summaryrefslogtreecommitdiffstats
path: root/fs/ocfs2/namei.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/ocfs2/namei.c')
-rw-r--r--fs/ocfs2/namei.c2264
1 files changed, 2264 insertions, 0 deletions
diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c
new file mode 100644
index 000000000000..f6b77ff1d2bf
--- /dev/null
+++ b/fs/ocfs2/namei.c
@@ -0,0 +1,2264 @@
+/* -*- mode: c; c-basic-offset: 8; -*-
+ * vim: noexpandtab sw=8 ts=8 sts=0:
+ *
+ * namei.c
+ *
+ * Create and rename file, directory, symlinks
+ *
+ * Copyright (C) 2002, 2004 Oracle. All rights reserved.
+ *
+ * Portions of this code from linux/fs/ext3/dir.c
+ *
+ * Copyright (C) 1992, 1993, 1994, 1995
+ * Remy Card (card@masi.ibp.fr)
+ * Laboratoire MASI - Institut Blaise pascal
+ * Universite Pierre et Marie Curie (Paris VI)
+ *
+ * from
+ *
+ * linux/fs/minix/dir.c
+ *
+ * Copyright (C) 1991, 1992 Linux Torvalds
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * 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 for more details.
+ *
+ * You should have received a copy of the GNU General Public
+ * License along with this program; if not, write to the
+ * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+ * Boston, MA 021110-1307, USA.
+ */
+
+#include <linux/fs.h>
+#include <linux/types.h>
+#include <linux/slab.h>
+#include <linux/highmem.h>
+
+#define MLOG_MASK_PREFIX ML_NAMEI
+#include <cluster/masklog.h>
+
+#include "ocfs2.h"
+
+#include "alloc.h"
+#include "dcache.h"
+#include "dir.h"
+#include "dlmglue.h"
+#include "extent_map.h"
+#include "file.h"
+#include "inode.h"
+#include "journal.h"
+#include "namei.h"
+#include "suballoc.h"
+#include "symlink.h"
+#include "sysfile.h"
+#include "uptodate.h"
+#include "vote.h"
+
+#include "buffer_head_io.h"
+
+#define NAMEI_RA_CHUNKS 2
+#define NAMEI_RA_BLOCKS 4
+#define NAMEI_RA_SIZE (NAMEI_RA_CHUNKS * NAMEI_RA_BLOCKS)
+#define NAMEI_RA_INDEX(c,b) (((c) * NAMEI_RA_BLOCKS) + (b))
+
+static int inline ocfs2_search_dirblock(struct buffer_head *bh,
+ struct inode *dir,
+ const char *name, int namelen,
+ unsigned long offset,
+ struct ocfs2_dir_entry **res_dir);
+
+static int ocfs2_delete_entry(struct ocfs2_journal_handle *handle,
+ struct inode *dir,
+ struct ocfs2_dir_entry *de_del,
+ struct buffer_head *bh);
+
+static int __ocfs2_add_entry(struct ocfs2_journal_handle *handle,
+ struct inode *dir,
+ const char *name, int namelen,
+ struct inode *inode, u64 blkno,
+ struct buffer_head *parent_fe_bh,
+ struct buffer_head *insert_bh);
+
+static int ocfs2_mknod_locked(struct ocfs2_super *osb,
+ struct inode *dir,
+ struct dentry *dentry, int mode,
+ dev_t dev,
+ struct buffer_head **new_fe_bh,
+ struct buffer_head *parent_fe_bh,
+ struct ocfs2_journal_handle *handle,
+ struct inode **ret_inode,
+ struct ocfs2_alloc_context *inode_ac);
+
+static int ocfs2_fill_new_dir(struct ocfs2_super *osb,
+ struct ocfs2_journal_handle *handle,
+ struct inode *parent,
+ struct inode *inode,
+ struct buffer_head *fe_bh,
+ struct ocfs2_alloc_context *data_ac);
+
+static int ocfs2_double_lock(struct ocfs2_super *osb,
+ struct ocfs2_journal_handle *handle,
+ struct buffer_head **bh1,
+ struct inode *inode1,
+ struct buffer_head **bh2,
+ struct inode *inode2);
+
+static int ocfs2_prepare_orphan_dir(struct ocfs2_super *osb,
+ struct ocfs2_journal_handle *handle,
+ struct inode *inode,
+ char *name,
+ struct buffer_head **de_bh);
+
+static int ocfs2_orphan_add(struct ocfs2_super *osb,
+ struct ocfs2_journal_handle *handle,
+ struct inode *inode,
+ struct ocfs2_dinode *fe,
+ char *name,
+ struct buffer_head *de_bh);
+
+static int ocfs2_create_symlink_data(struct ocfs2_super *osb,
+ struct ocfs2_journal_handle *handle,
+ struct inode *inode,
+ const char *symname);
+
+static inline int ocfs2_add_entry(struct ocfs2_journal_handle *handle,
