X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=blobdiff_plain;f=lustre%2Fllite%2Fnamei.c;h=29d4806be58e5cea153a1978f06c23629fc032ac;hp=5fc7c29d28c81ffe6abf11133fe9dcb5fcc5ffdb;hb=09c558d16f0a80f436522edde89367c088fe2055;hpb=ccabce23bd9e366c345c852f565766a799f61238 diff --git a/lustre/llite/namei.c b/lustre/llite/namei.c index 5fc7c29..29d4806 100644 --- a/lustre/llite/namei.c +++ b/lustre/llite/namei.c @@ -27,20 +27,17 @@ */ /* * This file is part of Lustre, http://www.lustre.org/ - * Lustre is a trademark of Sun Microsystems, Inc. */ #include #include #include +#include #include #include #include -#include #include -#ifdef HAVE_UIDGID_HEADER -# include -#endif +#include #define DEBUG_SUBSYSTEM S_LLITE @@ -51,7 +48,8 @@ static int ll_create_it(struct inode *dir, struct dentry *dentry, struct lookup_intent *it, - void *secctx, __u32 secctxlen); + void *secctx, __u32 secctxlen, bool encrypt, + void *encctx, __u32 encctxlen); /* called from iget5_locked->find_inode() under inode_lock spinlock */ static int ll_test_inode(struct inode *inode, void *opaque) @@ -156,27 +154,44 @@ struct inode *ll_iget(struct super_block *sb, ino_t hash, RETURN(inode); } -static void ll_invalidate_negative_children(struct inode *dir) +/* mark negative sub file dentries invalid and prune unused dentries */ +static void ll_prune_negative_children(struct inode *dir) { - struct dentry *dentry, *tmp_subdir; - DECLARE_LL_D_HLIST_NODE_PTR(p); + struct dentry *dentry; + struct dentry *child; + + ENTRY; - ll_lock_dcache(dir); - ll_d_hlist_for_each_entry(dentry, p, &dir->i_dentry) { +restart: + spin_lock(&dir->i_lock); + hlist_for_each_entry(dentry, &dir->i_dentry, d_alias) { spin_lock(&dentry->d_lock); - if (!list_empty(&dentry->d_subdirs)) { - struct dentry *child; - - list_for_each_entry_safe(child, tmp_subdir, - &dentry->d_subdirs, - d_child) { - if (child->d_inode == NULL) - d_lustre_invalidate(child, 1); + list_for_each_entry(child, &dentry->d_subdirs, d_child) { + if (child->d_inode) + continue; + + spin_lock_nested(&child->d_lock, DENTRY_D_LOCK_NESTED); + __d_lustre_invalidate(child); + if (!ll_d_count(child)) { + dget_dlock(child); + __d_drop(child); + spin_unlock(&child->d_lock); + spin_unlock(&dentry->d_lock); + spin_unlock(&dir->i_lock); + + CDEBUG(D_DENTRY, "prune negative dentry %pd\n", + child); + + dput(child); + goto restart; } + spin_unlock(&child->d_lock); } spin_unlock(&dentry->d_lock); } - ll_unlock_dcache(dir); + spin_unlock(&dir->i_lock); + + EXIT; } int ll_test_inode_by_fid(struct inode *inode, void *opaque) @@ -184,216 +199,310 @@ int ll_test_inode_by_fid(struct inode *inode, void *opaque) return lu_fid_eq(&ll_i2info(inode)->lli_fid, opaque); } -int ll_dom_lock_cancel(struct inode *inode, struct ldlm_lock *lock) +static int ll_dom_lock_cancel(struct inode *inode, struct ldlm_lock *lock) { struct lu_env *env; struct ll_inode_info *lli = ll_i2info(inode); - struct cl_layout clt = { .cl_layout_gen = 0, }; - int rc; __u16 refcheck; - - + int rc; ENTRY; - if (!lli->lli_clob) - RETURN(0); - env = cl_env_get(&refcheck); if (IS_ERR(env)) RETURN(PTR_ERR(env)); - rc = cl_object_layout_get(env, lli->lli_clob, &clt); - if (rc) { - CDEBUG(D_INODE, "Cannot get layout for "DFID"\n", - PFID(ll_inode2fid(inode))); - rc = -ENODATA; - } else if (clt.cl_size == 0 || clt.cl_dom_comp_size == 0) { - CDEBUG(D_INODE, "DOM lock without DOM layout for "DFID"\n", - PFID(ll_inode2fid(inode))); - } else { - enum cl_fsync_mode mode; - loff_t end = clt.cl_dom_comp_size - 1; + OBD_FAIL_TIMEOUT(OBD_FAIL_LDLM_REPLAY_PAUSE, cfs_fail_val); - mode = ldlm_is_discard_data(lock) ? - CL_FSYNC_DISCARD : CL_FSYNC_LOCAL; - rc = cl_sync_file_range(inode, 0, end, mode, 1); - truncate_inode_pages_range(inode->i_mapping, 0, end); + /* reach MDC layer to flush data under the DoM ldlm lock */ + rc = cl_object_flush(env, lli->lli_clob, lock); + if (rc == -ENODATA) { + CDEBUG(D_INODE, "inode "DFID" layout has no DoM stripe\n", + PFID(ll_inode2fid(inode))); + /* most likely result of layout change, do nothing */ + rc = 0; } + cl_env_put(env, &refcheck); RETURN(rc); } -int ll_md_blocking_ast(struct ldlm_lock *lock, struct ldlm_lock_desc *desc, - void *data, int flag) +static void ll_lock_cancel_bits(struct ldlm_lock *lock, __u64 to_cancel) { - struct lustre_handle lockh; + struct inode *inode = ll_inode_from_resource_lock(lock); + struct ll_inode_info *lli; + __u64 bits = to_cancel; int rc; + ENTRY; - switch (flag) { - case LDLM_CB_BLOCKING: - ldlm_lock2handle(lock, &lockh); - rc = ldlm_cli_cancel(&lockh, LCF_ASYNC); - if (rc < 0) { - CDEBUG(D_INODE, "ldlm_cli_cancel: rc = %d\n", rc); - RETURN(rc); - } - break; - case LDLM_CB_CANCELING: { - struct inode *inode = ll_inode_from_resource_lock(lock); - __u64 bits = lock->l_policy_data.l_inodebits.bits; + if (!inode) { + /* That means the inode is evicted most likely and may cause + * the skipping of lock cleanups below, so print the message + * about that in log. + */ + if (lock->l_resource->lr_lvb_inode) + LDLM_DEBUG(lock, + "can't take inode for the lock (%sevicted)\n", + lock->l_resource->lr_lvb_inode->i_state & + I_FREEING ? "" : "not "); + RETURN_EXIT; + } - if (inode == NULL) - break; + if (!fid_res_name_eq(ll_inode2fid(inode), + &lock->l_resource->lr_name)) { + LDLM_ERROR(lock, "data mismatch with object "DFID"(%p)", + PFID(ll_inode2fid(inode)), inode); + LBUG(); + } + + if (bits & MDS_INODELOCK_XATTR) { + ll_xattr_cache_destroy(inode); + bits &= ~MDS_INODELOCK_XATTR; + } + + /* For OPEN locks we differentiate between lock modes + * LCK_CR, LCK_CW, LCK_PR - bug 22891 */ + if (bits & MDS_INODELOCK_OPEN) + ll_have_md_lock(inode, &bits, lock->l_req_mode); - /* Invalidate all dentries associated with this inode */ - LASSERT(ldlm_is_canceling(lock)); + if (bits & MDS_INODELOCK_OPEN) { + fmode_t fmode; - if (!fid_res_name_eq(ll_inode2fid(inode), - &lock->l_resource->lr_name)) { - LDLM_ERROR(lock, "data mismatch with object "DFID"(%p)", - PFID(ll_inode2fid(inode)), inode); + switch (lock->l_req_mode) { + case LCK_CW: + fmode = FMODE_WRITE; + break; + case LCK_PR: + fmode = FMODE_EXEC; + break; + case LCK_CR: + fmode = FMODE_READ; + break; + default: + LDLM_ERROR(lock, "bad lock mode for OPEN lock"); LBUG(); } - if (bits & MDS_INODELOCK_XATTR) { - if (S_ISDIR(inode->i_mode)) - ll_i2info(inode)->lli_def_stripe_offset = -1; - ll_xattr_cache_destroy(inode); - bits &= ~MDS_INODELOCK_XATTR; - } + ll_md_real_close(inode, fmode); - /* For OPEN locks we differentiate between lock modes - * LCK_CR, LCK_CW, LCK_PR - bug 22891 */ - if (bits & MDS_INODELOCK_OPEN) - ll_have_md_lock(inode, &bits, lock->l_req_mode); - - if (bits & MDS_INODELOCK_OPEN) { - fmode_t fmode; - - switch (lock->l_req_mode) { - case LCK_CW: - fmode = FMODE_WRITE; - break; - case LCK_PR: - fmode = FMODE_EXEC; - break; - case LCK_CR: - fmode = FMODE_READ; - break; - default: - LDLM_ERROR(lock, "bad lock mode for OPEN lock"); - LBUG(); - } + bits &= ~MDS_INODELOCK_OPEN; + } - ll_md_real_close(inode, fmode); + if (bits & (MDS_INODELOCK_LOOKUP | MDS_INODELOCK_UPDATE | + MDS_INODELOCK_LAYOUT | MDS_INODELOCK_PERM | + MDS_INODELOCK_DOM)) + ll_have_md_lock(inode, &bits, LCK_MINMODE); - bits &= ~MDS_INODELOCK_OPEN; - } + if (bits & MDS_INODELOCK_DOM) { + rc = ll_dom_lock_cancel(inode, lock); + if (rc < 0) + CDEBUG(D_INODE, "cannot flush DoM data " + DFID": rc = %d\n", + PFID(ll_inode2fid(inode)), rc); + } - if (bits & (MDS_INODELOCK_LOOKUP | MDS_INODELOCK_UPDATE | - MDS_INODELOCK_LAYOUT | MDS_INODELOCK_PERM | - MDS_INODELOCK_DOM)) - ll_have_md_lock(inode, &bits, LCK_MINMODE); - - if (bits & MDS_INODELOCK_DOM) { - rc = ll_dom_lock_cancel(inode, lock); - if (rc < 0) - CDEBUG(D_INODE, "cannot flush DoM data " - DFID": rc = %d\n", - PFID(ll_inode2fid(inode)), rc); - lock_res_and_lock(lock); - ldlm_set_kms_ignore(lock); - unlock_res_and_lock(lock); - bits &= ~MDS_INODELOCK_DOM; - } + if (bits & MDS_INODELOCK_LAYOUT) { + struct cl_object_conf conf = { + .coc_opc = OBJECT_CONF_INVALIDATE, + .coc_inode = inode, + }; - if (bits & MDS_INODELOCK_LAYOUT) { - struct cl_object_conf conf = { - .coc_opc = OBJECT_CONF_INVALIDATE, - .coc_inode = inode, - }; - - rc = ll_layout_conf(inode, &conf); - if (rc < 0) - CDEBUG(D_INODE, "cannot invalidate layout of " - DFID": rc = %d\n", - PFID(ll_inode2fid(inode)), rc); - } + rc = ll_layout_conf(inode, &conf); + if (rc < 0) + CDEBUG(D_INODE, "cannot invalidate layout of " + DFID": rc = %d\n", + PFID(ll_inode2fid(inode)), rc); + } - if (bits & MDS_INODELOCK_UPDATE) { - struct ll_inode_info *lli = ll_i2info(inode); - lli->lli_update_atime = 1; - } + lli = ll_i2info(inode); - if ((bits & MDS_INODELOCK_UPDATE) && S_ISDIR(inode->i_mode)) { - struct ll_inode_info *lli = ll_i2info(inode); + if (bits & MDS_INODELOCK_UPDATE) + set_bit(LLIF_UPDATE_ATIME, &lli->lli_flags); - CDEBUG(D_INODE, "invalidating inode "DFID" lli = %p, " - "pfid = "DFID"\n", PFID(ll_inode2fid(inode)), - lli, PFID(&lli->lli_pfid)); - truncate_inode_pages(inode->i_mapping, 0); + if ((bits & MDS_INODELOCK_UPDATE) && S_ISDIR(inode->i_mode)) { + CDEBUG(D_INODE, "invalidating inode "DFID" lli = %p, " + "pfid = "DFID"\n", PFID(ll_inode2fid(inode)), + lli, PFID(&lli->lli_pfid)); + truncate_inode_pages(inode->i_mapping, 0); - if (unlikely(!fid_is_zero(&lli->lli_pfid))) { - struct inode *master_inode = NULL; - unsigned long hash; + if (unlikely(!fid_is_zero(&lli->lli_pfid))) { + struct inode *master_inode = NULL; + unsigned long hash; - /* This is slave inode, since all of the child - * dentry is connected on the master inode, so - * we have to invalidate the negative children - * on master inode */ - CDEBUG(D_INODE, "Invalidate s"DFID" m"DFID"\n", - PFID(ll_inode2fid(inode)), - PFID(&lli->lli_pfid)); + /* This is slave inode, since all of the child dentry + * is connected on the master inode, so we have to + * invalidate the negative children on master inode */ + CDEBUG(D_INODE, "Invalidate s"DFID" m"DFID"\n", + PFID(ll_inode2fid(inode)), PFID(&lli->lli_pfid)); - hash = cl_fid_build_ino(&lli->lli_pfid, + hash = cl_fid_build_ino(&lli->lli_pfid, ll_need_32bit_api(ll_i2sbi(inode))); - /* Do not lookup the inode with ilookup5, - * otherwise it will cause dead lock, - * - * 1. Client1 send chmod req to the MDT0, then - * on MDT0, it enqueues master and all of its - * slaves lock, (mdt_attr_set() -> - * mdt_lock_slaves()), after gets master and - * stripe0 lock, it will send the enqueue req - * (for stripe1) to MDT1, then MDT1 finds the - * lock has been granted to client2. Then MDT1 - * sends blocking ast to client2. - * - * 2. At the same time, client2 tries to unlink - * the striped dir (rm -rf striped_dir), and - * during lookup, it will hold the master inode - * of the striped directory, whose inode state - * is NEW, then tries to revalidate all of its - * slaves, (ll_prep_inode()->ll_iget()-> - * ll_read_inode2()-> ll_update_inode().). And - * it will be blocked on the server side because - * of 1. - * - * 3. Then the client get the blocking_ast req, - * cancel the lock, but being blocked if using - * ->ilookup5()), because master inode state is - * NEW. */ - master_inode = ilookup5_nowait(inode->i_sb, - hash, ll_test_inode_by_fid, + /* Do not lookup the inode with ilookup5, otherwise + * it will cause dead lock, + * 1. Client1 send chmod req to the MDT0, then on MDT0, + * it enqueues master and all of its slaves lock, + * (mdt_attr_set() -> mdt_lock_slaves()), after gets + * master and stripe0 lock, it will send the enqueue + * req (for stripe1) to MDT1, then MDT1 finds the lock + * has been granted to client2. Then MDT1 sends blocking + * ast to client2. + * 2. At the same time, client2 tries to unlink + * the striped dir (rm -rf striped_dir), and during + * lookup, it will hold the master inode of the striped + * directory, whose inode state is NEW, then tries to + * revalidate all of its slaves, (ll_prep_inode()-> + * ll_iget()->ll_read_inode2()-> ll_update_inode().). + * And it will be blocked on the server side because + * of 1. + * 3. Then the client get the blocking_ast req, cancel + * the lock, but being blocked if using ->ilookup5()), + * because master inode state is NEW. */ + master_inode = ilookup5_nowait(inode->i_sb, hash, + ll_test_inode_by_fid, (void *)&lli->lli_pfid); - if (master_inode) { - ll_invalidate_negative_children( - master_inode); - iput(master_inode); - } - } else { - ll_invalidate_negative_children(inode); + if (master_inode) { + ll_prune_negative_children(master_inode); + iput(master_inode); } + } else { + ll_prune_negative_children(inode); } + } - if ((bits & (MDS_INODELOCK_LOOKUP | MDS_INODELOCK_PERM)) && - inode->i_sb->s_root != NULL && - inode != inode->i_sb->s_root->d_inode) - ll_invalidate_aliases(inode); + if ((bits & (MDS_INODELOCK_LOOKUP | MDS_INODELOCK_PERM)) && + !is_root_inode(inode)) + ll_prune_aliases(inode); - iput(inode); + if (bits & (MDS_INODELOCK_LOOKUP | MDS_INODELOCK_PERM)) + forget_all_cached_acls(inode); + + iput(inode); + RETURN_EXIT; +} + +/* Check if the given lock may be downgraded instead of canceling and + * that convert is really needed. */ +int ll_md_need_convert(struct ldlm_lock *lock) +{ + struct ldlm_namespace *ns = ldlm_lock_to_ns(lock); + struct inode *inode; + __u64 wanted = lock->l_policy_data.l_inodebits.cancel_bits; + __u64 bits = lock->l_policy_data.l_inodebits.bits & ~wanted; + enum ldlm_mode mode = LCK_MINMODE; + + if (!lock->l_conn_export || + !exp_connect_lock_convert(lock->l_conn_export)) + return 0; + + if (!wanted || !bits || ldlm_is_cancel(lock)) + return 0; + + /* do not convert locks other than DOM for now */ + if (!((bits | wanted) & MDS_INODELOCK_DOM)) + return 0; + + /* We may have already remaining bits in some other lock so + * lock convert will leave us just extra lock for the same bit. + * Check if client has other lock with the same bits and the same + * or lower mode and don't convert if any. + */ + switch (lock->l_req_mode) { + case LCK_PR: + mode = LCK_PR; + /* fallthrough */ + case LCK_PW: + mode |= LCK_CR; + break; + case LCK_CW: + mode = LCK_CW; + /* fallthrough */ + case LCK_CR: + mode |= LCK_CR; + break; + default: + /* do not convert other modes */ + return 0; + } + + /* is lock is too old to be converted? */ + lock_res_and_lock(lock); + if (ktime_after(ktime_get(), + ktime_add(lock->l_last_used, ns->ns_dirty_age_limit))) { + unlock_res_and_lock(lock); + return 0; + } + unlock_res_and_lock(lock); + + inode = ll_inode_from_resource_lock(lock); + ll_have_md_lock(inode, &bits, mode); + iput(inode); + return !!(bits); +} + +int ll_md_blocking_ast(struct ldlm_lock *lock, struct ldlm_lock_desc *ld, + void *data, int flag) +{ + struct lustre_handle lockh; + int rc; + + ENTRY; + + switch (flag) { + case LDLM_CB_BLOCKING: + { + __u64 cancel_flags = LCF_ASYNC; + + /* if lock convert is not needed then still have to + * pass lock via ldlm_cli_convert() to keep all states + * correct, set cancel_bits to full lock bits to cause + * full cancel to happen. + */ + if (!ll_md_need_convert(lock)) { + lock_res_and_lock(lock); + lock->l_policy_data.l_inodebits.cancel_bits = + lock->l_policy_data.l_inodebits.bits; + unlock_res_and_lock(lock); + } + rc = ldlm_cli_convert(lock, cancel_flags); + if (!rc) + RETURN(0); + /* continue with cancel otherwise */ + ldlm_lock2handle(lock, &lockh); + rc = ldlm_cli_cancel(&lockh, cancel_flags); + if (rc < 0) { + CDEBUG(D_INODE, "ldlm_cli_cancel: rc = %d\n", rc); + RETURN(rc); + } + break; + } + case LDLM_CB_CANCELING: + { + __u64 to_cancel = lock->l_policy_data.l_inodebits.bits; + + /* Nothing to do for non-granted locks */ + if (!ldlm_is_granted(lock)) + break; + + /* If 'ld' is supplied then bits to be cancelled are passed + * implicitly by lock converting and cancel_bits from 'ld' + * should be used. Otherwise full cancel is being performed + * and lock inodebits are used. + * + * Note: we cannot rely on cancel_bits in lock itself at this + * moment