static int ocfs2_info_scan_inode_alloc(struct ocfs2_super *osb,
struct inode *inode_alloc, u64 blkno,
struct ocfs2_info_freeinode *fi,
u32 slot)
{
int status = 0, unlock = 0;
struct buffer_head *bh = NULL;
struct ocfs2_dinode *dinode_alloc = NULL;
if (inode_alloc)
mutex_lock(&inode_alloc->i_mutex);
if (o2info_coherent(&fi->ifi_req)) {
status = ocfs2_inode_lock(inode_alloc, &bh, 0);
if (status < 0) {
mlog_errno(status);
goto bail;
}
unlock = 1;
} else {
status = ocfs2_read_blocks_sync(osb, blkno, 1, &bh);
if (status < 0) {
mlog_errno(status);
goto bail;
}
}
dinode_alloc = (struct ocfs2_dinode *)bh->b_data;
fi->ifi_stat[slot].lfi_total =
le32_to_cpu(dinode_alloc->id1.bitmap1.i_total);
fi->ifi_stat[slot].lfi_free =
le32_to_cpu(dinode_alloc->id1.bitmap1.i_total) -
le32_to_cpu(dinode_alloc->id1.bitmap1.i_used);
bail:
if (unlock)
ocfs2_inode_unlock(inode_alloc, 0);
if (inode_alloc)
mutex_unlock(&inode_alloc->i_mutex);
brelse(bh);
return status;
}
static void ocfs2_update_super_and_backups(struct inode *inode,
int new_clusters)
{
int ret;
u32 clusters = 0;
struct buffer_head *super_bh = NULL;
struct ocfs2_dinode *super_di = NULL;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
/*
* update the superblock last.
* It doesn't matter if the write failed.
*/
ret = ocfs2_read_blocks_sync(osb, OCFS2_SUPER_BLOCK_BLKNO, 1,
&super_bh);
if (ret < 0) {
mlog_errno(ret);
goto out;
}
super_di = (struct ocfs2_dinode *)super_bh->b_data;
le32_add_cpu(&super_di->i_clusters, new_clusters);
clusters = le32_to_cpu(super_di->i_clusters);
ret = ocfs2_write_super_or_backup(osb, super_bh);
if (ret < 0) {
mlog_errno(ret);
goto out;
}
if (OCFS2_HAS_COMPAT_FEATURE(osb->sb, OCFS2_FEATURE_COMPAT_BACKUP_SB))
ret = update_backups(inode, clusters, super_bh->b_data);
out:
brelse(super_bh);
if (ret)
printk(KERN_WARNING "ocfs2: Failed to update super blocks on %s"
" during fs resize. This condition is not fatal,"
" but fsck.ocfs2 should be run to fix it\n",
osb->dev_str);
return;
}
static int update_backups(struct inode * inode, u32 clusters, char *data)
{
int i, ret = 0;
u32 cluster;
u64 blkno;
struct buffer_head *backup = NULL;
struct ocfs2_dinode *backup_di = NULL;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
/* calculate the real backups we need to update. */
for (i = 0; i < OCFS2_MAX_BACKUP_SUPERBLOCKS; i++) {
blkno = ocfs2_backup_super_blkno(inode->i_sb, i);
cluster = ocfs2_blocks_to_clusters(inode->i_sb, blkno);
if (cluster > clusters)
break;
ret = ocfs2_read_blocks_sync(osb, blkno, 1, &backup);
if (ret < 0) {
mlog_errno(ret);
break;
}
memcpy(backup->b_data, data, inode->i_sb->s_blocksize);
backup_di = (struct ocfs2_dinode *)backup->b_data;
backup_di->i_blkno = cpu_to_le64(blkno);
ret = ocfs2_write_super_or_backup(osb, backup);
brelse(backup);
backup = NULL;
if (ret < 0) {
mlog_errno(ret);
break;
}
}
return ret;
}
/* Add a new group descriptor to global_bitmap. */
int ocfs2_group_add(struct inode *inode, struct ocfs2_new_group_input *input)
{
int ret;
handle_t *handle;
struct buffer_head *main_bm_bh = NULL;
struct inode *main_bm_inode = NULL;
struct ocfs2_dinode *fe = NULL;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
struct buffer_head *group_bh = NULL;
struct ocfs2_group_desc *group = NULL;
struct ocfs2_chain_list *cl;
struct ocfs2_chain_rec *cr;
u16 cl_bpc;
if (ocfs2_is_hard_readonly(osb) || ocfs2_is_soft_readonly(osb))
return -EROFS;
main_bm_inode = ocfs2_get_system_file_inode(osb,
GLOBAL_BITMAP_SYSTEM_INODE,
OCFS2_INVALID_SLOT);
if (!main_bm_inode) {
ret = -EINVAL;
mlog_errno(ret);
goto out;
}
mutex_lock(&main_bm_inode->i_mutex);
ret = ocfs2_inode_lock(main_bm_inode, &main_bm_bh, 1);
if (ret < 0) {
mlog_errno(ret);
goto out_mutex;
}
fe = (struct ocfs2_dinode *)main_bm_bh->b_data;
if (le16_to_cpu(fe->id2.i_chain.cl_cpg) !=
ocfs2_group_bitmap_size(osb->sb, 0,
osb->s_feature_incompat) * 8) {
mlog(ML_ERROR, "The disk is too old and small."
