#include <linux/types.h>
#include <linux/slab.h>
#include <linux/highmem.h>
-#include <linux/utsname.h>
#include <linux/init.h>
#include <linux/random.h>
#include <linux/statfs.h>
#include <linux/mount.h>
#include <linux/seq_file.h>
#include <linux/quotaops.h>
+#include <linux/smp_lock.h>
#define MLOG_MASK_PREFIX ML_SUPER
#include <cluster/masklog.h>
#include "ver.h"
#include "xattr.h"
#include "quota.h"
+#include "refcounttree.h"
#include "buffer_head_io.h"
static int ocfs2_check_volume(struct ocfs2_super *osb);
static int ocfs2_verify_volume(struct ocfs2_dinode *di,
struct buffer_head *bh,
- u32 sectsize);
+ u32 sectsize,
+ struct ocfs2_blockcheck_stats *stats);
static int ocfs2_initialize_super(struct super_block *sb,
struct buffer_head *bh,
- int sector_size);
+ int sector_size,
+ struct ocfs2_blockcheck_stats *stats);
static int ocfs2_get_sector(struct super_block *sb,
struct buffer_head **bh,
int block,
int sect_size);
-static void ocfs2_write_super(struct super_block *sb);
static struct inode *ocfs2_alloc_inode(struct super_block *sb);
static void ocfs2_destroy_inode(struct inode *inode);
static int ocfs2_susp_quotas(struct ocfs2_super *osb, int unsuspend);
.clear_inode = ocfs2_clear_inode,
.delete_inode = ocfs2_delete_inode,
.sync_fs = ocfs2_sync_fs,
- .write_super = ocfs2_write_super,
.put_super = ocfs2_put_super,
.remount_fs = ocfs2_remount,
.show_options = ocfs2_show_options,
{Opt_err, NULL}
};
-/*
- * write_super and sync_fs ripped right out of ext3.
- */
-static void ocfs2_write_super(struct super_block *sb)
+#ifdef CONFIG_DEBUG_FS
+static int ocfs2_osb_dump(struct ocfs2_super *osb, char *buf, int len)
{
- if (mutex_trylock(&sb->s_lock) != 0)
- BUG();
- sb->s_dirt = 0;
+ struct ocfs2_cluster_connection *cconn = osb->cconn;
+ struct ocfs2_recovery_map *rm = osb->recovery_map;
+ struct ocfs2_orphan_scan *os = &osb->osb_orphan_scan;
+ int i, out = 0;
+
+ out += snprintf(buf + out, len - out,
+ "%10s => Id: %-s Uuid: %-s Gen: 0x%X Label: %-s\n",
+ "Device", osb->dev_str, osb->uuid_str,
+ osb->fs_generation, osb->vol_label);
+
+ out += snprintf(buf + out, len - out,
+ "%10s => State: %d Flags: 0x%lX\n", "Volume",
+ atomic_read(&osb->vol_state), osb->osb_flags);
+
+ out += snprintf(buf + out, len - out,
+ "%10s => Block: %lu Cluster: %d\n", "Sizes",
+ osb->sb->s_blocksize, osb->s_clustersize);
+
+ out += snprintf(buf + out, len - out,
+ "%10s => Compat: 0x%X Incompat: 0x%X "
+ "ROcompat: 0x%X\n",
+ "Features", osb->s_feature_compat,
+ osb->s_feature_incompat, osb->s_feature_ro_compat);
+
+ out += snprintf(buf + out, len - out,
+ "%10s => Opts: 0x%lX AtimeQuanta: %u\n", "Mount",
+ osb->s_mount_opt, osb->s_atime_quantum);
+
+ if (cconn) {
+ out += snprintf(buf + out, len - out,
+ "%10s => Stack: %s Name: %*s "
+ "Version: %d.%d\n", "Cluster",
+ (*osb->osb_cluster_stack == '\0' ?
+ "o2cb" : osb->osb_cluster_stack),
+ cconn->cc_namelen, cconn->cc_name,
+ cconn->cc_version.pv_major,
+ cconn->cc_version.pv_minor);
+ }
+
+ spin_lock(&osb->dc_task_lock);
+ out += snprintf(buf + out, len - out,
+ "%10s => Pid: %d Count: %lu WakeSeq: %lu "
+ "WorkSeq: %lu\n", "DownCnvt",
+ (osb->dc_task ? task_pid_nr(osb->dc_task) : -1),
+ osb->blocked_lock_count, osb->dc_wake_sequence,
+ osb->dc_work_sequence);
+ spin_unlock(&osb->dc_task_lock);
+
+ spin_lock(&osb->osb_lock);
+ out += snprintf(buf + out, len - out, "%10s => Pid: %d Nodes:",
+ "Recovery",
+ (osb->recovery_thread_task ?
