-/* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
- * vim:expandtab:shiftwidth=8:tabstop=8:
+/*
+ * GPL HEADER START
*
- * NFS export of Lustre Light File System
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
*
- * Copyright (c) 2002, 2003 Cluster File Systems, Inc.
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
*
- * This file is part of Lustre, http://www.lustre.org.
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
*
- * Lustre is free software; you can redistribute it and/or
- * modify it under the terms of version 2 of the GNU General Public
- * License as published by the Free Software Foundation.
+ * 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
*
- * Lustre is distributed in the hope that it will be useful,
- * but WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
- * GNU General Public License for more details.
+ * 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.
*
- * You should have received a copy of the GNU General Public License
- * along with Lustre; if not, write to the Free Software
- * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ *
+ * Copyright (c) 2011, Whamcloud, Inc.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ *
+ * lustre/lustre/llite/llite_nfs.c
+ *
+ * NFS export of Lustre Light File System
+ *
+ * Author: Yury Umanets <umka@clusterfs.com>
+ * Author: Huang Hua <huanghua@clusterfs.com>
*/
#define DEBUG_SUBSYSTEM S_LLITE
-#include <linux/lustre_lite.h>
+#include <lustre_lite.h>
#include "llite_internal.h"
+#ifdef HAVE_LINUX_EXPORTFS_H
+#include <linux/exportfs.h>
+#endif
__u32 get_uuid2int(const char *name, int len)
{
return (key0 << 1);
}
-static struct inode * search_inode_for_lustre(struct super_block *sb,
- unsigned long ino,
- unsigned long generation,
- int mode)
+static int ll_nfs_test_inode(struct inode *inode, void *opaque)
{
+ return lu_fid_eq(&ll_i2info(inode)->lli_fid,
+ (struct lu_fid *)opaque);
+}
+
+static struct inode *search_inode_for_lustre(struct super_block *sb,
+ const struct lu_fid *fid)
+{
+ struct ll_sb_info *sbi = ll_s2sbi(sb);
struct ptlrpc_request *req = NULL;
- struct ll_sb_info *sbi = ll_s2sbi(sb);
- struct lustre_id id;
- unsigned long valid = 0;
- int eadatalen = 0, rc;
- struct inode *inode = NULL;
+ struct inode *inode = NULL;
+ int eadatalen = 0;
+ unsigned long hash = (unsigned long) cl_fid_build_ino(fid, 0);
+ struct md_op_data *op_data;
+ int rc;
+ ENTRY;
- inode = ILOOKUP(sb, ino, NULL, NULL);
+ CDEBUG(D_INFO, "searching inode for:(%lu,"DFID")\n", hash, PFID(fid));
+ inode = ilookup5(sb, hash, ll_nfs_test_inode, (void *)fid);
if (inode)
- return inode;
- if (S_ISREG(mode)) {
- eadatalen = obd_size_diskmd(sbi->ll_lov_exp, NULL);
- valid |= OBD_MD_FLEASIZE;
- }
- id.li_stc.u.e3s.l3s_type = mode;
- id.li_stc.u.e3s.l3s_ino = (__u64)ino;
- id.li_stc.u.e3s.l3s_gen = generation;
+ RETURN(inode);
- rc = md_getattr(sbi->ll_lmv_exp, &id, valid, eadatalen, &req);
- if (rc) {
- CERROR("failure %d inode %lu\n", rc, ino);
- return ERR_PTR(rc);
- }
+ rc = ll_get_max_mdsize(sbi, &eadatalen);
+ if (rc)
+ RETURN(ERR_PTR(rc));
+
+ /* Because inode is NULL, ll_prep_md_op_data can not
+ * be used here. So we allocate op_data ourselves */
+ OBD_ALLOC_PTR(op_data);
+ if (op_data == NULL)
+ return ERR_PTR(-ENOMEM);
+
+ op_data->op_fid1 = *fid;
