is lu_site::ls_lru.next. Accessed object are + * moved to the lu_site::ls_lru.prev (this is due to the non-existence + * of list_for_each_entry_safe_reverse()). + */ + cfs_list_t lsb_lru; + /** + * Wait-queue signaled when an object in this site is ultimately + * destroyed (lu_object_free()). It is used by lu_object_find() to + * wait before re-trying when object in the process of destruction is + * found in the hash table. + * + * \see htable_lookup(). + */ + cfs_waitq_t lsb_marche_funebre; +}; + +enum { + LU_SS_CREATED = 0, + LU_SS_CACHE_HIT, + LU_SS_CACHE_MISS, + LU_SS_CACHE_RACE, + LU_SS_CACHE_DEATH_RACE, + LU_SS_LRU_PURGED, + LU_SS_LAST_STAT +}; + /** * lu_site is a "compartment" within which objects are unique, and LRU * discipline is maintained. @@ -563,106 +619,43 @@ struct fld; */ struct lu_site { /** - * Site-wide lock. - * - * lock protecting: - * - * - lu_site::ls_hash hash table (and its linkages in objects); - * - * - lu_site::ls_lru list (and its linkages in objects); - * - * - 0/1 transitions of object lu_object_header::loh_ref - * reference count; - * - * yes, it's heavy. + * objects hash table */ - rwlock_t ls_guard; + cfs_hash_t *ls_obj_hash; /** - * Hash-table where objects are indexed by fid. + * index of bucket on hash table while purging */ - struct hlist_head *ls_hash; + int ls_purge_start; /** - * Bit-mask for hash-table size. + * Top-level device for this stack. */ - int ls_hash_mask; + struct lu_device *ls_top_dev; /** - * Order of hash-table. + * Linkage into global list of sites. */ - int ls_hash_bits; + cfs_list_t ls_linkage; /** - * Number of buckets in the hash-table. - */ - int ls_hash_size; + * List for lu device for this site, protected + * by ls_ld_lock. + **/ + cfs_list_t ls_ld_linkage; + cfs_spinlock_t ls_ld_lock; /** - * LRU list, updated on each access to object. Protected by - * lu_site::ls_guard. - * - * "Cold" end of LRU is lu_site::ls_lru.next. Accessed object are - * moved to the lu_site::ls_lru.prev (this is due to the non-existence - * of list_for_each_entry_safe_reverse()). - */ - struct list_head ls_lru; - /** - * Total number of objects in this site. Protected by - * lu_site::ls_guard. + * lu_site stats */ - unsigned ls_total; - /** - * Total number of objects in this site with reference counter greater - * than 0. Protected by lu_site::ls_guard. - */ - unsigned ls_busy; - - /** - * Top-level device for this stack. - */ - struct lu_device *ls_top_dev; + struct lprocfs_stats *ls_stats; + struct lprocfs_stats *ls_time_stats; +}; - /** - * Wait-queue signaled when an object in this site is ultimately - * destroyed (lu_object_free()). It is used by lu_object_find() to - * wait before re-trying when object in the process of destruction is - * found in the hash table. - * - * If having a single wait-queue turns out to be a problem, a - * wait-queue per hash-table bucket can be easily implemented. - * - * \see htable_lookup(). - */ - cfs_waitq_t ls_marche_funebre; - - /** statistical counters. Protected by nothing, races are accepted. */ - struct { - __u32 s_created; - __u32 s_cache_hit; - __u32 s_cache_miss; - /** - * Number of hash-table entry checks made. - * - * ->s_cache_check / (->s_cache_miss + ->s_cache_hit) - * - * is an average number of hash slots inspected during single - * lookup. - */ - __u32 s_cache_check; - /** Races with cache insertions. */ - __u32 s_cache_race; - /** - * Races with object destruction. - * - * \see lu_site::ls_marche_funebre. - */ - __u32 s_cache_death_race; - __u32 s_lru_purged; - } ls_stats; +static inline struct lu_site_bkt_data * +lu_site_bkt_from_fid(struct lu_site *site, struct lu_fid *fid) +{ + cfs_hash_bd_t bd; - /** - * Linkage into global list