-/*
- * smfs/inode.c
+/* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
+ * vim:expandtab:shiftwidth=8:tabstop=8:
*
+ * lustre/smfs/inode.c
+ * Lustre filesystem abstraction routines
+ *
+ * Copyright (C) 2004 Cluster File Systems, Inc.
+ *
+ * This file is part of Lustre, http://www.lustre.org.
+ *
+ * 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.
+ *
+ * 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.
+ *
+ * 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.
*/
#define DEBUG_SUBSYSTEM S_SM
#include <linux/fs.h>
#include <linux/slab.h>
#include <linux/string.h>
-#include "smfs_internal.h"
+#include <linux/obd_class.h>
+#include <linux/obd_support.h>
+#include <linux/lustre_lib.h>
+#include <linux/lustre_idl.h>
+#include <linux/lustre_fsfilt.h>
+#include <linux/lustre_smfs.h>
+#include "smfs_internal.h"
+
+static void smfs_init_inode_info(struct inode *inode, void *opaque)
+{
+ if (!I2SMI(inode)) {
+ struct inode *cache_inode = NULL;
+ struct smfs_iget_args *sargs;
+ unsigned long ino;
+
+ sargs = (struct smfs_iget_args *)opaque;
+
+ /* getting backing fs inode. */
+ ino = sargs ? sargs->s_ino : inode->i_ino;
+ cache_inode = iget(S2CSB(inode->i_sb), ino);
+
+ OBD_ALLOC(inode->u.generic_ip,
+ sizeof(struct smfs_inode_info));
+
+ LASSERT(inode->u.generic_ip);
+ I2CI(inode) = cache_inode;
+
+ CDEBUG(D_INODE, "cache_inode i_count ino %lu i_count %d\n",
+ cache_inode->i_ino, atomic_read(&cache_inode->i_count));
+
+ post_smfs_inode(inode, cache_inode);
+ sm_set_inode_ops(cache_inode, inode);
+
+ if (sargs) {
+ struct inode *dir = sargs->s_inode;
+ if (dir)
+ I2SMI(inode)->smi_flags = I2SMI(dir)->smi_flags;
+ }
+ }
+}
-void duplicate_inode(struct inode *cache_inode, struct inode *inode)
+static void smfs_clear_inode_info(struct inode *inode)
{
-
- inode->i_mode = cache_inode->i_mode;
- inode->i_uid = cache_inode->i_uid;
- inode->i_gid = cache_inode->i_gid;
-
- inode->i_nlink = cache_inode->i_nlink;
- inode->i_size = cache_inode->i_size;
- inode->i_atime = cache_inode->i_atime;
- inode->i_ctime = cache_inode->i_ctime;
- inode->i_mtime = cache_inode->i_mtime;
- inode->i_blksize = cache_inode->i_blksize; /* This is the optimal IO size
- * (for stat), not the fs block
- * size */
- inode->i_blocks = cache_inode->i_blocks;
- inode->i_version = cache_inode->i_version;
- inode->i_state = cache_inode->i_state;
+ if (I2SMI(inode)) {
+ struct inode *cache_inode = I2CI(inode);
+
+ LASSERTF(((atomic_read(&cache_inode->i_count) == 1) ||
+ cache_inode == cache_inode->i_sb->s_root->d_inode),
+ "inode %p cache inode %p %lu i_count %d != 0 \n",
+ inode, cache_inode, cache_inode->i_ino,
+ atomic_read(&cache_inode->i_count));
+
+ if (cache_inode != cache_inode->i_sb->s_root->d_inode)
+ iput(cache_inode);
+
+ OBD_FREE(inode->u.generic_ip,
+ sizeof(struct smfs_inode_info));
+ inode->u.generic_ip = NULL;
+ }
}
