X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=blobdiff_plain;f=lustre%2Fllite%2Fnamei.c;h=9b7241a89cfa1886aacefcd1050cc473eadd2c10;hp=0464951e76cf523c85c34dd2d17241f53b362860;hb=a71586d4ee8d6f039a413e2a0fd791db847a3c19;hpb=2102c86e0d0ae735aed9ee8c1c6a77b63eda6037 diff --git a/lustre/llite/namei.c b/lustre/llite/namei.c index 0464951..9b7241a 100644 --- a/lustre/llite/namei.c +++ b/lustre/llite/namei.c @@ -38,9 +38,7 @@ #include #include #include -#ifdef HAVE_UIDGID_HEADER -# include -#endif +#include #define DEBUG_SUBSYSTEM S_LLITE @@ -51,7 +49,8 @@ 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) @@ -159,10 +158,9 @@ struct inode *ll_iget(struct super_block *sb, ino_t hash, static void ll_invalidate_negative_children(struct inode *dir) { struct dentry *dentry, *tmp_subdir; - DECLARE_LL_D_HLIST_NODE_PTR(p); - ll_lock_dcache(dir); - ll_d_hlist_for_each_entry(dentry, p, &dir->i_dentry) { + 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; @@ -176,7 +174,7 @@ static void ll_invalidate_negative_children(struct inode *dir) } spin_unlock(&dentry->d_lock); } - ll_unlock_dcache(dir); + spin_unlock(&dir->i_lock); } int ll_test_inode_by_fid(struct inode *inode, void *opaque) @@ -184,15 +182,12 @@ 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) { @@ -206,30 +201,23 @@ int ll_dom_lock_cancel(struct inode *inode, struct ldlm_lock *lock) 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", + /* 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))); - } else { - enum cl_fsync_mode mode; - loff_t end = clt.cl_dom_comp_size - 1; - - 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); + /* most likely result of layout change, do nothing */ + rc = 0; } + cl_env_put(env, &refcheck); RETURN(rc); } -void ll_lock_cancel_bits(struct ldlm_lock *lock, __u64 to_cancel) +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; @@ -256,8 +244,6 @@ void ll_lock_cancel_bits(struct ldlm_lock *lock, __u64 to_cancel) } 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; } @@ -301,9 +287,6 @@ void ll_lock_cancel_bits(struct ldlm_lock *lock, __u64 to_cancel) 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) { @@ -319,15 +302,12 @@ void ll_lock_cancel_bits(struct ldlm_lock *lock, __u64 to_cancel) PFID(ll_inode2fid(inode)), rc); } - if (bits & MDS_INODELOCK_UPDATE) { - struct ll_inode_info *lli = ll_i2info(inode); + lli = ll_i2info(inode); - lli->lli_update_atime = 1; - } + if (bits & MDS_INODELOCK_UPDATE) + ll_file_set_flag(lli, LLIF_UPDATE_ATIME); if ((bits & MDS_INODELOCK_UPDATE) && S_ISDIR(inode->i_mode)) { - struct ll_inode_info *lli = ll_i2info(inode); - CDEBUG(D_INODE, "invalidating inode "DFID" lli = %p, " "pfid = "DFID"\n", PFID(ll_inode2fid(inode)), lli, PFID(&lli->lli_pfid)); @@ -383,6 +363,9 @@ void ll_lock_cancel_bits(struct ldlm_lock *lock, __u64 to_cancel) inode != inode->i_sb->s_root->d_inode) ll_invalidate_aliases(inode); + if (bits & (MDS_INODELOCK_LOOKUP | MDS_INODELOCK_PERM)) + forget_all_cached_acls(inode); + iput(inode); RETURN_EXIT; } @@ -416,11 +399,13 @@ int ll_md_need_convert(struct ldlm_lock *lock) 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; @@ -432,8 +417,7 @@ int ll_md_need_convert(struct ldlm_lock *lock) /* 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)))) { + ktime_add(lock->l_last_used, ns->ns_dirty_age_limit))) { unlock_res_and_lock(lock); return 0; } @@ -445,11 +429,10 @@ int ll_md_need_convert(struct ldlm_lock *lock) return !!(bits); } -int ll_md_blocking_ast(struct ldlm_lock *lock, struct ldlm_lock_desc *desc, +int ll_md_blocking_ast(struct ldlm_lock *lock, struct ldlm_lock_desc *ld, void *data, int flag) { struct lustre_handle lockh; - __u64 bits = lock->l_policy_data.l_inodebits.bits; int rc; ENTRY; @@ -459,17 +442,21 @@ int ll_md_blocking_ast(struct ldlm_lock *lock, struct ldlm_lock_desc *desc, { __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); + /* 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) { @@ -479,24 +466,34 @@ int ll_md_blocking_ast(struct ldlm_lock *lock, struct ldlm_lock_desc *desc, 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 (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. + /* 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() */ - bits = lock->l_policy_data.l_inodebits.cancel_bits & - ~MDS_INODELOCK_DOM; - } else { - LASSERT(ldlm_is_canceling(lock)); + to_cancel &= ld->l_policy_data.l_inodebits.cancel_bits; } - ll_lock_cancel_bits(lock, bits); + ll_lock_cancel_bits(lock, to_cancel); break; + } default: LBUG(); } @@ -540,15 +537,14 @@ 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); @@ -573,7 +569,7 @@ 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; } @@ -615,7 +611,9 @@ 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, - void *secctx, __u32 secctxlen) + void *secctx, __u32 secctxlen, + void *encctx, __u32 encctxlen, + ktime_t kstart, bool encrypt) { struct inode *inode = NULL; __u64 bits = 0; @@ -636,6 +634,36 @@ static int ll_lookup_it_finish(struct ptlrpc_request *request, 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); + } + } + } + if (it->it_op & IT_OPEN) ll_dom_finish_open(inode, request, it); @@ -668,14 +696,17 @@ static int ll_lookup_it_finish(struct ptlrpc_request *request, } if (secctx != NULL && secctxlen != 0) { - inode_lock(inode); + /* 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) - CWARN("cannot set security context for " - DFID": rc = %d\n", - PFID(ll_inode2fid(inode)), rc); + CWARN("%s: cannot set security context for "DFID": rc = %d\n", + ll_i2sbi(inode)->ll_fsname, + PFID(ll_inode2fid(inode)), + rc); } } @@ -692,6 +723,14 @@ 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 was created on the server, the dentry is revalidated @@ -704,7 +743,7 @@ 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, @@ -720,25 +759,30 @@ 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) + if (rc != 0 && it->it_op & IT_OPEN) { + ll_intent_drop_lock(it); ll_open_cleanup((*de)->d_sb, request); + } 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, void **secctx, __u32 *secctxlen, - struct pcc_create_attach *pca) + 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; @@ -750,27 +794,26 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry, 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) @@ -806,6 +849,20 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry, if (secctxlen != NULL) *secctxlen = 0; } + if (it->it_op & IT_CREAT && encrypt) { + 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)) { @@ -835,7 +892,7 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry, 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_id; + op_data->op_archive_id = dataset->pccd_rwid; rc = obd_fid_alloc(NULL, ll_i2mdexp(parent), &op_data->op_fid2, op_data); @@ -885,18 +942,21 @@ static struct dentry *ll_lookup_it(struct inode *parent, struct dentry *dentry, 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)); - } - - 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); + 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); @@ -908,6 +968,14 @@ 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); } @@ -918,16 +986,14 @@ out: 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. @@ -942,7 +1008,8 @@ 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, NULL); + de = ll_lookup_it(parent, dentry, itp, NULL, NULL, NULL, false, + NULL, NULL); if (itp != NULL) ll_intent_release(itp); @@ -950,29 +1017,51 @@ 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; - struct ll_sb_info *sbi; - struct pcc_create_attach pca = {NULL, NULL}; - struct pcc_dataset *dataset = NULL; + 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); @@ -1004,9 +1093,15 @@ static int ll_atomic_open(struct inode *dir, struct dentry *dentry, /* 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)) { - dataset = pcc_dataset_get(&sbi->ll_pcc_super, - ll_i2info(dir)->lli_projid, - 0); + 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; } } @@ -1014,8 +1109,22 @@ static int ll_atomic_open(struct inode *dir, struct dentry *dentry, 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)) { + /* we know that we are going to create 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 (!llcrypt_has_encryption_key(dir)) + GOTO(out_release, rc = -ENOKEY); + encrypt = true; + rc = 0; + } + /* Dentry added to dcache tree in ll_lookup_it */ - de = ll_lookup_it(dir, dentry, it, &secctx, &secctxlen, &pca); + 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) @@ -1026,26 +1135,49 @@ 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); + rc = ll_create_it(dir, dentry, it, secctx, secctxlen, + encrypt, encctx, encctxlen); security_release_secctx(secctx, secctxlen); + llcrypt_free_ctx(encctx, encctxlen); if (rc) { /* We dget in ll_splice_alias. */ if (de != NULL) dput(de); goto out_release; } - 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); - } + + rc = pcc_inode_create_fini(dentry->d_inode, &pca); + if (rc) { + if (de != NULL) + dput(de); + GOTO(out_release, rc); } - *opened |= FILE_CREATED; + 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)) { @@ -1057,7 +1189,7 @@ 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. */ @@ -1067,122 +1199,17 @@ 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: - if (dataset != NULL) - pcc_dataset_put(dataset); ll_intent_release(it); OBD_FREE(it, sizeof(*it)); 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, 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, 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) { @@ -1232,16 +1259,16 @@ 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) @@ -1253,18 +1280,26 @@ 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) @@ -1302,16 +1337,17 @@ 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, @@ -1328,17 +1364,33 @@ 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; @@ -1350,24 +1402,54 @@ 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; + + 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); @@ -1378,20 +1460,29 @@ again: 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) @@ -1406,144 +1497,95 @@ 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, @@ -1554,39 +1596,46 @@ 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(); @@ -1595,7 +1644,8 @@ 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); } @@ -1603,35 +1653,38 @@ 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)); + 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) + ll_update_times(request, dir); - ptlrpc_req_finished(request); - RETURN(rc); + ptlrpc_req_finished(request); + if (!rc) + ll_stats_ops_tally(ll_i2sbi(dir), LPROC_LL_RMDIR, + ktime_us_delta(ktime_get(), kstart)); + RETURN(rc); } /** @@ -1641,6 +1694,7 @@ 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; @@ -1654,12 +1708,13 @@ 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); } @@ -1669,10 +1724,13 @@ 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, @@ -1687,7 +1745,11 @@ static int ll_unlink(struct inode *dir, struct dentry *dchild) 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); @@ -1703,10 +1765,12 @@ 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); } @@ -1722,6 +1786,7 @@ 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(); int err; ENTRY; @@ -1730,15 +1795,22 @@ 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); +#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); + op_data = ll_prep_md_op_data(NULL, src, tgt, NULL, 0, 0, LUSTRE_OPC_ANY, NULL); if (IS_ERR(op_data)) @@ -1750,29 +1822,29 @@ static int ll_rename(struct inode *src, struct dentry *src_dchild, if (tgt_dchild->d_inode != NULL) 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. */ @@ -1791,9 +1863,7 @@ 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 @@ -1809,9 +1879,7 @@ 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