" Force to do offline resize.");
ret = -EINVAL;
goto out_unlock;
}
ret = ocfs2_read_blocks_sync(osb, input->group, 1, &group_bh);
if (ret < 0) {
mlog(ML_ERROR, "Can't read the group descriptor # %llu "
"from the device.", (unsigned long long)input->group);
goto out_unlock;
}
ocfs2_set_new_buffer_uptodate(INODE_CACHE(inode), group_bh);
ret = ocfs2_verify_group_and_input(main_bm_inode, fe, input, group_bh);
if (ret) {
mlog_errno(ret);
goto out_unlock;
}
trace_ocfs2_group_add((unsigned long long)input->group,
input->chain, input->clusters, input->frees);
handle = ocfs2_start_trans(osb, OCFS2_GROUP_ADD_CREDITS);
if (IS_ERR(handle)) {
mlog_errno(PTR_ERR(handle));
ret = -EINVAL;
goto out_unlock;
}
cl_bpc = le16_to_cpu(fe->id2.i_chain.cl_bpc);
cl = &fe->id2.i_chain;
cr = &cl->cl_recs[input->chain];
ret = ocfs2_journal_access_gd(handle, INODE_CACHE(main_bm_inode),
group_bh, OCFS2_JOURNAL_ACCESS_WRITE);
if (ret < 0) {
mlog_errno(ret);
goto out_commit;
}
group = (struct ocfs2_group_desc *)group_bh->b_data;
group->bg_next_group = cr->c_blkno;
ocfs2_journal_dirty(handle, group_bh);
ret = ocfs2_journal_access_di(handle, INODE_CACHE(main_bm_inode),
main_bm_bh, OCFS2_JOURNAL_ACCESS_WRITE);
if (ret < 0) {
mlog_errno(ret);
goto out_commit;
}
if (input->chain == le16_to_cpu(cl->cl_next_free_rec)) {
le16_add_cpu(&cl->cl_next_free_rec, 1);
memset(cr, 0, sizeof(struct ocfs2_chain_rec));
}
cr->c_blkno = cpu_to_le64(input->group);
le32_add_cpu(&cr->c_total, input->clusters * cl_bpc);
le32_add_cpu(&cr->c_free, input->frees * cl_bpc);
le32_add_cpu(&fe->id1.bitmap1.i_total, input->clusters *cl_bpc);
le32_add_cpu(&fe->id1.bitmap1.i_used,
(input->clusters - input->frees) * cl_bpc);
le32_add_cpu(&fe->i_clusters, input->clusters);
ocfs2_journal_dirty(handle, main_bm_bh);
spin_lock(&OCFS2_I(main_bm_inode)->ip_lock);
OCFS2_I(main_bm_inode)->ip_clusters = le32_to_cpu(fe->i_clusters);
le64_add_cpu(&fe->i_size, input->clusters << osb->s_clustersize_bits);
spin_unlock(&OCFS2_I(main_bm_inode)->ip_lock);
i_size_write(main_bm_inode, le64_to_cpu(fe->i_size));
ocfs2_update_super_and_backups(main_bm_inode, input->clusters);
out_commit:
ocfs2_commit_trans(osb, handle);
out_unlock:
brelse(group_bh);
brelse(main_bm_bh);
ocfs2_inode_unlock(main_bm_inode, 1);
out_mutex:
mutex_unlock(&main_bm_inode->i_mutex);
iput(main_bm_inode);
out:
return ret;
}
static int ocfs2_info_freefrag_scan_bitmap(struct ocfs2_super *osb,
struct inode *gb_inode, u64 blkno,
struct ocfs2_info_freefrag *ffg)
{
u32 chunks_in_group;
int status = 0, unlock = 0, i;
struct buffer_head *bh = NULL;
struct ocfs2_chain_list *cl = NULL;
struct ocfs2_chain_rec *rec = NULL;
struct ocfs2_dinode *gb_dinode = NULL;
if (gb_inode)
mutex_lock(&gb_inode->i_mutex);
if (o2info_coherent(&ffg->iff_req)) {
status = ocfs2_inode_lock(gb_inode, &bh, 0);
if (status < 0) {
mlog_errno(status);
goto bail;
}
unlock = 1;
} else {
status = ocfs2_read_blocks_sync(osb, blkno, 1, &bh);
if (status < 0) {
mlog_errno(status);
goto bail;
}
}
gb_dinode = (struct ocfs2_dinode *)bh->b_data;
cl = &(gb_dinode->id2.i_chain);
/*
* Chunksize(in) clusters from userspace should be
* less than clusters in a group.