+ struct dentry *dentry,
+ struct inode *inode, u64 blkno,
+ struct buffer_head *parent_fe_bh,
+ struct buffer_head *insert_bh)
+{
+ return __ocfs2_add_entry(handle, dentry->d_parent->d_inode,
+ dentry->d_name.name, dentry->d_name.len,
+ inode, blkno, parent_fe_bh, insert_bh);
+}
+
+/* An orphan dir name is an 8 byte value, printed as a hex string */
+#define OCFS2_ORPHAN_NAMELEN ((int)(2 * sizeof(u64)))
+
+static struct dentry *ocfs2_lookup(struct inode *dir, struct dentry *dentry,
+ struct nameidata *nd)
+{
+ int status;
+ u64 blkno;
+ struct buffer_head *dirent_bh = NULL;
+ struct inode *inode = NULL;
+ struct dentry *ret;
+ struct ocfs2_dir_entry *dirent;
+ struct ocfs2_inode_info *oi;
+
+ mlog_entry("(0x%p, 0x%p, '%.*s')\n", dir, dentry,
+ dentry->d_name.len, dentry->d_name.name);
+
+ if (dentry->d_name.len > OCFS2_MAX_FILENAME_LEN) {
+ ret = ERR_PTR(-ENAMETOOLONG);
+ goto bail;
+ }
+
+ mlog(0, "find name %.*s in directory %"MLFu64"\n", dentry->d_name.len,
+ dentry->d_name.name, OCFS2_I(dir)->ip_blkno);
+
+ status = ocfs2_meta_lock(dir, NULL, NULL, 0);
+ if (status < 0) {
+ if (status != -ENOENT)
+ mlog_errno(status);
+ ret = ERR_PTR(status);
+ goto bail;
+ }
+
+ status = ocfs2_find_files_on_disk(dentry->d_name.name,
+ dentry->d_name.len, &blkno,
+ dir, &dirent_bh, &dirent);
+ if (status < 0)
+ goto bail_add;
+
+ inode = ocfs2_iget(OCFS2_SB(dir->i_sb), blkno);
+ if (IS_ERR(inode)) {
+ mlog(ML_ERROR, "Unable to create inode %"MLFu64"\n", blkno);
+ ret = ERR_PTR(-EACCES);
+ goto bail_unlock;
+ }
+
+ oi = OCFS2_I(inode);
+ /* Clear any orphaned state... If we were able to look up the
+ * inode from a directory, it certainly can't be orphaned. We
+ * might have the bad state from a node which intended to
+ * orphan this inode but crashed before it could commit the
+ * unlink. */
+ spin_lock(&oi->ip_lock);
+ oi->ip_flags &= ~OCFS2_INODE_MAYBE_ORPHANED;
+ oi->ip_orphaned_slot = OCFS2_INVALID_SLOT;
+ spin_unlock(&oi->ip_lock);
+
+bail_add:
+
+ dentry->d_op = &ocfs2_dentry_ops;
+ ret = d_splice_alias(inode, dentry);
+
+bail_unlock:
+ /* Don't drop the cluster lock until *after* the d_add --
+ * unlink on another node will message us to remove that
+ * dentry under this lock so otherwise we can race this with
+ * the vote thread and have a stale dentry. */
+ ocfs2_meta_unlock(dir, 0);
+
+bail:
+ if (dirent_bh)
+ brelse(dirent_bh);
+
+ mlog_exit_ptr(ret);
+
+ return ret;
+}
+
+static int ocfs2_fill_new_dir(struct ocfs2_super *osb,
+ struct ocfs2_journal_handle *handle,
+ struct inode *parent,
+ struct inode *inode,
+ struct buffer_head *fe_bh,
+ struct ocfs2_alloc_context *data_ac)
+{
+ int status;
+ struct buffer_head *new_bh = NULL;
+ struct ocfs2_dir_entry *de = NULL;
+
+ mlog_entry_void();
+
+ status = ocfs2_do_extend_dir(osb->sb, handle, inode, fe_bh,
+ data_ac, NULL, &new_bh);
+ if (status < 0) {
+ mlog_errno(status);
+ goto bail;
+ }
+
+ ocfs2_set_new_buffer_uptodate(inode, new_bh);
+
+ status = ocfs2_journal_access(handle, inode, new_bh,
+ OCFS2_JOURNAL_ACCESS_CREATE);
+ if (status < 0) {
+ mlog_errno(status);
+ goto bail;
+ }
+ memset(new_bh->b_data, 0, osb->sb->s_blocksize);
+
+ de = (struct ocfs2_dir_entry *) new_bh->b_data;
+ de->inode = cpu_to_le64(OCFS2_I(inode)->ip_blkno);
+ de->name_len = 1;
+ de->rec_len =
+ cpu_to_le16(OCFS2_DIR_REC_LEN(de->name_len));
+ strcpy(de->name, ".");
+ ocfs2_set_de_type(de, S_IFDIR);
+ de = (struct ocfs2_dir_entry *) ((char *)de + le16_to_cpu(de->rec_len));
+ de->inode = cpu_to_le64(OCFS2_I(parent)->ip_blkno);
+ de->rec_len = cpu_to_le16(inode->i_sb->s_blocksize -
+ OCFS2_DIR_REC_LEN(1));
+ de->name_len = 2;
+ strcpy(de->name, "..");
+ ocfs2_set_de_type(de, S_IFDIR);
+
+ status = ocfs2_journal_dirty(handle, new_bh);
+ if (status < 0) {
+ mlog_errno(status);
+ goto bail;