+ task_pid_nr(osb->recovery_thread_task) : -1));
+ if (rm->rm_used == 0)
+ out += snprintf(buf + out, len - out, " None\n");
+ else {
+ for (i = 0; i < rm->rm_used; i++)
+ out += snprintf(buf + out, len - out, " %d",
+ rm->rm_entries[i]);
+ out += snprintf(buf + out, len - out, "\n");
+ }
+ spin_unlock(&osb->osb_lock);
+
+ out += snprintf(buf + out, len - out,
+ "%10s => Pid: %d Interval: %lu Needs: %d\n", "Commit",
+ (osb->commit_task ? task_pid_nr(osb->commit_task) : -1),
+ osb->osb_commit_interval,
+ atomic_read(&osb->needs_checkpoint));
+
+ out += snprintf(buf + out, len - out,
+ "%10s => State: %d TxnId: %lu NumTxns: %d\n",
+ "Journal", osb->journal->j_state,
+ osb->journal->j_trans_id,
+ atomic_read(&osb->journal->j_num_trans));
+
+ out += snprintf(buf + out, len - out,
+ "%10s => GlobalAllocs: %d LocalAllocs: %d "
+ "SubAllocs: %d LAWinMoves: %d SAExtends: %d\n",
+ "Stats",
+ atomic_read(&osb->alloc_stats.bitmap_data),
+ atomic_read(&osb->alloc_stats.local_data),
+ atomic_read(&osb->alloc_stats.bg_allocs),
+ atomic_read(&osb->alloc_stats.moves),
+ atomic_read(&osb->alloc_stats.bg_extends));
+
+ out += snprintf(buf + out, len - out,
+ "%10s => State: %u Descriptor: %llu Size: %u bits "
+ "Default: %u bits\n",
+ "LocalAlloc", osb->local_alloc_state,
+ (unsigned long long)osb->la_last_gd,
+ osb->local_alloc_bits, osb->local_alloc_default_bits);
+
+ spin_lock(&osb->osb_lock);
+ out += snprintf(buf + out, len - out,
+ "%10s => Slot: %d NumStolen: %d\n", "Steal",
+ osb->s_inode_steal_slot,
+ atomic_read(&osb->s_num_inodes_stolen));
+ spin_unlock(&osb->osb_lock);
+
+ out += snprintf(buf + out, len - out, "OrphanScan => ");
+ out += snprintf(buf + out, len - out, "Local: %u Global: %u ",
+ os->os_count, os->os_seqno);
+ out += snprintf(buf + out, len - out, " Last Scan: ");
+ if (atomic_read(&os->os_state) == ORPHAN_SCAN_INACTIVE)
+ out += snprintf(buf + out, len - out, "Disabled\n");
+ else
+ out += snprintf(buf + out, len - out, "%lu seconds ago\n",
+ (get_seconds() - os->os_scantime.tv_sec));
+
+ out += snprintf(buf + out, len - out, "%10s => %3s %10s\n",
+ "Slots", "Num", "RecoGen");
+ for (i = 0; i < osb->max_slots; ++i) {
+ out += snprintf(buf + out, len - out,
+ "%10s %c %3d %10d\n",
+ " ",
+ (i == osb->slot_num ? '*' : ' '),
+ i, osb->slot_recovery_generations[i]);
+ }
+
+ return out;
}
+static int ocfs2_osb_debug_open(struct inode *inode, struct file *file)
+{
+ struct ocfs2_super *osb = inode->i_private;
+ char *buf = NULL;
+
+ buf = kmalloc(PAGE_SIZE, GFP_KERNEL);
+ if (!buf)
+ goto bail;
+
+ i_size_write(inode, ocfs2_osb_dump(osb, buf, PAGE_SIZE));
+
+ file->private_data = buf;
+
+ return 0;
+bail:
+ return -ENOMEM;
+}
+
+static int ocfs2_debug_release(struct inode *inode, struct file *file)
+{
+ kfree(file->private_data);
+ return 0;
+}
+
+static ssize_t ocfs2_debug_read(struct file *file, char __user *buf,
+ size_t nbytes, loff_t *ppos)
+{
+ return simple_read_from_buffer(buf, nbytes, ppos, file->private_data,
+ i_size_read(file->f_mapping->host));
+}
+#else
+static int ocfs2_osb_debug_open(struct inode *inode, struct file *file)
+{
+ return 0;
+}
+static int ocfs2_debug_release(struct inode *inode, struct file *file)
+{
+ return 0;
+}
+static ssize_t ocfs2_debug_read(struct file *file, char __user *buf,
+ size_t nbytes, loff_t *ppos)
+{
+ return 0;
+}
+#endif /* CONFIG_DEBUG_FS */
+
+static const struct file_operations ocfs2_osb_debug_fops = {
+ .open = ocfs2_osb_debug_open,
+ .release = ocfs2_debug_release,
+ .read = ocfs2_debug_read,
+ .llseek = generic_file_llseek,
+};
+
static int ocfs2_sync_fs(struct super_block *sb, int wait)
{
int status;
tid_t target;
struct ocfs2_super *osb = OCFS2_SB(sb);
- sb->s_dirt = 0;
-
if (ocfs2_is_hard_readonly(osb))
return -EROFS;
*/
#if BITS_PER_LONG == 32
-# if defined(CONFIG_LBD)
+# if defined(CONFIG_LBDAF)
BUILD_BUG_ON(sizeof(sector_t) != 8);
/*
* We might be limited by page cache size.