*/
if (ffg->iff_chunksize > le16_to_cpu(cl->cl_cpg)) {
status = -EINVAL;
goto bail;
}
memset(&ffg->iff_ffs, 0, sizeof(struct ocfs2_info_freefrag_stats));
ffg->iff_ffs.ffs_min = ~0U;
ffg->iff_ffs.ffs_clusters =
le32_to_cpu(gb_dinode->id1.bitmap1.i_total);
ffg->iff_ffs.ffs_free_clusters = ffg->iff_ffs.ffs_clusters -
le32_to_cpu(gb_dinode->id1.bitmap1.i_used);
chunks_in_group = le16_to_cpu(cl->cl_cpg) / ffg->iff_chunksize + 1;
for (i = 0; i < le16_to_cpu(cl->cl_next_free_rec); i++) {
rec = &(cl->cl_recs[i]);
status = ocfs2_info_freefrag_scan_chain(osb, gb_inode,
gb_dinode,
rec, ffg,
chunks_in_group);
if (status)
goto bail;
}
if (ffg->iff_ffs.ffs_free_chunks_real)
ffg->iff_ffs.ffs_avg = (ffg->iff_ffs.ffs_avg /
ffg->iff_ffs.ffs_free_chunks_real);
bail:
if (unlock)
ocfs2_inode_unlock(gb_inode, 0);
if (gb_inode)
mutex_unlock(&gb_inode->i_mutex);
if (gb_inode)
iput(gb_inode);
brelse(bh);
return status;
}
static int ocfs2_info_freefrag_scan_chain(struct ocfs2_super *osb,
struct inode *gb_inode,
struct ocfs2_dinode *gb_dinode,
struct ocfs2_chain_rec *rec,
struct ocfs2_info_freefrag *ffg,
u32 chunks_in_group)
{
int status = 0, used;
u64 blkno;
struct buffer_head *bh = NULL;
struct ocfs2_group_desc *bg = NULL;
unsigned int max_bits, num_clusters;
unsigned int offset = 0, cluster, chunk;
unsigned int chunk_free, last_chunksize = 0;
if (!le32_to_cpu(rec->c_free))
goto bail;
do {
if (!bg)
blkno = le64_to_cpu(rec->c_blkno);
else
blkno = le64_to_cpu(bg->bg_next_group);
if (bh) {
brelse(bh);
bh = NULL;
}
if (o2info_coherent(&ffg->iff_req))
status = ocfs2_read_group_descriptor(gb_inode,
gb_dinode,
blkno, &bh);
else
status = ocfs2_read_blocks_sync(osb, blkno, 1, &bh);
if (status < 0) {
mlog(ML_ERROR, "Can't read the group descriptor # "
"%llu from device.", (unsigned long long)blkno);
status = -EIO;
goto bail;
}
bg = (struct ocfs2_group_desc *)bh->b_data;
if (!le16_to_cpu(bg->bg_free_bits_count))
continue;
max_bits = le16_to_cpu(bg->bg_bits);
offset = 0;
for (chunk = 0; chunk < chunks_in_group; chunk++) {
/*
* last chunk may be not an entire one.
