X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=blobdiff_plain;f=lustre%2Fllite%2Fnamei.c;h=03911dc5ae8b0da23b34d83c0ef6154d6452ff96;hp=632955165113d32567b759d121d3083878d8c898;hb=707bab62f5d6c704b30e4ee9e769b5c9f026e1e7;hpb=34e7d46234e5c957e1e815c5267b13fe610a9d8d diff --git a/lustre/llite/namei.c b/lustre/llite/namei.c index 6329551..03911dc 100644 --- a/lustre/llite/namei.c +++ b/lustre/llite/namei.c @@ -15,11 +15,7 @@ * * 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 - * - * 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. + * http://www.gnu.org/licenses/gpl-2.0.html * * GPL HEADER END */ @@ -27,7 +23,7 @@ * Copyright (c) 2002, 2010, Oracle and/or its affiliates. All rights reserved. * Use is subject to license terms. * - * Copyright (c) 2011, 2014, Intel Corporation. + * Copyright (c) 2011, 2017, Intel Corporation. */ /* * This file is part of Lustre, http://www.lustre.org/ @@ -41,17 +37,21 @@ #include #include #include +#include +#ifdef HAVE_UIDGID_HEADER +# include +#endif #define DEBUG_SUBSYSTEM S_LLITE #include #include #include -#include #include "llite_internal.h" static int ll_create_it(struct inode *dir, struct dentry *dentry, - struct lookup_intent *it); + struct lookup_intent *it, + void *secctx, __u32 secctxlen); /* called from iget5_locked->find_inode() under inode_lock spinlock */ static int ll_test_inode(struct inode *inode, void *opaque) @@ -83,7 +83,7 @@ static int ll_set_inode(struct inode *inode, void *opaque) lli->lli_fid = body->mbo_fid1; if (unlikely(!(body->mbo_valid & OBD_MD_FLTYPE))) { CERROR("Can not initialize inode "DFID" without object type: " - "valid = "LPX64"\n", + "valid = %#llx\n", PFID(&lli->lli_fid), body->mbo_valid); return -EINVAL; } @@ -120,25 +120,34 @@ struct inode *ll_iget(struct super_block *sb, ino_t hash, if (inode->i_state & I_NEW) { rc = ll_read_inode2(inode, md); if (rc == 0 && S_ISREG(inode->i_mode) && - ll_i2info(inode)->lli_clob == NULL) { - CDEBUG(D_INODE, "%s: apply lsm %p to inode "DFID"\n", - ll_get_fsname(sb, NULL, 0), md->lsm, - PFID(ll_inode2fid(inode))); + ll_i2info(inode)->lli_clob == NULL) rc = cl_file_inode_init(inode, md); - } + if (rc != 0) { + /* Let's clear directory lsm here, otherwise + * make_bad_inode() will reset the inode mode + * to regular, then ll_clear_inode will not + * be able to clear lsm_md */ + if (S_ISDIR(inode->i_mode)) + ll_dir_clear_lsm_md(inode); make_bad_inode(inode); unlock_new_inode(inode); iput(inode); inode = ERR_PTR(rc); } else { + inode_has_no_xattr(inode); unlock_new_inode(inode); } + } else if (is_bad_inode(inode)) { + iput(inode); + inode = ERR_PTR(-ESTALE); } else if (!(inode->i_state & (I_FREEING | I_CLEAR))) { rc = ll_update_inode(inode, md); CDEBUG(D_VFSTRACE, "got inode: "DFID"(%p): rc = %d\n", PFID(&md->body->mbo_fid1), inode, rc); if (rc != 0) { + if (S_ISDIR(inode->i_mode)) + ll_dir_clear_lsm_md(inode); iput(inode); inode = ERR_PTR(rc); } @@ -153,14 +162,14 @@ static void ll_invalidate_negative_children(struct inode *dir) DECLARE_LL_D_HLIST_NODE_PTR(p); ll_lock_dcache(dir); - ll_d_hlist_for_each_entry(dentry, p, &dir->i_dentry, d_alias) { + ll_d_hlist_for_each_entry(dentry, p, &dir->i_dentry) { 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_u.d_child) { + d_child) { if (child->d_inode == NULL) d_lustre_invalidate(child, 1); } @@ -175,148 +184,298 @@ int ll_test_inode_by_fid(struct inode *inode, void *opaque) return lu_fid_eq(&ll_i2info(inode)->lli_fid, opaque); } -int ll_md_blocking_ast(struct ldlm_lock *lock, struct