because they can be changed by concurrent thread, + * so ldlm_cli_inodebits_convert() pass cancel bits implicitly + * in 'ld' parameter. + */ + if (ld) { + /* partial bits cancel allowed only during convert */ + LASSERT(ldlm_is_converting(lock)); + /* mask cancel bits by lock bits so only no any unused + * bits are passed to ll_lock_cancel_bits() + */ + to_cancel &= ld->l_policy_data.l_inodebits.cancel_bits; + } + ll_lock_cancel_bits(lock, to_cancel); break; } default: @@ -439,15 +548,14 @@ void ll_i2gids(__u32 *suppgids, struct inode *i1, struct inode *i2) static struct dentry *ll_find_alias(struct inode *inode, struct dentry *dentry) { struct dentry *alias, *discon_alias, *invalid_alias; - DECLARE_LL_D_HLIST_NODE_PTR(p); - if (ll_d_hlist_empty(&inode->i_dentry)) + if (hlist_empty(&inode->i_dentry)) return NULL; discon_alias = invalid_alias = NULL; - ll_lock_dcache(inode); - ll_d_hlist_for_each_entry(alias, p, &inode->i_dentry) { + spin_lock(&inode->i_lock); + hlist_for_each_entry(alias, &inode->i_dentry, d_alias) { LASSERT(alias != dentry); spin_lock(&alias->d_lock); @@ -472,7 +580,7 @@ static struct dentry *ll_find_alias(struct inode *inode, struct dentry *dentry) dget_dlock(alias); spin_unlock(&alias->d_lock); } - ll_unlock_dcache(inode); + spin_unlock(&inode->i_lock); return alias; } @@ -506,6 +614,27 @@ struct dentry *ll_splice_alias(struct inode *inode, struct dentry *de) if (rc < 0) return ERR_PTR(rc); d_add(de, inode); + + /* this needs only to be done for foreign symlink dirs as + * DCACHE_SYMLINK_TYPE is already set by d_flags_for_inode() + * kernel routine for files with symlink ops (ie, real symlink) + */ + if (inode && S_ISDIR(inode->i_mode) && + ll_sbi_has_foreign_symlink(ll_i2sbi(inode)) && +#ifdef HAVE_IOP_GET_LINK + inode->i_op->get_link) { +#else + inode->i_op->follow_link) { +#endif + CDEBUG(D_INFO, "%s: inode "DFID": faking foreign dir as a symlink\n", + ll_i2sbi(inode)->ll_fsname, PFID(ll_inode2fid(inode))); + spin_lock(&de->d_lock); + /* like d_flags_for_inode() already does for files */ + de->d_flags = (de->d_flags & ~DCACHE_ENTRY_TYPE) | + DCACHE_SYMLINK_TYPE; + spin_unlock(&de->d_lock); + } + CDEBUG(D_DENTRY, "Add dentry %p inode %p refc %d flags %#x\n", de, de->d_inode, ll_d_count(de), de->d_flags); return de; @@ -513,7 +642,10 @@ struct dentry *ll_splice_alias(struct inode *inode, struct dentry *de) static int ll_lookup_it_finish(struct ptlrpc_request *request, struct lookup_intent *it, - struct inode *parent, struct dentry **de) + struct inode *parent, struct dentry **de, + void *secctx, __u32 secctxlen, + void *encctx, __u32 encctxlen, + ktime_t kstart, bool encrypt) { struct inode *inode = NULL; __u64 bits = 0; @@ -526,20 +658,89 @@ static int ll_lookup_it_finish(struct ptlrpc_request *request, CDEBUG(D_DENTRY, "it %p it_disposition %x\n", it, it->it_disposition); if (!it_disposition(it, DISP_LOOKUP_NEG)) { - rc = ll_prep_inode(&inode, request, (*de)->d_sb, it); - if (rc) - RETURN(rc); - - ll_set_lock_data(ll_i2sbi(parent)->ll_md_exp, inode, it, &bits); - - /* We used to query real size from OSTs here, but actually - this is not needed. For stat() calls size would be updated - from subsequent do_revalidate()->ll_inode_revalidate_it() in - 2.4 and - vfs_getattr_it->ll_getattr()->ll_inode_revalidate_it() in 2.6 - Everybody else who needs correct file size would call - ll_glimpse_size or some equivalent themselves anyway. - Also see bug 7198. */ + struct req_capsule *pill = &request->rq_pill; + struct mdt_body *body = req_capsule_server_get(pill, + &RMF_MDT_BODY); + + rc = ll_prep_inode(&inode, &request->rq_pill, (*de)->d_sb, it); + if (rc) + RETURN(rc); + + /* If encryption context was returned by MDT, put it in + * inode now to save an extra getxattr and avoid deadlock. + */ + if (body->mbo_valid & OBD_MD_ENCCTX) { + encctx = req_capsule_server_get(pill, &RMF_FILE_ENCCTX); + encctxlen = req_capsule_get_size(pill, + &RMF_FILE_ENCCTX, + RCL_SERVER); + + if (encctxlen) { + CDEBUG(D_SEC, + "server returned encryption ctx for "DFID"\n", + PFID(ll_inode2fid(inode))); + rc = ll_xattr_cache_insert(inode, + LL_XATTR_NAME_ENCRYPTION_CONTEXT, + encctx, encctxlen); + if (rc) + CWARN("%s: cannot set enc ctx for "DFID": rc = %d\n", + ll_i2sbi(inode)->ll_fsname, + PFID(ll_inode2fid(inode)), rc); + else if (encrypt) { + rc = llcrypt_get_encryption_info(inode); + if (rc) + CDEBUG(D_SEC, + "cannot get enc info for "DFID": rc = %d\n", + PFID(ll_inode2fid(inode)), rc); + } + } + } + + ll_set_lock_data(ll_i2sbi(parent)->ll_md_exp, inode, it, &bits); + /* OPEN can return data if lock has DoM+LAYOUT bits set */ + if (it->it_op & IT_OPEN && + bits & MDS_INODELOCK_DOM && bits & MDS_INODELOCK_LAYOUT) + ll_dom_finish_open(inode, request); + + /* We used to query real size from OSTs here, but actually + * this is not needed. For stat() calls size would be updated + * from subsequent do_revalidate()->ll_inode_revalidate_it() in + * 2.4 and + * vfs_getattr_it->ll_getattr()->ll_inode_revalidate_it() in 2.6 + * Everybody else who needs correct file size would call + * ll_glimpse_size or some equivalent themselves anyway. + * Also see bug 7198. + */ + + /* If security context was returned by MDT, put it in + * inode now to save an extra getxattr from security hooks, + * and avoid deadlock. + */ + if (body->mbo_valid & OBD_MD_SECCTX) { + secctx = req_capsule_server_get(pill, &RMF_FILE_SECCTX); + secctxlen = req_capsule_get_size(pill, + &RMF_FILE_SECCTX, + RCL_SERVER); + + if (secctxlen) + CDEBUG(D_SEC, "server returned security context" + " for "DFID"\n", + PFID(ll_inode2fid(inode))); + } + + if (secctx != NULL && secctxlen != 0) { + /* no need to protect selinux_inode_setsecurity() by + * inode_lock. Taking it would lead to a client deadlock + * LU-13617 + */ + rc = security_inode_notifysecctx(inode, secctx, + secctxlen); + if (rc) + CWARN("%s: cannot set security context for "DFID": rc = %d\n", + ll_i2sbi(inode)->ll_fsname, + PFID(ll_inode2fid(inode)), + rc); + } } /* Only hash *de if it is unhashed (new dentry). @@ -555,10 +756,18 @@ static int ll_lookup_it_finish(struct ptlrpc_request *request, /* we have lookup look - unhide dentry */ if (bits & MDS_INODELOCK_LOOKUP) d_lustre_revalidate(*de); + + if (encrypt) { + rc = llcrypt_get_encryption_info(inode); + if (rc) + GOTO(out, rc); + if (!llcrypt_has_encryption_key(inode)) + GOTO(out, rc = -ENOKEY); + } } else if (!it_disposition(it, DISP_OPEN_CREATE)) { - /* If file created on server, don't depend on parent UPDATE - * lock to unhide it. It is left hidden and next lookup can - * find it in ll_splice_alias. + /* + * If file was created on the server, the dentry is revalidated + * in ll_create_it if the lock allows for it. */ /* Check that parent has UPDATE lock. */ struct lookup_intent parent_it = { @@ -567,7 +776,7 @@ static int ll_lookup_it_finish(struct ptlrpc_request *request, struct lu_fid fid = ll_i2info(parent)->lli_fid; /* If it is striped directory, get the real stripe parent */ - if (unlikely(ll_i2info(parent)->lli_lsm_md != NULL)) { + if (unlikely(ll_dir_striped(parent))) { rc = md_get_fid_from_lsm(ll_i2mdexp(parent), ll_i2info(parent)->lli_lsm_md, (*de)->d_name.name, @@ -583,48 +792,61 @@ static int ll_lookup_it_finish(struct ptlrpc_request *request, } } + if (it_disposition(it, DISP_OPEN_CREATE)) { + ll_stats_ops_tally(ll_i2sbi(parent), LPROC_LL_MKNOD, + ktime_us_delta(ktime_get(), kstart)); + } + GOTO(out, rc = 0); out: - if (rc != 0 && it->it_op & IT_OPEN) - ll_open_cleanup((*de)->d_sb, request); + if (rc != 0 && it->it_op & IT_OPEN) { + ll_intent_drop_lock(it); + ll_open_cleanup((*de)->d_sb, &request->rq_pill); + } return rc; } static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry, struct lookup_intent *it, - void **secctx, __u32 *secctxlen) + void **secctx, __u32 *secctxlen, + struct pcc_create_attach *pca, + bool encrypt, + void **encctx, __u32 *encctxlen) { + ktime_t kstart = ktime_get(); struct lookup_intent lookup_it = { .it_op = IT_LOOKUP }; struct dentry *save = dentry, *retval; struct ptlrpc_request *req = NULL; struct md_op_data *op_data = NULL; - __u32 opc; - int rc; - ENTRY; + struct lov_user_md *lum = NULL; + __u32 opc; + int rc; + char secctx_name[XATTR_NAME_MAX + 1]; + + ENTRY; - if (dentry->d_name.len > ll_i2sbi(parent)->ll_namelen) - RETURN(ERR_PTR(-ENAMETOOLONG)); + if (dentry->d_name.len > ll_i2sbi(parent)->ll_namelen) + RETURN(ERR_PTR(-ENAMETOOLONG)); - CDEBUG(D_VFSTRACE, "VFS Op:name=%.