*/
if ((offset + ffg->iff_chunksize) > max_bits)
num_clusters = max_bits - offset;
else
num_clusters = ffg->iff_chunksize;
chunk_free = 0;
for (cluster = 0; cluster < num_clusters; cluster++) {
used = ocfs2_test_bit(offset,
(unsigned long *)bg->bg_bitmap);
/*
* - chunk_free counts free clusters in #N chunk.
* - last_chunksize records the size(in) clusters
* for the last real free chunk being counted.
*/
if (!used) {
last_chunksize++;
chunk_free++;
}
if (used && last_chunksize) {
ocfs2_info_update_ffg(ffg,
last_chunksize);
last_chunksize = 0;
}
offset++;
}
if (chunk_free == ffg->iff_chunksize)
ffg->iff_ffs.ffs_free_chunks++;
}
/*
* need to update the info for last free chunk.
*/
if (last_chunksize)
ocfs2_info_update_ffg(ffg, last_chunksize);
} while (le64_to_cpu(bg->bg_next_group));
bail:
brelse(bh);
return status;
}
static int ocfs2_read_locked_inode(struct inode *inode,
struct ocfs2_find_inode_args *args)
{
struct super_block *sb;
struct ocfs2_super *osb;
struct ocfs2_dinode *fe;
struct buffer_head *bh = NULL;
int status, can_lock;
u32 generation = 0;
status = -EINVAL;
if (inode == NULL || inode->i_sb == NULL) {
mlog(ML_ERROR, "bad inode\n");
return status;
}
sb = inode->i_sb;
osb = OCFS2_SB(sb);
if (!args) {
mlog(ML_ERROR, "bad inode args\n");
make_bad_inode(inode);
return status;
}
/*
* To improve performance of cold-cache inode stats, we take
* the cluster lock here if possible.
*
* Generally, OCFS2 never trusts the contents of an inode
* unless it's holding a cluster lock, so taking it here isn't
* a correctness issue as much as it is a performance
* improvement.
*
* There are three times when taking the lock is not a good idea:
*
* 1) During startup, before we have initialized the DLM.
*
* 2) If we are reading certain system files which never get
* cluster locks (local alloc, truncate log).
*
* 3) If the process doing the iget() is responsible for
* orphan dir recovery. We're holding the orphan dir lock and
* can get into a deadlock with another process on another
* node in ->delete_inode().
*
* #1 and #2 can be simply solved by never taking the lock
* here for system files (which are the only type we read
* during mount). It's a heavier approach, but our main
* concern is user-accessible files anyway.
*
* #3 works itself out because we'll eventually take the
* cluster lock before trusting anything anyway.
*/
can_lock = !(args->fi_flags & OCFS2_FI_FLAG_SYSFILE)
&& !(args->fi_flags & OCFS2_FI_FLAG_ORPHAN_RECOVERY)
&& !ocfs2_mount_local(osb);
trace_ocfs2_read_locked_inode(
(unsigned long long)OCFS2_I(inode)->ip_blkno, can_lock);
/*
* To maintain backwards compatibility with older versions of
* ocfs2-tools, we still store the generation value for system
* files. The only ones that actually matter to userspace are
* the journals, but it's easier and inexpensive to just flag
* all system files similarly.
*/
if (args->fi_flags & OCFS2_FI_FLAG_SYSFILE)
generation = osb->fs_generation;
ocfs2_inode_lock_res_init(&OCFS2_I(inode)->ip_inode_lockres,
OCFS2_LOCK_TYPE_META,
generation, inode);
ocfs2_inode_lock_res_init(&OCFS2_I(inode)->ip_open_lockres,
OCFS2_LOCK_TYPE_OPEN,
0, inode);
if (can_lock) {
status = ocfs2_open_lock(inode);
if (status) {
make_bad_inode(inode);
mlog_errno(status);
return status;
}
status = ocfs2_inode_lock(inode, NULL, 0);
if (status) {
make_bad_inode(inode);
mlog_errno(status);
return status;
}
}
if (args->fi_flags & OCFS2_FI_FLAG_ORPHAN_RECOVERY) {
status = ocfs2_try_open_lock(inode, 0);
if (status) {
make_bad_inode(inode);
return status;
}
}
if (can_lock) {
status = ocfs2_read_inode_block_full(inode, &bh,
OCFS2_BH_IGNORE_CACHE);
} else {
status = ocfs2_read_blocks_sync(osb, args->fi_blkno, 1, &bh);
/*
* If buffer is in jbd, then its checksum may not have been
* computed as yet.