struct mount_options parsed_options;
struct ocfs2_super *osb = OCFS2_SB(sb);
+ lock_kernel();
+
if (!ocfs2_parse_options(sb, data, &parsed_options, 1)) {
ret = -EINVAL;
goto out;
ocfs2_set_journal_params(osb);
}
out:
+ unlock_kernel();
return ret;
}
static int ocfs2_sb_probe(struct super_block *sb,
struct buffer_head **bh,
- int *sector_size)
+ int *sector_size,
+ struct ocfs2_blockcheck_stats *stats)
{
int status, tmpstat;
struct ocfs1_vol_disk_hdr *hdr;
*bh = NULL;
/* may be > 512 */
- *sector_size = bdev_hardsect_size(sb->s_bdev);
+ *sector_size = bdev_logical_block_size(sb->s_bdev);
if (*sector_size > OCFS2_MAX_BLOCKSIZE) {
mlog(ML_ERROR, "Hardware sector size too large: %d (max=%d)\n",
*sector_size, OCFS2_MAX_BLOCKSIZE);
if (tmpstat < 0) {
status = tmpstat;
mlog_errno(status);
- goto bail;
+ break;
}
di = (struct ocfs2_dinode *) (*bh)->b_data;
- status = ocfs2_verify_volume(di, *bh, blksize);
- if (status >= 0)
- goto bail;
- brelse(*bh);
- *bh = NULL;
- if (status != -EAGAIN)
+ memset(stats, 0, sizeof(struct ocfs2_blockcheck_stats));
+ spin_lock_init(&stats->b_lock);
+ tmpstat = ocfs2_verify_volume(di, *bh, blksize, stats);
+ if (tmpstat < 0) {
+ brelse(*bh);
+ *bh = NULL;
+ }
+ if (tmpstat != -EAGAIN) {
+ status = tmpstat;
break;
+ }
}
bail:
return vfs_quota_disable(sb, type, DQUOT_LIMITS_ENABLED);
}
-static struct quotactl_ops ocfs2_quotactl_ops = {
+static const struct quotactl_ops ocfs2_quotactl_ops = {
.quota_on = ocfs2_quota_on,
.quota_off = ocfs2_quota_off,
.quota_sync = vfs_quota_sync,
struct ocfs2_super *osb = NULL;
struct buffer_head *bh = NULL;
char nodestr[8];
+ struct ocfs2_blockcheck_stats stats;
mlog_entry("%p, %p, %i", sb, data, silent);
}
/* probe for superblock */
- status = ocfs2_sb_probe(sb, &bh, §or_size);
+ status = ocfs2_sb_probe(sb, &bh, §or_size, &stats);
if (status < 0) {
mlog(ML_ERROR, "superblock probe failed!\n");
goto read_super_error;
}
- status = ocfs2_initialize_super(sb, bh, sector_size);
+ status = ocfs2_initialize_super(sb, bh, sector_size, &stats);
osb = OCFS2_SB(sb);
if (status < 0) {
mlog_errno(status);
goto read_super_error;
}
+ osb->osb_ctxt = debugfs_create_file("fs_state", S_IFREG|S_IRUSR,
+ osb->osb_debug_root,
+ osb,
+ &ocfs2_osb_debug_fops);
+ if (!osb->osb_ctxt) {
+ status = -EINVAL;
+ mlog_errno(status);
+ goto read_super_error;
+ }
+
+ if (ocfs2_meta_ecc(osb)) {
+ status = ocfs2_blockcheck_stats_debugfs_install(
+ &osb->osb_ecc_stats,
+ osb->osb_debug_root);
+ if (status) {
+ mlog(ML_ERROR,
+ "Unable to create blockcheck statistics "
+ "files\n");
+ goto read_super_error;
+ }
+ }
+
status = ocfs2_mount_volume(sb);
if (osb->root_inode)
inode = igrab(osb->root_inode);
atomic_set(&osb->vol_state, VOLUME_MOUNTED_QUOTAS);
wake_up(&osb->osb_mount_event);