+ }
+
+ i_size_write(inode, inode->i_sb->s_blocksize);
+ inode->i_nlink = 2;
+ inode->i_blocks = ocfs2_align_bytes_to_sectors(inode->i_sb->s_blocksize);
+ status = ocfs2_mark_inode_dirty(handle, inode, fe_bh);
+ if (status < 0) {
+ mlog_errno(status);
+ goto bail;
+ }
+
+ status = 0;
+bail:
+ if (new_bh)
+ brelse(new_bh);
+
+ mlog_exit(status);
+ return status;
+}
+
+static int ocfs2_mknod(struct inode *dir,
+ struct dentry *dentry,
+ int mode,
+ dev_t dev)
+{
+ int status = 0;
+ struct buffer_head *parent_fe_bh = NULL;
+ struct ocfs2_journal_handle *handle = NULL;
+ struct ocfs2_super *osb;
+ struct ocfs2_dinode *dirfe;
+ struct buffer_head *new_fe_bh = NULL;
+ struct buffer_head *de_bh = NULL;
+ struct inode *inode = NULL;
+ struct ocfs2_alloc_context *inode_ac = NULL;
+ struct ocfs2_alloc_context *data_ac = NULL;
+
+ mlog_entry("(0x%p, 0x%p, %d, %lu, '%.*s')\n", dir, dentry, mode,
+ (unsigned long)dev, dentry->d_name.len,
+ dentry->d_name.name);
+
+ /* get our super block */
+ osb = OCFS2_SB(dir->i_sb);
+
+ if (S_ISDIR(mode) && (dir->i_nlink >= OCFS2_LINK_MAX)) {
+ mlog(ML_ERROR, "inode %"MLFu64" has i_nlink of %u\n",
+ OCFS2_I(dir)->ip_blkno, dir->i_nlink);
+ status = -EMLINK;
+ goto leave;
+ }
+
+ handle = ocfs2_alloc_handle(osb);
+ if (handle == NULL) {
+ status = -ENOMEM;
+ mlog_errno(status);
+ goto leave;
+ }
+
+ status = ocfs2_meta_lock(dir, handle, &parent_fe_bh, 1);
+ if (status < 0) {
+ if (status != -ENOENT)
+ mlog_errno(status);
+ goto leave;
+ }
+
+ dirfe = (struct ocfs2_dinode *) parent_fe_bh->b_data;
+ if (!dirfe->i_links_count) {
+ /* can't make a file in a deleted directory. */
+ status = -ENOENT;
+ goto leave;
+ }
+
+ status = ocfs2_check_dir_for_entry(dir, dentry->d_name.name,
+ dentry->d_name.len);
+ if (status)
+ goto leave;
+
+ /* get a spot inside the dir. */
+ status = ocfs2_prepare_dir_for_insert(osb, dir, parent_fe_bh,
+ dentry->d_name.name,
+ dentry->d_name.len, &de_bh);
+ if (status < 0) {
+ mlog_errno(status);
+ goto leave;
+ }
+
+ /* reserve an inode spot */
+ status = ocfs2_reserve_new_inode(osb, handle, &inode_ac);
+ if (status < 0) {
+ if (status != -ENOSPC)
+ mlog_errno(status);
+ goto leave;
+ }
+
+ /* are we making a directory? If so, reserve a cluster for his
+ * 1st extent. */
+ if (S_ISDIR(mode)) {
+ status = ocfs2_reserve_clusters(osb, handle, 1, &data_ac);
+ if (status < 0) {
+ if (status != -ENOSPC)
+ mlog_errno(status);
+ goto leave;
+ }
+ }
+
+ handle = ocfs2_start_trans(osb, handle, OCFS2_MKNOD_CREDITS);
+ if (IS_ERR(handle)) {
+ status = PTR_ERR(handle);
+ handle = NULL;
+ mlog_errno(status);
+ goto leave;
+ }
+
+ /* do the real work now. */
+ status = ocfs2_mknod_locked(osb, dir, dentry, mode, dev,
+ &new_fe_bh, parent_fe_bh, handle,
+ &inode, inode_ac);
+ if (status < 0) {
+ mlog_errno(status);
+ goto leave;
+ }
+
+ if (S_ISDIR(mode)) {
+ status = ocfs2_fill_new_dir(osb, handle, dir, inode,
+ new_fe_bh, data_ac);
+ if (status < 0) {
+ mlog_errno(status);
+ goto leave;
+ }
+
+ status = ocfs2_journal_access(handle, dir, parent_fe_bh,
+ OCFS2_JOURNAL_ACCESS_WRITE);
+ if (status < 0) {
+ mlog_errno(status);
+ goto leave;
+ }
+ le16_add_cpu(&dirfe->i_links_count, 1);
+ status = ocfs2_journal_dirty(handle, parent_fe_bh);
+ if (status < 0) {
+ mlog_errno(status);
+ goto leave;
+ }
+ dir->i_nlink++;
+ }
+
+ status = ocfs2_add_entry(handle, dentry, inode,
+ OCFS2_I(inode)->ip_blkno, parent_fe_bh,
+ de_bh);
+ if (status < 0) {
+ mlog_errno(status);
+ goto leave;
+ }
+
+ insert_inode_hash(inode);
+ dentry->d_op = &ocfs2_dentry_ops;
+ d_instantiate(dentry, inode);
+ status = 0;
+leave:
+ if (handle)
+ ocfs2_commit_trans(handle);
+
+ if (status == -ENOSPC)
+ mlog(0, "Disk is full\n");
+
+ if (new_fe_bh)
+ brelse(new_fe_bh);
+
+ if (de_bh)
+ brelse(de_bh);
+
+ if (parent_fe_bh)