+ op_data->op_mode = eadatalen;
+ op_data->op_valid = OBD_MD_FLEASIZE;
- rc = ll_prep_inode(sbi->ll_lov_exp, sbi->ll_lmv_exp,
- &inode, req, 0, sb);
+ /* mds_fid2dentry ignores f_type */
+ rc = md_getattr(sbi->ll_md_exp, op_data, &req);
+ OBD_FREE_PTR(op_data);
if (rc) {
- ptlrpc_req_finished(req);
- return ERR_PTR(rc);
+ CERROR("can't get object attrs, fid "DFID", rc %d\n",
+ PFID(fid), rc);
+ RETURN(ERR_PTR(rc));
}
+ rc = ll_prep_inode(&inode, req, sb);
ptlrpc_req_finished(req);
+ if (rc)
+ RETURN(ERR_PTR(rc));
- return inode;
+ RETURN(inode);
}
-extern struct dentry_operations ll_d_ops;
+struct lustre_nfs_fid {
+ struct lu_fid lnf_child;
+ struct lu_fid lnf_parent;
+};
-static struct dentry *ll_iget_for_nfs(struct super_block *sb, unsigned long ino,
- __u32 generation, umode_t mode)
+static struct dentry *
+ll_iget_for_nfs(struct super_block *sb, struct lu_fid *fid, struct lu_fid *parent)
{
- struct inode *inode;
+ struct inode *inode;
struct dentry *result;
- struct list_head *lp;
+ ENTRY;
- if (ino == 0)
- return ERR_PTR(-ESTALE);
+ CDEBUG(D_INFO, "Get dentry for fid: "DFID"\n", PFID(fid));
+ if (!fid_is_sane(fid))
+ RETURN(ERR_PTR(-ESTALE));
- inode = search_inode_for_lustre(sb, ino, generation, mode);
- if (IS_ERR(inode)) {
- return ERR_PTR(PTR_ERR(inode));
- }
- if (is_bad_inode(inode)
- || (generation && inode->i_generation != generation)
- ){
+ inode = search_inode_for_lustre(sb, fid);
+ if (IS_ERR(inode))
+ RETURN(ERR_PTR(PTR_ERR(inode)));
+
+ if (is_bad_inode(inode)) {
/* we didn't find the right inode.. */
- CERROR(" Inode %lu, Bad count: %lu %d or version %u %u\n",
- inode->i_ino,
- (unsigned long)inode->i_nlink,
- atomic_read(&inode->i_count),
- inode->i_generation,
- generation);
iput(inode);
- return ERR_PTR(-ESTALE);
+ RETURN(ERR_PTR(-ESTALE));
}
-
- /* now to find a dentry.
- * If possible, get a well-connected one
+
+ /**
+ * It is an anonymous dentry without OST objects created yet.
+ * We have to find the parent to tell MDS how to init lov objects.
*/
- spin_lock(&dcache_lock);
- for (lp = inode->i_dentry.next; lp != &inode->i_dentry ; lp=lp->next) {
- result = list_entry(lp,struct dentry, d_alias);
- if (!(result->d_flags & DCACHE_DISCONNECTED)) {
- dget_locked(result);
- ll_set_dflags(result, DCACHE_REFERENCED);
- spin_unlock(&dcache_lock);
- iput(inode);
- return result;
- }
- }
- spin_unlock(&dcache_lock);
- result = d_alloc_root(inode);
- if (result == NULL) {
- iput(inode);
- return ERR_PTR(-ENOMEM);
- }
- result->d_flags |= DCACHE_DISCONNECTED;
-
- ll_set_dd(result);
- result->d_op = &ll_d_ops;
- return result;
+ if (S_ISREG(inode->i_mode) && !ll_i2info(inode)->lli_has_smd &&
+ parent != NULL) {
+ struct ll_inode_info *lli = ll_i2info(inode);
+
+ spin_lock(&lli->lli_lock);
+ lli->lli_pfid = *parent;
+ spin_unlock(&lli->lli_lock);
+ }
+
+ result = d_obtain_alias(inode);
+ if (IS_ERR(result))
+ RETURN(result);
+
+ ll_dops_init(result, 1, 0);
+
+ RETURN(result);
}
-struct dentry *ll_fh_to_dentry(struct super_block *sb, __u32 *data, int len,
- int fhtype, int parent)
+#define LUSTRE_NFS_FID 0x97
+
+/**
+ * \a connectable - is nfsd will connect himself or this should be done
+ * at lustre
+ *
+ * The return value is file handle type:
+ * 1 -- contains child file handle;
+ * 2 -- contains child file handle and parent file handle;
+ * 255 -- error.