+ /* Start this when the mount is almost sure of being successful */
+ ocfs2_orphan_scan_start(osb);
+
mlog_exit(status);
return status;
mnt);
}
+static void ocfs2_kill_sb(struct super_block *sb)
+{
+ struct ocfs2_super *osb = OCFS2_SB(sb);
+
+ /* Failed mount? */
+ if (!osb || atomic_read(&osb->vol_state) == VOLUME_DISABLED)
+ goto out;
+
+ /* Prevent further queueing of inode drop events */
+ spin_lock(&dentry_list_lock);
+ ocfs2_set_osb_flag(osb, OCFS2_OSB_DROP_DENTRY_LOCK_IMMED);
+ spin_unlock(&dentry_list_lock);
+ /* Wait for work to finish and/or remove it */
+ cancel_work_sync(&osb->dentry_lock_work);
+out:
+ kill_block_super(sb);
+}
+
static struct file_system_type ocfs2_fs_type = {
.owner = THIS_MODULE,
.name = "ocfs2",
.get_sb = ocfs2_get_sb, /* is this called when we mount
* the fs? */
- .kill_sb = kill_block_super, /* set to the generic one
- * right now, but do we
- * need to change that? */
+ .kill_sb = ocfs2_kill_sb,
+
.fs_flags = FS_REQUIRES_DEV|FS_RENAME_DOES_D_MOVE,
.next = NULL
};
{
mlog_entry("(0x%p)\n", sb);
+ lock_kernel();
+
ocfs2_sync_blockdev(sb);
ocfs2_dismount_volume(sb, 0);
+ unlock_kernel();
+
mlog_exit_void();
}
buf->f_bavail = buf->f_bfree;
buf->f_files = numbits;
buf->f_ffree = freebits;
+ buf->f_fsid.val[0] = crc32_le(0, osb->uuid_str, OCFS2_VOL_UUID_LEN)
+ & 0xFFFFFFFFUL;
+ buf->f_fsid.val[1] = crc32_le(0, osb->uuid_str + OCFS2_VOL_UUID_LEN,
+ OCFS2_VOL_UUID_LEN) & 0xFFFFFFFFUL;
brelse(bh);
spin_lock_init(&oi->ip_lock);
ocfs2_extent_map_init(&oi->vfs_inode);
INIT_LIST_HEAD(&oi->ip_io_markers);
- oi->ip_created_trans = 0;
- oi->ip_last_trans = 0;
oi->ip_dir_start_lookup = 0;
init_rwsem(&oi->ip_alloc_sem);
ocfs2_lock_res_init_once(&oi->ip_inode_lockres);
ocfs2_lock_res_init_once(&oi->ip_open_lockres);
- ocfs2_metadata_cache_init(&oi->vfs_inode);
+ ocfs2_metadata_cache_init(INODE_CACHE(&oi->vfs_inode),
+ &ocfs2_inode_caching_ops);
inode_init_once(&oi->vfs_inode);
}
unlock_buffer(*bh);
ll_rw_block(READ, 1, bh);
wait_on_buffer(*bh);
+ if (!buffer_uptodate(*bh)) {
+ mlog_errno(-EIO);
+ brelse(*bh);
+ *bh = NULL;
+ return -EIO;
+ }
+
return 0;
}
}
status = ocfs2_truncate_log_init(osb);
- if (status < 0) {
+ if (status < 0)
mlog_errno(status);
- goto leave;
- }
-
- if (ocfs2_mount_local(osb))
- goto leave;
leave:
if (unlock_super)
osb = OCFS2_SB(sb);
BUG_ON(!osb);
+ debugfs_remove(osb->osb_ctxt);
+
+ /*
+ * Flush inode dropping work queue so that deletes are
+ * performed while the filesystem is still working
+ */
+ ocfs2_drop_all_dl_inodes(osb);
+
+ /* Orphan scan should be stopped as early as possible */
+ ocfs2_orphan_scan_stop(osb);
+
ocfs2_disable_quotas(osb);
ocfs2_shutdown_local_alloc(osb);
ocfs2_sync_blockdev(sb);
+ ocfs2_purge_refcount_trees(osb);
