int ubifs_leb_write_cb(struct ubifs_info *c, int lnum, const void *buf, int offs,
int len, int dtype)
{
int err;
ubifs_assert(!c->ro_media && !c->ro_mount);
if (c->ro_error)
return -EROFS;
if (!dbg_is_tst_rcvry(c))
err = ubi_leb_write_cb(c->ubi, lnum, buf, offs, len, dtype);
else
err = dbg_leb_write_cb(c, lnum, buf, offs, len, dtype);
if (err) {
ubifs_err("writing %d bytes to LEB %d:%d failed, error %d",
len, lnum, offs, err);
ubifs_ro_mode(c, err);
dbg_dump_stack();
}
return err;
}
/**
* ubifs_load_znode - load znode to TNC cache.
* @c: UBIFS file-system description object
* @zbr: znode branch
* @parent: znode's parent
* @iip: index in parent
*
* This function loads znode pointed to by @zbr into the TNC cache and
* returns pointer to it in case of success and a negative error code in case
* of failure.
*/
struct ubifs_znode *ubifs_load_znode(struct ubifs_info *c,
struct ubifs_zbranch *zbr,
struct ubifs_znode *parent, int iip)
{
int err;
struct ubifs_znode *znode;
ubifs_assert(!zbr->znode);
/*
* A slab cache is not presently used for znodes because the znode size
* depends on the fanout which is stored in the superblock.
*/
znode = kzalloc(c->max_znode_sz, GFP_NOFS);
if (!znode)
return ERR_PTR(-ENOMEM);
err = read_znode(c, zbr->lnum, zbr->offs, zbr->len, znode);
if (err)
goto out;
atomic_long_inc(&c->clean_zn_cnt);
/*
* Increment the global clean znode counter as well. It is OK that
* global and per-FS clean znode counters may be inconsistent for some
* short time (because we might be preempted at this point), the global
* one is only used in shrinker.
*/
atomic_long_inc(&ubifs_clean_zn_cnt);
zbr->znode = znode;
znode->parent = parent;
znode->time = get_seconds();
znode->iip = iip;
return znode;
out:
kfree(znode);
return ERR_PTR(err);
}
/**
* ubifs_wbuf_init - initialize write-buffer.
* @c: UBIFS file-system description object
* @wbuf: write-buffer to initialize
*
* This function initializes write-buffer. Returns zero in case of success
* %-ENOMEM in case of failure.
*/
int ubifs_wbuf_init(struct ubifs_info *c, struct ubifs_wbuf *wbuf)
{
size_t size;
wbuf->buf = kmalloc(c->max_write_size, GFP_KERNEL);
if (!wbuf->buf)
return -ENOMEM;
size = (c->max_write_size / UBIFS_CH_SZ + 1) * sizeof(ino_t);
wbuf->inodes = kmalloc(size, GFP_KERNEL);
if (!wbuf->inodes) {
kfree(wbuf->buf);
wbuf->buf = NULL;
return -ENOMEM;
}
wbuf->used = 0;
wbuf->lnum = wbuf->offs = -1;
/*
* If the LEB starts at the max. write size aligned address, then
* write-buffer size has to be set to @c->max_write_size. Otherwise,
* set it to something smaller so that it ends at the closest max.
* write size boundary.
*/
size = c->max_write_size - (c->leb_start % c->max_write_size);
wbuf->avail = wbuf->size = size;
wbuf->dtype = UBI_UNKNOWN;
wbuf->sync_callback = NULL;
mutex_init(&wbuf->io_mutex);
spin_lock_init(&wbuf->lock);
wbuf->c = c;
wbuf->next_ino = 0;
hrtimer_init(&wbuf->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
wbuf->timer.function = wbuf_timer_callback_nolock;
wbuf->softlimit = ktime_set(WBUF_TIMEOUT_SOFTLIMIT, 0);
wbuf->delta = WBUF_TIMEOUT_HARDLIMIT - WBUF_TIMEOUT_SOFTLIMIT;
wbuf->delta *= 1000000000ULL;
ubifs_assert(wbuf->delta <= ULONG_MAX);
return 0;
}
/**
* ubifs_prepare_node - prepare node to be written to flash.
* @c: UBIFS file-system description object
* @node: the node to pad
* @len: node length
* @pad: if the buffer has to be padded
*
* This function prepares node at @node to be written to the media - it
* calculates node CRC, fills the common header, and adds proper padding up to
* the next minimum I/O unit if @pad is not zero.