-static void smfs_read_inode(struct inode *inode)
+
+static void smfs_read_inode2(struct inode *inode, void *opaque)
{
- struct super_block *cache_sb;
- struct inode *cache_inode;
- ENTRY;
-
- if (!inode)
- return;
-
- CDEBUG(D_INODE, "read_inode ino %lu\n", inode->i_ino);
- cache_sb = S2CSB(inode->i_sb);
-
- cache_inode = iget(cache_sb, inode->i_ino);
- I2CI(inode) = cache_inode;
-
- if(cache_sb && cache_sb->s_op->read_inode)
- cache_sb->s_op->read_inode(cache_inode);
-
- duplicate_inode(cache_inode, inode);
- sm_set_inode_ops(cache_inode, inode);
-
- CDEBUG(D_INODE, "read_inode ino %lu icount %d \n",
- inode->i_ino, atomic_read(&inode->i_count));
-
- iput(cache_inode);
- return;
+ ENTRY;
+
+ if (!inode) {
+ EXIT;
+ return;
+ }
+
+ CDEBUG(D_INODE, "read_inode ino %lu\n", inode->i_ino);
+ smfs_init_inode_info(inode, opaque);
+ CDEBUG(D_INODE, "read_inode ino %lu icount %d \n",
+ inode->i_ino, atomic_read(&inode->i_count));
+ EXIT;
}
-/* Although some filesystem(such as ext3) do not have
- * clear_inode method, but we need it to free the
- * cache inode
- */
-static void smfs_clear_inode(struct inode *inode)
+
+#if (LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0))
+static int smfs_test_inode(struct inode *inode, unsigned long ino,
+ void *opaque)
+#else
+static int smfs_test_inode(struct inode *inode, void *opaque)
+#endif
+{
+ struct smfs_iget_args *sargs = (struct smfs_iget_args*)opaque;
+
+ LASSERT(sargs);
+
+ if (!sargs)
+ return 1;
+
+ if (inode->i_ino != sargs->s_ino)
+ return 0;
+
+#ifdef CONFIG_SNAPFS
+ if (SMFS_DO_COW(S2SMI(inode->i_sb)) &&
+ !smfs_snap_test_inode(inode, opaque))
+ return 0;
+#endif
+ if (I2SMI(inode)) {
+ struct inode *cache_inode = I2CI(inode);
+
+ /* this is needed to make symatry between smfs_put_inode(). */
+ LASSERT(cache_inode != NULL);
+ igrab(cache_inode);
+
+ LASSERTF(cache_inode->i_ino == inode->i_ino,
+ "inode ino %lu != cache ino %lu",
+ cache_inode->i_ino, inode->i_ino);
+ } else {
+ smfs_init_inode_info(inode, opaque);
+ }
+
+ return 1;
+}
+
+#if (LINUX_VERSION_CODE >= KERNEL_VERSION(2,5,0))
+int smfs_set_inode(struct inode *inode, void *opaque)
+{
+ smfs_read_inode2(inode, opaque);
+ return 0;
+}
+
+struct inode *smfs_iget(struct super_block *sb, ino_t hash,
+ struct smfs_iget_args *sargs)
+{
+ struct inode *inode;
+ LASSERT(hash != 0);
+
+ inode = iget5_locked(sb, hash, smfs_test_inode,
+ smfs_set_inode, sargs);
+ if (inode) {
+ if (inode->i_state & I_NEW)
+ unlock_new_inode(inode);
+ CDEBUG(D_VFSTRACE, "inode: %lu/%u(%p) index %d "
+ "ino %d\n", inode->i_ino, inode->i_generation,
+ inode, sargs->s_index, sargs->s_ino);
+ inode->i_ino = hash;
+ }
+ return inode;
+}
+#else
+struct inode *smfs_iget(struct super_block *sb, ino_t hash,
+ struct smfs_iget_args *sargs)
+{
+ struct inode *inode;
+ LASSERT(hash != 0);
+
+ inode = iget4(sb, hash, smfs_test_inode, sargs);
+ if (inode) {