+ brelse(parent_fe_bh);
+
+ if ((status < 0) && inode)
+ iput(inode);
+
+ if (inode_ac)
+ ocfs2_free_alloc_context(inode_ac);
+
+ if (data_ac)
+ ocfs2_free_alloc_context(data_ac);
+
+ mlog_exit(status);
+
+ return status;
+}
+
+static int ocfs2_mknod_locked(struct ocfs2_super *osb,
+ struct inode *dir,
+ struct dentry *dentry, int mode,
+ dev_t dev,
+ struct buffer_head **new_fe_bh,
+ struct buffer_head *parent_fe_bh,
+ struct ocfs2_journal_handle *handle,
+ struct inode **ret_inode,
+ struct ocfs2_alloc_context *inode_ac)
+{
+ int status = 0;
+ struct ocfs2_dinode *fe = NULL;
+ struct ocfs2_extent_list *fel;
+ u64 fe_blkno = 0;
+ u16 suballoc_bit;
+ struct inode *inode = NULL;
+
+ mlog_entry("(0x%p, 0x%p, %d, %lu, '%.*s')\n", dir, dentry, mode,
+ (unsigned long)dev, dentry->d_name.len,
+ dentry->d_name.name);
+
+ *new_fe_bh = NULL;
+ *ret_inode = NULL;
+
+ status = ocfs2_claim_new_inode(osb, handle, inode_ac, &suballoc_bit,
+ &fe_blkno);
+ if (status < 0) {
+ mlog_errno(status);
+ goto leave;
+ }
+
+ inode = new_inode(dir->i_sb);
+ if (IS_ERR(inode)) {
+ status = PTR_ERR(inode);
+ mlog(ML_ERROR, "new_inode failed!\n");
+ goto leave;
+ }
+
+ /* populate as many fields early on as possible - many of
+ * these are used by the support functions here and in
+ * callers. */
+ inode->i_ino = ino_from_blkno(osb->sb, fe_blkno);
+ OCFS2_I(inode)->ip_blkno = fe_blkno;
+ if (S_ISDIR(mode))
+ inode->i_nlink = 2;
+ else
+ inode->i_nlink = 1;
+ inode->i_mode = mode;
+ spin_lock(&osb->osb_lock);
+ inode->i_generation = osb->s_next_generation++;
+ spin_unlock(&osb->osb_lock);
+
+ *new_fe_bh = sb_getblk(osb->sb, fe_blkno);
+ if (!*new_fe_bh) {
+ status = -EIO;
+ mlog_errno(status);
+ goto leave;
+ }
+ ocfs2_set_new_buffer_uptodate(inode, *new_fe_bh);
+
+ status = ocfs2_journal_access(handle, inode, *new_fe_bh,
+ OCFS2_JOURNAL_ACCESS_CREATE);
+ if (status < 0) {
+ mlog_errno(status);
+ goto leave;
+ }
+
+ fe = (struct ocfs2_dinode *) (*new_fe_bh)->b_data;
+ memset(fe, 0, osb->sb->s_blocksize);
+
+ fe->i_generation = cpu_to_le32(inode->i_generation);
+ fe->i_fs_generation = cpu_to_le32(osb->fs_generation);
+ fe->i_blkno = cpu_to_le64(fe_blkno);
+ fe->i_suballoc_bit = cpu_to_le16(suballoc_bit);
+ fe->i_suballoc_slot = cpu_to_le16(osb->slot_num);
+ fe->i_uid = cpu_to_le32(current->fsuid);
+ if (dir->i_mode & S_ISGID) {
+ fe->i_gid = cpu_to_le32(dir->i_gid);
+ if (S_ISDIR(mode))
+ mode |= S_ISGID;
+ } else
+ fe->i_gid = cpu_to_le32(current->fsgid);
+ fe->i_mode = cpu_to_le16(mode);
+ if (S_ISCHR(mode) || S_ISBLK(mode))
+ fe->id1.dev1.i_rdev = cpu_to_le64(huge_encode_dev(dev));
+
+ fe->i_links_count = cpu_to_le16(inode->i_nlink);
+
+ fe->i_last_eb_blk = 0;
+ strcpy(fe->i_signature, OCFS2_INODE_SIGNATURE);
+ le32_add_cpu(&fe->i_flags, OCFS2_VALID_FL);
+ fe->i_atime = fe->i_ctime = fe->i_mtime =
+ cpu_to_le64(CURRENT_TIME.tv_sec);
+ fe->i_mtime_nsec = fe->i_ctime_nsec = fe->i_atime_nsec =
+ cpu_to_le32(CURRENT_TIME.tv_nsec);
+ fe->i_dtime = 0;
+
+ fel = &fe->id2.i_list;
+ fel->l_tree_depth = 0;
+ fel->l_next_free_rec = 0;
+ fel->l_count = cpu_to_le16(ocfs2_extent_recs_per_inode(osb->sb));
+
+ status = ocfs2_journal_dirty(handle, *new_fe_bh);
+ if (status < 0) {
+ mlog_errno(status);
+ goto leave;
+ }
+
+ if (ocfs2_populate_inode(inode, fe, 1) < 0) {
+ mlog(ML_ERROR, "populate inode failed! bh->b_blocknr=%llu, "
+ "i_blkno=%"MLFu64", i_ino=%lu\n",
+ (unsigned long long) (*new_fe_bh)->b_blocknr,
+ fe->i_blkno, inode->i_ino);
+ BUG();
+ }
+
+ ocfs2_inode_set_new(osb, inode);
+ status = ocfs2_create_new_inode_locks(inode);
+ if (status < 0)
+ mlog_errno(status);
+
+ status = 0; /* error in ocfs2_create_new_inode_locks is not
+ * critical */
+
+ *ret_inode = inode;
+leave:
+ if (status < 0) {
+ if (*new_fe_bh) {
+ brelse(*new_fe_bh);
+ *new_fe_bh = NULL;