*/
if (!status && !buffer_jbd(bh))
status = ocfs2_validate_inode_block(osb->sb, bh);
}
if (status < 0) {
mlog_errno(status);
goto bail;
}
status = -EINVAL;
fe = (struct ocfs2_dinode *) bh->b_data;
/*
* This is a code bug. Right now the caller needs to
* understand whether it is asking for a system file inode or
* not so the proper lock names can be built.
*/
mlog_bug_on_msg(!!(fe->i_flags & cpu_to_le32(OCFS2_SYSTEM_FL)) !=
!!(args->fi_flags & OCFS2_FI_FLAG_SYSFILE),
"Inode %llu: system file state is ambigous\n",
(unsigned long long)args->fi_blkno);
if (S_ISCHR(le16_to_cpu(fe->i_mode)) ||
S_ISBLK(le16_to_cpu(fe->i_mode)))
inode->i_rdev = huge_decode_dev(le64_to_cpu(fe->id1.dev1.i_rdev));
ocfs2_populate_inode(inode, fe, 0);
BUG_ON(args->fi_blkno != le64_to_cpu(fe->i_blkno));
status = 0;
bail:
if (can_lock)
ocfs2_inode_unlock(inode, 0);
if (status < 0)
make_bad_inode(inode);
if (args && bh)
brelse(bh);
return status;
}
/*
* find the victim alloc group, where #blkno fits.
*/
static int ocfs2_find_victim_alloc_group(struct inode *inode,
u64 vict_blkno,
int type, int slot,
int *vict_bit,
struct buffer_head **ret_bh)
{
int ret, i, bits_per_unit = 0;
u64 blkno;
char namebuf[40];
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
struct buffer_head *ac_bh = NULL, *gd_bh = NULL;
struct ocfs2_chain_list *cl;
struct ocfs2_chain_rec *rec;
struct ocfs2_dinode *ac_dinode;
struct ocfs2_group_desc *bg;
ocfs2_sprintf_system_inode_name(namebuf, sizeof(namebuf), type, slot);
ret = ocfs2_lookup_ino_from_name(osb->sys_root_inode, namebuf,
strlen(namebuf), &blkno);
if (ret) {
ret = -ENOENT;
goto out;
}
ret = ocfs2_read_blocks_sync(osb, blkno, 1, &ac_bh);
if (ret) {
mlog_errno(ret);
goto out;
}
ac_dinode = (struct ocfs2_dinode *)ac_bh->b_data;
cl = &(ac_dinode->id2.i_chain);
rec = &(cl->cl_recs[0]);
if (type == GLOBAL_BITMAP_SYSTEM_INODE)
bits_per_unit = osb->s_clustersize_bits -
inode->i_sb->s_blocksize_bits;
/*
* 'vict_blkno' was out of the valid range.
*/
if ((vict_blkno < le64_to_cpu(rec->c_blkno)) ||
(vict_blkno >= ((u64)le32_to_cpu(ac_dinode->id1.bitmap1.i_total) <<
bits_per_unit))) {
ret = -EINVAL;
goto out;
}
for (i = 0; i < le16_to_cpu(cl->cl_next_free_rec); i++) {
rec = &(cl->cl_recs[i]);
if (!rec)
continue;
bg = NULL;
do {
if (!bg)
blkno = le64_to_cpu(rec->c_blkno);
else
blkno = le64_to_cpu(bg->bg_next_group);
if (gd_bh) {
brelse(gd_bh);
gd_bh = NULL;
}
ret = ocfs2_read_blocks_sync(osb, blkno, 1, &gd_bh);
if (ret) {
mlog_errno(ret);
goto out;
}
bg = (struct ocfs2_group_desc *)gd_bh->b_data;
if (vict_blkno < (le64_to_cpu(bg->bg_blkno) +
le16_to_cpu(bg->bg_bits))) {
*ret_bh = gd_bh;
*vict_bit = (vict_blkno - blkno) >>
bits_per_unit;
mlog(0, "find the victim group: #%llu, "
"total_bits: %u, vict_bit: %u\n",
blkno, le16_to_cpu(bg->bg_bits),
*vict_bit);
goto out;
}
} while (le64_to_cpu(bg->bg_next_group));
}
ret = -EINVAL;
out:
brelse(ac_bh);
/*
* caller has to release the gd_bh properly.
*/
return ret;
}