*s, dir="DFID"(%p), intent=%s\n", - dentry->d_name.len, dentry->d_name.name, - PFID(ll_inode2fid(parent)), parent, LL_IT2STR(it)); + CDEBUG(D_VFSTRACE, "VFS Op:name=%pd, dir="DFID"(%p), intent=%s\n", + dentry, PFID(ll_inode2fid(parent)), parent, LL_IT2STR(it)); - if (d_mountpoint(dentry)) - CERROR("Tell Peter, lookup on mtpt, it %s\n", LL_IT2STR(it)); + if (d_mountpoint(dentry)) + CERROR("Tell Peter, lookup on mtpt, it %s\n", LL_IT2STR(it)); if (it == NULL || it->it_op == IT_GETXATTR) it = &lookup_it; if (it->it_op == IT_GETATTR && dentry_may_statahead(parent, dentry)) { - rc = ll_statahead(parent, &dentry, 0); + rc = ll_revalidate_statahead(parent, &dentry, 0); if (rc == 1) RETURN(dentry == save ? NULL : dentry); } if (it->it_op & IT_OPEN && it->it_flags & FMODE_WRITE && - dentry->d_sb->s_flags & MS_RDONLY) + dentry->d_sb->s_flags & SB_RDONLY) RETURN(ERR_PTR(-EROFS)); if (it->it_op & IT_CREAT) @@ -654,6 +876,136 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry, *secctx = op_data->op_file_secctx; if (secctxlen != NULL) *secctxlen = op_data->op_file_secctx_size; + } else { + if (secctx != NULL) + *secctx = NULL; + if (secctxlen != NULL) + *secctxlen = 0; + } + if (it->it_op & IT_CREAT && encrypt) { + /* Volatile file name may look like: + * /LUSTRE_VOLATILE_HDR:::fd= + * where fd is opened descriptor of reference file. + */ + if (unlikely(filename_is_volatile(dentry->d_name.name, + dentry->d_name.len, NULL))) { + int ctx_size = LLCRYPT_ENC_CTX_SIZE; + struct lustre_sb_info *lsi; + struct file *ref_file; + struct inode *ref_inode; + char *p, *q, *fd_str; + void *ctx; + int fd; + + p = strnstr(dentry->d_name.name, ":fd=", + dentry->d_name.len); + if (!p || strlen(p + 4) == 0) + GOTO(out, retval = ERR_PTR(-EINVAL)); + + q = strchrnul(p + 4, ':'); + fd_str = kstrndup(p + 4, q - p - 4, GFP_NOFS); + if (!fd_str) + GOTO(out, retval = ERR_PTR(-ENOMEM)); + rc = kstrtouint(fd_str, 10, &fd); + kfree(fd_str); + if (rc) + GOTO(inherit, rc = -EINVAL); + + ref_file = fget(fd); + if (!ref_file) + GOTO(inherit, rc = -EINVAL); + + ref_inode = file_inode(ref_file); + if (!ref_inode) { + fput(ref_file); + GOTO(inherit, rc = -EINVAL); + } + + lsi = s2lsi(ref_inode->i_sb); + +getctx: + OBD_ALLOC(ctx, ctx_size); + if (!ctx) + GOTO(out, retval = ERR_PTR(-ENOMEM)); + +#ifdef CONFIG_LL_ENCRYPTION + rc = lsi->lsi_cop->get_context(ref_inode, + ctx, ctx_size); +#else + rc = -ENODATA; +#endif + if (rc == -ERANGE) { + OBD_FREE(ctx, ctx_size); + ctx_size *= 2; + goto getctx; + } + fput(ref_file); + if (rc < 0) { + OBD_FREE(ctx, ctx_size); + GOTO(inherit, rc); + } + + op_data->op_file_encctx_size = rc; + if (rc == ctx_size) { + op_data->op_file_encctx = ctx; + } else { + OBD_ALLOC(op_data->op_file_encctx, + op_data->op_file_encctx_size); + if (!op_data->op_file_encctx) { + OBD_FREE(ctx, ctx_size); + GOTO(out, retval = ERR_PTR(-ENOMEM)); + } + memcpy(op_data->op_file_encctx, ctx, + op_data->op_file_encctx_size); + OBD_FREE(ctx, ctx_size); + } + + } else { +inherit: + rc = llcrypt_inherit_context(parent, NULL, op_data, + false); + if (rc) + GOTO(out, retval = ERR_PTR(rc)); + } + if (encctx != NULL) + *encctx = op_data->op_file_encctx; + if (encctxlen != NULL) + *encctxlen = op_data->op_file_encctx_size; + } else { + if (encctx != NULL) + *encctx = NULL; + if (encctxlen != NULL) + *encctxlen = 0; + } + + /* ask for security context upon intent */ + if (it->it_op & (IT_LOOKUP | IT_GETATTR | IT_OPEN)) { + /* get name of security xattr to request to server */ + rc = ll_listsecurity(parent, secctx_name, + sizeof(secctx_name)); + if (rc < 0) { + CDEBUG(D_SEC, "cannot get security xattr name for " + DFID": rc = %d\n", + PFID(ll_inode2fid(parent)), rc); + } else if (rc > 0) { + op_data->op_file_secctx_name = secctx_name; + op_data->op_file_secctx_name_size = rc; + CDEBUG(D_SEC, "'%.*s' is security xattr for "DFID"\n", + rc, secctx_name, PFID(ll_inode2fid(parent))); + } + } + + if (pca && pca->pca_dataset) { + OBD_ALLOC_PTR(lum); + if (lum == NULL) + GOTO(out, retval = ERR_PTR(-ENOMEM)); + + lum->lmm_magic = LOV_USER_MAGIC_V1; + lum->lmm_pattern = LOV_PATTERN_F_RELEASED | LOV_PATTERN_RAID0; + op_data->op_data = lum; + op_data->op_data_size = sizeof(*lum); + op_data->op_archive_id = pca->pca_dataset->pccd_rwid; + it->it_flags |= MDS_OPEN_PCC; } rc = md_intent_lock(ll_i2mdexp(parent), op_data, it, &req, @@ -687,18 +1039,33 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry, if (rc < 0) GOTO(out, retval = ERR_PTR(rc)); - rc = ll_lookup_it_finish(req, it, parent, &dentry); - if (rc != 0) { - ll_intent_release(it); - GOTO(out, retval = ERR_PTR(rc)); - } + if (pca && pca->pca_dataset) { + rc = pcc_inode_create(parent->i_sb, pca->pca_dataset, + &op_data->op_fid2, + &pca->pca_dentry); + if (rc) + GOTO(out, retval = ERR_PTR(rc)); + } - if ((it->it_op & IT_OPEN) && dentry->d_inode && - !S_ISREG(dentry->d_inode->i_mode) && - !S_ISDIR(dentry->d_inode->i_mode)) { - ll_release_openhandle(dentry, it); - } - ll_lookup_finish_locks(it, dentry); + /* dir layout may change */ + ll_unlock_md_op_lsm(op_data); + rc = ll_lookup_it_finish(req, it, parent, &dentry, + secctx != NULL ? *secctx : NULL, + secctxlen != NULL ? *secctxlen : 0, + encctx != NULL ? *encctx : NULL, + encctxlen != NULL ? *encctxlen : 0, + kstart, encrypt); + if (rc != 0) { + ll_intent_release(it); + GOTO(out, retval = ERR_PTR(rc)); + } + + if ((it->it_op & IT_OPEN) && dentry->d_inode && + !S_ISREG(dentry->d_inode->i_mode) && + !S_ISDIR(dentry->d_inode->i_mode)) { + ll_release_openhandle(dentry, it); + } + ll_lookup_finish_locks(it, dentry); GOTO(out, retval = (dentry == save) ? NULL : dentry); @@ -710,23 +1077,32 @@ out: op_data->op_file_secctx = NULL; op_data->op_file_secctx_size = 0; } + if (encctx != NULL && encctxlen != NULL && + it->it_op & IT_CREAT && encrypt) { + /* caller needs enc ctx info, so reset it in op_data to + * prevent it from being freed + */ + op_data->op_file_encctx = NULL; + op_data->op_file_encctx_size = 0; + } ll_finish_md_op_data(op_data); } + if (lum != NULL) + OBD_FREE_PTR(lum); + ptlrpc_req_finished(req); return retval; } -#ifdef HAVE_IOP_ATOMIC_OPEN static struct dentry *ll_lookup_nd(struct inode *parent, struct dentry *dentry, unsigned int flags) { struct lookup_intent *itp, it = { .it_op = IT_GETATTR }; struct dentry *de; - CDEBUG(D_VFSTRACE, "VFS Op:name=%.*s, dir="DFID"(%p), flags=%u\n", - dentry->d_name.len, dentry->d_name.name, - PFID(ll_inode2fid(parent)), parent, flags); + CDEBUG(D_VFSTRACE, "VFS Op:name=%pd, dir="DFID"(%p), flags=%u\n", + dentry, PFID(ll_inode2fid(parent)), parent, flags); /* * Optimize away (CREATE && !OPEN). Let .create handle the race. @@ -741,7 +1117,8 @@ static struct dentry *ll_lookup_nd(struct inode *parent, struct dentry *dentry, itp = NULL; else itp = ⁢ - de = ll_lookup_it(parent, dentry, itp, NULL, NULL); + de = ll_lookup_it(parent, dentry, itp, NULL, NULL, NULL, false, + NULL, NULL); if (itp != NULL) ll_intent_release(itp); @@ -749,26 +1126,51 @@ static struct dentry *ll_lookup_nd(struct inode *parent, struct dentry *dentry, return de; } +#ifdef FMODE_CREATED /* added in Linux v4.18-rc1-20-g73a09dd */ +# define ll_is_opened(o, f) ((f)->f_mode & FMODE_OPENED) +# define ll_finish_open(f, d, o) finish_open((f), (d), NULL) +# define ll_last_arg +# define ll_set_created(o, f) \ +do { \ + (f)->f_mode |= FMODE_CREATED; \ +} while (0) + +#else +# define ll_is_opened(o, f) (*(o)) +# define ll_finish_open(f, d, o) finish_open((f), (d), NULL, (o)) +# define ll_last_arg , int *opened +# define ll_set_created(o, f) \ +do { \ + *(o) |= FILE_CREATED; \ +} while (0) + +#endif + /* * For cached negative dentry and new dentry, handle lookup/create/open * together. */ static int ll_atomic_open(struct inode *dir, struct dentry *dentry, struct file *file, unsigned open_flags, - umode_t mode, int *opened) + umode_t mode ll_last_arg) { struct lookup_intent *it; struct dentry *de; long long lookup_flags = LOOKUP_OPEN; void *secctx = NULL; __u32 secctxlen = 0; + void *encctx = NULL; + __u32 encctxlen = 0; + struct ll_sb_info *sbi = NULL; + struct pcc_create_attach pca = { NULL, NULL }; + bool encrypt = false; int rc = 0; ENTRY; - CDEBUG(D_VFSTRACE, "VFS Op:name=%.