X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=blobdiff_plain;f=lustre%2Fllite%2Fllite_lib.c;h=91993d737858701c0dea0eba8948a0755f6ad1b5;hp=8191558f4025761511ca92e208442717ca86b947;hb=26bcb8fcd034f2edc993fa27a9158479836523b8;hpb=f3fa8051dc74a775146b648e90cad74c26340866 diff --git a/lustre/llite/llite_lib.c b/lustre/llite/llite_lib.c index 8191558..91993d7 100644 --- a/lustre/llite/llite_lib.c +++ b/lustre/llite/llite_lib.c @@ -41,11 +41,12 @@ #define DEBUG_SUBSYSTEM S_LLITE #include +#include #include #include #include -#include +#include #include #include #include @@ -58,14 +59,8 @@ struct kmem_cache *ll_file_data_slab; -CFS_LIST_HEAD(ll_super_blocks); -DEFINE_SPINLOCK(ll_sb_lock); - -#ifndef MS_HAS_NEW_AOPS -extern struct address_space_operations ll_aops; -#else -extern struct address_space_operations_ext ll_aops; -#endif +static struct list_head ll_super_blocks = LIST_HEAD_INIT(ll_super_blocks); +static DEFINE_SPINLOCK(ll_sb_lock); #ifndef log2 #define log2(n) ffz(~(n)) @@ -76,14 +71,14 @@ static struct ll_sb_info *ll_init_sbi(void) struct ll_sb_info *sbi = NULL; unsigned long pages; unsigned long lru_page_max; - struct sysinfo si; - class_uuid_t uuid; - int i; - ENTRY; + struct sysinfo si; + class_uuid_t uuid; + int i; + ENTRY; - OBD_ALLOC(sbi, sizeof(*sbi)); - if (!sbi) - RETURN(NULL); + OBD_ALLOC_PTR(sbi); + if (sbi == NULL) + RETURN(NULL); spin_lock_init(&sbi->ll_lock); mutex_init(&sbi->ll_lco.lco_lock); @@ -93,33 +88,34 @@ static struct ll_sb_info *ll_init_sbi(void) si_meminfo(&si); pages = si.totalram - si.totalhigh; - if (pages >> (20 - PAGE_CACHE_SHIFT) < 512) { - lru_page_max = pages / 2; - } else { - lru_page_max = (pages / 4) * 3; - } + lru_page_max = pages / 2; - /* initialize lru data */ - cfs_atomic_set(&sbi->ll_cache.ccc_users, 0); + /* initialize ll_cache data */ + atomic_set(&sbi->ll_cache.ccc_users, 0); sbi->ll_cache.ccc_lru_max = lru_page_max; - cfs_atomic_set(&sbi->ll_cache.ccc_lru_left, lru_page_max); + atomic_long_set(&sbi->ll_cache.ccc_lru_left, lru_page_max); spin_lock_init(&sbi->ll_cache.ccc_lru_lock); - CFS_INIT_LIST_HEAD(&sbi->ll_cache.ccc_lru); + INIT_LIST_HEAD(&sbi->ll_cache.ccc_lru); + + /* turn unstable check off by default as it impacts performance */ + sbi->ll_cache.ccc_unstable_check = 0; + atomic_long_set(&sbi->ll_cache.ccc_unstable_nr, 0); + init_waitqueue_head(&sbi->ll_cache.ccc_unstable_waitq); - sbi->ll_ra_info.ra_max_pages_per_file = min(pages / 32, - SBI_DEFAULT_READAHEAD_MAX); - sbi->ll_ra_info.ra_max_pages = sbi->ll_ra_info.ra_max_pages_per_file; - sbi->ll_ra_info.ra_max_read_ahead_whole_pages = - SBI_DEFAULT_READAHEAD_WHOLE_MAX; - CFS_INIT_LIST_HEAD(&sbi->ll_conn_chain); - CFS_INIT_LIST_HEAD(&sbi->ll_orphan_dentry_list); + sbi->ll_ra_info.ra_max_pages_per_file = min(pages / 32, + SBI_DEFAULT_READAHEAD_MAX); + sbi->ll_ra_info.ra_max_pages = sbi->ll_ra_info.ra_max_pages_per_file; + sbi->ll_ra_info.ra_max_read_ahead_whole_pages = + SBI_DEFAULT_READAHEAD_WHOLE_MAX; + INIT_LIST_HEAD(&sbi->ll_conn_chain); + INIT_LIST_HEAD(&sbi->ll_orphan_dentry_list); ll_generate_random_uuid(uuid); class_uuid_unparse(uuid, &sbi->ll_sb_uuid); CDEBUG(D_CONFIG, "generated uuid: %s\n", sbi->ll_sb_uuid.uuid); spin_lock(&ll_sb_lock); - cfs_list_add_tail(&sbi->ll_list, &ll_super_blocks); + list_add_tail(&sbi->ll_list, &ll_super_blocks); spin_unlock(&ll_sb_lock); sbi->ll_flags |= LL_SBI_VERBOSE; @@ -138,25 +134,34 @@ static struct ll_sb_info *ll_init_sbi(void) pp_w_hist.oh_lock); } - /* metadata statahead is enabled by default */ - sbi->ll_sa_max = LL_SA_RPC_DEF; - cfs_atomic_set(&sbi->ll_sa_total, 0); - cfs_atomic_set(&sbi->ll_sa_wrong, 0); - cfs_atomic_set(&sbi->ll_agl_total, 0); - sbi->ll_flags |= LL_SBI_AGL_ENABLED; - - RETURN(sbi); + /* metadata statahead is enabled by default */ + sbi->ll_sa_max = LL_SA_RPC_DEF; + atomic_set(&sbi->ll_sa_total, 0); + atomic_set(&sbi->ll_sa_wrong, 0); + atomic_set(&sbi->ll_sa_running, 0); + atomic_set(&sbi->ll_agl_total, 0); + sbi->ll_flags |= LL_SBI_AGL_ENABLED; + + /* root squash */ + sbi->ll_squash.rsi_uid = 0; + sbi->ll_squash.rsi_gid = 0; + INIT_LIST_HEAD(&sbi->ll_squash.rsi_nosquash_nids); + init_rwsem(&sbi->ll_squash.rsi_sem); + + RETURN(sbi); } -void ll_free_sbi(struct super_block *sb) +static void ll_free_sbi(struct super_block *sb) { struct ll_sb_info *sbi = ll_s2sbi(sb); ENTRY; if (sbi != NULL) { spin_lock(&ll_sb_lock); - cfs_list_del(&sbi->ll_list); + list_del(&sbi->ll_list); spin_unlock(&ll_sb_lock); + if (!list_empty(&sbi->ll_squash.rsi_nosquash_nids)) + cfs_free_nidlist(&sbi->ll_squash.rsi_nosquash_nids); OBD_FREE(sbi, sizeof(*sbi)); } EXIT; @@ -208,7 +213,10 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt, OBD_CONNECT_JOBSTATS | OBD_CONNECT_LVB_TYPE | OBD_CONNECT_LAYOUTLOCK | OBD_CONNECT_PINGLESS | OBD_CONNECT_MAX_EASIZE | - OBD_CONNECT_FLOCK_DEAD; + OBD_CONNECT_FLOCK_DEAD | + OBD_CONNECT_DISP_STRIPE | OBD_CONNECT_LFSCK | + OBD_CONNECT_OPEN_BY_FID | + OBD_CONNECT_DIR_STRIPE; if (sbi->ll_flags & LL_SBI_SOM_PREVIEW) data->ocd_connect_flags |= OBD_CONNECT_SOM; @@ -375,14 +383,12 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt, sbi->ll_flags |= LL_SBI_64BIT_HASH; if (data->ocd_connect_flags & OBD_CONNECT_BRW_SIZE) - sbi->ll_md_brw_size = data->ocd_brw_size; + sbi->ll_md_brw_pages = data->ocd_brw_size >> PAGE_CACHE_SHIFT; else - sbi->ll_md_brw_size = PAGE_CACHE_SIZE; + sbi->ll_md_brw_pages = 1; - if (data->ocd_connect_flags & OBD_CONNECT_LAYOUTLOCK) { - LCONSOLE_INFO("Layout lock feature supported.\n"); + if (data->ocd_connect_flags & OBD_CONNECT_LAYOUTLOCK) sbi->ll_flags |= LL_SBI_LAYOUT_LOCK; - } if (data->ocd_ibits_known & MDS_INODELOCK_XATTR) { if (!(data->ocd_connect_flags & OBD_CONNECT_MAX_EASIZE)) { @@ -410,7 +416,8 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt, OBD_CONNECT_MAXBYTES | OBD_CONNECT_EINPROGRESS | OBD_CONNECT_JOBSTATS | OBD_CONNECT_LVB_TYPE | - OBD_CONNECT_LAYOUTLOCK | OBD_CONNECT_PINGLESS; + OBD_CONNECT_LAYOUTLOCK | + OBD_CONNECT_PINGLESS | OBD_CONNECT_LFSCK; if (sbi->ll_flags & LL_SBI_SOM_PREVIEW) data->ocd_connect_flags |= OBD_CONNECT_SOM; @@ -494,7 +501,8 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt, /* make root inode * XXX: move this to after cbd setup? */ - valid = OBD_MD_FLGETATTR | OBD_MD_FLBLOCKS | OBD_MD_FLMDSCAPA; + valid = OBD_MD_FLGETATTR | OBD_MD_FLBLOCKS | OBD_MD_FLMDSCAPA | + OBD_MD_FLMODEASIZE; if (sbi->ll_flags & LL_SBI_RMT_CLIENT) valid |= OBD_MD_FLRMTPERM; else if (sbi->ll_flags & LL_SBI_ACL) @@ -534,9 +542,9 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt, md_free_lustre_md(sbi->ll_md_exp, &lmd); ptlrpc_req_finished(request); - if (root == NULL || IS_ERR(root)) { - if (lmd.lsm) - obd_free_memmd(sbi->ll_dt_exp, &lmd.lsm); + if (IS_ERR(root)) { + if (lmd.lsm) + obd_free_memmd(sbi->ll_dt_exp, &lmd.lsm); #ifdef CONFIG_FS_POSIX_ACL if (lmd.posix_acl) { posix_acl_release(lmd.posix_acl); @@ -590,10 +598,8 @@ static int client_common_fill_super(struct super_block *sb, char *md, char *dt, /* s_dev is also used in lt_compare() to compare two fs, but that is * only a node-local comparison. */ uuid = obd_get_uuid(sbi->ll_md_exp); - if (uuid != NULL) { + if (uuid != NULL) sb->s_dev = get_uuid2int(uuid->uuid, strlen(uuid->uuid)); - get_uuid2fsid(uuid->uuid, strlen(uuid->uuid), &sbi->ll_fsid); - } if (data != NULL) OBD_FREE_PTR(data); @@ -635,19 +641,58 @@ out: int ll_get_max_mdsize(struct ll_sb_info *sbi, int *lmmsize) { - int size, rc; + int size, rc; - *lmmsize = obd_size_diskmd(sbi->ll_dt_exp, NULL); - size = sizeof(int); - rc = obd_get_info(NULL, sbi->ll_md_exp, sizeof(KEY_MAX_EASIZE), - KEY_MAX_EASIZE, &size, lmmsize, NULL); - if (rc) - CERROR("Get max mdsize error rc %d \n", rc); + *lmmsize = obd_size_diskmd(sbi->ll_dt_exp, NULL); + size = sizeof(int); + rc = obd_get_info(NULL, sbi->ll_md_exp, sizeof(KEY_MAX_EASIZE), + KEY_MAX_EASIZE, &size, lmmsize, NULL); + if (rc) + CERROR("Get max mdsize error rc %d\n", rc); - RETURN(rc); + RETURN(rc); +} + +int ll_get_default_mdsize(struct ll_sb_info *sbi, int *lmmsize) +{ + int size, rc; + + size = sizeof(int); + rc = obd_get_info(NULL, sbi->ll_md_exp, sizeof(KEY_DEFAULT_EASIZE), + KEY_DEFAULT_EASIZE, &size, lmmsize, NULL); + if (rc) + CERROR("Get default mdsize error rc %d\n", rc); + + RETURN(rc); } -void ll_dump_inode(struct inode *inode) +int ll_get_max_cookiesize(struct ll_sb_info *sbi, int *lmmsize) +{ + int size, rc; + + size = sizeof(int); + rc = obd_get_info(NULL, sbi->ll_md_exp, sizeof(KEY_MAX_COOKIESIZE), + KEY_MAX_COOKIESIZE, &size, lmmsize, NULL); + if (rc) + CERROR("Get max cookiesize error rc %d\n", rc); + + RETURN(rc); +} + +int ll_get_default_cookiesize(struct ll_sb_info *sbi, int *lmmsize) +{ + int size, rc; + + size = sizeof(int); + rc = obd_get_info(NULL, sbi->ll_md_exp, sizeof(KEY_DEFAULT_COOKIESIZE), + KEY_DEFAULT_COOKIESIZE, &size, lmmsize, NULL); + if (rc) + CERROR("Get default cookiesize error rc %d\n", rc); + + RETURN(rc); +} + +static void ll_dump_inode(struct inode *inode) { struct ll_d_hlist_node *tmp; int dentry_count = 0; @@ -657,8 +702,10 @@ void ll_dump_inode(struct inode *inode) ll_d_hlist_for_each(tmp, &inode->i_dentry) dentry_count++; - CERROR("inode %p dump: dev=%s ino=%lu mode=%o count=%u, %d dentries\n", - inode, ll_i2mdexp(inode)->exp_obd->obd_name, inode->i_ino, + CERROR("%s: inode %p dump: dev=%s fid="DFID + " mode=%o count=%u, %d dentries\n", + ll_get_fsname(inode->i_sb, NULL, 0), inode, + ll_i2mdexp(inode)->exp_obd->obd_name, PFID(ll_inode2fid(inode)), inode->i_mode, atomic_read(&inode->i_count), dentry_count); } @@ -676,7 +723,7 @@ void lustre_dump_dentry(struct dentry *dentry, int recur) " flags=0x%x, fsdata=%p, %d subdirs\n", dentry, dentry->d_name.len, dentry->d_name.name, dentry->d_parent->d_name.len, dentry->d_parent->d_name.name, - dentry->d_parent, dentry->d_inode, d_refcount(dentry), + dentry->d_parent, dentry->d_inode, ll_d_count(dentry), dentry->d_flags, dentry->d_fsdata, subdirs); if (dentry->d_inode != NULL) ll_dump_inode(dentry->d_inode); @@ -690,7 +737,7 @@ void lustre_dump_dentry(struct dentry *dentry, int recur) } } -void client_common_put_super(struct super_block *sb) +static void client_common_put_super(struct super_block *sb) { struct ll_sb_info *sbi = ll_s2sbi(sb); ENTRY; @@ -706,7 +753,7 @@ void client_common_put_super(struct super_block *sb) cl_sb_fini(sb); - cfs_list_del(&sbi->ll_conn_chain); + list_del(&sbi->ll_conn_chain); obd_fid_fini(sbi->ll_dt_exp->exp_obd); obd_disconnect(sbi->ll_dt_exp); @@ -726,47 +773,28 @@ void client_common_put_super(struct super_block *sb) void ll_kill_super(struct super_block *sb) { - struct ll_sb_info *sbi; - - ENTRY; + struct ll_sb_info *sbi; + ENTRY; /* not init sb ?*/ - if (!(sb->s_flags & MS_ACTIVE)) - return; + if (!(sb->s_flags & MS_ACTIVE)) + return; - sbi = ll_s2sbi(sb); - /* we need restore s_dev from changed for clustred NFS before put_super - * because new kernels have cached s_dev and change sb->s_dev in - * put_super not affected real removing devices */ + sbi = ll_s2sbi(sb); + /* we need restore s_dev from changed for clustred NFS before put_super + * because new kernels have cached s_dev and change sb->s_dev in + * put_super not affected real removing devices */ if (sbi) { sb->s_dev = sbi->ll_sdev_orig; sbi->ll_umounting = 1; - } - EXIT; -} -char *ll_read_opt(const char *opt, char *data) -{ - char *value; - char *retval; - ENTRY; - - CDEBUG(D_SUPER, "option: %s, data %s\n", opt, data); - if (strncmp(opt, data, strlen(opt))) - RETURN(NULL); - if ((value = strchr(data, '=')) == NULL) - RETURN(NULL); - - value++; - OBD_ALLOC(retval, strlen(value) + 1); - if (!retval) { - CERROR("out of memory!\n"); - RETURN(NULL); - } + /* wait running statahead threads to quit */ + while (atomic_read(&sbi->ll_sa_running) > 0) + schedule_timeout_and_set_state(TASK_UNINTERRUPTIBLE, + msecs_to_jiffies(MSEC_PER_SEC >> 3)); + } - memcpy(retval, value, strlen(value)+1); - CDEBUG(D_SUPER, "Assigned option: %s, value %s\n", opt, retval); - RETURN(retval); + EXIT; } static inline int ll_set_opt(const char *opt, char *data, int fl) @@ -821,30 +849,11 @@ static int ll_options(char *options, int *flags) *flags &= ~tmp; goto next; } -#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(2, 5, 50, 0) - tmp = ll_set_opt("acl", s1, LL_SBI_ACL); - if (tmp) { - /* Ignore deprecated mount option. The client will - * always try to mount with ACL support, whether this - * is used depends on whether server supports it. */ - LCONSOLE_ERROR_MSG(0x152, "Ignoring deprecated " - "mount option 'acl'.\n"); - goto next; - } - tmp = ll_set_opt("noacl", s1, LL_SBI_ACL); - if (tmp) { - LCONSOLE_ERROR_MSG(0x152, "Ignoring deprecated " - "mount option 'noacl'.\n"); - goto next; - } -#else -#warning "{no}acl options have been deprecated since 1.8, please remove them" -#endif - tmp = ll_set_opt("remote_client", s1, LL_SBI_RMT_CLIENT); - if (tmp) { - *flags |= tmp; - goto next; - } + tmp = ll_set_opt("remote_client", s1, LL_SBI_RMT_CLIENT); + if (tmp) { + *flags |= tmp; + goto next; + } tmp = ll_set_opt("user_fid2path", s1, LL_SBI_USER_FID2PATH); if (tmp) { *flags |= tmp; @@ -930,15 +939,15 @@ void ll_lli_init(struct ll_inode_info *lli) lli->lli_posix_acl = NULL; lli->lli_remote_perms = NULL; mutex_init(&lli->lli_rmtperm_mutex); - /* Do not set lli_fid, it has been initialized already. */ - fid_zero(&lli->lli_pfid); - CFS_INIT_LIST_HEAD(&lli->lli_close_list); - CFS_INIT_LIST_HEAD(&lli->lli_oss_capas); - cfs_atomic_set(&lli->lli_open_count, 0); - lli->lli_mds_capa = NULL; - lli->lli_rmtperm_time = 0; - lli->lli_pending_och = NULL; - lli->lli_mds_read_och = NULL; + /* Do not set lli_fid, it has been initialized already. */ + fid_zero(&lli->lli_pfid); + INIT_LIST_HEAD(&lli->lli_close_list); + INIT_LIST_HEAD(&lli->lli_oss_capas); + atomic_set(&lli->lli_open_count, 0); + lli->lli_mds_capa = NULL; + lli->lli_rmtperm_time = 0; + lli->lli_pending_och = NULL; + lli->lli_mds_read_och = NULL; lli->lli_mds_write_och = NULL; lli->lli_mds_exec_och = NULL; lli->lli_open_fd_read_count = 0; @@ -947,7 +956,8 @@ void ll_lli_init(struct ll_inode_info *lli) mutex_init(&lli->lli_och_mutex); spin_lock_init(&lli->lli_agl_lock); lli->lli_has_smd = false; - lli->lli_layout_gen = LL_LAYOUT_GEN_NONE; + spin_lock_init(&lli->lli_layout_lock); + ll_layout_version_set(lli, LL_LAYOUT_GEN_NONE); lli->lli_clob = NULL; init_rwsem(&lli->lli_xattrs_list_rwsem); @@ -958,32 +968,30 @@ void ll_lli_init(struct ll_inode_info *lli) mutex_init(&lli->lli_readdir_mutex); lli->lli_opendir_key = NULL; lli->lli_sai = NULL; - lli->lli_def_acl = NULL; spin_lock_init(&lli->lli_sa_lock); lli->lli_opendir_pid = 0; + lli->lli_sa_enabled = 0; } else { - sema_init(&lli->lli_size_sem, 1); - lli->lli_size_sem_owner = NULL; + mutex_init(&lli->lli_size_mutex); lli->lli_symlink_name = NULL; init_rwsem(&lli->lli_trunc_sem); - mutex_init(&lli->lli_write_mutex); + range_lock_tree_init(&lli->lli_write_tree); init_rwsem(&lli->lli_glimpse_sem); lli->lli_glimpse_time = 0; - CFS_INIT_LIST_HEAD(&lli->lli_agl_list); + INIT_LIST_HEAD(&lli->lli_agl_list); lli->lli_agl_index = 0; lli->lli_async_rc = 0; - lli->lli_volatile = false; } mutex_init(&lli->lli_layout_mutex); } static inline int ll_bdi_register(struct backing_dev_info *bdi) { - static atomic_t ll_bdi_num = ATOMIC_INIT(0); + static atomic_t ll_bdi_num = ATOMIC_INIT(0); - bdi->name = "lustre"; - return