*/
void ubifs_prepare_node(struct ubifs_info *c, void *node, int len, int pad)
{
uint32_t crc;
struct ubifs_ch *ch = node;
unsigned long long sqnum = next_sqnum(c);
ubifs_assert(len >= UBIFS_CH_SZ);
ch->magic = cpu_to_le32(UBIFS_NODE_MAGIC);
ch->len = cpu_to_le32(len);
ch->group_type = UBIFS_NO_NODE_GROUP;
ch->sqnum = cpu_to_le64(sqnum);
ch->padding[0] = ch->padding[1] = 0;
crc = crc32(UBIFS_CRC32_INIT, node + 8, len - 8);
ch->crc = cpu_to_le32(crc);
if (pad) {
len = ALIGN(len, 8);
pad = ALIGN(len, c->min_io_size) - len;
ubifs_pad(c, node + len, pad);
}
}
int ubifs_decrypt(const struct inode *inode, struct ubifs_data_node *dn,
unsigned int *out_len, int block)
{
struct ubifs_info *c = inode->i_sb->s_fs_info;
int err;
unsigned int clen = le16_to_cpu(dn->compr_size);
unsigned int dlen = *out_len;
if (clen <= 0 || clen > UBIFS_BLOCK_SIZE || clen > dlen) {
ubifs_err(c, "bad compr_size: %i", clen);
return -EINVAL;
}
ubifs_assert(dlen <= UBIFS_BLOCK_SIZE);
err = fscrypt_decrypt_page(inode, virt_to_page(&dn->data), dlen,
offset_in_page(&dn->data), block);
if (err) {
ubifs_err(c, "fscrypt_decrypt_page failed: %i", err);
return err;
}
*out_len = clen;
return 0;
}
/**
* ubifs_sync_wbufs_by_inodes - synchronize write-buffers which have data.
* belonging to specified inodes.
* @c: UBIFS file-system description object
* @inodes: array of inodes
* @count: number of elements in @inodes
*
* This function synchronizes write-buffers which contain nodes belonging to
* any inode specified in @inodes array. Returns zero in case of success and a
* negative error code in case of failure.
*/
int ubifs_sync_wbufs_by_inodes(struct ubifs_info *c,
struct inode * const *inodes, int count)
{
int i, j, err = 0;
ubifs_assert(count);
for (i = 0; i < c->jhead_cnt; i++) {
struct ubifs_wbuf *wbuf = &c->jheads[i].wbuf;
if (i == GCHD)
/*
* GC head is special, do not look at it. Even if the
* head contains something related to this inode, it is
* a _copy_ of corresponding on-flash node which sits
* somewhere else.
*/
continue;
for (j = 0; j < count && !err; j++)
if (wbuf_has_ino(wbuf, inodes[j]->i_ino)) {
mutex_lock_nested(&wbuf->io_mutex, wbuf->jhead);
if (wbuf_has_ino(wbuf, inodes[j]->i_ino))
err = ubifs_wbuf_sync_nolock(wbuf);
mutex_unlock(&wbuf->io_mutex);
break;
}
if (err) {
ubifs_ro_mode(c, err);
break;
}
}
return err;
}
static int ubifs_create(struct inode *dir, struct dentry *dentry, umode_t mode,
bool excl)
{
struct inode *inode;
struct ubifs_info *c = dir->i_sb->s_fs_info;
int err, sz_change = CALC_DENT_SIZE(dentry->d_name.len);
struct ubifs_budget_req req = { .new_ino = 1, .new_dent = 1,
.dirtied_ino = 1 };
struct ubifs_inode *dir_ui = ubifs_inode(dir);
/*
* Budget request settings: new inode, new direntry, changing the
* parent directory inode.
*/
dbg_gen("dent '%pd', mode %#hx in dir ino %lu",
dentry, mode, dir->i_ino);
err = ubifs_budget_space(c, &req);
if (err)
return err;
inode = ubifs_new_inode(c, dir, mode);
if (IS_ERR(inode)) {
err = PTR_ERR(inode);
goto out_budg;
}
err = ubifs_init_security(dir, inode, &dentry->d_name);
if (err)
goto out_inode;
mutex_lock(&dir_ui->ui_mutex);
dir->i_size += sz_change;
dir_ui->ui_size = dir->i_size;
dir->i_mtime = dir->i_ctime = inode->i_ctime;
err = ubifs_jnl_update(c, dir, &dentry->d_name, inode, 0, 0);
if (err)
goto out_cancel;
mutex_unlock(&dir_ui->ui_mutex);
ubifs_release_budget(c, &req);
insert_inode_hash(inode);
d_instantiate(dentry, inode);
return 0;
out_cancel:
dir->i_size -= sz_change;
dir_ui->ui_size = dir->i_size;
mutex_unlock(&dir_ui->ui_mutex);
out_inode:
make_bad_inode(inode);
iput(inode);
out_budg:
ubifs_release_budget(c, &req);
ubifs_err(c, "cannot create regular file, error %d", err);
return err;
}
/**
* vfs_dent_type - get VFS directory entry type.
* @type: UBIFS directory entry type
*
* This function converts UBIFS directory entry type into VFS directory entry
* type.
