X-Git-Url: https://git.whamcloud.com/?a=blobdiff_plain;f=lustre%2Fobdclass%2Fcapa.c;h=421df58daad66dce490bc989c84e113ee534eb79;hb=fbf5870b9848929d352460f1f005b79c0b5ccc5a;hp=5646ec13a4970c448819f5a2086e3db2cf402502;hpb=1c9cd4f867aa891ec579c32ccd24b707079468e6;p=fs%2Flustre-release.git diff --git a/lustre/obdclass/capa.c b/lustre/obdclass/capa.c index 5646ec1..421df58 100644 --- a/lustre/obdclass/capa.c +++ b/lustre/obdclass/capa.c @@ -1,27 +1,45 @@ /* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*- * vim:expandtab:shiftwidth=8:tabstop=8: * - * lustre/obdclass/capa.c - * Lustre Capability Cache Management + * GPL HEADER START * - * Copyright (c) 2001-2003 Cluster File Systems, Inc. - * Author: Lai Siyao + * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. * - * This file is part of Lustre, http://www.lustre.org. + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License version 2 only, + * as published by the Free Software Foundation. * - * Lustre is free software; you can redistribute it and/or - * modify it under the terms of version 2 of the GNU General Public - * License as published by the Free Software Foundation. + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License version 2 for more details (a copy is included + * in the LICENSE file that accompanied this code). * - * Lustre is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. + * You should have received a copy of the GNU General Public License + * version 2 along with this program; If not, see + * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf * - * You should have received a copy of the GNU General Public License - * along with Lustre; if not, write to the Free Software - * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara, + * CA 95054 USA or visit www.sun.com if you need additional information or + * have any questions. + * + * GPL HEADER END + */ +/* + * Copyright 2008 Sun Microsystems, Inc. All rights reserved + * Use is subject to license terms. */ +/* + * This file is part of Lustre, http://www.lustre.org/ + * Lustre is a trademark of Sun Microsystems, Inc. + * + * lustre/obdclass/capa.c + * + * Lustre Capability Hash Management + * + * Author: Lai Siyao + */ + #ifndef EXPORT_SYMTAB # define EXPORT_SYMTAB #endif @@ -36,429 +54,255 @@ #include #include -#include -#include -#include -#include +#include +#include +#include #else #include #endif #include -#include +#include -kmem_cache_t *capa_cachep = NULL; +#define NR_CAPAHASH 32 +#define CAPA_HASH_SIZE 3000 /* for MDS & OSS */ -/* capa_lock protect capa hash, list and content. */ +cfs_mem_cache_t *capa_cachep = NULL; + +#ifdef __KERNEL__ +/* lock for capa hash/capa_list/fo_capa_keys */ spinlock_t capa_lock = SPIN_LOCK_UNLOCKED; -struct hlist_head *capa_hash; -struct list_head capa_list[3]; -static int capa_count[3] = { 0 }; -static char *capa_type_name[] = { "client", "mds", "filter" }; +struct list_head capa_list[CAPA_SITE_MAX]; -/* TODO: mdc and llite all need this, so define it here. - * in the future it will be moved to ll_sb_info to support multi- - * mount point */ -struct timer_list ll_capa_timer; +static struct capa_hmac_alg capa_hmac_algs[] = { + DEF_CAPA_HMAC_ALG("sha1", SHA1, 20, 20), +}; +#endif +/* capa count */ +int capa_count[CAPA_SITE_MAX] = { 0, }; -EXPORT_SYMBOL(capa_lock); -EXPORT_SYMBOL(capa_hash); +EXPORT_SYMBOL(capa_cachep); EXPORT_SYMBOL(capa_list); -EXPORT_SYMBOL(ll_capa_timer); - -static