ldlm_lock_desc *desc, - void *data, int flag) +static int ll_dom_lock_cancel(struct inode *inode, struct ldlm_lock *lock) { - struct lustre_handle lockh; + struct lu_env *env; + struct ll_inode_info *lli = ll_i2info(inode); + __u16 refcheck; 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 (!lli->lli_clob) { + /* due to DoM read on open, there may exist pages for Lustre + * regular file even though cl_object is not set up yet. */ + truncate_inode_pages(inode->i_mapping, 0); + RETURN(0); + } - /* Inode is set to lock->l_resource->lr_lvb_inode - * for mdc - bug 24555 */ - LASSERT(lock->l_ast_data == NULL); + env = cl_env_get(&refcheck); + if (IS_ERR(env)) + RETURN(PTR_ERR(env)); - if (inode == NULL) - break; + /* reach MDC layer to flush data under the DoM ldlm lock */ + rc = cl_object_flush(env, lli->lli_clob, lock); - /* Invalidate all dentries associated with this inode */ - LASSERT(ldlm_is_canceling(lock)); + cl_env_put(env, &refcheck); + RETURN(rc); +} - 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(); - } +static void ll_lock_cancel_bits(struct ldlm_lock *lock, __u64 to_cancel) +{ + struct inode *inode = ll_inode_from_resource_lock(lock); + struct ll_inode_info *lli; + __u64 bits = to_cancel; + int rc; - 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; - } + ENTRY; - /* 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(); - } + 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; + } - ll_md_real_close(inode, 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); + LBUG(); + } - bits &= ~MDS_INODELOCK_OPEN; - } + if (bits & MDS_INODELOCK_XATTR) { + ll_xattr_cache_destroy(inode); + bits &= ~MDS_INODELOCK_XATTR; + } - if (bits & (MDS_INODELOCK_LOOKUP | MDS_INODELOCK_UPDATE | - MDS_INODELOCK_LAYOUT | MDS_INODELOCK_PERM)) - ll_have_md_lock(inode, &bits, LCK_MINMODE); - - 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); - } + /* 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_UPDATE) { - struct ll_inode_info *lli = ll_i2info(inode); + if (bits & MDS_INODELOCK_OPEN) { + fmode_t fmode; - spin_lock(&lli->lli_lock); - lli->lli_flags &= ~LLIF_MDS_SIZE_LOCK; - spin_unlock(&lli->lli_lock); + 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_UPDATE) && S_ISDIR(inode->i_mode)) { - struct ll_inode_info *lli = ll_i2info(inode); + ll_md_real_close(inode, fmode); + + bits &= ~MDS_INODELOCK_OPEN; + } + + 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); + } + + 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); + } + + lli = ll_i2info(inode); + + if (bits & MDS_INODELOCK_UPDATE) + lli->lli_update_atime = 1; - 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))); - master_inode = ilookup5(inode->i_sb, hash, + /* 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 != NULL && - !IS_ERR(master_inode)) { - ll_invalidate_negative_children( - master_inode); - iput(master_inode); - } - } else { - ll_invalidate_negative_children(inode); + if (master_inode) { + ll_invalidate_negative_children(master_inode); + iput(master_inode); } + } else { + ll_invalidate_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)) && + inode->i_sb->s_root != NULL && + inode != inode->i_sb->s_root->d_inode) + ll_invalidate_aliases(inode); - iput(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; + case LCK_PW: + mode |= LCK_CR; + break; + case LCK_CW: + mode = LCK_CW; + 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, + ktime_set(ns->ns_dirty_age_limit, 0)))) { + 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 *desc, + void *data, int flag) +{ + struct lustre_handle lockh; + __u64 bits = lock->l_policy_data.l_inodebits.bits; + int rc; + + ENTRY; + + switch (flag) { + case LDLM_CB_BLOCKING: + { + __u64 cancel_flags = LCF_ASYNC; + + if (ll_md_need_convert(lock)) { + cancel_flags |= LCF_CONVERT; + /* For lock convert some cancel actions may require + * this lock with non-dropped canceled bits, e.g. page + * flush for DOM lock. So call ll_lock_cancel_bits() + * here while canceled bits are still set. + */ + bits = lock->l_policy_data.l_inodebits.cancel_bits; + if (bits & MDS_INODELOCK_DOM) + ll_lock_cancel_bits(lock, MDS_INODELOCK_DOM); + } + 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: + /* Nothing to do for non-granted locks */ + if (!ldlm_is_granted(lock)) + break; + + if (ldlm_is_converting(lock)) { + /* this is called on already converted lock, so + * ibits has remained bits only and cancel_bits + * are bits that were dropped. + * Note that DOM lock is handled prior lock convert + * and is excluded here. + */ + bits = lock->l_policy_data.l_inodebits.cancel_bits & + ~MDS_INODELOCK_DOM; + } else { + LASSERT(ldlm_is_canceling(lock)); + } + ll_lock_cancel_bits(lock, bits); + break; default: LBUG(); } @@ -368,11 +527,12 @@ static struct dentry *ll_find_alias(struct inode *inode, struct dentry *dentry) discon_alias = invalid_alias = NULL; ll_lock_dcache(inode); - ll_d_hlist_for_each_entry(alias, p, &inode->i_dentry, d_alias) { + ll_d_hlist_for_each_entry(alias, p, &inode->i_dentry) { LASSERT(alias != dentry); spin_lock(&alias->d_lock); - if (alias->d_flags & DCACHE_DISCONNECTED) + if ((alias->d_flags & DCACHE_DISCONNECTED) && + S_ISDIR(inode->i_mode)) /* LASSERT(last_discon == NULL); LU-405, bz 20055 */ discon_alias = alias; else if (alias->d_parent == dentry->d_parent && @@ -433,75 +593,103 @@ 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) { struct inode *inode = NULL; __u64 bits = 0; int rc; + struct dentry *alias; ENTRY; /* NB 1 request reference will be taken away by ll_intent_lock() * when I return */ CDEBUG(D_DENTRY, "it %p it_disposition %x\n", it, - it->d.lustre.it_disposition); + 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, (*de)->d_sb, it); + if (rc) + RETURN(rc); + + if (it->it_op & IT_OPEN) + ll_dom_finish_open(inode, request, it); + + 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. + */ + + /* 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) { + inode_lock(inode); + rc = security_inode_notifysecctx(inode, secctx, + secctxlen); + inode_unlock(inode); + if (rc) + CWARN("cannot set security context for " + DFID": rc = %d\n", + PFID(ll_inode2fid(inode)), rc); + } } /* Only hash *de if it is unhashed (new dentry). * Atoimc_open may passin hashed dentries for open. */ - if (d_unhashed(*de)) { - struct dentry *alias; - - alias = ll_splice_alias(inode, *de); - if (IS_ERR(alias)) - RETURN(PTR_ERR(alias)); - *de = alias; - } else if (!it_disposition(it, DISP_LOOKUP_NEG) && - !it_disposition(it, DISP_OPEN_CREATE)) { - /* With DISP_OPEN_CREATE dentry will - instantiated in ll_create_it. */ - LASSERT((*de)->d_inode == NULL); - d_instantiate(*de, inode); - } + alias = ll_splice_alias(inode, *de); + if (IS_ERR(alias)) + GOTO(out, rc = PTR_ERR(alias)); + + *de = alias; if (!it_disposition(it, DISP_LOOKUP_NEG)) { /* we have lookup look - unhide dentry */ if (bits & MDS_INODELOCK_LOOKUP) d_lustre_revalidate(*de); } 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 = { .it_op = IT_GETATTR, - .d.lustre.it_lock_handle = 0 }; + .it_lock_handle = 0 }; 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, (*de)->d_name.len, &fid); if (rc != 0) - RETURN(rc); + GOTO(out, rc); } if (md_revalidate_lock(ll_i2mdexp(parent), &parent_it, &fid, @@ -509,21 +697,37 @@ static int ll_lookup_it_finish(struct ptlrpc_request *request, d_lustre_revalidate(*de); ll_intent_release(&parent_it); } - } + } + + GOTO(out, rc = 0); + +out: + if (rc != 0 && it->it_op & IT_OPEN) + ll_open_cleanup((*de)->d_sb, request); - RETURN(0); + return rc; } +struct pcc_create_attach { + struct pcc_dataset *pca_dataset; + struct dentry *pca_dentry; +}; + static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry, - struct lookup_intent *it) + struct lookup_intent *it, + void **secctx, __u32 *secctxlen, + struct pcc_create_attach *pca) { 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)); @@ -544,6 +748,10 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry, RETURN(dentry == save ? NULL : dentry); } + if (it->it_op & IT_OPEN && it->it_flags & FMODE_WRITE && + dentry->d_sb->s_flags & SB_RDONLY) + RETURN(ERR_PTR(-EROFS)); + if (it->it_op & IT_CREAT) opc = LUSTRE_OPC_CREATE; else @@ -552,12 +760,75 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry, op_data = ll_prep_md_op_data(NULL, parent, NULL, dentry->d_name.name, dentry->d_name.len, 0, opc, NULL); if (IS_ERR(op_data)) - RETURN((void *)op_data); + GOTO(out, retval = ERR_CAST(op_data)); /* enforce umask if acl disabled or MDS doesn't support umask */ if (!IS_POSIXACL(parent) || !exp_connect_umask(ll_i2mdexp(parent))) it->it_create_mode &= ~current_umask(); + if (it->it_op & IT_CREAT && + ll_i2sbi(parent)->ll_flags & LL_SBI_FILE_SECCTX) { + rc = ll_dentry_init_security(dentry, it->it_create_mode, + &dentry->d_name, + &op_data->op_file_secctx_name, + &op_data->op_file_secctx, + &op_data->op_file_secctx_size); + if (rc < 0) + GOTO(out, retval = ERR_PTR(rc)); + if (secctx != NULL) + *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; + } + + /* 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) { + struct pcc_dataset *dataset = 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 = dataset->pccd_rwid; + + rc = obd_fid_alloc(NULL, ll_i2mdexp(parent), &op_data->op_fid2, + op_data); + if (rc) + GOTO(out, retval = ERR_PTR(rc)); + + rc = pcc_inode_create(parent->i_sb, dataset, &op_data->op_fid2, + &pca->pca_dentry); + if (rc) + GOTO(out, retval = ERR_PTR(rc)); + + it->it_flags |= MDS_OPEN_PCC; + } + rc = md_intent_lock(ll_i2mdexp(parent), op_data, it, &req, &ll_md_blocking_ast, 0); /* If the MDS allows the client to chgrp (CFS_SETGRP_PERM), but the @@ -589,7 +860,11 @@ 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); + /* 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); if (rc != 0) { ll_intent_release(it); GOTO(out, retval = ERR_PTR(rc)); @@ -605,8 +880,18 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry, GOTO(out, retval = (dentry == save) ? NULL : dentry); out: - if (op_data != NULL && !IS_ERR(op_data)) + if (op_data != NULL && !IS_ERR(op_data)) { + if (secctx != NULL && secctxlen != NULL) { + /* caller needs sec ctx info, so reset it in op_data to + * prevent it from being freed */ + op_data->op_file_secctx = NULL; + op_data->op_file_secctx_size = 0; + } ll_finish_md_op_data(op_data); + } + + if (lum != NULL) + OBD_FREE_PTR(lum); ptlrpc_req_finished(req); return retval; @@ -623,15 +908,20 @@ static struct dentry *ll_lookup_nd(struct inode *parent, struct dentry *dentry, dentry->d_name.len, dentry->d_name.name, PFID(ll_inode2fid(parent)), parent, flags); - /* Optimize away (CREATE && !OPEN). Let .create handle the race. */ - if ((flags & LOOKUP_CREATE) && !