*/
static unsigned int vfs_dent_type(uint8_t type)
{
switch (type) {
case UBIFS_ITYPE_REG:
return DT_REG;
case UBIFS_ITYPE_DIR:
return DT_DIR;
case UBIFS_ITYPE_LNK:
return DT_LNK;
case UBIFS_ITYPE_BLK:
return DT_BLK;
case UBIFS_ITYPE_CHR:
return DT_CHR;
case UBIFS_ITYPE_FIFO:
return DT_FIFO;
case UBIFS_ITYPE_SOCK:
return DT_SOCK;
default:
BUG();
}
return 0;
}
/*
* The classical Unix view for directory is that it is a linear array of
* (name, inode number) entries. Linux/VFS assumes this model as well.
* Particularly, 'readdir()' call wants us to return a directory entry offset
* which later may be used to continue 'readdir()'ing the directory or to
* 'seek()' to that specific direntry. Obviously UBIFS does not really fit this
* model because directory entries are identified by keys, which may collide.
*
* UBIFS uses directory entry hash value for directory offsets, so
* 'seekdir()'/'telldir()' may not always work because of possible key
* collisions. But UBIFS guarantees that consecutive 'readdir()' calls work
* properly by means of saving full directory entry name in the private field
* of the file description object.
*
* This means that UBIFS cannot support NFS which requires full
* 'seekdir()'/'telldir()' support.
*/
static int ubifs_readdir(struct file *file, struct dir_context *ctx)
{
int err;
struct qstr nm;
union ubifs_key key;
struct ubifs_dent_node *dent;
struct inode *dir = file_inode(file);
struct ubifs_info *c = dir->i_sb->s_fs_info;
dbg_gen("dir ino %lu, f_pos %#llx", dir->i_ino, ctx->pos);
if (ctx->pos > UBIFS_S_KEY_HASH_MASK || ctx->pos == 2)
/*
* The directory was seek'ed to a senseless position or there
* are no more entries.
*/
return 0;
if (file->f_version == 0) {
/*
* The file was seek'ed, which means that @file->private_data
* is now invalid. This may also be just the first
* 'ubifs_readdir()' invocation, in which case
* @file->private_data is NULL, and the below code is
* basically a no-op.
*/
kfree(file->private_data);
file->private_data = NULL;
}
/*
* 'generic_file_llseek()' unconditionally sets @file->f_version to
* zero, and we use this for detecting whether the file was seek'ed.
*/
file->f_version = 1;
/* File positions 0 and 1 correspond to "." and ".." */
if (ctx->pos < 2) {
ubifs_assert(!file->private_data);
if (!dir_emit_dots(file, ctx))
return 0;
/* Find the first entry in TNC and save it */
lowest_dent_key(c, &key, dir->i_ino);
nm.name = NULL;
dent = ubifs_tnc_next_ent(c, &key, &nm);
if (IS_ERR(dent)) {
err = PTR_ERR(dent);
goto out;
}
ctx->pos = key_hash_flash(c, &dent->key);
file->private_data = dent;
}
dent = file->private_data;
if (!dent) {
/*
* The directory was seek'ed to and is now readdir'ed.
* Find the entry corresponding to @ctx->pos or the closest one.
*/
dent_key_init_hash(c, &key, dir->i_ino, ctx->pos);
nm.name = NULL;
dent = ubifs_tnc_next_ent(c, &key, &nm);
if (IS_ERR(dent)) {
err = PTR_ERR(dent);
goto out;
}
ctx->pos = key_hash_flash(c, &dent->key);
file->private_data = dent;
}
while (1) {
dbg_gen("feed '%s', ino %llu, new f_pos %#x",
dent->name, (unsigned long long)le64_to_cpu(dent->inum),
key_hash_flash(c, &dent->key));
ubifs_assert(le64_to_cpu(dent->ch.sqnum) >
ubifs_inode(dir)->creat_sqnum);
nm.len = le16_to_cpu(dent->nlen);
if (!dir_emit(ctx, dent->name, nm.len,
le64_to_cpu(dent->inum),
vfs_dent_type(dent->type)))
return 0;
/* Switch to the next entry */
key_read(c, &dent->key, &key);
nm.name = dent->name;
dent = ubifs_tnc_next_ent(c, &key, &nm);
if (IS_ERR(dent)) {
err = PTR_ERR(dent);
goto out;
}
kfree(file->private_data);
ctx->pos = key_hash_flash(c, &dent->key);
file->private_data = dent;
cond_resched();
}
out:
if (err != -ENOENT) {
ubifs_err(c, "cannot find next direntry, error %d", err);
return err;
}
kfree(file->private_data);
file->private_data = NULL;
/* 2 is a special value indicating that there are no more direntries */
ctx->pos = 2;
return 0;
}
/**
* ubifs_budget_space - ensure there is enough space to complete an operation.