of sites. - */ - struct list_head ls_linkage; - struct lprocfs_stats *ls_time_stats; -}; + cfs_hash_bd_get(site->ls_obj_hash, fid, &bd); + return cfs_hash_bd_extra_get(site->ls_obj_hash, &bd); +} /** \name ctors * Constructors/destructors. @@ -707,8 +700,8 @@ void lu_types_stop(void); */ static inline void lu_object_get(struct lu_object *o) { - LASSERT(atomic_read(&o->lo_header->loh_ref) > 0); - atomic_inc(&o->lo_header->loh_ref); + LASSERT(cfs_atomic_read(&o->lo_header->loh_ref) > 0); + cfs_atomic_inc(&o->lo_header->loh_ref); } /** @@ -717,7 +710,7 @@ static inline void lu_object_get(struct lu_object *o) */ static inline int lu_object_is_dying(const struct lu_object_header *h) { - return test_bit(LU_OBJECT_HEARD_BANSHEE, &h->loh_flags); + return cfs_test_bit(LU_OBJECT_HEARD_BANSHEE, &h->loh_flags); } void lu_object_put(const struct lu_env *env, struct lu_object *o); @@ -749,7 +742,7 @@ struct lu_object *lu_object_find_slice(const struct lu_env *env, */ static inline struct lu_object *lu_object_top(struct lu_object_header *h) { - LASSERT(!list_empty(&h->loh_layers)); + LASSERT(!cfs_list_empty(&h->loh_layers)); return container_of0(h->loh_layers.next, struct lu_object, lo_linkage); } @@ -815,7 +808,7 @@ int lu_cdebug_printer(const struct lu_env *env, do { \ static DECLARE_LU_CDEBUG_PRINT_INFO(__info, mask); \ \ - if (cdebug_show(mask, DEBUG_SUBSYSTEM)) { \ + if (cfs_cdebug_show(mask, DEBUG_SUBSYSTEM)) { \ lu_object_print(env, &__info, lu_cdebug_printer, object); \ CDEBUG(mask, format , ## __VA_ARGS__); \ } \ @@ -828,7 +821,7 @@ do { \ do { \ static DECLARE_LU_CDEBUG_PRINT_INFO(__info, mask); \ \ - if (cdebug_show(mask, DEBUG_SUBSYSTEM)) { \ + if (cfs_cdebug_show(mask, DEBUG_SUBSYSTEM)) { \ lu_object_header_print(env, &__info, lu_cdebug_printer, \ (object)->lo_header); \ lu_cdebug_printer(env, &__info, "\n"); \ @@ -847,11 +840,11 @@ void lu_object_header_print(const struct lu_env *env, void *cookie, */ int lu_object_invariant(const struct lu_object *o); -void lu_stack_fini(const struct lu_env *env, struct lu_device *top); /** - * Returns 1 iff object @o exists on the stable storage, - * returns -1 iff object @o is on remote server. + * \retval 1 iff object \a o exists on stable storage, + * \retval 0 iff object \a o not exists on stable storage. + * \retval -1 iff object \a o is on remote server. */ static inline int lu_object_exists(const struct lu_object *o) { @@ -905,12 +898,18 @@ static inline void lu_object_ref_del_at(struct lu_object *o, lu_ref_del_at(&o->lo_header->loh_reference, link, scope, source); } +/** input params, should be filled out by mdt */ struct lu_rdpg { - /* input params, should be filled out by mdt */ - __u64 rp_hash; /* hash */ - int rp_count; /* count in bytes */ - int rp_npages; /* number of pages */ - struct page **rp_pages; /* pointers to pages */ + /** hash */ + __u64 rp_hash; + /** count in bytes */ + unsigned int rp_count; + /** number of pages */ + unsigned int rp_npages; + /** requested attr */ + __u32 rp_attrs; + /** pointers to pages */ + struct page **rp_pages; }; enum lu_xattr_flags { @@ -980,7 +979,7 @@ struct lu_context { * `non-transient' contexts, i.e., ones created for service threads * are placed here. */ - struct list_head lc_remember; + cfs_list_t lc_remember; /** * Version counter used to skip calls to lu_context_refill() when no * keys were registered. @@ -1114,11 +1113,11 @@ struct lu_context_key { * Internal implementation detail: number of values created for this * key. */ - atomic_t lct_used; + cfs_atomic_t lct_used; /** * Internal implementation detail: module for this key. */ - struct module *lct_owner; + cfs_module_t *lct_owner; /** * References to this key. For debugging. */