inline int const -capa_hashfn(unsigned int uid, __u64 mdsid, unsigned long ino) -{ - return (ino ^ uid) * (unsigned long)(mdsid + 1) % NR_CAPAHASH; -} - -int capa_op(int flags) -{ - if (flags & (FMODE_WRITE|MDS_OPEN_TRUNC)) - return CAPA_WRITE; - else if (flags & FMODE_READ) - return CAPA_READ; - - LBUG(); /* should be either MAY_READ or MAY_WRITE */ - return 0; -} +EXPORT_SYMBOL(capa_lock); +EXPORT_SYMBOL(capa_count); -static struct obd_capa * -find_capa(struct hlist_head *head, uid_t uid, int capa_op, __u64 mdsid, - unsigned long ino, __u32 igen, int type) +struct hlist_head *init_capa_hash(void) { - struct hlist_node *pos; - struct obd_capa *ocapa; - uid_t ouid; - - CDEBUG(D_INODE, "find capa for (uid %u, op %d, mdsid "LPU64", ino %lu" - " igen %u, type %d\n", (unsigned) uid, capa_op, mdsid, ino, igen, type); - hlist_for_each_entry(ocapa, pos, head, c_hash) { - if (ocapa->c_capa.lc_ino != ino) - continue; - if (ocapa->c_capa.lc_igen != igen) - continue; - if (ocapa->c_capa.lc_mdsid != mdsid) - continue; - if ((ocapa->c_capa.lc_op & capa_op) != ocapa->c_capa.lc_op) - continue; - if (ocapa->c_type != type) - continue; - - if (ocapa->c_type == CLIENT_CAPA) - ouid = ocapa->c_capa.lc_ruid; - else - ouid = ocapa->c_capa.lc_uid; - - if (ouid != uid) - continue; + struct hlist_head *hash; + int nr_hash, i; - DEBUG_CAPA(D_INODE, &ocapa->c_capa, "found %s", - capa_type_name[ocapa->c_type]); + OBD_ALLOC(hash, CFS_PAGE_SIZE); + if (!hash) + return NULL; - return ocapa; - } + nr_hash = CFS_PAGE_SIZE / sizeof(struct hlist_head); + LASSERT(nr_hash > NR_CAPAHASH); - return NULL; + for (i = 0; i < NR_CAPAHASH; i++) + INIT_HLIST_HEAD(hash + i); + return hash; } -static struct obd_capa * -filter_find_capa(struct hlist_head *head, struct lustre_capa *capa) +#ifdef __KERNEL__ +static inline int capa_on_server(struct obd_capa *ocapa) { - struct hlist_node *pos; - struct obd_capa *ocapa; - - hlist_for_each_entry(ocapa, pos, head, c_hash) { - if (ocapa->c_type != FILTER_CAPA) - continue; - if (!memcmp(&ocapa->c_capa, capa, - sizeof(struct lustre_capa_data))) { - - DEBUG_CAPA(D_INODE, &ocapa->c_capa, "found %s", - capa_type_name[ocapa->c_type]); - return ocapa; - } - } - - return NULL; + return ocapa->c_site == CAPA_SITE_SERVER; } -inline void __capa_get(struct obd_capa *ocapa) +static inline void capa_delete(struct obd_capa *ocapa) { - if (ocapa->c_type != CLIENT_CAPA) - atomic_inc(&ocapa->c_refc); + LASSERT(capa_on_server(ocapa)); + hlist_del(&ocapa->u.tgt.c_hash); + list_del(&ocapa->c_list); + capa_count[ocapa->c_site]--; + free_capa(ocapa); } -static struct obd_capa * -find_capa_locked(struct hlist_head *head, uid_t uid, int capa_op, __u64 mdsid, - unsigned long ino, __u32 igen, int type) +void cleanup_capa_hash(struct hlist_head *hash) { - struct obd_capa *ocapa; + int i; + struct hlist_node *pos, *next; + struct obd_capa *oc; spin_lock(&capa_lock); - ocapa = find_capa(head, uid, capa_op, mdsid, ino, igen, type); - if (ocapa) - __capa_get(ocapa); - spin_unlock(&capa_lock); - - return ocapa; -} - -static struct obd_capa *alloc_capa(void) -{ - struct obd_capa *ocapa; - - OBD_SLAB_ALLOC(ocapa, capa_cachep, SLAB_NOFS, sizeof(*ocapa)); - if (ocapa) { - INIT_HLIST_NODE(&ocapa->c_hash); - INIT_LIST_HEAD(&ocapa->c_list); + for (i = 0; i < NR_CAPAHASH; i++) { + hlist_for_each_entry_safe(oc, pos, next, hash + i, u.tgt.c_hash) + capa_delete(oc); } + spin_unlock(&capa_lock); - return ocapa; + OBD_FREE(hash, CFS_PAGE_SIZE); } -static void __capa_put(struct obd_capa *ocapa) +static inline int capa_hashfn(struct lu_fid *fid) { - hlist_del_init(&ocapa->c_hash); - list_del_init(&ocapa->c_list); - capa_count[ocapa->c_type]--; + return (fid_oid(fid) ^ fid_ver(fid)) * + (unsigned long)(fid_seq(fid) + 1) % NR_CAPAHASH; } -static void destroy_capa(struct obd_capa *ocapa) +/* capa renewal time check is earlier than that on client, which is to prevent + * client renew right after obtaining it. */ +static inline int capa_is_to_expire(struct obd_capa *oc) { - OBD_SLAB_FREE(ocapa, capa_cachep, sizeof(*ocapa)); + return cfs_time_before(cfs_time_sub(oc->c_expiry, + cfs_time_seconds(oc->c_capa.lc_timeout)*2/3), + cfs_time_current()); } -int capa_cache_init(void) +static struct obd_capa *find_capa(struct lustre_capa *capa, + struct hlist_head *head, int alive) { - int nr_hash, i; - - OBD_ALLOC(capa_hash, PAGE_SIZE); - if (!capa_hash) - return -ENOMEM; - - nr_hash = PAGE_SIZE / sizeof(struct hlist_head); - LASSERT(nr_hash > NR_CAPAHASH); - - for (i = 0; i < NR_CAPAHASH; i++) - INIT_HLIST_HEAD(capa_hash + i); - - for (i = 0; i < 3; i++) - INIT_LIST_HEAD(&capa_list[i]); + struct hlist_node *pos; + struct obd_capa *ocapa; + int len = alive ? offsetof(struct lustre_capa, lc_keyid):sizeof(*capa); - return 0; -} + hlist_for_each_entry(ocapa, pos, head, u.tgt.c_hash) { + if (memcmp(&ocapa->c_capa, capa, len)) + continue; + /* don't return one that will expire soon in this case */ + if (alive && capa_is_to_expire(ocapa)) + continue; -void capa_cache_cleanup(void) -{ - struct obd_capa *ocapa, *tmp; - int i; + LASSERT(capa_on_server(ocapa)); - for (i = MDS_CAPA; i <= FILTER_CAPA; i++) { - list_for_each_entry_safe(ocapa, tmp, &capa_list[i], c_list) { - __capa_put(ocapa); - destroy_capa(ocapa); - } + DEBUG_CAPA(D_SEC, &ocapa->c_capa, "found"); + return ocapa; } - OBD_FREE(capa_hash, PAGE_SIZE); + return NULL; } - -static inline void list_add_capa(struct obd_capa *ocapa, struct list_head *head) +#define LRU_CAPA_DELETE_COUNT 12 +static inline void capa_delete_lru(struct list_head *head) { - struct obd_capa *tmp; - - /* XXX: capa is sorted in client, this could be optimized */ - if (ocapa->c_type == CLIENT_CAPA) { - list_for_each_entry_reverse(tmp, head, c_list) { - if (ocapa->c_capa.lc_expiry > tmp->c_capa.lc_expiry) { - list_add(&ocapa->c_list, &tmp->c_list); - return; - } - } - list_add(&ocapa->c_list, head); - return; - } - - list_add_tail(&ocapa->c_list, head); -} + struct obd_capa *ocapa; + struct list_head *node = head->next; + int count = 0; + + /* free LRU_CAPA_DELETE_COUNT unused capa from head */ + while (count++ < LRU_CAPA_DELETE_COUNT) { + ocapa = list_entry(node, struct obd_capa, c_list); + node = node->next; + if (atomic_read(&ocapa->c_refc)) + continue; -static inline void do_update_capa(struct obd_capa *ocapa, struct lustre_capa *capa) -{ - memcpy(&ocapa->c_capa, capa, sizeof(*capa)); + DEBUG_CAPA(D_SEC, &ocapa->c_capa, "free lru"); + capa_delete(ocapa); + } } -static struct obd_capa * -get_new_capa_locked(struct hlist_head *head, int type, struct lustre_capa *capa) +/* add or update */ +struct obd_capa *capa_add(struct hlist_head *hash, struct lustre_capa *capa) { - uid_t uid = capa->lc_uid; - int capa_op = capa->lc_op; - __u64 mdsid = capa->lc_mdsid; - unsigned long ino = capa->lc_ino; - struct