* @c: UBIFS file-system description object
* @req: budget request
*
* This function allocates budget for an operation. It uses pessimistic
* approximation of how much flash space the operation needs. The goal of this
* function is to make sure UBIFS always has flash space to flush all dirty
* pages, dirty inodes, and dirty znodes (liability). This function may force
* commit, garbage-collection or write-back. Returns zero in case of success,
* %-ENOSPC if there is no free space and other negative error codes in case of
* failures.
*/
int ubifs_budget_space(struct ubifs_info *c, struct ubifs_budget_req *req)
{
int err, idx_growth, data_growth, dd_growth, retried = 0;
ubifs_assert(req->new_page <= 1);
ubifs_assert(req->dirtied_page <= 1);
ubifs_assert(req->new_dent <= 1);
ubifs_assert(req->mod_dent <= 1);
ubifs_assert(req->new_ino <= 1);
ubifs_assert(req->new_ino_d <= UBIFS_MAX_INO_DATA);
ubifs_assert(req->dirtied_ino <= 4);
ubifs_assert(req->dirtied_ino_d <= UBIFS_MAX_INO_DATA * 4);
ubifs_assert(!(req->new_ino_d & 7));
ubifs_assert(!(req->dirtied_ino_d & 7));
data_growth = calc_data_growth(c, req);
dd_growth = calc_dd_growth(c, req);
if (!data_growth && !dd_growth)
return 0;
idx_growth = calc_idx_growth(c, req);
again:
spin_lock(&c->space_lock);
ubifs_assert(c->bi.idx_growth >= 0);
ubifs_assert(c->bi.data_growth >= 0);
ubifs_assert(c->bi.dd_growth >= 0);
if (unlikely(c->bi.nospace) && (c->bi.nospace_rp || !can_use_rp(c))) {
dbg_budg("no space");
spin_unlock(&c->space_lock);
return -ENOSPC;
}
c->bi.idx_growth += idx_growth;
c->bi.data_growth += data_growth;
c->bi.dd_growth += dd_growth;
err = do_budget_space(c);
if (likely(!err)) {
req->idx_growth = idx_growth;
req->data_growth = data_growth;
req->dd_growth = dd_growth;
spin_unlock(&c->space_lock);
return 0;
}
/* Restore the old values */
c->bi.idx_growth -= idx_growth;
c->bi.data_growth -= data_growth;
c->bi.dd_growth -= dd_growth;
spin_unlock(&c->space_lock);
if (req->fast) {
dbg_budg("no space for fast budgeting");
return err;
}
err = make_free_space(c);
cond_resched();
if (err == -EAGAIN) {
dbg_budg("try again");
goto again;
} else if (err == -ENOSPC) {
if (!retried) {
retried = 1;
dbg_budg("-ENOSPC, but anyway try once again");
goto again;
}
dbg_budg("FS is full, -ENOSPC");
c->bi.nospace = 1;
if (can_use_rp(c) || c->rp_size == 0)
c->bi.nospace_rp = 1;
smp_wmb();
} else
ubifs_err(c, "cannot budget space, error %d", err);
return err;
}
/**
* ubifs_release_budget - release budgeted free space.
* @c: UBIFS file-system description object
* @req: budget request
*
* This function releases the space budgeted by 'ubifs_budget_space()'. Note,
* since the index changes (which were budgeted for in @req->idx_growth) will
* only be written to the media on commit, this function moves the index budget
* from @c->bi.idx_growth to @c->bi.uncommitted_idx. The latter will be zeroed
* by the commit operation.
*/
void ubifs_release_budget(struct ubifs_info *c, struct ubifs_budget_req *req)
{
ubifs_assert(req->new_page <= 1);
ubifs_assert(req->dirtied_page <= 1);
ubifs_assert(req->new_dent <= 1);
ubifs_assert(req->mod_dent <= 1);
ubifs_assert(req->new_ino <= 1);
ubifs_assert(req->new_ino_d <= UBIFS_MAX_INO_DATA);
ubifs_assert(req->dirtied_ino <= 4);
ubifs_assert(req->dirtied_ino_d <= UBIFS_MAX_INO_DATA * 4);
ubifs_assert(!(req->new_ino_d & 7));
ubifs_assert(!(req->dirtied_ino_d & 7));
if (!req->recalculate) {
ubifs_assert(req->idx_growth >= 0);
ubifs_assert(req->data_growth >= 0);
ubifs_assert(req->dd_growth >= 0);
}
if (req->recalculate) {
req->data_growth = calc_data_growth(c, req);
req->dd_growth = calc_dd_growth(c, req);
req->idx_growth = calc_idx_growth(c, req);
}
if (!req->data_growth && !req->dd_growth)
return;
c->bi.nospace = c->bi.nospace_rp = 0;
smp_wmb();
spin_lock(&c->space_lock);
c->bi.idx_growth -= req->idx_growth;
c->bi.uncommitted_idx += req->idx_growth;
c->bi.data_growth -= req->data_growth;
c->bi.dd_growth -= req->dd_growth;
c->bi.min_idx_lebs = ubifs_calc_min_idx_lebs(c);
ubifs_assert(c->bi.idx_growth >= 0);
ubifs_assert(c->bi.data_growth >= 0);
ubifs_assert(c->bi.dd_growth >= 0);
ubifs_assert(c->bi.min_idx_lebs < c->main_lebs);
ubifs_assert(!(c->bi.idx_growth & 7));
ubifs_assert(!(c->bi.data_growth & 7));
ubifs_assert(!(c->bi.dd_growth & 7));
spin_unlock(&c->space_lock);
}
/**
* ubifs_tnc_levelorder_next - next TNC tree element in levelorder traversal.