bdi_register(bdi, NULL, "lustre-%d", - atomic_inc_return(&ll_bdi_num)); + bdi->name = "lustre"; + return bdi_register(bdi, NULL, "lustre-%d", + atomic_inc_return(&ll_bdi_num)); } int ll_fill_super(struct super_block *sb, struct vfsmount *mnt) @@ -1005,15 +1013,15 @@ int ll_fill_super(struct super_block *sb, struct vfsmount *mnt) if (cfg == NULL) RETURN(-ENOMEM); - cfs_module_get(); + try_module_get(THIS_MODULE); - /* client additional sb info */ - lsi->lsi_llsbi = sbi = ll_init_sbi(); - if (!sbi) { - cfs_module_put(THIS_MODULE); - OBD_FREE_PTR(cfg); - RETURN(-ENOMEM); - } + /* client additional sb info */ + lsi->lsi_llsbi = sbi = ll_init_sbi(); + if (!sbi) { + module_put(THIS_MODULE); + OBD_FREE_PTR(cfg); + RETURN(-ENOMEM); + } err = ll_options(lsi->lsi_lmd->lmd_opts, &sbi->ll_flags); if (err) @@ -1042,10 +1050,8 @@ int ll_fill_super(struct super_block *sb, struct vfsmount *mnt) cfg->cfg_callback = class_config_llog_handler; /* set up client obds */ err = lustre_process_log(sb, profilenm, cfg); - if (err < 0) { - CERROR("Unable to process log: %d\n", err); - GOTO(out_free, err); - } + if (err < 0) + GOTO(out_free, err); /* Profile set with LCFG_MOUNTOPT so we can find our mdc and osc obds */ lprof = class_get_profile(profilenm); @@ -1085,9 +1091,6 @@ out_free: RETURN(err); } /* ll_fill_super */ - -void lu_context_keys_dump(void); - void ll_put_super(struct super_block *sb) { struct config_llog_instance cfg, params_cfg; @@ -1095,7 +1098,8 @@ void ll_put_super(struct super_block *sb) struct lustre_sb_info *lsi = s2lsi(sb); struct ll_sb_info *sbi = ll_s2sbi(sb); char *profilenm = get_profile_name(sb); - int force = 1, next; + long ccc_count; + int next, force = 1, rc = 0; ENTRY; CDEBUG(D_VFSTRACE, "VFS Op: sb %p - %s\n", sb, profilenm); @@ -1114,6 +1118,19 @@ void ll_put_super(struct super_block *sb) force = obd->obd_force; } + /* Wait for unstable pages to be committed to stable storage */ + if (force == 0) { + struct l_wait_info lwi = LWI_INTR(LWI_ON_SIGNAL_NOOP, NULL); + rc = l_wait_event(sbi->ll_cache.ccc_unstable_waitq, + atomic_long_read(&sbi->ll_cache.ccc_unstable_nr) == 0, + &lwi); + } + + ccc_count = atomic_long_read(&sbi->ll_cache.ccc_unstable_nr); + if (force == 0 && rc != -EINTR) + LASSERTF(ccc_count == 0, "count: %li\n", ccc_count); + + /* We need to set force before the lov_disconnect in lustre_common_put_super, since l_d cleans up osc's as well. */ if (force) { @@ -1148,11 +1165,13 @@ void ll_put_super(struct super_block *sb) ll_free_sbi(sb); lsi->lsi_llsbi = NULL; - lustre_common_put_super(sb); + lustre_common_put_super(sb); - cfs_module_put(THIS_MODULE); + cl_env_cache_purge(~0); - EXIT; + module_put(THIS_MODULE); + + EXIT; } /* client_put_super */ struct inode *ll_inode_from_resource_lock(struct ldlm_lock *lock) @@ -1180,26 +1199,209 @@ struct inode *ll_inode_from_resource_lock(struct ldlm_lock *lock) return inode; } -struct inode *ll_inode_from_lock(struct ldlm_lock *lock) +static void ll_dir_clear_lsm_md(struct inode *inode) { - struct inode *inode = NULL; - /* NOTE: we depend on atomic igrab() -bzzz */ - lock_res_and_lock(lock); - if (lock->l_ast_data) { - struct ll_inode_info *lli = ll_i2info(lock->l_ast_data); - if (lli->lli_inode_magic == LLI_INODE_MAGIC) { - inode = igrab(lock->l_ast_data); - } else { - inode = lock->l_ast_data; - LDLM_DEBUG_LIMIT(inode->i_state & I_FREEING ? D_INFO : - D_WARNING, lock, "l_ast_data %p is " - "bogus: magic %08x", lock->l_ast_data, - lli->lli_inode_magic); - inode = NULL; - } - } - unlock_res_and_lock(lock); - return inode; + struct ll_inode_info *lli = ll_i2info(inode); + + LASSERT(S_ISDIR(inode->i_mode)); + + if (lli->lli_lsm_md != NULL) { + lmv_free_memmd(lli->lli_lsm_md); + lli->lli_lsm_md = NULL; + } +} + +static struct inode *ll_iget_anon_dir(struct super_block *sb, + const struct lu_fid *fid, + struct lustre_md *md) +{ + struct ll_sb_info *sbi = ll_s2sbi(sb); + struct mdt_body *body = md->body; + struct inode *inode; + ino_t ino; + ENTRY; + + ino = cl_fid_build_ino(fid, sbi->ll_flags & LL_SBI_32BIT_API); + inode = iget_locked(sb, ino); + if (inode == NULL) { + CERROR("%s: failed get simple inode "DFID": rc = -ENOENT\n", + ll_get_fsname(sb, NULL, 0), PFID(fid)); + RETURN(ERR_PTR(-ENOENT)); + } + + if (inode->i_state & I_NEW) { + struct ll_inode_info *lli = ll_i2info(inode); + struct lmv_stripe_md *lsm = md->lmv; + + inode->i_mode = (inode->i_mode & ~S_IFMT) | + (body->mbo_mode & S_IFMT); + LASSERTF(S_ISDIR(inode->i_mode), "Not slave inode "DFID"\n", + PFID(fid)); + + LTIME_S(inode->i_mtime) = 0; + LTIME_S(inode->i_atime) = 0; + LTIME_S(inode->i_ctime) = 0; + inode->i_rdev = 0; + + /* initializing backing dev info. */ + inode->i_mapping->backing_dev_info = + &s2lsi(inode->i_sb)->lsi_bdi; + inode->i_op = &ll_dir_inode_operations; + inode->i_fop = &ll_dir_operations; + lli->lli_fid = *fid; + ll_lli_init(lli); + + LASSERT(lsm != NULL); + /* master object FID */ + lli->lli_pfid = body->mbo_fid1; + CDEBUG(D_INODE, "lli %p slave "DFID" master "DFID"\n", + lli, PFID(fid), PFID(&lli->lli_pfid)); + unlock_new_inode(inode); + } + + RETURN(inode); +} + +static int ll_init_lsm_md(struct inode *inode, struct lustre_md *md) +{ + struct lu_fid *fid; + struct lmv_stripe_md *lsm = md->lmv; + int i; + + LASSERT(lsm != NULL); + /* XXX sigh, this lsm_root initialization should be in + * LMV layer, but it needs ll_iget right now, so we + * put this here right now. */ + for (i = 0; i < lsm->lsm_md_stripe_count; i++) { + fid = &lsm->lsm_md_oinfo[i].lmo_fid; + LASSERT(lsm->lsm_md_oinfo[i].lmo_root == NULL); + /* Unfortunately ll_iget will call ll_update_inode, + * where the initialization of slave inode is slightly + * different, so it reset lsm_md to NULL to avoid + * initializing lsm for slave inode. */ + /* For migrating inode, master stripe and master object will + * be same, so we only need assign this inode */ + if (lsm->lsm_md_hash_type & LMV_HASH_FLAG_MIGRATION && i == 0) + lsm->lsm_md_oinfo[i].lmo_root = inode; + else + lsm->lsm_md_oinfo[i].lmo_root = + ll_iget_anon_dir(inode->i_sb, fid, md); + + if (IS_ERR(lsm->lsm_md_oinfo[i].lmo_root)) { + int rc = PTR_ERR(lsm->lsm_md_oinfo[i].lmo_root); + + lsm->lsm_md_oinfo[i].lmo_root = NULL; + return rc; + } + } + + /* Here is where the lsm is being initialized(fill lmo_info) after + * client retrieve MD stripe information from MDT. */ + return md_update_lsm_md(ll_i2mdexp(inode), lsm, md->body, + ll_md_blocking_ast); +} + +static inline int lli_lsm_md_eq(const struct lmv_stripe_md *lsm_md1, + const struct lmv_stripe_md *lsm_md2) +{ + return lsm_md1->lsm_md_magic == lsm_md2->lsm_md_magic && + lsm_md1->lsm_md_stripe_count == lsm_md2->lsm_md_stripe_count && + lsm_md1->lsm_md_master_mdt_index == + lsm_md2->lsm_md_master_mdt_index && + lsm_md1->lsm_md_hash_type == lsm_md2->lsm_md_hash_type && + lsm_md1->lsm_md_layout_version == + lsm_md2->lsm_md_layout_version && + strcmp(lsm_md1->lsm_md_pool_name, + lsm_md2->lsm_md_pool_name) == 0; +} + +static int ll_update_lsm_md(struct inode *inode, struct lustre_md *md) +{ + struct ll_inode_info *lli = ll_i2info(inode); + struct lmv_stripe_md *lsm = md->lmv; + int rc; + ENTRY; + + LASSERT(S_ISDIR(inode->i_mode)); + CDEBUG(D_INODE, "update lsm %p of "DFID"\n", lli->lli_lsm_md, + PFID(ll_inode2fid(inode))); + + /* no striped information from request. */ + if (lsm == NULL) { + if (lli->lli_lsm_md == NULL) { + RETURN(0); + } else if (lli->lli_lsm_md->lsm_md_hash_type & + LMV_HASH_FLAG_MIGRATION) { + /* migration is done, the temporay MIGRATE layout has + * been removed */ + CDEBUG(D_INODE, DFID" finish migration.