(flags & LOOKUP_OPEN)) + /* + * 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 ((flags & LOOKUP_CREATE) && !(flags & LOOKUP_OPEN) && + (inode_permission(parent, MAY_WRITE | MAY_EXEC) == 0)) return NULL; if (flags & (LOOKUP_PARENT|LOOKUP_OPEN|LOOKUP_CREATE)) itp = NULL; else itp = ⁢ - de = ll_lookup_it(parent, dentry, itp); + de = ll_lookup_it(parent, dentry, itp, NULL, NULL, NULL); if (itp != NULL) ll_intent_release(itp); @@ -639,24 +929,68 @@ 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; + struct ll_sb_info *sbi; + struct pcc_create_attach pca = {NULL, NULL}; + struct pcc_dataset *dataset = NULL; 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); + PFID(ll_inode2fid(dir)), dir, file, open_flags, mode, + ll_is_opened(opened, file)); + + /* Only negative dentries enter here */ + LASSERT(dentry->d_inode == NULL); + + if (!d_unhashed(dentry)) { + /* A valid negative dentry that just passed revalidation, + * there's little point to try and open it server-side, + * even though there's a minuscule chance it might succeed. + * Either way it's a valid race to just return -ENOENT here. + */ + if (!(open_flags & O_CREAT)) + return -ENOENT; + + /* Otherwise we just unhash it to be rehashed afresh via + * lookup if necessary + */ + d_drop(dentry); + } OBD_ALLOC(it, sizeof(*it)); if (!it) @@ -666,45 +1000,68 @@ 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; + + 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; /* Dentry added to dcache tree in ll_lookup_it */ - de = ll_lookup_it(dir, dentry, it); + de = ll_lookup_it(dir, dentry, it, &secctx, &secctxlen, &pca); if (IS_ERR(de)) rc = PTR_ERR(de); else if (de != NULL) dentry = de; + CFS_FAIL_TIMEOUT(OBD_FAIL_LLITE_CREATE_FILE_PAUSE, cfs_fail_val); + if (!rc) { if (it_disposition(it, DISP_OPEN_CREATE)) { /* Dentry instantiated in ll_create_it. */ - rc = ll_create_it(dir, dentry, it); + rc = ll_create_it(dir, dentry, it, secctx, secctxlen); + security_release_secctx(secctx, secctxlen); if (rc) { /* We dget in ll_splice_alias. */ if (de != NULL) dput(de); goto out_release; } - - *opened |= FILE_CREATED; + if (dataset != NULL && dentry->d_inode) { + rc = pcc_inode_create_fini(dataset, + dentry->d_inode, + pca.pca_dentry); + if (rc) { + if (de != NULL) + dput(de); + GOTO(out_release, rc); + } + } + ll_set_created(opened, file); } + if (dentry->d_inode && it_disposition(it, DISP_OPEN_OPEN)) { /* Open dentry. */ if (S_ISFIFO(dentry->d_inode->i_mode)) { - /* We cannot call open here as it would - * deadlock. - */ - if (it_disposition(it, DISP_ENQ_OPEN_REF)) - ptlrpc_req_finished( - (struct ptlrpc_request *) - it->d.lustre.it_data); + /* We cannot call open here as it might + * deadlock. This case is unreachable in + * practice because of OBD_CONNECT_NODEVOH. */ 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. */ @@ -717,6 +1074,8 @@ static int ll_atomic_open(struct inode *dir, struct dentry *dentry, } out_release: + if (dataset != NULL) + pcc_dataset_put(dataset); ll_intent_release(it); OBD_FREE(it, sizeof(*it)); @@ -725,7 +1084,7 @@ out_release: #else /* !HAVE_IOP_ATOMIC_OPEN */ static struct lookup_intent * -ll_convert_intent(struct open_intent *oit, int lookup_flags) +ll_convert_intent(struct open_intent *oit, int lookup_flags, bool is_readonly) { struct lookup_intent *it; @@ -735,10 +1094,12 @@ ll_convert_intent(struct open_intent *oit, int lookup_flags) if (lookup_flags & LOOKUP_OPEN) { it->it_op = IT_OPEN; - if (lookup_flags & LOOKUP_CREATE) + /* 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); + 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; @@ -750,38 +1111,47 @@ ll_convert_intent(struct open_intent *oit, int lookup_flags) static struct dentry *ll_lookup_nd(struct inode *parent, struct dentry *dentry, struct nameidata *nd) { - struct dentry *de; - ENTRY; + struct dentry *de; + ENTRY; - if (nd && !