obd_capa *ocapa, *old; + struct hlist_head *head = hash + capa_hashfn(&capa->lc_fid); + struct obd_capa *ocapa, *old = NULL; + struct list_head *list = &capa_list[CAPA_SITE_SERVER]; - ocapa = alloc_capa(); + ocapa = alloc_capa(CAPA_SITE_SERVER); if (!ocapa) return NULL; spin_lock(&capa_lock); + old = find_capa(capa, head, 0); + if (!old) { + ocapa->c_capa = *capa; + set_capa_expiry(ocapa); + hlist_add_head(&ocapa->u.tgt.c_hash, head); + list_add_tail(&ocapa->c_list, list); + capa_count[CAPA_SITE_SERVER]++; + capa_get(ocapa); - if (type == FILTER_CAPA) - old = filter_find_capa(head, capa); - else - old = find_capa(head, uid, capa_op, mdsid, ino, - capa->lc_igen, type); + if (capa_count[CAPA_SITE_SERVER] > CAPA_HASH_SIZE) + capa_delete_lru(list); - if (!old) { - do_update_capa(ocapa, capa); - ocapa->c_type = type; - list_add_capa(ocapa, &capa_list[type]); - hlist_add_head(&ocapa->c_hash, head); - if (type == CLIENT_CAPA) - INIT_LIST_HEAD(&ocapa->c_lli_list); - __capa_get(ocapa); - - capa_count[type]++; - - DEBUG_CAPA(D_INODE, &ocapa->c_capa, "new %s", - capa_type_name[type]); - - if (type != CLIENT_CAPA && capa_count[type] > CAPA_CACHE_SIZE) { - struct list_head *node = capa_list[type].next; - struct obd_capa *tcapa; - int count = 0; - - /* free 12 unused capa from head */ - while (node->next != &capa_list[type] && count < 12) { - tcapa = list_entry(node, struct obd_capa, - c_list); - node = node->next; - if (atomic_read(&tcapa->c_refc) > 0) - continue; - DEBUG_CAPA(D_INODE, &tcapa->c_capa, - "free unused %s", - capa_type_name[type]); - __capa_put(tcapa); - destroy_capa(tcapa); - count++; - } - } + DEBUG_CAPA(D_SEC, &ocapa->c_capa, "new"); spin_unlock(&capa_lock); return ocapa; } - spin_unlock(&capa_lock); - destroy_capa(ocapa); - return old; -} - -struct obd_capa * -capa_get(uid_t uid, int capa_op,__u64 mdsid, unsigned long ino, - __u32 igen, int type) -{ - struct hlist_head *head = capa_hash + capa_hashfn(uid, mdsid, ino); - struct obd_capa *ocapa; - - ocapa = find_capa_locked(head, uid, capa_op, mdsid, ino, igen, type); - - return ocapa; -} - -struct obd_capa * filter_capa_get(struct lustre_capa *capa) -{ - struct hlist_head *head = capa_hash + - capa_hashfn(capa->lc_uid, capa->lc_mdsid, capa->lc_ino); - struct obd_capa *ocapa; - - spin_lock(&capa_lock); - ocapa = filter_find_capa(head, capa); - if (ocapa) - __capa_get(ocapa); + capa_get(old); spin_unlock(&capa_lock); - return ocapa; -} -void capa_put(struct obd_capa *ocapa) -{ - if (!ocapa) - return; + DEBUG_CAPA(D_SEC, &old->c_capa, "update"); - DEBUG_CAPA(D_INODE, &ocapa->c_capa, "put %s", - capa_type_name[ocapa->c_type]); - spin_lock(&capa_lock); - if (ocapa->c_type == CLIENT_CAPA) { - list_del_init(&ocapa->c_lli_list); - __capa_put(ocapa); - destroy_capa(ocapa); - } else { - atomic_dec(&ocapa->c_refc); - } - spin_unlock(&capa_lock); + free_capa(ocapa); + return old; } -struct obd_capa *capa_renew(struct lustre_capa *capa, int type) +struct obd_capa *capa_lookup(struct hlist_head *hash, struct lustre_capa *capa, + int alive) { - uid_t uid = capa->lc_uid; - int capa_op = capa->lc_op; - __u64 mdsid = capa->lc_mdsid; - unsigned long ino = capa->lc_ino; - struct hlist_head *head = capa_hash + - capa_hashfn(uid, mdsid, ino); struct obd_capa *ocapa; spin_lock(&capa_lock); - - if (type == FILTER_CAPA) - ocapa = filter_find_capa(head, capa); - else - ocapa = find_capa(head, uid, capa_op, mdsid, ino, - capa->lc_igen, type); + ocapa = find_capa(capa, hash + capa_hashfn(&capa->lc_fid), alive); if (ocapa) { - DEBUG_CAPA(D_INFO, capa, "renew %s", capa_type_name[type]); - do_update_capa(ocapa, capa); + list_move_tail(&ocapa->c_list, &capa_list[CAPA_SITE_SERVER]); + capa_get(ocapa); } - spin_unlock(&capa_lock); - if (!ocapa) - ocapa = get_new_capa_locked(head, type, capa); - return ocapa; } -void capa_hmac(__u8 *key, struct lustre_capa *capa) +int capa_hmac(__u8 *hmac, struct lustre_capa *capa, __u8 *key) { - struct crypto_tfm *tfm; - int keylen = CAPA_KEY_LEN; + struct ll_crypto_hash *tfm; + struct capa_hmac_alg *alg; + int keylen; struct scatterlist sl = { .page = virt_to_page(capa), - .offset = (unsigned long)(capa) % PAGE_SIZE, - .length = sizeof(struct lustre_capa_data), + .offset = (unsigned long)(capa) % CFS_PAGE_SIZE, + .length = offsetof(struct lustre_capa, lc_hmac), }; - tfm = crypto_alloc_tfm(CAPA_HMAC_ALG, 0); - LASSERT(tfm); - crypto_hmac(tfm, key, &keylen, &sl, 1, capa->lc_hmac); - crypto_free_tfm(tfm); -} + if (capa_alg(capa) != CAPA_HMAC_ALG_SHA1) { + CERROR("unknown capability hmac algorithm!\n"); + return -EFAULT; + } -void capa_dup(void *dst, struct obd_capa *ocapa) -{ - spin_lock(&capa_lock); - memcpy(dst, &ocapa->c_capa, sizeof(ocapa->c_capa)); - spin_unlock(&capa_lock); -} + alg = &capa_hmac_algs[capa_alg(capa)]; -void capa_dup2(void *dst, struct lustre_capa *capa) -{ - spin_lock(&capa_lock); - memcpy(dst, capa, sizeof(*capa)); - spin_unlock(&capa_lock); -} + tfm = ll_crypto_alloc_hash(alg->ha_name, 0, 0); + if (!tfm) { + CERROR("crypto_alloc_tfm failed, check whether your kernel" + "has crypto support!\n"); + return -ENOMEM; + } + keylen = alg->ha_keylen; -int capa_expired(struct lustre_capa *capa) -{ - struct timeval tv; + ll_crypto_hmac(tfm, key, &keylen, &sl, sl.length, hmac); + ll_crypto_free_hash(tfm); - do_gettimeofday(&tv); - return ((unsigned long )capa->lc_expiry <= tv.tv_sec) ? 1 : 0; + return 0; } +#endif -int __capa_is_to_expire(struct obd_capa *ocapa, struct timeval *tv) +void capa_cpy(void *capa, struct obd_capa *ocapa) { - int pre_expiry = capa_pre_expiry(&ocapa->c_capa); - - /* XXX: in case the clock is inaccurate, minus one more - * pre_expiry to make sure the expiry won't miss */ - return ((unsigned long)ocapa->c_capa.lc_expiry - - 2 * pre_expiry <= tv->tv_sec)? 1 : 0; + spin_lock(&ocapa->c_lock); + *(struct lustre_capa *)capa = ocapa->c_capa; + spin_unlock(&ocapa->c_lock); } -int capa_is_to_expire(struct obd_capa *ocapa) +char *dump_capa_content(char *buf, char *key, int len) { - struct timeval tv; - int rc; - - do_gettimeofday(&tv); - spin_lock(&capa_lock); - rc = __capa_is_to_expire(ocapa, &tv); - spin_unlock(&capa_lock); + int i, n = 0; - return rc; + for (i = 0; i < len; i++) + n += sprintf(buf + n, "%02x", (unsigned char) key[i]); + return buf; } -EXPORT_SYMBOL(capa_op); -EXPORT_SYMBOL(capa_get); -EXPORT_SYMBOL(filter_capa_get); -EXPORT_SYMBOL(capa_put); -EXPORT_SYMBOL(capa_renew); -EXPORT_SYMBOL(__capa_get); +EXPORT_SYMBOL(init_capa_hash); +EXPORT_SYMBOL(cleanup_capa_hash); + +EXPORT_SYMBOL(capa_add); +EXPORT_SYMBOL(capa_lookup); + EXPORT_SYMBOL(capa_hmac); -EXPORT_SYMBOL(capa_dup); -EXPORT_SYMBOL(capa_dup2); -EXPORT_SYMBOL(capa_expired); -EXPORT_SYMBOL(__capa_is_to_expire); -EXPORT_SYMBOL(capa_is_to_expire); +EXPORT_SYMBOL(capa_cpy); + +EXPORT_SYMBOL(dump_capa_content);