* @zr: root of the subtree to traverse
* @znode: previous znode
*
* This function implements levelorder TNC traversal. The LNC is ignored.
* Returns the next element or %NULL if @znode is already the last one.
*/
struct ubifs_znode *ubifs_tnc_levelorder_next(struct ubifs_znode *zr,
struct ubifs_znode *znode)
{
int level, iip, level_search = 0;
struct ubifs_znode *zn;
ubifs_assert(zr);
if (unlikely(!znode))
return zr;
if (unlikely(znode == zr)) {
if (znode->level == 0)
return NULL;
return ubifs_tnc_find_child(zr, 0);
}
level = znode->level;
iip = znode->iip;
while (1) {
ubifs_assert(znode->level <= zr->level);
/*
* First walk up until there is a znode with next branch to
* look at.
*/
while (znode->parent != zr && iip >= znode->parent->child_cnt) {
znode = znode->parent;
iip = znode->iip;
}
if (unlikely(znode->parent == zr &&
iip >= znode->parent->child_cnt)) {
/* This level is done, switch to the lower one */
level -= 1;
if (level_search || level < 0)
/*
* We were already looking for znode at lower
* level ('level_search'). As we are here
* again, it just does not exist. Or all levels
* were finished ('level < 0').
*/
return NULL;
level_search = 1;
iip = -1;
znode = ubifs_tnc_find_child(zr, 0);
ubifs_assert(znode);
}
/* Switch to the next index */
zn = ubifs_tnc_find_child(znode->parent, iip + 1);
if (!zn) {
/* No more children to look at, we have walk up */
iip = znode->parent->child_cnt;
continue;
}
/* Walk back down to the level we came from ('level') */
while (zn->level != level) {
znode = zn;
zn = ubifs_tnc_find_child(zn, 0);
if (!zn) {
/*
* This path is not too deep so it does not
* reach 'level'. Try next path.
*/
iip = znode->iip;
break;
}
}
if (zn) {
ubifs_assert(zn->level >= 0);
return zn;
}
}
}
static int ubifs_finddir(struct super_block *sb, char *dirname,
unsigned long root_inum, unsigned long *inum)
{
int err;
struct qstr nm;
union ubifs_key key;
struct ubifs_dent_node *dent;
struct ubifs_info *c;
struct file *file;
struct dentry *dentry;
struct inode *dir;
file = kzalloc(sizeof(struct file), 0);
dentry = kzalloc(sizeof(struct dentry), 0);
dir = kzalloc(sizeof(struct inode), 0);
if (!file || !dentry || !dir) {
printf("%s: Error, no memory for malloc!\n", __func__);
err = -ENOMEM;
goto out;
}
dir->i_sb = sb;
file->f_path.dentry = dentry;
file->f_path.dentry->d_parent = dentry;
file->f_path.dentry->d_inode = dir;
file->f_path.dentry->d_inode->i_ino = root_inum;
c = sb->s_fs_info;
dbg_gen("dir ino %lu, f_pos %#llx", dir->i_ino, file->f_pos);
/* Find the first entry in TNC and save it */
lowest_dent_key(c, &key, dir->i_ino);
nm.name = NULL;
dent = ubifs_tnc_next_ent(c, &key, &nm);
if (IS_ERR(dent)) {
err = PTR_ERR(dent);
goto out;
}
file->f_pos = key_hash_flash(c, &dent->key);
file->private_data = dent;
while (1) {
dbg_gen("feed '%s', ino %llu, new f_pos %#x",
dent->name, (unsigned long long)le64_to_cpu(dent->inum),
key_hash_flash(c, &dent->key));
ubifs_assert(le64_to_cpu(dent->ch.sqnum) > ubifs_inode(dir)->creat_sqnum);
nm.len = le16_to_cpu(dent->nlen);
if ((strncmp(dirname, (char *)dent->name, nm.len) == 0) &&
(strlen(dirname) == nm.len)) {
*inum = le64_to_cpu(dent->inum);
return 1;
}
/* Switch to the next entry */
key_read(c, &dent->key, &key);
nm.name = (char *)dent->name;
dent = ubifs_tnc_next_ent(c, &key, &nm);
if (IS_ERR(dent)) {
err = PTR_ERR(dent);
goto out;
}
kfree(file->private_data);