*s, dir="DFID"(%p), file %p," - "open_flags %x, mode %x opened %d\n", - dentry->d_name.len, dentry->d_name.name, - PFID(ll_inode2fid(dir)), dir, file, open_flags, mode, *opened); + CDEBUG(D_VFSTRACE, + "VFS Op:name=%pd, dir="DFID"(%p), file %p, open_flags %x, mode %x opened %d\n", + dentry, PFID(ll_inode2fid(dir)), dir, file, open_flags, mode, + ll_is_opened(opened, file)); /* Only negative dentries enter here */ LASSERT(dentry->d_inode == NULL); @@ -796,13 +1198,54 @@ static int ll_atomic_open(struct inode *dir, struct dentry *dentry, if (open_flags & O_CREAT) { it->it_op |= IT_CREAT; lookup_flags |= LOOKUP_CREATE; + sbi = ll_i2sbi(dir); + /* Volatile file is used for HSM restore, so do not use PCC */ + if (!filename_is_volatile(dentry->d_name.name, + dentry->d_name.len, NULL)) { + struct pcc_matcher item; + struct pcc_dataset *dataset; + + item.pm_uid = from_kuid(&init_user_ns, current_uid()); + item.pm_gid = from_kgid(&init_user_ns, current_gid()); + item.pm_projid = ll_i2info(dir)->lli_projid; + item.pm_name = &dentry->d_name; + dataset = pcc_dataset_match_get(&sbi->ll_pcc_super, + &item); + pca.pca_dataset = dataset; + } } it->it_create_mode = (mode & S_IALLUGO) | S_IFREG; it->it_flags = (open_flags & ~O_ACCMODE) | OPEN_FMODE(open_flags); it->it_flags &= ~MDS_OPEN_FL_INTERNAL; + if (ll_sbi_has_encrypt(ll_i2sbi(dir)) && IS_ENCRYPTED(dir)) { + /* in case of create, this is going to be a regular file because + * we set S_IFREG bit on it->it_create_mode above + */ + rc = llcrypt_get_encryption_info(dir); + if (rc) + GOTO(out_release, rc); + if (open_flags & O_CREAT) { + if (!llcrypt_has_encryption_key(dir)) + GOTO(out_release, rc = -ENOKEY); + encrypt = true; + } + } + + OBD_FAIL_TIMEOUT(OBD_FAIL_LLITE_CREATE_FILE_PAUSE2, cfs_fail_val); + + /* We can only arrive at this path when we have no inode, so + * we only need to request open lock if it was requested + * for every open + */ + if (ll_i2sbi(dir)->ll_oc_thrsh_count == 1 && + exp_connect_flags2(ll_i2mdexp(dir)) & + OBD_CONNECT2_ATOMIC_OPEN_LOCK) + it->it_flags |= MDS_OPEN_LOCK; + /* Dentry added to dcache tree in ll_lookup_it */ - de = ll_lookup_it(dir, dentry, it, &secctx, &secctxlen); + de = ll_lookup_it(dir, dentry, it, &secctx, &secctxlen, &pca, encrypt, + &encctx, &encctxlen); if (IS_ERR(de)) rc = PTR_ERR(de); else if (de != NULL) @@ -813,8 +1256,10 @@ static int ll_atomic_open(struct inode *dir, struct dentry *dentry, if (!rc) { if (it_disposition(it, DISP_OPEN_CREATE)) { /* Dentry instantiated in ll_create_it. */ - rc = ll_create_it(dir, dentry, it, secctx, secctxlen); - security_release_secctx(secctx, secctxlen); + rc = ll_create_it(dir, dentry, it, secctx, secctxlen, + encrypt, encctx, encctxlen); + ll_security_release_secctx(secctx, secctxlen); + llcrypt_free_ctx(encctx, encctxlen); if (rc) { /* We dget in ll_splice_alias. */ if (de != NULL) @@ -822,9 +1267,43 @@ static int ll_atomic_open(struct inode *dir, struct dentry *dentry, goto out_release; } - *opened |= FILE_CREATED; + rc = pcc_inode_create_fini(dentry->d_inode, &pca); + if (rc) { + if (de != NULL) + dput(de); + GOTO(out_release, rc); + } + + ll_set_created(opened, file); + } else { + /* Open the file with O_CREAT, but the file already + * existed on MDT. This may happend in the case that + * the LOOKUP ibits lock is revoked and the + * corresponding dentry cache is deleted. + * i.e. In the current Lustre, the truncate operation + * will revoke the LOOKUP ibits lock, and the file + * dentry cache will be invalidated. The following open + * with O_CREAT flag will call into ->atomic_open, the + * file was wrongly though as newly created file and + * try to auto cache the file. So after client knows it + * is not a DISP_OPEN_CREATE, it should cleanup the + * already created PCC copy. + */ + pcc_create_attach_cleanup(dir->i_sb, &pca); + + if (open_flags & O_CREAT && encrypt && + dentry->d_inode) { + rc = ll_set_encflags(dentry->d_inode, encctx, + encctxlen, true); + llcrypt_free_ctx(encctx, encctxlen); + if (rc) + GOTO(out_release, rc); + } } - if (dentry->d_inode && it_disposition(it, DISP_OPEN_OPEN)) { + + /* check also if a foreign file is openable */ + if (dentry->d_inode && it_disposition(it, DISP_OPEN_OPEN) && + ll_foreign_is_openable(dentry, open_flags)) { /* Open dentry. */ if (S_ISFIFO(dentry->d_inode->i_mode)) { /* We cannot call open here as it might @@ -833,7 +1312,7 @@ static int ll_atomic_open(struct inode *dir, struct dentry *dentry, rc = finish_no_open(file, de); } else { file->private_data = it; - rc = finish_open(file, dentry, NULL, opened); + rc = ll_finish_open(file, dentry, opened); /* We dget in ll_splice_alias. finish_open takes * care of dget for fd open. */ @@ -843,6 +1322,8 @@ static int ll_atomic_open(struct inode *dir, struct dentry *dentry, } else { rc = finish_no_open(file, de); } + } else { + pcc_create_attach_cleanup(dir->i_sb, &pca); } out_release: @@ -852,111 +1333,6 @@ out_release: RETURN(rc); } -#else /* !HAVE_IOP_ATOMIC_OPEN */ -static struct lookup_intent * -ll_convert_intent(struct open_intent *oit, int lookup_flags, bool is_readonly) -{ - struct lookup_intent *it; - - OBD_ALLOC_PTR(it); - if (!it) - return ERR_PTR(-ENOMEM); - - if (lookup_flags & LOOKUP_OPEN) { - it->it_op = IT_OPEN; - /* Avoid file creation for ro bind mount point(is_readonly) */ - if ((lookup_flags & LOOKUP_CREATE) && !is_readonly) - it->it_op |= IT_CREAT; - it->it_create_mode = (oit->create_mode & S_IALLUGO) | S_IFREG; - it->it_flags = ll_namei_to_lookup_intent_flag(oit->flags & - ~(is_readonly ? O_CREAT : 0)); - it->it_flags &= ~MDS_OPEN_FL_INTERNAL; - } else { - it->it_op = IT_GETATTR; - } - - return it; -} - -static struct dentry *ll_lookup_nd(struct inode *parent, struct dentry *dentry, - struct nameidata *nd) -{ - struct dentry *de; - ENTRY; - - if (nd && !(nd->flags & (LOOKUP_CONTINUE|LOOKUP_PARENT))) { - struct lookup_intent *it; - - if (ll_d2d(dentry) && ll_d2d(dentry)->lld_it) { - it = ll_d2d(dentry)->lld_it; - ll_d2d(dentry)->lld_it = NULL; - } else { - /* - * Optimize away (CREATE && !OPEN). Let .create handle - * the race. But only if we have write permissions - * there, otherwise we need to proceed with lookup. - * LU-4185 - */ - if ((nd->flags & LOOKUP_CREATE) && - !(nd->flags & LOOKUP_OPEN) && - (inode_permission(parent, - MAY_WRITE | MAY_EXEC) == 0)) - RETURN(NULL); - - it = ll_convert_intent(&nd->intent.open, nd->flags, - (nd->path.mnt->mnt_flags & MNT_READONLY) || - (nd->path.mnt->mnt_sb->s_flags & MS_RDONLY)); - if (IS_ERR(it)) - RETURN((struct dentry *)it); - } - - de = ll_lookup_it(parent, dentry, it, NULL, NULL); - if (de) - dentry = de; - if ((nd->flags & LOOKUP_OPEN) && !IS_ERR(dentry)) { /* Open */ - if (dentry->d_inode && - it_disposition(it, DISP_OPEN_OPEN)) { /* nocreate */ - if (S_ISFIFO(dentry->d_inode->i_mode)) { - /* We cannot call open here as it might - * deadlock. This case is unreachable in - * practice because of - * OBD_CONNECT_NODEVOH. */ - } else { - struct file *filp; - - nd->intent.open.file->private_data = it; - filp = lookup_instantiate_filp(nd, - dentry, - NULL); - if (IS_ERR(filp)) { - if (de) - dput(de); - de = (struct dentry *)filp; - } - } - } else if (it_disposition(it, DISP_OPEN_CREATE)) { - /* XXX This can only reliably work on assumption - * that there are NO hashed negative dentries.