\n", + PFID(ll_inode2fid(inode))); + lmv_free_memmd(lli->lli_lsm_md); + lli->lli_lsm_md = NULL; + RETURN(0); + } else { + /* The lustre_md from req does not include stripeEA, + * see ll_md_setattr */ + RETURN(0); + } + } + + /* set the directory layout */ + if (lli->lli_lsm_md == NULL) { + + rc = ll_init_lsm_md(inode, md); + if (rc != 0) + RETURN(rc); + + lli->lli_lsm_md = lsm; + /* set lsm_md to NULL, so the following free lustre_md + * will not free this lsm */ + md->lmv = NULL; + CDEBUG(D_INODE, "Set lsm %p magic %x to "DFID"\n", lsm, + lsm->lsm_md_magic, PFID(ll_inode2fid(inode))); + RETURN(0); + } + + /* Compare the old and new stripe information */ + if (!lsm_md_eq(lli->lli_lsm_md, lsm)) { + struct lmv_stripe_md *old_lsm = lli->lli_lsm_md; + int idx; + + CERROR("%s: inode "DFID"(%p)'s lmv layout mismatch (%p)/(%p)" + "magic:0x%x/0x%x stripe count: %d/%d master_mdt: %d/%d" + "hash_type:0x%x/0x%x layout: 0x%x/0x%x pool:%s/%s\n", + ll_get_fsname(inode->i_sb, NULL, 0), PFID(&lli->lli_fid), + inode, lsm, old_lsm, + lsm->lsm_md_magic, old_lsm->lsm_md_magic, + lsm->lsm_md_stripe_count, + old_lsm->lsm_md_stripe_count, + lsm->lsm_md_master_mdt_index, + old_lsm->lsm_md_master_mdt_index, + lsm->lsm_md_hash_type, old_lsm->lsm_md_hash_type, + lsm->lsm_md_layout_version, + old_lsm->lsm_md_layout_version, + lsm->lsm_md_pool_name, + old_lsm->lsm_md_pool_name); + + for (idx = 0; idx < old_lsm->lsm_md_stripe_count; idx++) { + CERROR("%s: sub FIDs in old lsm idx %d, old: "DFID"\n", + ll_get_fsname(inode->i_sb, NULL, 0), idx, + PFID(&old_lsm->lsm_md_oinfo[idx].lmo_fid)); + } + + for (idx = 0; idx < lsm->lsm_md_stripe_count; idx++) { + CERROR("%s: sub FIDs in new lsm idx %d, new: "DFID"\n", + ll_get_fsname(inode->i_sb, NULL, 0), idx, + PFID(&lsm->lsm_md_oinfo[idx].lmo_fid)); + } + + RETURN(-EIO); + } + + rc = md_update_lsm_md(ll_i2mdexp(inode), ll_i2info(inode)->lli_lsm_md, + md->body, ll_md_blocking_ast); + + RETURN(rc); } void ll_clear_inode(struct inode *inode) @@ -1208,8 +1410,8 @@ void ll_clear_inode(struct inode *inode) struct ll_sb_info *sbi = ll_i2sbi(inode); ENTRY; - CDEBUG(D_VFSTRACE, "VFS Op:inode=%lu/%u(%p)\n", inode->i_ino, - inode->i_generation, inode); + CDEBUG(D_VFSTRACE, "VFS Op:inode="DFID"(%p)\n", + PFID(ll_inode2fid(inode)), inode); if (S_ISDIR(inode->i_mode)) { /* these should have been cleared in ll_file_release */ @@ -1242,39 +1444,41 @@ void ll_clear_inode(struct inode *inode) ll_xattr_cache_destroy(inode); - if (sbi->ll_flags & LL_SBI_RMT_CLIENT) { - LASSERT(lli->lli_posix_acl == NULL); - if (lli->lli_remote_perms) { - free_rmtperm_hash(lli->lli_remote_perms); - lli->lli_remote_perms = NULL; - } - } + if (sbi->ll_flags & LL_SBI_RMT_CLIENT) { + LASSERT(lli->lli_posix_acl == NULL); + if (lli->lli_remote_perms) { + free_rmtperm_hash(lli->lli_remote_perms); + lli->lli_remote_perms = NULL; + } + } #ifdef CONFIG_FS_POSIX_ACL - else if (lli->lli_posix_acl) { - LASSERT(cfs_atomic_read(&lli->lli_posix_acl->a_refcount) == 1); - LASSERT(lli->lli_remote_perms == NULL); - posix_acl_release(lli->lli_posix_acl); - lli->lli_posix_acl = NULL; - } + else if (lli->lli_posix_acl) { + LASSERT(atomic_read(&lli->lli_posix_acl->a_refcount) == 1); + LASSERT(lli->lli_remote_perms == NULL); + posix_acl_release(lli->lli_posix_acl); + lli->lli_posix_acl = NULL; + } #endif - lli->lli_inode_magic = LLI_INODE_DEAD; + lli->lli_inode_magic = LLI_INODE_DEAD; - ll_clear_inode_capas(inode); - if (!S_ISDIR(inode->i_mode)) - LASSERT(cfs_list_empty(&lli->lli_agl_list)); + ll_clear_inode_capas(inode); + if (S_ISDIR(inode->i_mode)) + ll_dir_clear_lsm_md(inode); + else if (S_ISREG(inode->i_mode) && !is_bad_inode(inode)) + LASSERT(list_empty(&lli->lli_agl_list)); - /* - * XXX This has to be done before lsm is freed below, because - * cl_object still uses inode lsm. - */ - cl_inode_fini(inode); + /* + * XXX This has to be done before lsm is freed below, because + * cl_object still uses inode lsm. + */ + cl_inode_fini(inode); lli->lli_has_smd = false; EXIT; } -int ll_md_setattr(struct dentry *dentry, struct md_op_data *op_data, - struct md_open_data **mod) +static int ll_md_setattr(struct dentry *dentry, struct md_op_data *op_data, + struct md_open_data **mod) { struct lustre_md md; struct inode *inode = dentry->d_inode; @@ -1324,13 +1528,13 @@ int ll_md_setattr(struct dentry *dentry, struct md_op_data *op_data, op_data->op_attr.ia_valid = ia_valid; /* Extract epoch data if obtained. */ - op_data->op_handle = md.body->handle; - op_data->op_ioepoch = md.body->ioepoch; + op_data->op_handle = md.body->mbo_handle; + op_data->op_ioepoch = md.body->mbo_ioepoch; - ll_update_inode(inode, &md); - ptlrpc_req_finished(request); + rc = ll_update_inode(inode, &md); + ptlrpc_req_finished(request); - RETURN(rc); + RETURN(rc); } /* Close IO epoch and send Size-on-MDS attribute update. */ @@ -1359,8 +1563,9 @@ static int ll_setattr_done_writing(struct inode *inode, * from OSTs and send setattr to back to MDS. */ rc = ll_som_update(inode, op_data); } else if (rc) { - CERROR("inode %lu mdc truncate failed: rc = %d\n", - inode->i_ino, rc); + CERROR("%s: inode "DFID" mdc truncate failed: rc = %d\n", + ll_i2sbi(inode)->ll_md_exp->exp_obd->obd_name, + PFID(ll_inode2fid(inode)), rc); } RETURN(rc); } @@ -1407,14 +1612,14 @@ int ll_setattr_raw(struct dentry *dentry, struct iattr *attr, bool hsm_import) struct md_op_data *op_data = NULL; struct md_open_data *mod = NULL; bool file_is_released = false; - int rc = 0, rc1 = 0; - ENTRY; + int rc = 0, rc1 = 0; + ENTRY; - CDEBUG(D_VFSTRACE, "%s: setattr inode %p/fid:"DFID" from %llu to %llu, " - "valid %x, hsm_import %d\n", - ll_get_fsname(inode->i_sb, NULL, 0), inode, - PFID(&lli->lli_fid), i_size_read(inode), attr->ia_size, - attr->ia_valid, hsm_import); + CDEBUG(D_VFSTRACE, "%s: setattr inode "DFID"(%p) from %llu to %llu, " + "valid %x, hsm_import %d\n", + ll_get_fsname(inode->i_sb, NULL, 0), PFID(&lli->lli_fid), + inode, i_size_read(inode), attr->ia_size, attr->ia_valid, + hsm_import); if (attr->ia_valid & ATTR_SIZE) { /* Check new size against VFS/VM file size limit and rlimit */ @@ -1435,12 +1640,12 @@ int ll_setattr_raw(struct dentry *dentry, struct iattr *attr, bool hsm_import) attr->ia_valid |= ATTR_MTIME | ATTR_CTIME; } - /* POSIX: check before ATTR_*TIME_SET set (from inode_change_ok) */ + /* POSIX: check before ATTR_*TIME_SET set (from inode_change_ok) */ if (attr->ia_valid & TIMES_SET_FLAGS) { - if (cfs_curproc_fsuid() != inode->i_uid && - !cfs_capable(CFS_CAP_FOWNER)) - RETURN(-EPERM); - } + if ((!uid_eq(current_fsuid(), inode->i_uid)) && + !cfs_capable(CFS_CAP_FOWNER)) + RETURN(-EPERM); + } /* We mark all of the fields "set" so MDS/OST does not re-set them */ if (attr->ia_valid & ATTR_CTIME) { @@ -1463,14 +1668,6 @@ int ll_setattr_raw(struct dentry *dentry, struct iattr *attr, bool hsm_import) LTIME_S(attr->ia_mtime), LTIME_S(attr->ia_ctime), cfs_time_current_sec()); - /* If we are changing file size, file content is modified, flag it. */ - if (attr->ia_valid & ATTR_SIZE) { - attr->ia_valid |= MDS_OPEN_OWNEROVERRIDE; - spin_lock(&lli->lli_lock); - lli->lli_flags |= LLIF_DATA_MODIFIED; - spin_unlock(&lli->lli_lock); - } - /* We always do an MDS RPC, even if we're only changing the size; * only the MDS knows whether truncate() should fail with -ETXTBUSY */ @@ -1482,16 +1679,8 @@ int ll_setattr_raw(struct dentry *dentry, struct iattr *attr, bool hsm_import) if (attr->ia_valid & ATTR_SIZE) inode_dio_write_done(inode); mutex_unlock(&inode->i_mutex); - down_write(&lli->lli_trunc_sem); } - memcpy(&op_data->op_attr, attr, sizeof(*attr)); - - /* Open epoch for truncate. */ - if (exp_connect_som(ll_i2mdexp(inode)) && - (attr->ia_valid & (ATTR_SIZE | ATTR_MTIME | ATTR_MTIME_SET))) - op_data->op_flags = MF_EPOCH_OPEN; - /* truncate on