+ }
+ if (inode)
+ iput(inode);
+ }
+
+ mlog_exit(status);
+ return status;
+}
+
+static int ocfs2_mkdir(struct inode *dir,
+ struct dentry *dentry,
+ int mode)
+{
+ int ret;
+
+ mlog_entry("(0x%p, 0x%p, %d, '%.*s')\n", dir, dentry, mode,
+ dentry->d_name.len, dentry->d_name.name);
+ ret = ocfs2_mknod(dir, dentry, mode | S_IFDIR, 0);
+ mlog_exit(ret);
+
+ return ret;
+}
+
+static int ocfs2_create(struct inode *dir,
+ struct dentry *dentry,
+ int mode,
+ struct nameidata *nd)
+{
+ int ret;
+
+ mlog_entry("(0x%p, 0x%p, %d, '%.*s')\n", dir, dentry, mode,
+ dentry->d_name.len, dentry->d_name.name);
+ ret = ocfs2_mknod(dir, dentry, mode | S_IFREG, 0);
+ mlog_exit(ret);
+
+ return ret;
+}
+
+static int ocfs2_link(struct dentry *old_dentry,
+ struct inode *dir,
+ struct dentry *dentry)
+{
+ struct ocfs2_journal_handle *handle = NULL;
+ struct inode *inode = old_dentry->d_inode;
+ int err;
+ struct buffer_head *fe_bh = NULL;
+ struct buffer_head *parent_fe_bh = NULL;
+ struct buffer_head *de_bh = NULL;
+ struct ocfs2_dinode *fe = NULL;
+ struct ocfs2_super *osb = OCFS2_SB(dir->i_sb);
+
+ mlog_entry("(inode=%lu, old='%.*s' new='%.*s')\n", inode->i_ino,
+ old_dentry->d_name.len, old_dentry->d_name.name,
+ dentry->d_name.len, dentry->d_name.name);
+
+ if (S_ISDIR(inode->i_mode)) {
+ err = -EPERM;
+ goto bail;
+ }
+
+ if (inode->i_nlink >= OCFS2_LINK_MAX) {
+ err = -EMLINK;
+ goto bail;
+ }
+
+ handle = ocfs2_alloc_handle(osb);
+ if (handle == NULL) {
+ err = -ENOMEM;
+ goto bail;
+ }
+
+ err = ocfs2_meta_lock(dir, handle, &parent_fe_bh, 1);
+ if (err < 0) {
+ if (err != -ENOENT)
+ mlog_errno(err);
+ goto bail;
+ }
+
+ err = ocfs2_check_dir_for_entry(dir, dentry->d_name.name,
+ dentry->d_name.len);
+ if (err)
+ goto bail;
+
+ err = ocfs2_prepare_dir_for_insert(osb, dir, parent_fe_bh,
+ dentry->d_name.name,
+ dentry->d_name.len, &de_bh);
+ if (err < 0) {
+ mlog_errno(err);
+ goto bail;
+ }
+
+ err = ocfs2_meta_lock(inode, handle, &fe_bh, 1);
+ if (err < 0) {
+ if (err != -ENOENT)
+ mlog_errno(err);
+ goto bail;
+ }
+
+ fe = (struct ocfs2_dinode *) fe_bh->b_data;
+ if (le16_to_cpu(fe->i_links_count) >= OCFS2_LINK_MAX) {
+ err = -EMLINK;
+ goto bail;
+ }
+
+ handle = ocfs2_start_trans(osb, handle, OCFS2_LINK_CREDITS);
+ if (IS_ERR(handle)) {
+ err = PTR_ERR(handle);
+ handle = NULL;
+ mlog_errno(err);
+ goto bail;
+ }
+
+ err = ocfs2_journal_access(handle, inode, fe_bh,
+ OCFS2_JOURNAL_ACCESS_WRITE);
+ if (err < 0) {
+ mlog_errno(err);
+ goto bail;
+ }
+
+ inode->i_nlink++;
+ inode->i_ctime = CURRENT_TIME;
+ fe->i_links_count = cpu_to_le16(inode->i_nlink);
+ fe->i_ctime = cpu_to_le64(inode->i_ctime.tv_sec);
+ fe->i_ctime_nsec = cpu_to_le32(inode->i_ctime.tv_nsec);
+
+ err = ocfs2_journal_dirty(handle, fe_bh);
+ if (err < 0) {
+ le16_add_cpu(&fe->i_links_count, -1);
+ inode->i_nlink--;
+ mlog_errno(err);
+ goto bail;
+ }
+
+ err = ocfs2_add_entry(handle, dentry, inode,
+ OCFS2_I(inode)->ip_blkno,
+ parent_fe_bh, de_bh);
+ if (err) {
+ le16_add_cpu(&fe->i_links_count, -1);
+ inode->i_nlink--;
+ mlog_errno(err);
+ goto bail;
+ }
+
+ atomic_inc(&inode->i_count);
+ dentry->d_op = &ocfs2_dentry_ops;
+ d_instantiate(dentry, inode);
+bail:
+ if (handle)
+ ocfs2_commit_trans(handle);
+ if (de_bh)
+ brelse(de_bh);
+ if (fe_bh)
+ brelse(fe_bh);
+ if (parent_fe_bh)
+ brelse(parent_fe_bh);
+
+ mlog_exit(err);
+
+ return err;
+}
+
+static int ocfs2_unlink(struct inode *dir,
+ struct dentry *dentry)
+{
+ int status;
+ unsigned int saved_nlink = 0;
+ struct inode *inode = dentry->d_inode;
+ struct ocfs2_super *osb = OCFS2_SB(dir->i_sb);
+ u64 blkno;
+ struct ocfs2_dinode *fe = NULL;
+ struct buffer_head *fe_bh = NULL;
+ struct buffer_head *parent_node_bh = NULL;
+ struct ocfs2_journal_handle *handle = NULL;
+ struct ocfs2_dir_entry *dirent = NULL;