+ */
+static int ll_encode_fh(struct dentry *de, __u32 *fh, int *plen,
+ int connectable)
{
- switch (fhtype) {
- case 2:
- if (len < 5)
- break;
- if (parent)
- return ll_iget_for_nfs(sb, data[3], 0, data[4]);
- case 1:
- if (len < 3)
- break;
- if (parent)
- break;
- return ll_iget_for_nfs(sb, data[0], data[1], data[2]);
- default: break;
- }
- return ERR_PTR(-EINVAL);
+ struct inode *inode = de->d_inode;
+ struct inode *parent = de->d_parent->d_inode;
+ struct lustre_nfs_fid *nfs_fid = (void *)fh;
+ ENTRY;
+
+ CDEBUG(D_INFO, "encoding for (%lu,"DFID") maxlen=%d minlen=%d\n",
+ inode->i_ino, PFID(ll_inode2fid(inode)), *plen,
+ (int)sizeof(struct lustre_nfs_fid));
+
+ if (*plen < sizeof(struct lustre_nfs_fid)/4)
+ RETURN(255);
+
+ nfs_fid->lnf_child = *ll_inode2fid(inode);
+ nfs_fid->lnf_parent = *ll_inode2fid(parent);
+ *plen = sizeof(struct lustre_nfs_fid)/4;
+
+ RETURN(LUSTRE_NFS_FID);
}
-int ll_dentry_to_fh(struct dentry *dentry, __u32 *datap, int *lenp,
- int need_parent)
+static int ll_nfs_get_name_filldir(void *cookie, const char *name, int namelen,
+ loff_t hash, u64 ino, unsigned type)
{
- if (*lenp < 3)
- return 255;
- *datap++ = dentry->d_inode->i_ino;
- *datap++ = dentry->d_inode->i_generation;
- *datap++ = (__u32)(S_IFMT & dentry->d_inode->i_mode);
-
- if (*lenp == 3 || S_ISDIR(dentry->d_inode->i_mode)) {
- *lenp = 3;
- return 1;
+ /* It is hack to access lde_fid for comparison with lgd_fid.
+ * So the input 'name' must be part of the 'lu_dirent'. */
+ struct lu_dirent *lde = container_of0(name, struct lu_dirent, lde_name);
+ struct ll_getname_data *lgd = cookie;
+ struct lu_fid fid;
+
+ fid_le_to_cpu(&fid, &lde->lde_fid);
+ if (lu_fid_eq(&fid, &lgd->lgd_fid)) {
+ memcpy(lgd->lgd_name, name, namelen);
+ lgd->lgd_name[namelen] = 0;
+ lgd->lgd_found = 1;
}
- if (dentry->d_parent) {
- *datap++ = dentry->d_parent->d_inode->i_ino;
- *datap++ = (__u32)(S_IFMT & dentry->d_parent->d_inode->i_mode);
-
- *lenp = 5;
- return 2;
+ return lgd->lgd_found;
+}
+
+static int ll_get_name(struct dentry *dentry, char *name,
+ struct dentry *child)
+{
+ struct inode *dir = dentry->d_inode;
+ struct ll_getname_data lgd;
+ __u64 offset = 0;
+ int rc;
+ ENTRY;
+
+ if (!dir || !S_ISDIR(dir->i_mode))
+ GOTO(out, rc = -ENOTDIR);
+
+ if (!dir->i_fop)
+ GOTO(out, rc = -EINVAL);
+
+ lgd.lgd_name = name;
+ lgd.lgd_fid = ll_i2info(child->d_inode)->lli_fid;
+ lgd.lgd_found = 0;
+
+ mutex_lock(&dir->i_mutex);
+ rc = ll_dir_read(dir, &offset, &lgd, ll_nfs_get_name_filldir);
+ mutex_unlock(&dir->i_mutex);
+ if (!rc && !lgd.lgd_found)
+ rc = -ENOENT;
+ EXIT;
+
+out:
+ return rc;
+}
+
+#ifdef HAVE_FH_TO_DENTRY
+static struct dentry *ll_fh_to_dentry(struct super_block *sb, struct fid *fid,
+ int fh_len, int fh_type)
+{
+ struct lustre_nfs_fid *nfs_fid = (struct lustre_nfs_fid *)fid;