*/ - ll_d2d(dentry)->lld_it = it; - it = NULL; /* Will be freed in ll_create_nd */ - /* We absolutely depend on ll_create_nd to be - * called to not leak this intent and possible - * data attached to it */ - } - } - - if (it) { - ll_intent_release(it); - OBD_FREE(it, sizeof(*it)); - } - } else { - de = ll_lookup_it(parent, dentry, NULL, NULL, NULL); - } - - RETURN(de); -} -#endif /* HAVE_IOP_ATOMIC_OPEN */ - /* We depend on "mode" being set with the proper file type/umask by now */ static struct inode *ll_create_node(struct inode *dir, struct lookup_intent *it) { @@ -970,10 +1346,10 @@ static struct inode *ll_create_node(struct inode *dir, struct lookup_intent *it) LASSERT(it_disposition(it, DISP_ENQ_CREATE_REF)); request = it->it_request; - it_clear_disposition(it, DISP_ENQ_CREATE_REF); - rc = ll_prep_inode(&inode, request, dir->i_sb, it); - if (rc) - GOTO(out, inode = ERR_PTR(rc)); + it_clear_disposition(it, DISP_ENQ_CREATE_REF); + rc = ll_prep_inode(&inode, &request->rq_pill, dir->i_sb, it); + if (rc) + GOTO(out, inode = ERR_PTR(rc)); /* Pause to allow for a race with concurrent access by fid */ OBD_FAIL_TIMEOUT(OBD_FAIL_LLITE_CREATE_NODE_PAUSE, cfs_fail_val); @@ -1006,15 +1382,16 @@ static struct inode *ll_create_node(struct inode *dir, struct lookup_intent *it) */ static int ll_create_it(struct inode *dir, struct dentry *dentry, struct lookup_intent *it, - void *secctx, __u32 secctxlen) + void *secctx, __u32 secctxlen, bool encrypt, + void *encctx, __u32 encctxlen) { struct inode *inode; + __u64 bits = 0; int rc = 0; ENTRY; - CDEBUG(D_VFSTRACE, "VFS Op:name=%.*s, dir="DFID"(%p), intent=%s\n", - dentry->d_name.len, dentry->d_name.name, - PFID(ll_inode2fid(dir)), dir, LL_IT2STR(it)); + CDEBUG(D_VFSTRACE, "VFS Op:name=%pd, dir="DFID"(%p), intent=%s\n", + dentry, PFID(ll_inode2fid(dir)), dir, LL_IT2STR(it)); rc = it_open_error(DISP_OPEN_CREATE, it); if (rc) @@ -1026,24 +1403,36 @@ static int ll_create_it(struct inode *dir, struct dentry *dentry, if ((ll_i2sbi(inode)->ll_flags & LL_SBI_FILE_SECCTX) && secctx != NULL) { - inode_lock(inode); /* must be done before d_instantiate, because it calls * security_d_instantiate, which means a getxattr if security * context is not set yet */ + /* no need to protect selinux_inode_setsecurity() by + * inode_lock. Taking it would lead to a client deadlock + * LU-13617 + */ rc = security_inode_notifysecctx(inode, secctx, secctxlen); - inode_unlock(inode); if (rc) RETURN(rc); } d_instantiate(dentry, inode); + if (encrypt) { + rc = ll_set_encflags(inode, encctx, encctxlen, true); + if (rc) + RETURN(rc); + } + if (!(ll_i2sbi(inode)->ll_flags & LL_SBI_FILE_SECCTX)) { rc = ll_inode_init_security(dentry, inode, dir); if (rc) RETURN(rc); } + ll_set_lock_data(ll_i2sbi(dir)->ll_md_exp, inode, it, &bits); + if (bits & MDS_INODELOCK_LOOKUP) + d_lustre_revalidate(dentry); + RETURN(0); } @@ -1054,32 +1443,34 @@ void ll_update_times(struct ptlrpc_request *request, struct inode *inode) LASSERT(body); if (body->mbo_valid & OBD_MD_FLMTIME && - body->mbo_mtime > LTIME_S(inode->i_mtime)) { - CDEBUG(D_INODE, "setting fid "DFID" mtime from %lu to %llu" - "\n", PFID(ll_inode2fid(inode)), - LTIME_S(inode->i_mtime), body->mbo_mtime); - LTIME_S(inode->i_mtime) = body->mbo_mtime; + body->mbo_mtime > inode->i_mtime.tv_sec) { + CDEBUG(D_INODE, + "setting fid " DFID " mtime from %lld to %llu\n", + PFID(ll_inode2fid(inode)), + (s64)inode->i_mtime.tv_sec, body->mbo_mtime); + inode->i_mtime.tv_sec = body->mbo_mtime; } if (body->mbo_valid & OBD_MD_FLCTIME && - body->mbo_ctime > LTIME_S(inode->i_ctime)) - LTIME_S(inode->i_ctime) = body->mbo_ctime; + body->mbo_ctime > inode->i_ctime.tv_sec) + inode->i_ctime.tv_sec = body->mbo_ctime; } static int ll_new_node(struct inode *dir, struct dentry *dchild, const char *tgt, umode_t mode, int rdev, __u32 opc) { struct qstr *name = &dchild->d_name; - struct ptlrpc_request *request = NULL; - struct md_op_data *op_data; - struct inode *inode = NULL; - struct ll_sb_info *sbi = ll_i2sbi(dir); - int tgt_len = 0; - int err; + struct ptlrpc_request *request = NULL; + struct md_op_data *op_data = NULL; + struct inode *inode = NULL; + struct ll_sb_info *sbi = ll_i2sbi(dir); + int tgt_len = 0; + bool encrypt = false; + int err; - ENTRY; - if (unlikely(tgt != NULL)) - tgt_len = strlen(tgt) + 1; + ENTRY; + if (unlikely(tgt != NULL)) + tgt_len = strlen(tgt) + 1; again: op_data = ll_prep_md_op_data(NULL, dir, NULL, name->name, @@ -1096,17 +1487,33 @@ again: GOTO(err_exit, err); } + if (ll_sbi_has_encrypt(sbi) && + ((IS_ENCRYPTED(dir) && + (S_ISREG(mode) || S_ISDIR(mode) || S_ISLNK(mode))) || + (unlikely(llcrypt_dummy_context_enabled(dir)) && S_ISDIR(mode)))) { + err = llcrypt_get_encryption_info(dir); + if (err) + GOTO(err_exit, err); + if (!llcrypt_has_encryption_key(dir)) + GOTO(err_exit, err = -ENOKEY); + encrypt = true; + } + + if (encrypt) { + err = llcrypt_inherit_context(dir, NULL, op_data, false); + if (err) + GOTO(err_exit, err); + } + err = md_create(sbi->ll_md_exp, op_data, tgt, tgt_len, mode, from_kuid(&init_user_ns, current_fsuid()), from_kgid(&init_user_ns, current_fsgid()), cfs_curproc_cap_pack(), rdev, &request); - if (err < 0 && err != -EREMOTE) - GOTO(err_exit, err); - - /* If the client doesn't know where to create a subdirectory (or - * in case of a race that sends the RPC to the wrong MDS), the - * MDS will return -EREMOTE and the client will fetch the layout - * of the directory, then create the directory on the right MDT. */ +#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(2, 14, 58, 0) + /* + * server < 2.12.58 doesn't pack default LMV in intent_getattr reply, + * fetch default LMV here. + */ if (unlikely(err == -EREMOTE)) { struct ll_inode_info *lli = ll_i2info(dir); struct lmv_user_md *lum; @@ -1118,48 +1525,91 @@ again: err2 = ll_dir_getstripe(dir, (void **)&lum, &lumsize, &request, OBD_MD_DEFAULT_MEA); + ll_finish_md_op_data(op_data); + op_data = NULL; if (err2 == 0) { - /* Update stripe_offset and retry */ - lli->lli_def_stripe_offset = lum->lum_stripe_offset; - } else if (err2 == -ENODATA && - lli->lli_def_stripe_offset != -1) { - /* If there are no default stripe EA on the MDT, but the + struct lustre_md md = { NULL }; + + md.body = req_capsule_server_get(&request->rq_pill, + &RMF_MDT_BODY); + if (!md.body) + GOTO(err_exit, err = -EPROTO); + + OBD_ALLOC_PTR(md.default_lmv); + if (!md.default_lmv) + GOTO(err_exit, err = -ENOMEM); + + md.default_lmv->lsm_md_magic = lum->lum_magic; + md.default_lmv->lsm_md_stripe_count = + lum->lum_stripe_count; + md.default_lmv->lsm_md_master_mdt_index = + lum->lum_stripe_offset; + md.default_lmv->lsm_md_hash_type = lum->lum_hash_type; + md.default_lmv->lsm_md_max_inherit = + lum->lum_max_inherit; + md.default_lmv->lsm_md_max_inherit_rr = + lum->lum_max_inherit_rr; + + err = ll_update_inode(dir, &md); + md_free_lustre_md(sbi->ll_md_exp, &md); + if (err) + GOTO(err_exit, err); + } else if (err2 == -ENODATA && lli->lli_default_lsm_md) { + /* + * If there are no default stripe EA on the MDT, but the * client has default stripe, then it probably means - * default stripe EA has just been deleted. */ - lli->lli_def_stripe_offset = -1; + * default stripe EA has just been deleted. + */ + down_write(&lli->lli_lsm_sem); + if (lli->lli_default_lsm_md) + OBD_FREE_PTR(lli->lli_default_lsm_md); + lli->lli_default_lsm_md = NULL; + up_write(&lli->lli_lsm_sem); } else { GOTO(err_exit, err); } ptlrpc_req_finished(request); request = NULL; - ll_finish_md_op_data(op_data); goto again; } +#endif + + if (err < 0) + GOTO(err_exit, err); ll_update_times(request, dir); CFS_FAIL_TIMEOUT(OBD_FAIL_LLITE_NEWNODE_PAUSE, cfs_fail_val); - err = ll_prep_inode(&inode, request, dchild->d_sb, NULL); + err = ll_prep_inode(&inode, &request->rq_pill, dchild->d_sb, NULL); if (err) GOTO(err_exit, err); if (sbi->ll_flags & LL_SBI_FILE_SECCTX) { - inode_lock(inode); /* must be done before d_instantiate, because it calls * security_d_instantiate, which means a getxattr if security * context is not set yet */ + /* no need to protect selinux_inode_setsecurity() by + * inode_lock. Taking it would lead to a client deadlock + * LU-13617 + */ err = security_inode_notifysecctx(inode, op_data->op_file_secctx, op_data->op_file_secctx_size); - inode_unlock(inode); if (err) GOTO(err_exit, err); } d_instantiate(dchild, inode); + if (encrypt) { + err = ll_set_encflags(inode, op_data->op_file_encctx, + op_data->op_file_encctx_size, true); + if (err) + GOTO(err_exit, err); + } + if (!