+ struct buffer_head *dirent_bh = NULL;
+ char orphan_name[OCFS2_ORPHAN_NAMELEN + 1];
+ struct buffer_head *orphan_entry_bh = NULL;
+
+ mlog_entry("(0x%p, 0x%p, '%.*s')\n", dir, dentry,
+ dentry->d_name.len, dentry->d_name.name);
+
+ BUG_ON(dentry->d_parent->d_inode != dir);
+
+ mlog(0, "ino = %"MLFu64"\n", OCFS2_I(inode)->ip_blkno);
+
+ if (inode == osb->root_inode) {
+ mlog(0, "Cannot delete the root directory\n");
+ status = -EPERM;
+ goto leave;
+ }
+
+ handle = ocfs2_alloc_handle(osb);
+ if (handle == NULL) {
+ status = -ENOMEM;
+ mlog_errno(status);
+ goto leave;
+ }
+
+ status = ocfs2_meta_lock(dir, handle, &parent_node_bh, 1);
+ if (status < 0) {
+ if (status != -ENOENT)
+ mlog_errno(status);
+ goto leave;
+ }
+
+ status = ocfs2_find_files_on_disk(dentry->d_name.name,
+ dentry->d_name.len, &blkno,
+ dir, &dirent_bh, &dirent);
+ if (status < 0) {
+ if (status != -ENOENT)
+ mlog_errno(status);
+ goto leave;
+ }
+
+ if (OCFS2_I(inode)->ip_blkno != blkno) {
+ status = -ENOENT;
+
+ mlog(0, "ip_blkno (%"MLFu64") != dirent blkno (%"MLFu64") "
+ "ip_flags = %x\n", OCFS2_I(inode)->ip_blkno, blkno,
+ OCFS2_I(inode)->ip_flags);
+ goto leave;
+ }
+
+ status = ocfs2_meta_lock(inode, handle, &fe_bh, 1);
+ if (status < 0) {
+ if (status != -ENOENT)
+ mlog_errno(status);
+ goto leave;
+ }
+
+ if (S_ISDIR(inode->i_mode)) {
+ if (!ocfs2_empty_dir(inode)) {
+ status = -ENOTEMPTY;
+ goto leave;
+ } else if (inode->i_nlink != 2) {
+ status = -ENOTEMPTY;
+ goto leave;
+ }
+ }
+
+ /* There are still a few steps left until we can consider the
+ * unlink to have succeeded. Save off nlink here before
+ * modification so we can set it back in case we hit an issue
+ * before commit. */
+ saved_nlink = inode->i_nlink;
+ if (S_ISDIR(inode->i_mode))
+ inode->i_nlink = 0;
+ else
+ inode->i_nlink--;
+
+ status = ocfs2_request_unlink_vote(inode, dentry,
+ (unsigned int) inode->i_nlink);
+ if (status < 0) {
+ /* This vote should succeed under all normal
+ * circumstances. */
+ mlog_errno(status);
+ goto leave;
+ }
+
+ if (!inode->i_nlink) {
+ status = ocfs2_prepare_orphan_dir(osb, handle, inode,
+ orphan_name,
+ &orphan_entry_bh);
+ if (status < 0) {
+ mlog_errno(status);
+ goto leave;
+ }
+ }
+
+ handle = ocfs2_start_trans(osb, handle, OCFS2_UNLINK_CREDITS);
+ if (IS_ERR(handle)) {
+ status = PTR_ERR(handle);
+ handle = NULL;
+ mlog_errno(status);
+ goto leave;
+ }
+
+ status = ocfs2_journal_access(handle, inode, fe_bh,
+ OCFS2_JOURNAL_ACCESS_WRITE);
+ if (status < 0) {
+ mlog_errno(status);
+ goto leave;
+ }
+
+ fe = (struct ocfs2_dinode *) fe_bh->b_data;
+
+ if (!inode->i_nlink) {
+ status = ocfs2_orphan_add(osb, handle, inode, fe, orphan_name,
+ orphan_entry_bh);
+ if (status < 0) {
+ mlog_errno(status);
+ goto leave;
+ }
+ }
+
+ /* delete the name from the parent dir */
+ status = ocfs2_delete_entry(handle, dir, dirent, dirent_bh);
+ if (status < 0) {
+ mlog_errno(status);
+ goto leave;
+ }
+
+ /* We can set nlink on the dinode now. clear the saved version
+ * so that it doesn't get set later. */
+ fe->i_links_count = cpu_to_le16(inode->i_nlink);
+ saved_nlink = 0;
+
+ status = ocfs2_journal_dirty(handle, fe_bh);
+ if (status < 0) {
+ mlog_errno(status);
+ goto leave;
+ }
+
+ if (S_ISDIR(inode->i_mode)) {
+ dir->i_nlink--;
+ status = ocfs2_mark_inode_dirty(handle, dir,
+ parent_node_bh);
+ if (status < 0) {
+ mlog_errno(status);
+ dir->i_nlink++;
+ }
+ }
+
+leave:
+ if (status < 0 && saved_nlink)
+ inode->i_nlink = saved_nlink;
+
+ if (handle)
+ ocfs2_commit_trans(handle);
+
+ if (fe_bh)
+ brelse(fe_bh);
+
+ if (dirent_bh)
+ brelse(dirent_bh);
+
+ if (parent_node_bh)
+ brelse(parent_node_bh);
+
+ if (orphan_entry_bh)
+ brelse(orphan_entry_bh);
+
+ mlog_exit(status);
+
+ return status;
+}
+
+/*
+ * The only place this should be used is rename!