file->f_pos = key_hash_flash(c, &dent->key);
file->private_data = dent;
cond_resched();
}
out:
if (err != -ENOENT) {
ubifs_err("cannot find next direntry, error %d", err);
return err;
}
if (file)
free(file);
if (dentry)
free(dentry);
if (dir)
free(dir);
if (file->private_data)
kfree(file->private_data);
file->private_data = NULL;
file->f_pos = 2;
return 0;
}
int ubifs_jrn_delete_xattr(struct ubifs_info *c, const struct inode *host,
const struct inode *inode, const struct qstr *nm,
int sync)
{
int err, xlen, hlen, len, lnum, xent_offs, aligned_xlen;
struct ubifs_dent_node *xent;
struct ubifs_ino_node *ino;
union ubifs_key xent_key, key1, key2;
dbg_jrn("host %lu, xattr ino %lu, name '%s', data len %d",
host->i_ino, inode->i_ino, nm->name,
ubifs_inode(inode)->data_len);
ubifs_assert(inode->i_nlink == 0);
/*
* Since we are deleting the inode, we do not bother to attach any data
* to it and assume its length is %UBIFS_INO_NODE_SZ.
*/
xlen = UBIFS_DENT_NODE_SZ + nm->len + 1;
aligned_xlen = ALIGN(xlen, 8);
hlen = ubifs_inode(host)->data_len + UBIFS_INO_NODE_SZ;
len = aligned_xlen + UBIFS_INO_NODE_SZ + ALIGN(hlen, 8);
xent = kmalloc(len, GFP_NOFS);
if (!xent)
return -ENOMEM;
xent->ch.node_type = UBIFS_XENT_NODE;
xent_key_init(c, &xent_key, host->i_ino, nm);
key_write(c, &xent_key, xent->key);
xent->inum = 0;
xent->type = get_dent_type(inode->i_mode);
xent->nlen = cpu_to_le16(nm->len);
memcpy(xent->name, nm->name, nm->len);
xent->name[nm->len] = '\0';
zero_dent_node_unused(xent);
ubifs_prep_grp_node(c, xent, xlen, 0);
ino = (void *)xent + aligned_xlen;
pack_inode(c, ino, inode, 0, 1);
ino = (void *)ino + UBIFS_INO_NODE_SZ;
pack_inode(c, ino, host, 1, 0);
err = make_reservation(c, BASEHD, len);
if (err) {
kfree(xent);
return err;
}
err = write_head(c, BASEHD, xent, len, &lnum, &xent_offs, sync);
if (!sync && !err)
ubifs_wbuf_add_ino_nolock(&c->jheads[BASEHD].wbuf, host->i_ino);
release_head(c, BASEHD);
kfree(xent);
if (err)
goto out_ro;
/* Remove the extended attribute entry from TNC */
err = ubifs_tnc_remove_nm(c, &xent_key, nm);
if (err)
goto out_ro;
err = ubifs_add_dirt(c, lnum, xlen);
if (err)
goto out_ro;
/*
* Remove all nodes belonging to the extended attribute inode from TNC.
* Well, there actually must be only one node - the inode itself.
*/
lowest_ino_key(c, &key1, inode->i_ino);
highest_ino_key(c, &key2, inode->i_ino);
err = ubifs_tnc_remove_range(c, &key1, &key2);
if (err)
goto out_ro;
err = ubifs_add_dirt(c, lnum, UBIFS_INO_NODE_SZ);
if (err)
goto out_ro;
/* And update TNC with the new host inode position */
ino_key_init(c, &key1, host->i_ino);
err = ubifs_tnc_add(c, &key1, lnum, xent_offs + len - hlen, hlen);
if (err)
goto out_ro;
finish_reservation(c);
return 0;
out_ro:
ubifs_ro_mode(c, err);
finish_reservation(c);
return err;
}
/**
* ubifs_jrn_update - update inode.
* @c: UBIFS file-system description object
* @dir: parent inode or host inode in case of extended attributes
* @nm: directory entry name
* @inode: inode
* @deletion: indicates a directory entry deletion i.e unlink or rmdir
* @sync: non-zero if the write-buffer has to be synchronized
* @xent: non-zero if the directory entry is an extended attribute entry
*
* This function updates an inode by writing a directory entry (or extended
* attribute entry), the inode itself, and the parent directory inode (or the
* host inode) to the journal.