+
+ if (fh_type != LUSTRE_NFS_FID)
+ RETURN(ERR_PTR(-EPROTO));
+
+ RETURN(ll_iget_for_nfs(sb, &nfs_fid->lnf_child, &nfs_fid->lnf_parent));
+}
+
+static struct dentry *ll_fh_to_parent(struct super_block *sb, struct fid *fid,
+ int fh_len, int fh_type)
+{
+ struct lustre_nfs_fid *nfs_fid = (struct lustre_nfs_fid *)fid;
+
+ if (fh_type != LUSTRE_NFS_FID)
+ RETURN(ERR_PTR(-EPROTO));
+
+ RETURN(ll_iget_for_nfs(sb, &nfs_fid->lnf_parent, NULL));
+}
+
+#else
+
+/*
+ * This length is counted as amount of __u32,
+ * It is composed of a fid and a mode
+ */
+static struct dentry *ll_decode_fh(struct super_block *sb, __u32 *fh, int fh_len,
+ int fh_type,
+ int (*acceptable)(void *, struct dentry *),
+ void *context)
+{
+ struct lustre_nfs_fid *nfs_fid = (void *)fh;
+ struct dentry *entry;
+ ENTRY;
+
+ CDEBUG(D_INFO, "decoding for "DFID" fh_len=%d fh_type=%x\n",
+ PFID(&nfs_fid->lnf_child), fh_len, fh_type);
+
+ if (fh_type != LUSTRE_NFS_FID)
+ RETURN(ERR_PTR(-EPROTO));
+
+ entry = sb->s_export_op->find_exported_dentry(sb, &nfs_fid->lnf_child,
+ &nfs_fid->lnf_parent,
+ acceptable, context);
+ RETURN(entry);
+}
+
+static struct dentry *ll_get_dentry(struct super_block *sb, void *data)
+{
+ struct dentry *entry;
+ ENTRY;
+
+ entry = ll_iget_for_nfs(sb, data, NULL);
+ RETURN(entry);
+}
+#endif
+static struct dentry *ll_get_parent(struct dentry *dchild)
+{
+ struct ptlrpc_request *req = NULL;
+ struct inode *dir = dchild->d_inode;
+ struct ll_sb_info *sbi;
+ struct dentry *result = NULL;
+ struct mdt_body *body;
+ static char dotdot[] = "..";
+ struct md_op_data *op_data;
+ int rc;
+ ENTRY;
+
+ LASSERT(dir && S_ISDIR(dir->i_mode));
+
+ sbi = ll_s2sbi(dir->i_sb);
+
+ CDEBUG(D_INFO, "getting parent for (%lu,"DFID")\n",
+ dir->i_ino, PFID(ll_inode2fid(dir)));
+
+ op_data = ll_prep_md_op_data(NULL, dir, NULL, dotdot,
+ strlen(dotdot), 0,
+ LUSTRE_OPC_ANY, NULL);
+ if (IS_ERR(op_data))
+ RETURN((void *)op_data);
+
+ rc = md_getattr_name(sbi->ll_md_exp, op_data, &req);
+ ll_finish_md_op_data(op_data);
+ if (rc) {
+ CERROR("failure %d inode %lu get parent\n", rc, dir->i_ino);
+ RETURN(ERR_PTR(rc));
}
- *lenp = 3;
- return 1;
+ body = req_capsule_server_get(&req->rq_pill, &RMF_MDT_BODY);
+ LASSERT(body->valid & OBD_MD_FLID);
+
+ CDEBUG(D_INFO, "parent for "DFID" is "DFID"\n",
+ PFID(ll_inode2fid(dir)), PFID(&body->fid1));
+
+ result = ll_iget_for_nfs(dir->i_sb, &body->fid1, NULL);
+
+ ptlrpc_req_finished(req);
+ RETURN(result);
}
+
+struct export_operations lustre_export_operations = {
+ .get_parent = ll_get_parent,
+ .encode_fh = ll_encode_fh,
+ .get_name = ll_get_name,
+#ifdef HAVE_FH_TO_DENTRY
+ .fh_to_dentry = ll_fh_to_dentry,
+ .fh_to_parent = ll_fh_to_parent,
+#else
+ .get_dentry = ll_get_dentry,
+ .decode_fh = ll_decode_fh,
+#endif
+};