+ * if they have the same id, then the 1st one is the only one locked.
+ */
+static int ocfs2_double_lock(struct ocfs2_super *osb,
+ struct ocfs2_journal_handle *handle,
+ struct buffer_head **bh1,
+ struct inode *inode1,
+ struct buffer_head **bh2,
+ struct inode *inode2)
+{
+ int status;
+ struct ocfs2_inode_info *oi1 = OCFS2_I(inode1);
+ struct ocfs2_inode_info *oi2 = OCFS2_I(inode2);
+ struct buffer_head **tmpbh;
+ struct inode *tmpinode;
+
+ mlog_entry("(inode1 = %"MLFu64", inode2 = %"MLFu64")\n",
+ oi1->ip_blkno, oi2->ip_blkno);
+
+ BUG_ON(!handle);
+
+ if (*bh1)
+ *bh1 = NULL;
+ if (*bh2)
+ *bh2 = NULL;
+
+ /* we always want to lock the one with the lower lockid first. */
+ if (oi1->ip_blkno != oi2->ip_blkno) {
+ if (oi1->ip_blkno < oi2->ip_blkno) {
+ /* switch id1 and id2 around */
+ mlog(0, "switching them around...\n");
+ tmpbh = bh2;
+ bh2 = bh1;
+ bh1 = tmpbh;
+
+ tmpinode = inode2;
+ inode2 = inode1;
+ inode1 = tmpinode;
+ }
+ /* lock id2 */
+ status = ocfs2_meta_lock(inode2, handle, bh2, 1);
+ if (status < 0) {
+ if (status != -ENOENT)
+ mlog_errno(status);
+ goto bail;
+ }
+ }
+ /* lock id1 */
+ status = ocfs2_meta_lock(inode1, handle, bh1, 1);
+ if (status < 0) {
+ if (status != -ENOENT)
+ mlog_errno(status);
+ goto bail;
+ }
+bail:
+ mlog_exit(status);
+ return status;
+}
+
+#define PARENT_INO(buffer) \
+ ((struct ocfs2_dir_entry *) \
+ ((char *)buffer + \
+ le16_to_cpu(((struct ocfs2_dir_entry *)buffer)->rec_len)))->inode
+
+static int ocfs2_rename(struct inode *old_dir,
+ struct dentry *old_dentry,
+ struct inode *new_dir,
+ struct dentry *new_dentry)
+{
+ int status = 0, rename_lock = 0;
+ struct inode *old_inode = old_dentry->d_inode;
+ struct inode *new_inode = new_dentry->d_inode;
+ struct ocfs2_dinode *newfe = NULL;
+ char orphan_name[OCFS2_ORPHAN_NAMELEN + 1];
+ struct buffer_head *orphan_entry_bh = NULL;
+ struct buffer_head *newfe_bh = NULL;
+ struct buffer_head *insert_entry_bh = NULL;
+ struct ocfs2_super *osb = NULL;
+ u64 newfe_blkno;
+ struct ocfs2_journal_handle *handle = NULL;
+ struct buffer_head *old_dir_bh = NULL;
+ struct buffer_head *new_dir_bh = NULL;
+ struct ocfs2_dir_entry *old_de = NULL, *new_de = NULL; // dirent for old_dentry
+ // and new_dentry
+ struct buffer_head *new_de_bh = NULL, *old_de_bh = NULL; // bhs for above
+ struct buffer_head *old_inode_de_bh = NULL; // if old_dentry is a dir,
+ // this is the 1st dirent bh
+ nlink_t old_dir_nlink = old_dir->i_nlink, new_dir_nlink = new_dir->i_nlink;
+ unsigned int links_count;
+
+ /* At some point it might be nice to break this function up a
+ * bit. */
+
+ mlog_entry("(0x%p, 0x%p, 0x%p, 0x%p, from='%.*s' to='%.*s')\n",
+ old_dir, old_dentry, new_dir, new_dentry,
+ old_dentry->d_name.len, old_dentry->d_name.name,
+ new_dentry->d_name.len, new_dentry->d_name.name);
+
+ osb = OCFS2_SB(old_dir->i_sb);
+
+ if (new_inode) {
+ if (!igrab(new_inode))
+ BUG();
+ }
+
+ if (atomic_read(&old_dentry->d_count) > 2) {
+ shrink_dcache_parent(old_dentry);
+ if (atomic_read(&old_dentry->d_count) > 2) {
+ status = -EBUSY;
+ goto bail;
+ }
+ }
+
+ /* Assume a directory heirarchy thusly:
+ * a/b/c
+ * a/d
+ * a,b,c, and d are all directories.