+ struct inode *cache_inode = I2CI(inode);
+ LASSERTF((inode->i_ino == cache_inode->i_ino),
+ "inode %p ino %lu != cache inode %p ino %lu",
+ inode, inode->i_ino, cache_inode, cache_inode->i_ino);
+ CDEBUG(D_VFSTRACE, "inode: %lu/%u(%p)\n", inode->i_ino,
+ inode->i_generation, inode);
+ }
+ return inode;
+}
+#endif
+
+struct inode *smfs_get_inode(struct super_block *sb, ino_t hash,
+ struct inode *dir, int index)
{
- struct super_block *cache_sb;
- struct inode *cache_inode;
-
- ENTRY;
-
- if (!inode) return;
-
- cache_sb = S2CSB(inode->i_sb);
- cache_inode = I2CI(inode);
- iput(cache_inode);
-
- return;
+ struct smfs_iget_args sargs;
+ struct inode *inode;
+ ENTRY;
+
+ sargs.s_ino = hash;
+ sargs.s_inode = dir;
+ sargs.s_index = index;
+ inode = smfs_iget(sb, hash, &sargs);
+
+ RETURN(inode);
}
+
static void smfs_delete_inode(struct inode *inode)
{
- struct inode *cache_inode;
- struct super_block *cache_sb;
-
- ENTRY;
- cache_inode = I2CI(inode);
- cache_sb = S2CSB(inode->i_sb);
-
- if (!cache_inode || !cache_sb)
- return;
-
- /*FIXME: because i_count of cache_inode may not
- * be 0 or 1 in before smfs_delete inode, So we
- * need to dec it to 1 before we call delete_inode
- * of the bellow cache filesystem Check again latter*/
-
- if (atomic_read(&cache_inode->i_count) < 1)
- BUG();
-
- while (atomic_read(&cache_inode->i_count) != 1) {
- atomic_dec(&cache_inode->i_count);
- }
-
- duplicate_inode(inode, cache_inode);
-
- list_del(&cache_inode->i_hash);
- INIT_LIST_HEAD(&cache_inode->i_hash);
- list_del(&cache_inode->i_list);
- INIT_LIST_HEAD(&cache_inode->i_list);
-
- if (cache_inode->i_data.nrpages)
- truncate_inode_pages(&cache_inode->i_data, 0);
-
- if (cache_sb->s_op->delete_inode)
- cache_sb->s_op->delete_inode(cache_inode);
-
- duplicate_inode(cache_inode, inode);
-
- return;
+ ENTRY;
+ clear_inode(inode);
+ EXIT;
}
+
static void smfs_write_inode(struct inode *inode, int wait)
{
- struct inode *cache_inode;
- struct super_block *cache_sb;
-
- ENTRY;
- cache_inode = I2CI(inode);
- cache_sb = S2CSB(inode->i_sb);
-
- if (!cache_inode || !cache_sb)
- return;
-
- if (cache_sb->s_op->write_inode)
- cache_sb->s_op->write_inode(cache_inode, wait);
-
- duplicate_inode(cache_inode, inode);
-
- return;
+ struct inode *cache_inode;
+ ENTRY;
+
+ cache_inode = I2CI(inode);
+ LASSERT(cache_inode != NULL);
+
+ pre_smfs_inode(inode, cache_inode);
+ if (S2CSB(inode->i_sb)->s_op->write_inode)
+ S2CSB(inode->i_sb)->s_op->write_inode(cache_inode, wait);
+
+ post_smfs_inode(inode, cache_inode);
+ EXIT;
}
+
static void smfs_dirty_inode(struct inode *inode)
{
- struct inode *cache_inode;
- struct super_block *cache_sb;
+ struct inode *cache_inode;
+ ENTRY;
- ENTRY;
- cache_inode = I2CI(inode);
- cache_sb = S2CSB(inode->i_sb);
+ cache_inode = I2CI(inode);
+ if (!cache_inode || !S2CSB(inode->i_sb)) {
+ EXIT;
+ return;
+ }