(sbi->ll_flags & LL_SBI_FILE_SECCTX)) { err = ll_inode_init_security(dchild, inode, dir); if (err) @@ -1174,144 +1624,95 @@ err_exit: if (!IS_ERR_OR_NULL(op_data)) ll_finish_md_op_data(op_data); - return err; + RETURN(err); } -static int ll_mknod(struct inode *dir, struct dentry *dchild, ll_umode_t mode, +static int ll_mknod(struct inode *dir, struct dentry *dchild, umode_t mode, dev_t rdev) { - struct qstr *name = &dchild->d_name; + ktime_t kstart = ktime_get(); int err; - ENTRY; + ENTRY; - CDEBUG(D_VFSTRACE, "VFS Op:name=%.*s, dir="DFID"(%p) mode %o dev %x\n", - name->len, name->name, PFID(ll_inode2fid(dir)), dir, - mode, rdev); + CDEBUG(D_VFSTRACE, "VFS Op:name=%pd, dir="DFID"(%p) mode %o dev %x\n", + dchild, PFID(ll_inode2fid(dir)), dir, mode, rdev); if (!IS_POSIXACL(dir) || !exp_connect_umask(ll_i2mdexp(dir))) mode &= ~current_umask(); - switch (mode & S_IFMT) { - case 0: - mode |= S_IFREG; /* for mode = 0 case, fallthrough */ - case S_IFREG: - case S_IFCHR: - case S_IFBLK: - case S_IFIFO: - case S_IFSOCK: + switch (mode & S_IFMT) { + case 0: + mode |= S_IFREG; + /* fallthrough */ + case S_IFREG: + case S_IFCHR: + case S_IFBLK: + case S_IFIFO: + case S_IFSOCK: err = ll_new_node(dir, dchild, NULL, mode, old_encode_dev(rdev), LUSTRE_OPC_MKNOD); - break; - case S_IFDIR: - err = -EPERM; - break; - default: - err = -EINVAL; - } - - if (!err) - ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_MKNOD, 1); - - RETURN(err); + break; + case S_IFDIR: + err = -EPERM; + break; + default: + err = -EINVAL; + } + + if (!err) + ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_MKNOD, + ktime_us_delta(ktime_get(), kstart)); + + RETURN(err); } -#ifdef HAVE_IOP_ATOMIC_OPEN /* * Plain create. Intent create is handled in atomic_open. */ static int ll_create_nd(struct inode *dir, struct dentry *dentry, umode_t mode, bool want_excl) { + ktime_t kstart = ktime_get(); int rc; CFS_FAIL_TIMEOUT(OBD_FAIL_LLITE_CREATE_FILE_PAUSE, cfs_fail_val); - CDEBUG(D_VFSTRACE, "VFS Op:name=%.*s, dir="DFID"(%p), " - "flags=%u, excl=%d\n", dentry->d_name.len, - dentry->d_name.name, PFID(ll_inode2fid(dir)), - dir, mode, want_excl); + CDEBUG(D_VFSTRACE, + "VFS Op:name=%pd, dir="DFID"(%p), flags=%u, excl=%d\n", + dentry, PFID(ll_inode2fid(dir)), dir, mode, want_excl); /* Using mknod(2) to create a regular file is designed to not recognize * volatile file name, so we use ll_mknod() here. */ rc = ll_mknod(dir, dentry, mode, 0); - ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_CREATE, 1); + CDEBUG(D_VFSTRACE, "VFS Op:name=%pd, unhashed %d\n", + dentry, d_unhashed(dentry)); - CDEBUG(D_VFSTRACE, "VFS Op:name=%.*s, unhashed %d\n", - dentry->d_name.len, dentry->d_name.name, d_unhashed(dentry)); + if (!rc) + ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_CREATE, + ktime_us_delta(ktime_get(), kstart)); return rc; } -#else /* !HAVE_IOP_ATOMIC_OPEN */ -static int ll_create_nd(struct inode *dir, struct dentry *dentry, - ll_umode_t mode, struct nameidata *nd) -{ - struct ll_dentry_data *lld = ll_d2d(dentry); - struct lookup_intent *it = NULL; - int rc; - - CFS_FAIL_TIMEOUT(OBD_FAIL_LLITE_CREATE_FILE_PAUSE, cfs_fail_val); - - if (lld != NULL) - it = lld->lld_it; - - if (!it) { - /* LU-8559: use LUSTRE_OPC_CREATE for non atomic open case - * so that volatile file name is recoginized. - * Mknod(2), however, is designed to not recognize volatile - * file name to avoid inode leak under orphan directory until - * MDT reboot */ - return ll_new_node(dir, dentry, NULL, mode, 0, - LUSTRE_OPC_CREATE); - } - - lld->lld_it = NULL; - - /* Was there an error? Propagate it! */ - if (it->it_status) { - rc = it->it_status; - goto out; - } - - rc = ll_create_it(dir, dentry, it, NULL, 0); - if (nd && (nd->flags & LOOKUP_OPEN) && dentry->d_inode) { /* Open */ - struct file *filp; - - nd->intent.open.file->private_data = it; - filp = lookup_instantiate_filp(nd, dentry, NULL); - if (IS_ERR(filp)) - rc = PTR_ERR(filp); - } - -out: - ll_intent_release(it); - OBD_FREE(it, sizeof(*it)); - - if (!rc) - ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_CREATE, 1); - - return rc; -} -#endif /* HAVE_IOP_ATOMIC_OPEN */ static int ll_symlink(struct inode *dir, struct dentry *dchild, const char *oldpath) { - struct qstr *name = &dchild->d_name; + ktime_t kstart = ktime_get(); int err; ENTRY; - CDEBUG(D_VFSTRACE, "VFS Op:name=%.*s, dir="DFID"(%p), target=%.*s\n", - name->len, name->name, PFID(ll_inode2fid(dir)), - dir, 3000, oldpath); + CDEBUG(D_VFSTRACE, "VFS Op:name=%pd, dir="DFID"(%p), target=%.*s\n", + dchild, PFID(ll_inode2fid(dir)), dir, 3000, oldpath); err = ll_new_node(dir, dchild, oldpath, S_IFLNK | S_IRWXUGO, 0, LUSTRE_OPC_SYMLINK); - if (!err) - ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_SYMLINK, 1); + if (!err) + ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_SYMLINK, + ktime_us_delta(ktime_get(), kstart)); - RETURN(err); + RETURN(err); } static int ll_link(struct dentry *old_dentry, struct inode *dir, @@ -1322,39 +1723,46 @@ static int ll_link(struct dentry *old_dentry, struct inode *dir, struct ll_sb_info *sbi = ll_i2sbi(dir); struct ptlrpc_request *request = NULL; struct md_op_data *op_data; + ktime_t kstart = ktime_get(); int err; ENTRY; - CDEBUG(D_VFSTRACE, "VFS Op: inode="DFID"(%p), dir="DFID"(%p), " - "target=%.*s\n", PFID(ll_inode2fid(src)), src, - PFID(ll_inode2fid(dir)), dir, name->len, name->name); - - op_data = ll_prep_md_op_data(NULL, src, dir, name->name, name->len, - 0, LUSTRE_OPC_ANY, NULL); - if (IS_ERR(op_data)) - RETURN(PTR_ERR(op_data)); - - err = md_link(sbi->ll_md_exp, op_data, &request); - ll_finish_md_op_data(op_data); - if (err) - GOTO(out, err); - - ll_update_times(request, dir); - ll_stats_ops_tally(sbi, LPROC_LL_LINK, 1); - EXIT; + CDEBUG(D_VFSTRACE, + "VFS Op: inode="DFID"(%p), dir="DFID"(%p), target=%pd\n", + PFID(ll_inode2fid(src)), src, + PFID(ll_inode2fid(dir)), dir, new_dentry); + + err = llcrypt_prepare_link(old_dentry, dir, new_dentry); + if (err) + RETURN(err); + + op_data = ll_prep_md_op_data(NULL, src, dir, name->name, name->len, + 0, LUSTRE_OPC_ANY, NULL); + if (IS_ERR(op_data)) + RETURN(PTR_ERR(op_data)); + + err = md_link(sbi->ll_md_exp, op_data, &request); + ll_finish_md_op_data(op_data); + if (err) + GOTO(out, err); + + ll_update_times(request, dir); + ll_stats_ops_tally(sbi, LPROC_LL_LINK, + ktime_us_delta(ktime_get(), kstart)); + EXIT; out: - ptlrpc_req_finished(request); - RETURN(err); + ptlrpc_req_finished(request); + RETURN(err); } -static int ll_mkdir(struct inode *dir, struct dentry *dchild, ll_umode_t mode) +static int ll_mkdir(struct inode *dir, struct dentry *dchild, umode_t mode) { - struct qstr *name = &dchild->d_name; - int err; - ENTRY; + ktime_t kstart = ktime_get(); + int err; + ENTRY; - CDEBUG(D_VFSTRACE, "VFS Op:name=%.*s, dir="DFID"(%p)\n", - name->len, name->name, PFID(ll_inode2fid(dir)), dir); + CDEBUG(D_VFSTRACE, "VFS Op:name=%pd, dir="DFID"(%p)\n", + dchild, PFID(ll_inode2fid(dir)), dir); if (!IS_POSIXACL(dir) || !exp_connect_umask(ll_i2mdexp(dir))) mode &= ~current_umask(); @@ -1363,7 +1771,8 @@ static int ll_mkdir(struct inode *dir, struct dentry *dchild, ll_umode_t mode) err = ll_new_node(dir, dchild, NULL, mode, 0, LUSTRE_OPC_MKDIR); if (err == 0) - ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_MKDIR, 1); + ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_MKDIR, + ktime_us_delta(ktime_get(), kstart)); RETURN(err); } @@ -1371,35 +1780,54 @@ static int ll_mkdir(struct inode *dir, struct dentry *dchild, ll_umode_t mode) static int ll_rmdir(struct inode *dir, struct dentry *dchild) { struct qstr *name = &dchild->d_name; - struct ptlrpc_request *request = NULL; - struct md_op_data *op_data; - int rc; - ENTRY; + struct ptlrpc_request *request = NULL; + struct md_op_data *op_data; + ktime_t kstart = ktime_get(); + int rc; - CDEBUG(D_VFSTRACE, "VFS Op:name=%.