(nd->flags & (LOOKUP_CONTINUE|LOOKUP_PARENT))) { - struct lookup_intent *it; + 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 { + 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)) - RETURN(NULL); - - it = ll_convert_intent(&nd->intent.open, nd->flags); - if (IS_ERR(it)) - RETURN((struct dentry *)it); - } - - de = ll_lookup_it(parent, dentry, it); - 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 would - // deadlock. - ptlrpc_req_finished( - (struct ptlrpc_request *) - it->d.lustre.it_data); - } else { + !(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 & SB_RDONLY)); + if (IS_ERR(it)) + RETURN((struct dentry *)it); + } + + de = ll_lookup_it(parent, dentry, it, NULL, 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; @@ -793,24 +1163,24 @@ static struct dentry *ll_lookup_nd(struct inode *parent, struct dentry *dentry, 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); + } + } 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, NULL); } RETURN(de); @@ -826,16 +1196,17 @@ static struct inode *ll_create_node(struct inode *dir, struct lookup_intent *it) int rc; ENTRY; - LASSERT(it && it->d.lustre.it_disposition); + LASSERT(it && it->it_disposition); - LASSERT(it_disposition(it, DISP_ENQ_CREATE_REF)); - request = it->d.lustre.it_data; + 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)); - LASSERT(ll_d_hlist_empty(&inode->i_dentry)); + /* Pause to allow for a race with concurrent access by fid */ + OBD_FAIL_TIMEOUT(OBD_FAIL_LLITE_CREATE_NODE_PAUSE, cfs_fail_val); /* We asked for a lock on the directory, but were granted a * lock on the inode. Since we finally have an inode pointer, @@ -864,9 +1235,11 @@ static struct inode *ll_create_node(struct inode *dir, struct lookup_intent *it) * with d_instantiate(). */ static int ll_create_it(struct inode *dir, struct dentry *dentry, - struct lookup_intent *it) + struct lookup_intent *it, + void *secctx, __u32 secctxlen) { struct inode *inode; + __u64 bits = 0; int rc = 0; ENTRY; @@ -882,7 +1255,30 @@ static int ll_create_it(struct inode *dir, struct dentry *dentry, if (IS_ERR(inode)) RETURN(PTR_ERR(inode)); + 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 */ + rc = security_inode_notifysecctx(inode, secctx, secctxlen); + inode_unlock(inode); + if (rc) + RETURN(rc); + } + d_instantiate(dentry, inode); + + 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); } @@ -893,16 +1289,17 @@ 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 "LPU64 - "\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, @@ -921,66 +1318,127 @@ static int ll_new_node(struct inode *dir, struct dentry *dchild, tgt_len = strlen(tgt) + 1; again: - op_data = ll_prep_md_op_data(NULL, dir, NULL, name->name, - name->len, 0, opc, NULL); - if (IS_ERR(op_data)) - GOTO(err_exit, err = PTR_ERR(op_data)); + op_data = ll_prep_md_op_data(NULL, dir, NULL, name->name, + name->len, 0, opc, NULL); + if (IS_ERR(op_data)) + GOTO(err_exit, err = PTR_ERR(op_data)); + + if (sbi->ll_flags & LL_SBI_FILE_SECCTX) { + err = ll_dentry_init_security(dchild, mode, &dchild->d_name, + &op_data->op_file_secctx_name, + &op_data->op_file_secctx, + &op_data->op_file_secctx_size); + if (err < 0) + 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); - ll_finish_md_op_data(op_data); - if (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 - * for the directory, either from the local xattr cache or the - * MDS, then create the directory on the right MDT. */ - if (err == -EREMOTE) { - struct lmv_user_md *lum; - int rc; - - ptlrpc_req_finished(request); - request = NULL; - - OBD_ALLOC_PTR(lum); - if (lum == NULL) - GOTO(err_exit, err = -ENOMEM); +#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; + int lumsize; + int err2; - rc = ll_getxattr_common(dir, XATTR_NAME_DEFAULT_LMV, - lum, sizeof(*lum), - OBD_MD_FLXATTR); - if (rc < 0) { - OBD_FREE_PTR(lum); - if (rc == -ENODATA) - GOTO(err_exit, err); - else - GOTO(err_exit, rc); - } + ptlrpc_req_finished(request); + request = NULL; + + 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) { + 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); - ll_i2info(dir)->lli_def_stripe_offset = - le32_to_cpu(lum->lum_stripe_offset); - OBD_FREE_PTR(lum); - goto again; + 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; + + 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. + */ + 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); } - GOTO(err_exit, err); + + ptlrpc_req_finished(request); + request = NULL; + goto again; } +#endif - ll_update_times(request, dir); + 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); 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 */ + 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); - EXIT; + if (!(sbi->ll_flags & LL_SBI_FILE_SECCTX)) { + err = ll_inode_init_security(dchild, inode, dir); + if (err) + GOTO(err_exit, err); + } + + EXIT; err_exit: if (request != NULL) ptlrpc_req_finished(request); - return err; + if (!IS_ERR_OR_NULL(op_data)) + ll_finish_md_op_data(op_data); + + return err; } static int ll_mknod(struct inode *dir, struct dentry *dchild, ll_umode_t mode, @@ -1030,11 +1488,15 @@ static int ll_create_nd(struct inode *dir, struct dentry *dentry, { 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); + /* 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); @@ -1050,24 +1512,33 @@ static int ll_create_nd(struct inode *dir, struct dentry *dentry, { struct ll_dentry_data *lld = ll_d2d(dentry); struct lookup_intent *it = NULL; - int rc; + int rc; + + CFS_FAIL_TIMEOUT(OBD_FAIL_LLITE_CREATE_FILE_PAUSE, cfs_fail_val); if (lld != NULL) it = lld->lld_it; - if (!it) - return ll_mknod(dir, dentry, mode, 0); + 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->d.lustre.it_status) { - rc = it->d.lustre.it_status; - goto out; - } + /* Was there an error? Propagate it! */ + if (it->it_status) { + rc = it->it_status; + goto out; + } - rc = ll_create_it(dir, dentry, it); - if (nd && (nd->flags & LOOKUP_OPEN) && dentry->d_inode) { /* Open */ + 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; @@ -1224,24 +1695,21 @@ int ll_rmdir_entry(struct inode *dir, char *name, int namelen) RETURN(rc); } -/* ll_unlink() doesn't update the inode with the new link count. - * Instead, ll_ddelete() and ll_d_iput() will update it based upon if - * there is any lock existing. They will recycle dentries and inodes - * based upon locks too. b=20433 */ static int ll_unlink(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; + struct mdt_body *body; + int rc; + ENTRY; CDEBUG(D_VFSTRACE, "VFS Op:name=%.