- if (!cache_inode || !cache_sb)
- return;
-
- if (cache_sb->s_op->dirty_inode)
- cache_sb->s_op->dirty_inode(cache_inode);
+ pre_smfs_inode(inode, cache_inode);
+ if (S2CSB(inode->i_sb)->s_op->dirty_inode)
+ S2CSB(inode->i_sb)->s_op->dirty_inode(cache_inode);
- duplicate_inode(cache_inode, inode);
- return;
+ post_smfs_inode(inode, cache_inode);
+ EXIT;
}
static void smfs_put_inode(struct inode *inode)
{
- struct inode *cache_inode;
- struct super_block *cache_sb;
-
- ENTRY;
- cache_inode = I2CI(inode);
- cache_sb = S2CSB(inode->i_sb);
-
- if (!cache_inode || !cache_sb)
- return;
- if (cache_sb->s_op->put_inode)
- cache_sb->s_op->put_inode(cache_inode);
+ struct inode *cache_inode;
+ ENTRY;
+
+ cache_inode = I2CI(inode);
+ LASSERT(cache_inode != NULL);
+
+ if (atomic_read(&cache_inode->i_count) > 1)
+ iput(cache_inode);
+
+ /*
+ * check if we have to release backend inode. As iput() calls fs
+ * specific ->put_inode() first and than decrements inode rerfcount, we
+ * should check here ->i_count for 1, not for 0.
+
+ * In principle we could rely on ->clear_inode(), but it will be called
+ * later, and thus we will have iputing of @inode and @cache_inode in
+ * different control paths, this is why we do it here. --umka
+ */
+ if (atomic_read(&inode->i_count) == 1 &&
+ cache_inode->i_sb->s_root->d_inode != cache_inode)
+ smfs_clear_inode_info(inode);
+
+ EXIT;
+}
- return;
+static void smfs_clear_inode(struct inode *inode)
+{
+ ENTRY;
+ smfs_clear_inode_info(inode);
+ EXIT;
}
static void smfs_write_super(struct super_block *sb)
{
- struct super_block *cache_sb;
-
- ENTRY;
- cache_sb = S2CSB(sb);
+ ENTRY;
- if (!cache_sb)
- return;
-
- if (cache_sb->s_op->write_super)
- cache_sb->s_op->write_super(cache_sb);
+ if (!S2CSB(sb)) {
+ EXIT;
+ return;
+ }
- duplicate_sb(cache_sb, sb);
- return;
+ if (S2CSB(sb)->s_op->write_super)
+ S2CSB(sb)->s_op->write_super(S2CSB(sb));
+ duplicate_sb(sb, S2CSB(sb));
+ EXIT;
}
static void smfs_write_super_lockfs(struct super_block *sb)
{
- struct super_block *cache_sb;
+ struct super_block *cache_sb;
+ ENTRY;
- ENTRY;
- cache_sb = S2CSB(sb);
+ cache_sb = S2CSB(sb);
+ if (!cache_sb) {
+ EXIT;
+ return;
+ }
- if (!cache_sb)
- return;
-
- if (cache_sb->s_op->write_super_lockfs)
- cache_sb->s_op->write_super_lockfs(cache_sb);
+ if (cache_sb->s_op->write_super_lockfs)
+ cache_sb->s_op->write_super_lockfs(cache_sb);
- duplicate_sb(cache_sb, sb);
- return;
+ duplicate_sb(sb, cache_sb);
+ EXIT;
}
static void smfs_unlockfs(struct super_block *sb)
{
- struct super_block *cache_sb;
+ struct super_block *cache_sb;
+ ENTRY;
- ENTRY;
- cache_sb = S2CSB(sb);
+ cache_sb = S2CSB(sb);
+ if (!cache_sb) {
+ EXIT;
+ return;
+ }
- if (!cache_sb)
- return;
-
- if (cache_sb->s_op->unlockfs)
- cache_sb->s_op->unlockfs(cache_sb);
+ if (cache_sb->s_op->unlockfs)
+ cache_sb->s_op->unlockfs(cache_sb);
- duplicate_sb(cache_sb, sb);