+
/* No cluster connection means we've failed during mount, so skip
* all the steps which depended on that to complete. */
if (osb->cconn) {
if (osb->cconn)
ocfs2_dlm_shutdown(osb, hangup_needed);
+ ocfs2_blockcheck_stats_debugfs_remove(&osb->osb_ecc_stats);
debugfs_remove(osb->osb_debug_root);
if (hangup_needed)
static int ocfs2_initialize_super(struct super_block *sb,
struct buffer_head *bh,
- int sector_size)
+ int sector_size,
+ struct ocfs2_blockcheck_stats *stats)
{
int status;
int i, cbits, bbits;
bbits = le32_to_cpu(di->id2.i_super.s_blocksize_bits);
sb->s_maxbytes = ocfs2_max_file_offset(bbits, cbits);
+ osb->osb_dx_mask = (1 << (cbits - bbits)) - 1;
+
+ for (i = 0; i < 3; i++)
+ osb->osb_dx_seed[i] = le32_to_cpu(di->id2.i_super.s_dx_seed[i]);
+ osb->osb_dx_seed[3] = le32_to_cpu(di->id2.i_super.s_uuid_hash);
+
osb->sb = sb;
/* Save off for ocfs2_rw_direct */
osb->s_sectsize_bits = blksize_bits(sector_size);
INIT_LIST_HEAD(&osb->blocked_lock_list);
osb->blocked_lock_count = 0;
spin_lock_init(&osb->osb_lock);
+ spin_lock_init(&osb->osb_xattr_lock);
ocfs2_init_inode_steal_slot(osb);
atomic_set(&osb->alloc_stats.moves, 0);
atomic_set(&osb->alloc_stats.bg_allocs, 0);
atomic_set(&osb->alloc_stats.bg_extends, 0);
+ /* Copy the blockcheck stats from the superblock probe */
+ osb->osb_ecc_stats = *stats;
+
ocfs2_init_node_maps(osb);
snprintf(osb->dev_str, sizeof(osb->dev_str), "%u,%u",
MAJOR(osb->sb->s_dev), MINOR(osb->sb->s_dev));
+ ocfs2_orphan_scan_init(osb);
+
status = ocfs2_recovery_init(osb);
if (status) {
mlog(ML_ERROR, "Unable to initialize recovery state\n");
goto bail;
}
+ osb->osb_rf_lock_tree = RB_ROOT;
+
osb->s_feature_compat =
le32_to_cpu(OCFS2_RAW_SB(di)->s_feature_compat);
osb->s_feature_ro_compat =
*/
static int ocfs2_verify_volume(struct ocfs2_dinode *di,
struct buffer_head *bh,
- u32 blksz)
+ u32 blksz,
+ struct ocfs2_blockcheck_stats *stats)
{
int status = -EAGAIN;
OCFS2_FEATURE_INCOMPAT_META_ECC) {
status = ocfs2_block_check_validate(bh->b_data,
bh->b_size,
- &di->i_check);
+ &di->i_check,
+ stats);
if (status)
goto out;
}
* lock, and it's marked as dirty, set the bit in the recover
* map and launch a recovery thread for it. */
status = ocfs2_mark_dead_nodes(osb);
+ if (status < 0) {
+ mlog_errno(status);
+ goto finally;
+ }
+
+ status = ocfs2_compute_replay_slots(osb);
if (status < 0)
mlog_errno(status);
/* Force a panic(). This stinks, but it's better than letting
* things continue without having a proper hard readonly
* here. */
- OCFS2_SB(sb)->s_mount_opt |= OCFS2_MOUNT_ERRORS_PANIC;
+ if (!ocfs2_mount_local(OCFS2_SB(sb)))
+ OCFS2_SB(sb)->s_mount_opt |= OCFS2_MOUNT_ERRORS_PANIC;
ocfs2_handle_error(sb);
}
git://ftp.safe.ca / safe / jmp / linux-2.6 / blobdiff