+ *
+ * from cwd of 'a' on both nodes:
+ * node1: mv b/c d
+ * node2: mv d b/c
+ *
+ * And that's why, just like the VFS, we need a file system
+ * rename lock. */
+ if (old_dentry != new_dentry) {
+ status = ocfs2_rename_lock(osb);
+ if (status < 0) {
+ mlog_errno(status);
+ goto bail;
+ }
+ rename_lock = 1;
+ }
+
+ handle = ocfs2_alloc_handle(osb);
+ if (handle == NULL) {
+ status = -ENOMEM;
+ mlog_errno(status);
+ goto bail;
+ }
+
+ /* if old and new are the same, this'll just do one lock. */
+ status = ocfs2_double_lock(osb, handle,
+ &old_dir_bh, old_dir,
+ &new_dir_bh, new_dir);
+ if (status < 0) {
+ mlog_errno(status);
+ goto bail;
+ }
+
+ /* make sure both dirs have bhs
+ * get an extra ref on old_dir_bh if old==new */
+ if (!new_dir_bh) {
+ if (old_dir_bh) {
+ new_dir_bh = old_dir_bh;
+ get_bh(new_dir_bh);
+ } else {
+ mlog(ML_ERROR, "no old_dir_bh!\n");
+ status = -EIO;
+ goto bail;
+ }
+ }
+
+ if (S_ISDIR(old_inode->i_mode)) {
+ /* Directories actually require metadata updates to
+ * the directory info so we can't get away with not
+ * doing node locking on it. */
+ status = ocfs2_meta_lock(old_inode, handle, NULL, 1);
+ if (status < 0) {
+ if (status != -ENOENT)
+ mlog_errno(status);
+ goto bail;
+ }
+
+ status = ocfs2_request_rename_vote(old_inode, old_dentry);
+ if (status < 0) {
+ mlog_errno(status);
+ goto bail;
+ }
+
+ status = -EIO;
+ old_inode_de_bh = ocfs2_bread(old_inode, 0, &status, 0);
+ if (!old_inode_de_bh)
+ goto bail;
+
+ status = -EIO;
+ if (le64_to_cpu(PARENT_INO(old_inode_de_bh->b_data)) !=
+ OCFS2_I(old_dir)->ip_blkno)
+ goto bail;
+ status = -EMLINK;
+ if (!new_inode && new_dir!=old_dir &&
+ new_dir->i_nlink >= OCFS2_LINK_MAX)
+ goto bail;
+ } else {
+ /* Ah, the simple case - we're a file so just send a
+ * message. */
+ status = ocfs2_request_rename_vote(old_inode, old_dentry);
+ if (status < 0) {
+ mlog_errno(status);
+ goto bail;
+ }
+ }
+
+ status = -ENOENT;
+ old_de_bh = ocfs2_find_entry(old_dentry->d_name.name,
+ old_dentry->d_name.len,
+ old_dir, &old_de);
+ if (!old_de_bh)
+ goto bail;
+
+ /*
+ * Check for inode number is _not_ due to possible IO errors.
+ * We might rmdir the source, keep it as pwd of some process
+ * and merrily kill the link to whatever was created under the
+ * same name. Goodbye sticky bit ;-<
+ */
+ if (le64_to_cpu(old_de->inode) != OCFS2_I(old_inode)->ip_blkno)
+ goto bail;
+
+ /* check if the target already exists (in which case we need
+ * to delete it */
+ status = ocfs2_find_files_on_disk(new_dentry->d_name.name,
+ new_dentry->d_name.len,
+ &newfe_blkno, new_dir, &new_de_bh,
+ &new_de);
+ /* The only error we allow here is -ENOENT because the new
+ * file not existing is perfectly valid. */
+ if ((status < 0) && (status != -ENOENT)) {
+ /* If we cannot find the file specified we should just */
+ /* return the error... */
+ mlog_errno(status);
+ goto bail;
+ }
+
+ if (!new_de && new_inode)
+ mlog(ML_ERROR, "inode %lu does not exist in it's parent "
+ "directory!", new_inode->i_ino);
+
+ /* In case we need to overwrite an existing file, we blow it
+ * away first */
+ if (new_de) {
+ /* VFS didn't think there existed an inode here, but
+ * someone else in the cluster must have raced our
+ * rename to create one. Today we error cleanly, in
+ * the future we should consider calling iget to build
+ * a new struct inode for this entry. */
+ if (!new_inode) {
+ status = -EACCES;
+
+ mlog(0, "We found an inode for name %.*s but VFS "
+ "didn't give us one.\n", new_dentry->d_name.len,
+ new_dentry->d_name.name);
+ goto bail;
+ }
+
+ if (OCFS2_I(new_inode)->ip_blkno != newfe_blkno) {
+ status = -EACCES;
+
+ mlog(0, "Inode blkno (%"MLFu64") and dir (%"MLFu64") "
+ "disagree. ip_flags = %x\n",
+ OCFS2_I(new_inode)->ip_blkno, newfe_blkno,
+ OCFS2_I(new_inode)->ip_flags);
+ goto bail;
+ }
+
+ status = ocfs2_meta_lock(new_inode, handle, &newfe_bh, 1);
+ if (status < 0) {
+ if (status != -ENOENT)
+ mlog_errno(status);
+ goto bail;
+ }
+
+ if (S_ISDIR(new_inode->i_mode))
+ links_count = 0;
+ else
+ links_count = (unsigned int) (new_inode->i_nlink - 1);
+
+ status = ocfs2_request_unlink_vote(new_inode, new_dentry,
+ links_count);
+ if (status < 0) {
+ mlog_errno(status);
+ goto bail;
+ }
+
+ newfe = (struct ocfs2_dinode *) newfe_bh->b_data;
+
+ mlog(0, "aha rename over existing... new_de=%p "
+ "new_blkno=%"MLFu64" newfebh=%p bhblocknr=%llu\n",
+ new_de, newfe_blkno, newfe_bh, newfe_bh ?
+ (unsigned long long)newfe_bh->b_blocknr : 0ULL);
+
+ if (S_ISDIR(new_inode->i_mode) || (new_inode->i_nlink == 1)) {
+ status = ocfs2_prepare_orphan_dir(osb, handle,
+ new_inode,
+ orphan_name,