- return;
+ duplicate_sb(sb, cache_sb);
+ EXIT;
}
-static int smfs_statfs(struct super_block * sb, struct statfs * buf)
+
+#if (LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0))
+static int smfs_statfs(struct super_block *sb, struct statfs *buf)
+#else
+static int smfs_statfs(struct super_block *sb, struct kstatfs *buf)
+#endif
{
- struct super_block *cache_sb;
- int rc = 0;
+ struct super_block *cache_sb;
+ int rc = 0;
+ ENTRY;
- ENTRY;
- cache_sb = S2CSB(sb);
+ cache_sb = S2CSB(sb);
+ if (!cache_sb)
+ RETURN(-EINVAL);
- if (!cache_sb)
- RETURN(-EINVAL);
-
- if (cache_sb->s_op->statfs)
- rc = cache_sb->s_op->statfs(cache_sb, buf);
+ if (cache_sb->s_op->statfs)
+ rc = cache_sb->s_op->statfs(cache_sb, buf);
- duplicate_sb(cache_sb, sb);
-
- return rc;
+ duplicate_sb(sb, cache_sb);
+ RETURN(rc);
}
-static int smfs_remount(struct super_block * sb, int * flags, char * data)
+static int smfs_remount(struct super_block *sb, int *flags, char *data)
{
- struct super_block *cache_sb;
- int rc = 0;
+ struct super_block *cache_sb;
+ int rc = 0;
+ ENTRY;
+
+ cache_sb = S2CSB(sb);
- ENTRY;
- cache_sb = S2CSB(sb);
+ if (!cache_sb)
+ RETURN(-EINVAL);
- if (!cache_sb)
- RETURN(-EINVAL);
-
- if (cache_sb->s_op->remount_fs)
- rc = cache_sb->s_op->remount_fs(cache_sb, flags, data);
+ if (cache_sb->s_op->remount_fs)
+ rc = cache_sb->s_op->remount_fs(cache_sb, flags, data);
- duplicate_sb(cache_sb, sb);
- RETURN(rc);
+ duplicate_sb(sb, cache_sb);
+ RETURN(rc);
}
-struct super_operations smfs_super_ops = {
- read_inode: smfs_read_inode,
- clear_inode: smfs_clear_inode,
- put_super: smfs_put_super,
- delete_inode: smfs_delete_inode,
- write_inode: smfs_write_inode,
- dirty_inode: smfs_dirty_inode, /* BKL not held. We take it */
- put_inode: smfs_put_inode, /* BKL not held. Don't need */
-
- write_super: smfs_write_super, /* BKL held */
- write_super_lockfs: smfs_write_super_lockfs, /* BKL not held. Take it */
- unlockfs: smfs_unlockfs, /* BKL not held. We take it */
- statfs: smfs_statfs, /* BKL held */
- remount_fs: smfs_remount, /* BKL held */
+struct super_operations smfs_super_ops = {
+#if (LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0))
+ .read_inode2 = smfs_read_inode2,
+#endif
+ .clear_inode = smfs_clear_inode,
+ .put_super = smfs_put_super,
+ .delete_inode = smfs_delete_inode,
+ .write_inode = smfs_write_inode,
+ .dirty_inode = smfs_dirty_inode, /* BKL not held. */
+ .put_inode = smfs_put_inode, /* BKL not held. */
+ .write_super = smfs_write_super, /* BKL held */
+ .write_super_lockfs = smfs_write_super_lockfs, /* BKL not held. */
+ .unlockfs = smfs_unlockfs, /* BKL not held. */
+ .statfs = smfs_statfs, /* BKL held */
+ .remount_fs = smfs_remount, /* BKL held */
};
-
-
-
-
+int is_smfs_sb(struct super_block *sb)
+{
+ return (sb->s_op->put_super == smfs_super_ops.put_super);
+}
+EXPORT_SYMBOL(is_smfs_sb);