*
* The function writes the host inode @dir last, which is important in case of
* extended attributes. Indeed, then we guarantee that if the host inode gets
* synchronized, and the write-buffer it sits in gets flushed, the extended
* attribute inode gets flushed too. And this is exactly what the user expects -
* synchronizing the host inode synchronizes its extended attributes.
* Similarly, this guarantees that if @dir is synchronized, its directory entry
* corresponding to @nm gets synchronized too.
*
* This function returns %0 on success and a negative error code on failure.
*/
int ubifs_jrn_update(struct ubifs_info *c, const struct inode *dir,
const struct qstr *nm, const struct inode *inode,
int deletion, int sync, int xent)
{
int err, dlen, ilen, len, lnum, ino_offs, dent_offs;
int aligned_dlen, aligned_ilen;
int last_reference = !!(deletion && inode->i_nlink == 0);
struct ubifs_dent_node *dent;
struct ubifs_ino_node *ino;
union ubifs_key dent_key, ino_key;
dbg_jrn("ino %lu, dent '%.*s', data len %d in dir ino %lu",
inode->i_ino, nm->len, nm->name, ubifs_inode(inode)->data_len,
dir->i_ino);
ubifs_assert(ubifs_inode(dir)->data_len == 0);
dlen = UBIFS_DENT_NODE_SZ + nm->len + 1;
ilen = UBIFS_INO_NODE_SZ;
/*
* If the last reference to the inode is being deleted, then there is no
* need to attach and write inode data, it is being deleted anyway.
*/
if (!last_reference)
ilen += ubifs_inode(inode)->data_len;
aligned_dlen = ALIGN(dlen, 8);
aligned_ilen = ALIGN(ilen, 8);
len = aligned_dlen + aligned_ilen + UBIFS_INO_NODE_SZ;
dent = kmalloc(len, GFP_NOFS);
if (!dent)
return -ENOMEM;
if (!xent) {
dent->ch.node_type = UBIFS_DENT_NODE;
dent_key_init(c, &dent_key, dir->i_ino, nm);
} else {
dent->ch.node_type = UBIFS_XENT_NODE;
xent_key_init(c, &dent_key, dir->i_ino, nm);
}
key_write(c, &dent_key, dent->key);
dent->inum = deletion ? 0 : cpu_to_le64(inode->i_ino);
dent->type = get_dent_type(inode->i_mode);
dent->nlen = cpu_to_le16(nm->len);
memcpy(dent->name, nm->name, nm->len);
dent->name[nm->len] = '\0';
zero_dent_node_unused(dent);
ubifs_prep_grp_node(c, dent, dlen, 0);
ino = (void *)dent + aligned_dlen;
pack_inode(c, ino, inode, 0, last_reference);
ino = (void *)ino + aligned_ilen;
pack_inode(c, ino, dir, 1, 0);
err = make_reservation(c, BASEHD, len);
if (err)
goto out_free;
if (last_reference) {
err = ubifs_add_orphan(c, inode->i_ino);
if (err) {
release_head(c, BASEHD);
goto out_finish;
}
}
err = write_head(c, BASEHD, dent, len, &lnum, &dent_offs, sync);
if (!sync && !err) {
struct ubifs_wbuf *wbuf = &c->jheads[BASEHD].wbuf;
ubifs_wbuf_add_ino_nolock(wbuf, inode->i_ino);
ubifs_wbuf_add_ino_nolock(wbuf, dir->i_ino);
}
release_head(c, BASEHD);
kfree(dent);
if (err)
goto out_ro;
if (deletion) {
err = ubifs_tnc_remove_nm(c, &dent_key, nm);
if (err)
goto out_ro;
err = ubifs_add_dirt(c, lnum, dlen);
} else
err = ubifs_tnc_add_nm(c, &dent_key, lnum, dent_offs, dlen, nm);
if (err)
goto out_ro;
/*
* Note, we do not remove the inode from TNC even if the last reference
* to it has just been deleted, because the inode may still be opened.
* Instead, the inode has been added to orphan lists and the orphan
* subsystem will take further care about it.
*/
ino_key_init(c, &ino_key, inode->i_ino);
ino_offs = dent_offs + aligned_dlen;
err = ubifs_tnc_add(c, &ino_key, lnum, ino_offs, ilen);
if (err)
goto out_ro;
ino_key_init(c, &ino_key, dir->i_ino);
ino_offs += aligned_ilen;
err = ubifs_tnc_add(c, &ino_key, lnum, ino_offs, UBIFS_INO_NODE_SZ);
if (err)
goto out_ro;
finish_reservation(c);
return 0;
out_finish:
finish_reservation(c);
out_free:
kfree(dent);
return err;
out_ro:
ubifs_ro_mode(c, err);
if (last_reference)
ubifs_delete_orphan(c, inode->i_ino);
finish_reservation(c);
return err;
}
/**
* ubifs_read_node_wbuf - read node from the media or write-buffer.