*s, dir="DFID"(%p)\n", - name->len, name->name, PFID(ll_inode2fid(dir)), dir); + ENTRY; + + CDEBUG(D_VFSTRACE, "VFS Op:name=%pd, dir="DFID"(%p)\n", + dchild, PFID(ll_inode2fid(dir)), dir); if (unlikely(d_mountpoint(dchild))) RETURN(-EBUSY); - op_data = ll_prep_md_op_data(NULL, dir, NULL, name->name, name->len, - S_IFDIR, LUSTRE_OPC_ANY, NULL); - if (IS_ERR(op_data)) - RETURN(PTR_ERR(op_data)); + /* some foreign dir may not be allowed to be removed */ + if (!ll_foreign_is_removable(dchild, false)) + RETURN(-EPERM); + + op_data = ll_prep_md_op_data(NULL, dir, NULL, name->name, name->len, + S_IFDIR, LUSTRE_OPC_ANY, NULL); + if (IS_ERR(op_data)) + RETURN(PTR_ERR(op_data)); if (dchild->d_inode != NULL) op_data->op_fid3 = *ll_inode2fid(dchild->d_inode); op_data->op_fid2 = op_data->op_fid3; - rc = md_unlink(ll_i2sbi(dir)->ll_md_exp, op_data, &request); - ll_finish_md_op_data(op_data); - if (rc == 0) { - ll_update_times(request, dir); - ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_RMDIR, 1); - } + rc = md_unlink(ll_i2sbi(dir)->ll_md_exp, op_data, &request); + ll_finish_md_op_data(op_data); + if (!rc) { + struct mdt_body *body; - ptlrpc_req_finished(request); - RETURN(rc); + ll_update_times(request, dir); + ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_RMDIR, + ktime_us_delta(ktime_get(), kstart)); + + /* + * The server puts attributes in on the last unlink, use them + * to update the link count so the inode can be freed + * immediately. + */ + body = req_capsule_server_get(&request->rq_pill, &RMF_MDT_BODY); + if (body->mbo_valid & OBD_MD_FLNLINK) + set_nlink(dchild->d_inode, body->mbo_nlink); + } + + ptlrpc_req_finished(request); + + RETURN(rc); } /** @@ -1409,6 +1837,7 @@ int ll_rmdir_entry(struct inode *dir, char *name, int namelen) { struct ptlrpc_request *request = NULL; struct md_op_data *op_data; + ktime_t kstart = ktime_get(); int rc; ENTRY; @@ -1422,12 +1851,13 @@ int ll_rmdir_entry(struct inode *dir, char *name, int namelen) op_data->op_cli_flags |= CLI_RM_ENTRY; rc = md_unlink(ll_i2sbi(dir)->ll_md_exp, op_data, &request); ll_finish_md_op_data(op_data); - if (rc == 0) { + if (!rc) ll_update_times(request, dir); - ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_RMDIR, 1); - } ptlrpc_req_finished(request); + if (!rc) + ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_RMDIR, + ktime_us_delta(ktime_get(), kstart)); RETURN(rc); } @@ -1437,10 +1867,13 @@ static int ll_unlink(struct inode *dir, struct dentry *dchild) struct ptlrpc_request *request = NULL; struct md_op_data *op_data; struct mdt_body *body; + ktime_t kstart = ktime_get(); int rc; + ENTRY; - CDEBUG(D_VFSTRACE, "VFS Op:name=%.*s, dir="DFID"(%p)\n", - name->len, name->name, PFID(ll_inode2fid(dir)), dir); + + CDEBUG(D_VFSTRACE, "VFS Op:name=%pd, dir="DFID"(%p)\n", + dchild, PFID(ll_inode2fid(dir)), dir); /* * XXX: unlink bind mountpoint maybe call to here, @@ -1449,13 +1882,21 @@ static int ll_unlink(struct inode *dir, struct dentry *dchild) if (unlikely(d_mountpoint(dchild))) RETURN(-EBUSY); + /* some foreign file/dir may not be allowed to be unlinked */ + if (!ll_foreign_is_removable(dchild, false)) + RETURN(-EPERM); + op_data = ll_prep_md_op_data(NULL, dir, NULL, name->name, name->len, 0, LUSTRE_OPC_ANY, NULL); if (IS_ERR(op_data)) RETURN(PTR_ERR(op_data)); op_data->op_fid3 = *ll_inode2fid(dchild->d_inode); - + /* notify lower layer if inode has dirty pages */ + if (S_ISREG(dchild->d_inode->i_mode) && + ll_i2info(dchild->d_inode)->lli_clob && + dirty_cnt(dchild->d_inode)) + op_data->op_cli_flags |= CLI_DIRTY_DATA; op_data->op_fid2 = op_data->op_fid3; rc = md_unlink(ll_i2sbi(dir)->ll_md_exp, op_data, &request); ll_finish_md_op_data(op_data); @@ -1471,10 +1912,12 @@ static int ll_unlink(struct inode *dir, struct dentry *dchild) set_nlink(dchild->d_inode, body->mbo_nlink); ll_update_times(request, dir); - ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_UNLINK, 1); out: ptlrpc_req_finished(request); + if (!rc) + ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_UNLINK, + ktime_us_delta(ktime_get(), kstart)); RETURN(rc); } @@ -1490,6 +1933,8 @@ static int ll_rename(struct inode *src, struct dentry *src_dchild, struct ptlrpc_request *request = NULL; struct ll_sb_info *sbi = ll_i2sbi(src); struct md_op_data *op_data; + ktime_t kstart = ktime_get(); + umode_t mode = 0; int err; ENTRY; @@ -1498,49 +1943,67 @@ static int ll_rename(struct inode *src, struct dentry *src_dchild, return -EINVAL; #endif - CDEBUG(D_VFSTRACE, "VFS Op:oldname=%.*s, src_dir="DFID - "(%p), newname=%.*s, tgt_dir="DFID"(%p)\n", - src_name->len, src_name->name, - PFID(ll_inode2fid(src)), src, tgt_name->len, - tgt_name->name, PFID(ll_inode2fid(tgt)), tgt); + CDEBUG(D_VFSTRACE, + "VFS Op:oldname=%pd, src_dir="DFID"(%p), newname=%pd, tgt_dir="DFID"(%p)\n", + src_dchild, PFID(ll_inode2fid(src)), src, + tgt_dchild, PFID(ll_inode2fid(tgt)), tgt); if (unlikely(d_mountpoint(src_dchild) || d_mountpoint(tgt_dchild))) RETURN(-EBUSY); - op_data = ll_prep_md_op_data(NULL, src, tgt, NULL, 0, 0, +#ifdef HAVE_IOPS_RENAME_WITH_FLAGS + err = llcrypt_prepare_rename(src, src_dchild, tgt, tgt_dchild, flags); +#else + err = llcrypt_prepare_rename(src, src_dchild, tgt, tgt_dchild, 0); +#endif + if (err) + RETURN(err); + /* we prevent an encrypted file from being renamed + * into an unencrypted dir + */ + if (IS_ENCRYPTED(src) && !IS_ENCRYPTED(tgt)) + RETURN(-EXDEV); + + if (src_dchild->d_inode) + mode = src_dchild->d_inode->i_mode; + + if (tgt_dchild->d_inode) + mode = tgt_dchild->d_inode->i_mode; + + op_data = ll_prep_md_op_data(NULL, src, tgt, NULL, 0, mode, LUSTRE_OPC_ANY, NULL); if (IS_ERR(op_data)) RETURN(PTR_ERR(op_data)); - if (src_dchild->d_inode != NULL) + if (src_dchild->d_inode) op_data->op_fid3 = *ll_inode2fid(src_dchild->d_inode); - if (tgt_dchild->d_inode != NULL) + if (tgt_dchild->d_inode) op_data->op_fid4 = *ll_inode2fid(tgt_dchild->d_inode); - err = md_rename(sbi->ll_md_exp, op_data, - src_name->name, src_name->len, - tgt_name->name, tgt_name->len, &request); - ll_finish_md_op_data(op_data); - if (!err) { - ll_update_times(request, src); - ll_update_times(request, tgt); - ll_stats_ops_tally(sbi, LPROC_LL_RENAME, 1); - } + err = md_rename(sbi->ll_md_exp, op_data, + src_name->name, src_name->len, + tgt_name->name, tgt_name->len, &request); + ll_finish_md_op_data(op_data); + if (!err) { + ll_update_times(request, src); + ll_update_times(request, tgt); + } - ptlrpc_req_finished(request); + ptlrpc_req_finished(request); - if (err == 0) + if (!err) { d_move(src_dchild, tgt_dchild); + ll_stats_ops_tally(sbi, LPROC_LL_RENAME, + ktime_us_delta(ktime_get(), kstart)); + } RETURN(err); } const struct inode_operations ll_dir_inode_operations = { .mknod = ll_mknod, -#ifdef HAVE_IOP_ATOMIC_OPEN .atomic_open = ll_atomic_open, -#endif .lookup = ll_lookup_nd, .create = ll_create_nd, /* We need all these non-raw things for NFSD, to not patch it. */ @@ -1559,9 +2022,7 @@ const struct inode_operations ll_dir_inode_operations = { .removexattr = ll_removexattr, #endif .listxattr = ll_listxattr, -#ifdef HAVE_IOP_GET_ACL .get_acl = ll_get_acl, -#endif #ifdef HAVE_IOP_SET_ACL .set_acl = ll_set_acl, #endif @@ -1577,9 +2038,7 @@ const struct inode_operations ll_special_inode_operations = { .removexattr = ll_removexattr, #endif .listxattr = ll_listxattr, -#ifdef HAVE_IOP_GET_ACL .get_acl = ll_get_acl, -#endif #ifdef HAVE_IOP_SET_ACL .set_acl = ll_set_acl, #endif