a released file must failed with -ENODATA, * so size must not be set on MDS for released file * but other attributes must be set @@ -1505,27 +1694,38 @@ int ll_setattr_raw(struct dentry *dentry, struct iattr *attr, bool hsm_import) if (lsm && lsm->lsm_pattern & LOV_PATTERN_F_RELEASED) file_is_released = true; ccc_inode_lsm_put(inode, lsm); - } - /* if not in HSM import mode, clear size attr for released file - * we clear the attribute send to MDT in op_data, not the original - * received from caller in attr which is used later to - * decide return code */ - if (file_is_released && (attr->ia_valid & ATTR_SIZE) && !hsm_import) - op_data->op_attr.ia_valid &= ~ATTR_SIZE; + if (!hsm_import && attr->ia_valid & ATTR_SIZE) { + if (file_is_released) { + rc = ll_layout_restore(inode, 0, attr->ia_size); + if (rc < 0) + GOTO(out, rc); - rc = ll_md_setattr(dentry, op_data, &mod); - if (rc) - GOTO(out, rc); + file_is_released = false; + ll_layout_refresh(inode, &gen); + } - /* truncate failed (only when non HSM import), others succeed */ - if (file_is_released) { - if ((attr->ia_valid & ATTR_SIZE) && !hsm_import) - GOTO(out, rc = -ENODATA); - else - GOTO(out, rc = 0); + /* If we are changing file size, file content is + * modified, flag it. */ + attr->ia_valid |= MDS_OPEN_OWNEROVERRIDE; + spin_lock(&lli->lli_lock); + lli->lli_flags |= LLIF_DATA_MODIFIED; + spin_unlock(&lli->lli_lock); + op_data->op_bias |= MDS_DATA_MODIFIED; + } } + memcpy(&op_data->op_attr, attr, sizeof(*attr)); + + /* Open epoch for truncate. */ + if (exp_connect_som(ll_i2mdexp(inode)) && !hsm_import && + (attr->ia_valid & (ATTR_SIZE | ATTR_MTIME | ATTR_MTIME_SET))) + op_data->op_flags = MF_EPOCH_OPEN; + + rc = ll_md_setattr(dentry, op_data, &mod); + if (rc) + GOTO(out, rc); + /* RPC to MDT is sent, cancel data modification flag */ if (rc == 0 && (op_data->op_bias & MDS_DATA_MODIFIED)) { spin_lock(&lli->lli_lock); @@ -1533,33 +1733,36 @@ int ll_setattr_raw(struct dentry *dentry, struct iattr *attr, bool hsm_import) spin_unlock(&lli->lli_lock); } - ll_ioepoch_open(lli, op_data->op_ioepoch); - if (!S_ISREG(inode->i_mode)) - GOTO(out, rc = 0); + ll_ioepoch_open(lli, op_data->op_ioepoch); + if (!S_ISREG(inode->i_mode) || file_is_released) + GOTO(out, rc = 0); if (attr->ia_valid & (ATTR_SIZE | ATTR_ATIME | ATTR_ATIME_SET | ATTR_MTIME | ATTR_MTIME_SET)) { - /* For truncate and utimes sending attributes to OSTs, setting - * mtime/atime to the past will be performed under PW [0:EOF] - * extent lock (new_size:EOF for truncate). It may seem - * excessive to send mtime/atime updates to OSTs when not - * setting times to past, but it is necessary due to possible - * time de-synchronization between MDT inode and OST objects */ - rc = ll_setattr_ost(inode, attr); + /* For truncate and utimes sending attributes to OSTs, setting + * mtime/atime to the past will be performed under PW [0:EOF] + * extent lock (new_size:EOF for truncate). It may seem + * excessive to send mtime/atime updates to OSTs when not + * setting times to past, but it is necessary due to possible + * time de-synchronization between MDT inode and OST objects */ + if (attr->ia_valid & ATTR_SIZE) + down_write(&lli->lli_trunc_sem); + rc = ll_setattr_ost(inode, attr); + if (attr->ia_valid & ATTR_SIZE) + up_write(&lli->lli_trunc_sem); } - EXIT; + EXIT; out: - if (op_data) { - if (op_data->op_ioepoch) { - rc1 = ll_setattr_done_writing(inode, op_data, mod); - if (!rc) - rc = rc1; - } - ll_finish_md_op_data(op_data); - } + if (op_data) { + if (op_data->op_ioepoch) { + rc1 = ll_setattr_done_writing(inode, op_data, mod); + if (!rc) + rc = rc1; + } + ll_finish_md_op_data(op_data); + } if (!S_ISDIR(inode->i_mode)) { - up_write(&lli->lli_trunc_sem); mutex_lock(&inode->i_mutex); if ((attr->ia_valid & ATTR_SIZE) && !hsm_import) inode_dio_wait(inode); @@ -1586,12 +1789,14 @@ int ll_setattr(struct dentry *de, struct iattr *attr) !(attr->ia_mode & S_ISGID)))) attr->ia_valid |= ATTR_FORCE; - if ((mode & S_ISUID) && + if ((attr->ia_valid & ATTR_MODE) && + (mode & S_ISUID) && !(attr->ia_mode & S_ISUID) && !(attr->ia_valid & ATTR_KILL_SUID)) attr->ia_valid |= ATTR_KILL_SUID; - if (((mode & (S_ISGID|S_IXGRP)) == (S_ISGID|S_IXGRP)) && + if ((attr->ia_valid & ATTR_MODE) && + ((mode & (S_ISGID|S_IXGRP)) == (S_ISGID|S_IXGRP)) && !(attr->ia_mode & S_ISGID) && !(attr->ia_valid & ATTR_KILL_SGID)) attr->ia_valid |= ATTR_KILL_SGID; @@ -1652,6 +1857,7 @@ int ll_statfs(struct dentry *de, struct kstatfs *sfs) { struct super_block *sb = de->d_sb; struct obd_statfs osfs; + __u64 fsid = huge_encode_dev(sb->s_dev); int rc; CDEBUG(D_VFSTRACE, "VFS Op: at "LPU64" jiffies\n", get_jiffies_64()); @@ -1683,41 +1889,37 @@ int ll_statfs(struct dentry *de, struct kstatfs *sfs) sfs->f_blocks = osfs.os_blocks; sfs->f_bfree = osfs.os_bfree; sfs->f_bavail = osfs.os_bavail; - sfs->f_fsid = ll_s2sbi(sb)->ll_fsid; + sfs->f_fsid.val[0] = (__u32)fsid; + sfs->f_fsid.val[1] = (__u32)(fsid >> 32); return 0; } void ll_inode_size_lock(struct inode *inode) { - struct ll_inode_info *lli; + struct ll_inode_info *lli; - LASSERT(!S_ISDIR(inode->i_mode)); + LASSERT(!S_ISDIR(inode->i_mode)); - lli = ll_i2info(inode); - LASSERT(lli->lli_size_sem_owner != current); - down(&lli->lli_size_sem); - LASSERT(lli->lli_size_sem_owner == NULL); - lli->lli_size_sem_owner = current; + lli = ll_i2info(inode); + mutex_lock(&lli->lli_size_mutex); } void ll_inode_size_unlock(struct inode *inode) { - struct ll_inode_info *lli; + struct ll_inode_info *lli; - lli = ll_i2info(inode); - LASSERT(lli->lli_size_sem_owner == current); - lli->lli_size_sem_owner = NULL; - up(&lli->lli_size_sem); + lli = ll_i2info(inode); + mutex_unlock(&lli->lli_size_mutex); } -void ll_update_inode(struct inode *inode, struct lustre_md *md) +int ll_update_inode(struct inode *inode, struct lustre_md *md) { struct ll_inode_info *lli = ll_i2info(inode); struct mdt_body *body = md->body; struct lov_stripe_md *lsm = md->lsm; struct ll_sb_info *sbi = ll_i2sbi(inode); - LASSERT ((lsm != NULL) == ((body->valid & OBD_MD_FLEASIZE) != 0)); + LASSERT((lsm != NULL) == ((body->mbo_valid & OBD_MD_FLEASIZE) != 0)); if (lsm != NULL) { if (!lli->lli_has_smd && !(sbi->ll_flags & LL_SBI_LAYOUT_LOCK)) @@ -1728,12 +1930,20 @@ void ll_update_inode(struct inode *inode, struct lustre_md *md) lli->lli_maxbytes = MAX_LFS_FILESIZE; } - if (sbi->ll_flags & LL_SBI_RMT_CLIENT) { - if (body->valid & OBD_MD_FLRMTPERM) - ll_update_remote_perm(inode, md->remote_perm); - } + if (S_ISDIR(inode->i_mode)) { + int rc; + + rc = ll_update_lsm_md(inode, md); + if (rc != 0) + return rc; + } + + if (sbi->ll_flags & LL_SBI_RMT_CLIENT) { + if (body->mbo_valid & OBD_MD_FLRMTPERM) + ll_update_remote_perm(inode, md->remote_perm); + } #ifdef CONFIG_FS_POSIX_ACL - else if (body->valid & OBD_MD_FLACL) { + else if (body->mbo_valid & OBD_MD_FLACL) { spin_lock(&lli->lli_lock); if (lli->lli_posix_acl) posix_acl_release(lli->lli_posix_acl); @@ -1741,65 +1951,74 @@ void ll_update_inode(struct inode *inode, struct lustre_md *md) spin_unlock(&lli->lli_lock); } #endif - inode->i_ino = cl_fid_build_ino(&body->fid1, + inode->i_ino = cl_fid_build_ino(&body->mbo_fid1, sbi->ll_flags & LL_SBI_32BIT_API); - inode->i_generation = cl_fid_build_gen(&body->fid1); + inode->i_generation = cl_fid_build_gen(&body->mbo_fid1); - if (body->valid & OBD_MD_FLATIME) { - if (body->atime > LTIME_S(inode->i_atime)) - LTIME_S(inode->i_atime) = body->atime; - lli->lli_lvb.lvb_atime = body->atime; - } - if (body->valid & OBD_MD_FLMTIME) { - if (body->mtime > LTIME_S(inode->i_mtime)) { - CDEBUG(D_INODE, "setting ino %lu mtime from %lu " - "to "LPU64"\n", inode->i_ino, - LTIME_S(inode->i_mtime), body->mtime); - LTIME_S(inode->i_mtime) = body->mtime; - } - lli->lli_lvb.lvb_mtime = body->mtime; - } - if (body->valid & OBD_MD_FLCTIME) { - if (body->ctime > LTIME_S(inode->i_ctime)) - LTIME_S(inode->i_ctime) = body->ctime; - lli->lli_lvb.lvb_ctime = body->ctime; - } - if (body->valid & OBD_MD_FLMODE) - inode->i_mode = (inode->i_mode & S_IFMT)|(body->mode & ~S_IFMT); - if (body->valid & OBD_MD_FLTYPE) - inode->i_mode = (inode->i_mode & ~S_IFMT)|(body->mode & S_IFMT); - LASSERT(inode->i_mode != 0); - if (S_ISREG(inode->i_mode)) { - inode->i_blkbits = min(PTLRPC_MAX_BRW_BITS + 1, LL_MAX_BLKSIZE_BITS); - } else { - inode->i_blkbits = inode->i_sb->s_blocksize_bits; - } - if (body->valid & OBD_MD_FLUID) - inode->i_uid = body->uid; - if (body->valid & OBD_MD_FLGID) - inode->i_gid = body->gid; - if (body->valid & OBD_MD_FLFLAGS) - inode->i_flags = ll_ext_to_inode_flags(body->flags); - if (body->valid & OBD_MD_FLNLINK) - set_nlink(inode, body->nlink); - if (body->valid & OBD_MD_FLRDEV) - inode->i_rdev = old_decode_dev(body->rdev); - - if (body->valid & OBD_MD_FLID) { - /* FID shouldn't be changed! */ - if (fid_is_sane(&lli->lli_fid)) { - LASSERTF(lu_fid_eq(&lli->lli_fid, &body->fid1), - "Trying to change FID "DFID - " to the "DFID", inode %lu/%u(%p)\n", - PFID(&lli->lli_fid), PFID(&body->fid1), - inode->i_ino, inode->i_generation, inode); - } else - lli->lli_fid = body->fid1; - } + if (body->mbo_valid & OBD_MD_FLATIME) { + if (body->mbo_atime > LTIME_S(inode->i_atime)) + LTIME_S(inode->i_atime) = body->mbo_atime; + lli->lli_lvb.lvb_atime = body->mbo_atime; + } + + if (body->mbo_valid & OBD_MD_FLMTIME) { + if (body->mbo_mtime > LTIME_S(inode->i_mtime)) { + CDEBUG(D_INODE, "setting ino %lu mtime from %lu " + "to "LPU64"\n", inode->i_ino, + LTIME_S(inode->i_mtime), body->mbo_mtime); + LTIME_S(inode->i_mtime) = body->mbo_mtime; + } + lli->lli_lvb.lvb_mtime = body->mbo_mtime; + } + + if (body->mbo_valid & OBD_MD_FLCTIME) { + if (body->mbo_ctime > LTIME_S(inode->i_ctime)) + LTIME_S(inode->i_ctime) = body->mbo_ctime; + lli->lli_lvb.lvb_ctime = body->mbo_ctime; + } + + if (body->mbo_valid & OBD_MD_FLMODE) + inode->i_mode = (inode->i_mode & S_IFMT) | + (body->mbo_mode & ~S_IFMT); + + if (body->mbo_valid & OBD_MD_FLTYPE) + inode->i_mode = (inode->i_mode & ~S_IFMT) | + (body->mbo_mode & S_IFMT); + + LASSERT(inode->i_mode != 0); + if (S_ISREG(inode->i_mode)) + inode->i_blkbits = min(PTLRPC_MAX_BRW_BITS + 1, + LL_MAX_BLKSIZE_BITS); + else + inode->i_blkbits = inode->i_sb->s_blocksize_bits; + + if (body->mbo_valid & OBD_MD_FLUID) + inode->i_uid = make_kuid(&init_user_ns, body->mbo_uid); + if (body->mbo_valid & OBD_MD_FLGID) + inode->i_gid = make_kgid(&init_user_ns, body->mbo_gid); + if (body->mbo_valid & OBD_MD_FLFLAGS) + inode->i_flags = ll_ext_to_inode_flags(body->mbo_flags); + if (body->mbo_valid & OBD_MD_FLNLINK) + set_nlink(inode, body->mbo_nlink); + if (body->mbo_valid & OBD_MD_FLRDEV) + inode->i_rdev = old_decode_dev(body->mbo_rdev); + + if (body->mbo_valid & OBD_MD_FLID) { + /* FID shouldn't be changed! */ + if (fid_is_sane(&lli->lli_fid)) { + LASSERTF(lu_fid_eq(&lli->lli_fid, &body->mbo_fid1), + "Trying to change FID "DFID + " to the "DFID", inode "DFID"(%p)\n", + PFID(&lli->lli_fid), PFID(&body->mbo_fid1), + PFID(ll_inode2fid(inode)), inode); + } else { + lli->lli_fid = body->mbo_fid1; + } + } - LASSERT(fid_seq(&lli->lli_fid) != 0); + LASSERT(fid_seq(&lli->lli_fid) != 0); - if (body->valid & OBD_MD_FLSIZE) { + if (body->mbo_valid & OBD_MD_FLSIZE) { if (exp_connect_som(ll_i2mdexp(inode)) && S_ISREG(inode->i_mode)) { struct lustre_handle lockh; @@ -1817,14 +2036,16 @@ void ll_update_inode(struct inode *inode, struct lustre_md *md) if (lli->lli_flags & (LLIF_DONE_WRITING | LLIF_EPOCH_PENDING | LLIF_SOM_DIRTY)) { - CERROR("ino %lu flags %u still has " - "size authority! do not trust " - "the size got from MDS\n", - inode->i_ino, lli->lli_flags); + CERROR("%s: inode "DFID" flags %u still" + " has size authority! do not " + "trust the size from MDS\n", + sbi->ll_md_exp->exp_obd->obd_name, + PFID(ll_inode2fid(inode)), + lli->lli_flags); } else { /* Use old size assignment to avoid * deadlock bz14138 & bz14326 */ - i_size_write(inode, body->size); + i_size_write(inode, body->mbo_size); spin_lock(&lli->lli_lock); lli->lli_flags |= LLIF_MDS_SIZE_LOCK; spin_unlock(&lli->lli_lock); @@ -1834,35 +2055,41 @@ void ll_update_inode(struct inode *inode, struct lustre_md *md) } else { /* Use old size assignment to avoid * deadlock bz14138 & bz14326 */ - i_size_write(inode, body->size); + i_size_write(inode, body->mbo_size); - CDEBUG(D_VFSTRACE, "inode=%lu, updating i_size %llu\n", - inode->i_ino, (unsigned long long)body->size); - } + CDEBUG(D_VFSTRACE, + "inode="DFID", updating i_size %llu\n", + PFID(ll_inode2fid(inode)), + (unsigned long long)body->mbo_size); + } - if (body->valid & OBD_MD_FLBLOCKS) - inode->i_blocks = body->blocks; - } + if (body->mbo_valid & OBD_MD_FLBLOCKS) + inode->i_blocks = body->mbo_blocks; + } - if (body->valid & OBD_MD_FLMDSCAPA) { - LASSERT(md->mds_capa); - ll_add_capa(inode, md->mds_capa); - } - if (body->valid & OBD_MD_FLOSSCAPA) { - LASSERT(md->oss_capa); - ll_add_capa(inode, md->oss_capa); - } + if (body->mbo_valid & OBD_MD_FLMDSCAPA) { + LASSERT(md->mds_capa); + ll_add_capa(inode, md->mds_capa); + } - if (body->valid & OBD_MD_TSTATE) { - if (body->t_state & MS_RESTORE) + if (body->mbo_valid & OBD_MD_FLOSSCAPA) { + LASSERT(md->oss_capa); + ll_add_capa(inode, md->oss_capa); + } + + if (body->mbo_valid & OBD_MD_TSTATE) { + if (body->mbo_t_state & MS_RESTORE) lli->lli_flags |= LLIF_FILE_RESTORING; } + + return 0; } -void ll_read_inode2(struct inode *inode, void *opaque) +int ll_read_inode2(struct inode *inode, void *opaque) { struct lustre_md *md = opaque; struct ll_inode_info *lli = ll_i2info(inode); + int rc; ENTRY; CDEBUG(D_VFSTRACE, "VFS Op:inode="DFID"(%p)\n", @@ -1878,7 +2105,9 @@ void ll_read_inode2(struct inode *inode, void *opaque) LTIME_S(inode->i_atime) = 0; LTIME_S(inode->i_ctime) = 0; inode->i_rdev = 0; - ll_update_inode(inode, md); + rc = ll_update_inode(inode, md); + if (rc != 0) + RETURN(rc); /* OIDEBUG(inode); */ @@ -1907,6 +2136,8 @@ void ll_read_inode2(struct inode *inode, void *opaque) EXIT; } + + return 0; } void ll_delete_inode(struct inode *inode) @@ -1915,24 +2146,15 @@ void ll_delete_inode(struct inode *inode) ENTRY; if (S_ISREG(inode->i_mode) && lli->lli_clob != NULL) - /* discard all dirty pages before truncating them, required by - * osc_extent implementation at LU-1030. */ - cl_sync_file_range(inode, 0, OBD_OBJECT_EOF, - CL_FSYNC_DISCARD, 1); - - truncate_inode_pages(&inode->i_data, 0); - - /* Workaround for LU-118 */ - if (inode->i_data.nrpages) { - TREE_READ_LOCK_IRQ(&inode->i_data); - TREE_READ_UNLOCK_IRQ(&inode->i_data); - LASSERTF(inode->i_data.nrpages == 0, - "inode=%lu/%u(%p) nrpages=%lu, see " - "http://jira.whamcloud.com/browse/LU-118\n", - inode->i_ino, inode->i_generation, inode, - inode->i_data.nrpages); - } - /* Workaround end */ + /* It is last chance to write out dirty pages, + * otherwise we may lose data while umount */ + cl_sync_file_range(inode, 0, OBD_OBJECT_EOF, CL_FSYNC_LOCAL, 1); + + truncate_inode_pages_final(&inode->i_data); + + LASSERTF(inode->i_data.nrpages == 0, "inode="DFID"(%p) nrpages=%lu, " + "see https://jira.hpdd.intel.com/browse/LU-118\n", + PFID(ll_inode2fid(inode)), inode, inode->i_data.nrpages); #ifdef HAVE_SBOPS_EVICT_INODE ll_clear_inode(inode); @@ -1965,32 +2187,34 @@ int ll_iocontrol(struct inode *inode, struct file *file, rc = md_getattr(sbi->ll_md_exp, op_data, &req); ll_finish_md_op_data(op_data); if (rc) { - CERROR("failure %d inode %lu\n", rc, inode->i_ino); + CERROR("%s: failure inode "DFID": rc = %d\n", + sbi->ll_md_exp->exp_obd->obd_name, + PFID(ll_inode2fid(inode)), rc); RETURN(-abs(rc)); } body = req_capsule_server_get(&req->rq_pill, &RMF_MDT_BODY); - flags = body->flags; + flags = body->mbo_flags; ptlrpc_req_finished(req); - RETURN(put_user(flags, (int *)arg)); + RETURN(put_user(flags, (int __user *)arg)); } case FSFILT_IOC_SETFLAGS: { struct lov_stripe_md *lsm; struct obd_info oinfo = { { { 0 } } }; struct md_op_data *op_data; - if (get_user(flags, (int *)arg)) - RETURN(-EFAULT); + if (get_user(flags, (int __user *)arg)) + RETURN(-EFAULT); op_data = ll_prep_md_op_data(NULL, inode, NULL, NULL, 0, 0, LUSTRE_OPC_ANY, NULL); if (IS_ERR(op_data)) RETURN(PTR_ERR(op_data)); - ((struct ll_iattr *)&op_data->op_attr)->ia_attr_flags = flags; + op_data->op_attr_flags = flags; op_data->op_attr.ia_valid |= ATTR_ATTR_FLAG; rc = md_setattr(sbi->ll_md_exp, op_data, NULL, 0, NULL, 0, &req, NULL); @@ -2038,38 +2262,39 @@ int ll_iocontrol(struct inode *inode, struct file *file, int ll_flush_ctx(struct inode *inode) { - struct ll_sb_info *sbi = ll_i2sbi(inode); + struct ll_sb_info *sbi = ll_i2sbi(inode); - CDEBUG(D_SEC, "flush context for user %d\n", cfs_curproc_uid()); + CDEBUG(D_SEC, "flush context for user %d\n", + from_kuid(&init_user_ns, current_uid())); - obd_set_info_async(NULL, sbi->ll_md_exp, - sizeof(KEY_FLUSH_CTX), KEY_FLUSH_CTX, - 0, NULL, NULL); - obd_set_info_async(NULL, sbi->ll_dt_exp, - sizeof(KEY_FLUSH_CTX), KEY_FLUSH_CTX, - 0, NULL, NULL); - return 0; + obd_set_info_async(NULL, sbi->ll_md_exp, + sizeof(KEY_FLUSH_CTX), KEY_FLUSH_CTX, + 0, NULL, NULL); + obd_set_info_async(NULL, sbi->ll_dt_exp, + sizeof(KEY_FLUSH_CTX), KEY_FLUSH_CTX, + 0, NULL, NULL); + return 0; } /* umount -f client means force down, don't save state */ void ll_umount_begin(struct super_block *sb) { - struct ll_sb_info *sbi = ll_s2sbi(sb); - struct obd_device *obd; - struct obd_ioctl_data *ioc_data; - ENTRY; + struct ll_sb_info *sbi = ll_s2sbi(sb); + struct obd_device *obd; + struct obd_ioctl_data *ioc_data; + ENTRY; - CDEBUG(D_VFSTRACE, "VFS Op: superblock %p count %d active %d\n", sb, - sb->s_count, atomic_read(&sb->s_active)); + CDEBUG(D_VFSTRACE, "VFS Op: superblock %p count %d active %d\n", sb, + sb->s_count, atomic_read(&sb->s_active)); - obd = class_exp2obd(sbi->ll_md_exp); - if (obd == NULL) { - CERROR("Invalid MDC connection handle "LPX64"\n", - sbi->ll_md_exp->exp_handle.h_cookie); - EXIT; - return; - } - obd->obd_force = 1; + obd = class_exp2obd(sbi->ll_md_exp); + if (obd == NULL) { + CERROR("Invalid MDC connection handle "LPX64"\n", + sbi->ll_md_exp->exp_handle.h_cookie); + EXIT; + return; + } + obd->obd_force = 1; obd = class_exp2obd(sbi->ll_dt_exp); if (obd == NULL) { @@ -2091,12 +2316,12 @@ void ll_umount_begin(struct super_block *sb) OBD_FREE_PTR(ioc_data); } - /* Really, we'd like to wait until there are no requests outstanding, - * and then continue. For now, we just invalidate the requests, - * schedule() and sleep one second if needed, and hope. - */ - cfs_schedule(); - EXIT; + /* Really, we'd like to wait until there are no requests outstanding, + * and then continue. For now, we just invalidate the requests, + * schedule() and sleep one second if needed, and hope. + */ + schedule(); + EXIT; } int ll_remount_fs(struct super_block *sb, int *flags, char *data) @@ -2135,9 +2360,9 @@ int ll_prep_inode(struct inode **inode, struct ptlrpc_request *req, struct super_block *sb, struct lookup_intent *it) { struct ll_sb_info *sbi = NULL; - struct lustre_md md; - int rc; - ENTRY; + struct lustre_md md = { 0 }; + int rc; + ENTRY; LASSERT(*inode || sb); sbi = sb ? ll_s2sbi(sb) : ll_i2sbi(*inode); @@ -2146,21 +2371,23 @@ int ll_prep_inode(struct inode **inode, struct ptlrpc_request *req, if (rc) RETURN(rc); - if (*inode) { - ll_update_inode(*inode, &md); - } else { - LASSERT(sb != NULL); + if (*inode) { + rc = ll_update_inode(*inode, &md); + if (rc != 0) + GOTO(out, rc); + } else { + LASSERT(sb != NULL); /* * At this point server returns to client's same fid as client * generated for creating. So using ->fid1 is okay here. */ - LASSERT(fid_is_sane(&md.body->fid1)); + LASSERT(fid_is_sane(&md.body->mbo_fid1)); - *inode = ll_iget(sb, cl_fid_build_ino(&md.body->fid1, + *inode = ll_iget(sb, cl_fid_build_ino(&md.body->mbo_fid1, sbi->ll_flags & LL_SBI_32BIT_API), &md); - if (*inode == NULL || IS_ERR(*inode)) { + if (IS_ERR(*inode)) { #ifdef CONFIG_FS_POSIX_ACL if (md.posix_acl) { posix_acl_release(md.posix_acl); @@ -2209,14 +2436,14 @@ out: RETURN(rc); } -int ll_obd_statfs(struct inode *inode, void *arg) +int ll_obd_statfs(struct inode *inode, void __user *arg) { struct ll_sb_info *sbi = NULL; struct obd_export *exp; char *buf = NULL; struct obd_ioctl_data *data = NULL; __u32 type; - __u32 flags; + __u32 __user flags; /* not user, but obd_iocontrol is abused */ int len = 0, rc; if (!inode || !(sbi = ll_i2sbi(inode))) @@ -2257,43 +2484,49 @@ out_statfs: int ll_process_config(struct lustre_cfg *lcfg) { - char *ptr; - void *sb; - struct lprocfs_static_vars lvars; - unsigned long x; - int rc = 0; + struct super_block *sb; + unsigned long x; + int rc = 0; + char *ptr; - lprocfs_llite_init_vars(&lvars); - - /* The instance name contains the sb: lustre-client-aacfe000 */ - ptr = strrchr(lustre_cfg_string(lcfg, 0), '-'); - if (!ptr || !*(++ptr)) - return -EINVAL; - if (sscanf(ptr, "%lx", &x) != 1) - return -EINVAL; - sb = (void *)x; - /* This better be a real Lustre superblock! */ - LASSERT(s2lsi((struct super_block *)sb)->lsi_lmd->lmd_magic == LMD_MAGIC); - - /* Note we have not called client_common_fill_super yet, so - proc fns must be able to handle that! */ - rc = class_process_proc_param(PARAM_LLITE, lvars.obd_vars, - lcfg, sb); - if (rc > 0) - rc = 0; - return(rc); + /* The instance name contains the sb: lustre-client-aacfe000 */ + ptr = strrchr(lustre_cfg_string(lcfg, 0), '-'); + if (!ptr || !*(++ptr)) + return -EINVAL; + if (sscanf(ptr, "%lx", &x) != 1) + return -EINVAL; + sb = (struct super_block *)x; + /* This better be a real Lustre superblock! */ + LASSERT(s2lsi(sb)->lsi_lmd->lmd_magic == LMD_MAGIC); + + /* Note we have not called client_common_fill_super yet, so + proc fns must be able to handle that! */ + rc = class_process_proc_param(PARAM_LLITE, lprocfs_llite_obd_vars, + lcfg, sb); + if (rc > 0) + rc = 0; + return rc; } /* this function prepares md_op_data hint for passing ot down to MD stack. */ struct md_op_data * ll_prep_md_op_data(struct md_op_data *op_data, - struct inode *i1, struct inode *i2, - const char *name, int namelen, - int mode, __u32 opc, void *data) + struct inode *i1, struct inode *i2, + const char *name, size_t namelen, + __u32 mode, __u32 opc, void *data) { LASSERT(i1 != NULL); - if (namelen > ll_i2sbi(i1)->ll_namelen) - return ERR_PTR(-ENAMETOOLONG); + if (name == NULL) { + /* Do not reuse namelen for something else. */ + if (namelen != 0) + return ERR_PTR(-EINVAL); + } else { + if (namelen > ll_i2sbi(i1)->ll_namelen) + return ERR_PTR(-ENAMETOOLONG); + + if (!lu_name_is_valid_2(name, namelen)) + return ERR_PTR(-EINVAL); + } if (op_data == NULL) OBD_ALLOC_PTR(op_data); @@ -2301,47 +2534,44 @@ struct md_op_data * ll_prep_md_op_data(struct md_op_data *op_data, if (op_data == NULL) return ERR_PTR(-ENOMEM); - ll_i2gids(op_data->op_suppgids, i1, i2); - op_data->op_fid1 = *ll_inode2fid(i1); - op_data->op_capa1 = ll_mdscapa_get(i1); + ll_i2gids(op_data->op_suppgids, i1, i2); + op_data->op_fid1 = *ll_inode2fid(i1); + op_data->op_capa1 = ll_mdscapa_get(i1); + if (S_ISDIR(i1->i_mode)) + op_data->op_mea1 = ll_i2info(i1)->lli_lsm_md; + + if (i2) { + op_data->op_fid2 = *ll_inode2fid(i2); + op_data->op_capa2 = ll_mdscapa_get(i2); + if (S_ISDIR(i2->i_mode)) + op_data->op_mea2 = ll_i2info(i2)->lli_lsm_md; + } else { + fid_zero(&op_data->op_fid2); + op_data->op_capa2 = NULL; + } + + if (ll_i2sbi(i1)->ll_flags & LL_SBI_64BIT_HASH) + op_data->op_cli_flags |= CLI_HASH64; - if (i2) { - op_data->op_fid2 = *ll_inode2fid(i2); - op_data->op_capa2 = ll_mdscapa_get(i2); - } else { - fid_zero(&op_data->op_fid2); - op_data->op_capa2 = NULL; - } + if (ll_need_32bit_api(ll_i2sbi(i1))) + op_data->op_cli_flags |= CLI_API32; op_data->op_name = name; op_data->op_namelen = namelen; op_data->op_mode = mode; op_data->op_mod_time = cfs_time_current_sec(); - op_data->op_fsuid = cfs_curproc_fsuid(); - op_data->op_fsgid = cfs_curproc_fsgid(); + op_data->op_fsuid = from_kuid(&init_user_ns, current_fsuid()); + op_data->op_fsgid = from_kgid(&init_user_ns, current_fsgid()); op_data->op_cap = cfs_curproc_cap_pack(); op_data->op_bias = 0; op_data->op_cli_flags = 0; if ((opc == LUSTRE_OPC_CREATE) && (name != NULL) && - filename_is_volatile(name, namelen, NULL)) + filename_is_volatile(name, namelen, &op_data->op_mds)) { op_data->op_bias |= MDS_CREATE_VOLATILE; - op_data->op_opc = opc; - op_data->op_mds = 0; - op_data->op_data = data; - - /* If the file is being opened after mknod() (normally due to NFS) - * try to use the default stripe data from parent directory for - * allocating OST objects. Try to pass the parent FID to MDS. */ - if (opc == LUSTRE_OPC_CREATE && i1 == i2 && S_ISREG(i2->i_mode) && - !ll_i2info(i2)->lli_has_smd) { - struct ll_inode_info *lli = ll_i2info(i2); - - spin_lock(&lli->lli_lock); - if (likely(!lli->lli_has_smd && !fid_is_zero(&lli->lli_pfid))) - op_data->op_fid1 = lli->lli_pfid; - spin_unlock(&lli->lli_lock); - /** We ignore parent's capability temporary. */ + } else { + op_data->op_mds = 0; } + op_data->op_data = data; /* When called by ll_setattr_raw, file is i1. */ if (LLIF_DATA_MODIFIED & ll_i2info(i1)->lli_flags) @@ -2413,7 +2643,7 @@ int ll_get_obd_name(struct inode *inode, unsigned int cmd, unsigned long arg) if (!obd) RETURN(-ENOENT); - if (copy_to_user((void *)arg, obd->obd_name, + if (copy_to_user((void __user *)arg, obd->obd_name, strlen(obd->obd_name) + 1)) RETURN(-EFAULT); @@ -2481,11 +2711,12 @@ void ll_dirty_page_discard_warn(struct page *page, int ioret) path = ll_d_path(dentry, buf, PAGE_SIZE); } - CWARN("%s: dirty page discard: %s/fid: "DFID"/%s may get corrupted " - "(rc %d)\n", ll_get_fsname(page->mapping->host->i_sb, NULL, 0), - s2lsi(page->mapping->host->i_sb)->lsi_lmd->lmd_dev, - PFID(&obj->cob_header.coh_lu.loh_fid), - (path && !IS_ERR(path)) ? path : "", ioret); + CDEBUG(D_WARNING, + "%s: dirty page discard: %s/fid: "DFID"/%s may get corrupted " + "(rc %d)\n", ll_get_fsname(page->mapping->host->i_sb, NULL, 0), + s2lsi(page->mapping->host->i_sb)->lsi_lmd->lmd_dev, + PFID(&obj->cob_header.coh_lu.loh_fid), + (path && !IS_ERR(path)) ? path : "", ioret); if (dentry != NULL) dput(dentry); @@ -2493,3 +2724,193 @@ void ll_dirty_page_discard_warn(struct page *page, int ioret) if (buf != NULL) free_page((unsigned long)buf); } + +ssize_t ll_copy_user_md(const struct lov_user_md __user *md, + struct lov_user_md **kbuf) +{ + struct lov_user_md lum; + ssize_t lum_size; + ENTRY; + + if (copy_from_user(&lum, md, sizeof(lum))) + RETURN(-EFAULT); + + lum_size = ll_lov_user_md_size(&lum); + if (lum_size < 0) + RETURN(lum_size); + + OBD_ALLOC(*kbuf, lum_size); + if (*kbuf == NULL) + RETURN(-ENOMEM); + + if (copy_from_user(*kbuf, md, lum_size) != 0) { + OBD_FREE(*kbuf, lum_size); + RETURN(-EFAULT); + } + + RETURN(lum_size); +} + +/* + * Compute llite root squash state after a change of root squash + * configuration setting or add/remove of a lnet nid + */ +void ll_compute_rootsquash_state(struct ll_sb_info *sbi) +{ + struct root_squash_info *squash = &sbi->ll_squash; + int i; + bool matched; + lnet_process_id_t id; + + /* Update norootsquash flag */ + down_write(&squash->rsi_sem); + if (list_empty(&squash->rsi_nosquash_nids)) + sbi->ll_flags &= ~LL_SBI_NOROOTSQUASH; + else { + /* Do not apply root squash as soon as one of our NIDs is + * in the nosquash_nids list */ + matched = false; + i = 0; + while (LNetGetId(i++, &id) != -ENOENT) { + if (LNET_NETTYP(LNET_NIDNET(id.nid)) == LOLND) + continue; + if (cfs_match_nid(id.nid, &squash->rsi_nosquash_nids)) { + matched = true; + break; + } + } + if (matched) + sbi->ll_flags |= LL_SBI_NOROOTSQUASH; + else + sbi->ll_flags &= ~LL_SBI_NOROOTSQUASH; + } + up_write(&squash->rsi_sem); +} + +/** + * Parse linkea content to extract information about a given hardlink + * + * \param[in] ldata - Initialized linkea data + * \param[in] linkno - Link identifier + * \param[out] gpout - Destination structure to fill with linkno, + * parent FID and entry name + * \param[in] size - Size of the gp_name buffer in gpout + * + * \retval 0 on success + * \retval Appropriate negative error code on failure + */ +static int ll_linkea_decode(struct linkea_data *ldata, unsigned int linkno, + struct getparent *gpout, size_t name_size) +{ + unsigned int idx; + struct lu_name ln; + int rc; + ENTRY; + + rc = linkea_init(ldata); + if (rc < 0) + RETURN(rc); + + if (linkno >= ldata->ld_leh->leh_reccount) + /* beyond last link */ + RETURN(-ENODATA); + + linkea_first_entry(ldata); + idx = 0; + while (ldata->ld_lee != NULL) { + linkea_entry_unpack(ldata->ld_lee, &ldata->ld_reclen, &ln, + &gpout->gp_fid); + if (idx == linkno) + break; + + linkea_next_entry(ldata); + idx++; + } + + if (idx < linkno) + RETURN(-ENODATA); + + if (ln.ln_namelen >= name_size) + RETURN(-EOVERFLOW); + + gpout->gp_linkno = linkno; + strlcpy(gpout->gp_name, ln.ln_name, name_size); + RETURN(0); +} + +/** + * Get parent FID and name of an identified link. Operation is performed for + * a given link number, letting the caller iterate over linkno to list one or + * all links of an entry. + * + * \param[in] file - File descriptor against which to perform the operation + * \param[in,out] arg - User-filled structure containing the linkno to operate + * on and the available size. It is eventually filled with + * the requested information or left untouched on error + * + * \retval - 0 on success + * \retval - Appropriate negative error code on failure + */ +int ll_getparent(struct file *file, struct getparent __user *arg) +{ + struct dentry *dentry = file->f_dentry; + struct inode *inode = file->f_dentry->d_inode; + struct linkea_data *ldata; + struct lu_buf buf = LU_BUF_NULL; + struct getparent *gpout; + __u32 linkno; + __u32 name_size; + size_t out_size; + int rc; + + ENTRY; + + if (!cfs_capable(CFS_CAP_DAC_READ_SEARCH) && + !(ll_i2sbi(inode)->ll_flags & LL_SBI_USER_FID2PATH)) + RETURN(-EPERM); + + if (get_user(name_size, &arg->gp_name_size)) + RETURN(-EFAULT); + + if (get_user(linkno, &arg->gp_linkno)) + RETURN(-EFAULT); + + if (name_size > PATH_MAX) + RETURN(-EINVAL); + + OBD_ALLOC(ldata, sizeof(*ldata)); + if (ldata == NULL) + RETURN(-ENOMEM); + + rc = linkea_data_new(ldata, &buf); + if (rc < 0) + GOTO(ldata_free, rc); + + out_size = sizeof(*gpout) + name_size; + OBD_ALLOC(gpout, out_size); + if (gpout == NULL) + GOTO(lb_free, rc = -ENOMEM); + + if (copy_from_user(gpout, arg, sizeof(*gpout))) + GOTO(gp_free, rc = -EFAULT); + + rc = ll_getxattr(dentry, XATTR_NAME_LINK, buf.lb_buf, buf.lb_len); + if (rc < 0) + GOTO(gp_free, rc); + + rc = ll_linkea_decode(ldata, linkno, gpout, name_size); + if (rc < 0) + GOTO(gp_free, rc); + + if (copy_to_user(arg, gpout, out_size)) + GOTO(gp_free, rc = -EFAULT); + +gp_free: + OBD_FREE(gpout, out_size); +lb_free: + lu_buf_free(&buf); +ldata_free: + OBD_FREE(ldata, sizeof(*ldata)); + + RETURN(rc); +}