*s, dir="DFID"(%p)\n", name->len, name->name, PFID(ll_inode2fid(dir)), dir); - /* - * XXX: unlink bind mountpoint maybe call to here, - * just check it as vfs_unlink does. - */ + /* + * XXX: unlink bind mountpoint maybe call to here, + * just check it as vfs_unlink does. + */ if (unlikely(d_mountpoint(dchild))) RETURN(-EBUSY); @@ -1250,8 +1718,7 @@ static int ll_unlink(struct inode *dir, struct dentry *dchild) 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_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); @@ -1259,24 +1726,42 @@ static int ll_unlink(struct inode *dir, struct dentry *dchild) if (rc) GOTO(out, rc); - ll_update_times(request, dir); - ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_UNLINK, 1); + /* + * 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); - out: - ptlrpc_req_finished(request); - RETURN(rc); + ll_update_times(request, dir); + ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_UNLINK, 1); + +out: + ptlrpc_req_finished(request); + RETURN(rc); } static int ll_rename(struct inode *src, struct dentry *src_dchild, - struct inode *tgt, struct dentry *tgt_dchild) + struct inode *tgt, struct dentry *tgt_dchild +#ifdef HAVE_IOPS_RENAME_WITH_FLAGS + , unsigned int flags +#endif + ) { struct qstr *src_name = &src_dchild->d_name; struct qstr *tgt_name = &tgt_dchild->d_name; - struct ptlrpc_request *request = NULL; - struct ll_sb_info *sbi = ll_i2sbi(src); - struct md_op_data *op_data; - int err; - ENTRY; + struct ptlrpc_request *request = NULL; + struct ll_sb_info *sbi = ll_i2sbi(src); + struct md_op_data *op_data; + int err; + ENTRY; + +#ifdef HAVE_IOPS_RENAME_WITH_FLAGS + if (flags) + 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, @@ -1316,28 +1801,33 @@ static int ll_rename(struct inode *src, struct dentry *src_dchild, } const struct inode_operations ll_dir_inode_operations = { - .mknod = ll_mknod, + .mknod = ll_mknod, #ifdef HAVE_IOP_ATOMIC_OPEN - .atomic_open = ll_atomic_open, + .atomic_open = ll_atomic_open, #endif - .lookup = ll_lookup_nd, - .create = ll_create_nd, + .lookup = ll_lookup_nd, + .create = ll_create_nd, /* We need all these non-raw things for NFSD, to not patch it. */ - .unlink = ll_unlink, - .mkdir = ll_mkdir, - .rmdir = ll_rmdir, - .symlink = ll_symlink, - .link = ll_link, - .rename = ll_rename, - .setattr = ll_setattr, - .getattr = ll_getattr, - .permission = ll_inode_permission, - .setxattr = ll_setxattr, - .getxattr = ll_getxattr, - .listxattr = ll_listxattr, - .removexattr = ll_removexattr, + .unlink = ll_unlink, + .mkdir = ll_mkdir, + .rmdir = ll_rmdir, + .symlink = ll_symlink, + .link = ll_link, + .rename = ll_rename, + .setattr = ll_setattr, + .getattr = ll_getattr, + .permission = ll_inode_permission, +#ifdef HAVE_IOP_XATTR + .setxattr = ll_setxattr, + .getxattr = ll_getxattr, + .removexattr = ll_removexattr, +#endif + .listxattr = ll_listxattr, #ifdef HAVE_IOP_GET_ACL - .get_acl = ll_get_acl, + .get_acl = ll_get_acl, +#endif +#ifdef HAVE_IOP_SET_ACL + .set_acl = ll_set_acl, #endif }; @@ -1345,11 +1835,16 @@ const struct inode_operations ll_special_inode_operations = { .setattr = ll_setattr, .getattr = ll_getattr, .permission = ll_inode_permission, - .setxattr = ll_setxattr, - .getxattr = ll_getxattr, - .listxattr = ll_listxattr, +#ifdef HAVE_IOP_XATTR + .setxattr = ll_setxattr, + .getxattr = ll_getxattr, .removexattr = ll_removexattr, +#endif + .listxattr = ll_listxattr, #ifdef HAVE_IOP_GET_ACL - .get_acl = ll_get_acl, + .get_acl = ll_get_acl, +#endif +#ifdef HAVE_IOP_SET_ACL + .set_acl = ll_set_acl, #endif };