* @wbuf: wbuf to check for un-written data
* @buf: buffer to read to
* @type: node type
* @len: node length
* @lnum: logical eraseblock number
* @offs: offset within the logical eraseblock
*
* This function reads a node of known type and length, checks it and stores
* in @buf. If the node partially or fully sits in the write-buffer, this
* function takes data from the buffer, otherwise it reads the flash media.
* Returns zero in case of success, %-EUCLEAN if CRC mismatched and a negative
* error code in case of failure.
*/
int ubifs_read_node_wbuf(struct ubifs_wbuf *wbuf, void *buf, int type, int len,
int lnum, int offs)
{
const struct ubifs_info *c = wbuf->c;
int err, rlen, overlap;
struct ubifs_ch *ch = buf;
#if defined(FEATURE_UBIFS_PERF_INDEX)
unsigned long long time1 = sched_clock();
int log_len = 0;
#endif
dbg_io("LEB %d:%d, %s, length %d, jhead %s", lnum, offs,
dbg_ntype(type), len, dbg_jhead(wbuf->jhead));
ubifs_assert(wbuf && lnum >= 0 && lnum < c->leb_cnt && offs >= 0);
ubifs_assert(!(offs & 7) && offs < c->leb_size);
ubifs_assert(type >= 0 && type < UBIFS_NODE_TYPES_CNT);
spin_lock(&wbuf->lock);
overlap = (lnum == wbuf->lnum && offs + len > wbuf->offs);
if (!overlap) {
/* We may safely unlock the write-buffer and read the data */
spin_unlock(&wbuf->lock);
return ubifs_read_node(c, buf, type, len, lnum, offs);
}
/* Don't read under wbuf */
rlen = wbuf->offs - offs;
if (rlen < 0)
rlen = 0;
/* Copy the rest from the write-buffer */
memcpy(buf + rlen, wbuf->buf + offs + rlen - wbuf->offs, len - rlen);
spin_unlock(&wbuf->lock);
if (rlen > 0) {
/* Read everything that goes before write-buffer */
err = ubifs_leb_read(c, lnum, buf, offs, rlen, 0);
if (err && err != -EBADMSG)
return err;
}
if (type != ch->node_type) {
ubifs_err("bad node type (%d but expected %d)",
ch->node_type, type);
goto out;
}
err = ubifs_check_node(c, buf, lnum, offs, 0, 0);
if (err) {
ubifs_err("expected node type %d", type);
return err;
}
rlen = le32_to_cpu(ch->len);
if (rlen != len) {
ubifs_err("bad node length %d, expected %d", rlen, len);
goto out;
}
#if defined(FEATURE_UBIFS_PERF_INDEX)
if(log_len > 0) {
ubifs_perf_lrcount(sched_clock() - time1, log_len);
}
#endif
return 0;
out:
ubifs_err("bad node at LEB %d:%d", lnum, offs);
dbg_dump_node(c, buf);
dbg_dump_stack();
return -EINVAL;
}
/**
* ubifs_jrn_rename - rename a directory entry.
* @c: UBIFS file-system description object
* @old_dir: parent inode of directory entry to rename
* @old_dentry: directory entry to rename
* @new_dir: parent inode of directory entry to rename
* @new_dentry: new directory entry (or directory entry to replace)
* @sync: non-zero if the write-buffer has to be synchronized
*
* Returns zero in case of success and a negative error code in case of failure.
*/
int ubifs_jrn_rename(struct ubifs_info *c, const struct inode *old_dir,
const struct dentry *old_dentry,
const struct inode *new_dir,
const struct dentry *new_dentry, int sync)
{
const struct inode *old_inode = old_dentry->d_inode;
const struct inode *new_inode = new_dentry->d_inode;
int err, dlen1, dlen2, ilen, lnum, offs, len;
int aligned_dlen1, aligned_dlen2, plen = UBIFS_INO_NODE_SZ;
int last_reference = !!(new_inode && new_inode->i_nlink == 0);
struct ubifs_dent_node *dent, *dent2;
void *p;
union ubifs_key key;
dbg_jrn("dent '%.*s' in dir ino %lu to dent '%.*s' in dir ino %lu",
old_dentry->d_name.len, old_dentry->d_name.name,
old_dir->i_ino, new_dentry->d_name.len,
new_dentry->d_name.name, new_dir->i_ino);
ubifs_assert(ubifs_inode(old_dir)->data_len == 0);
ubifs_assert(ubifs_inode(new_dir)->data_len == 0);
dlen1 = UBIFS_DENT_NODE_SZ + new_dentry->d_name.len + 1;
dlen2 = UBIFS_DENT_NODE_SZ + old_dentry->d_name.len + 1;
if (new_inode) {
ilen = UBIFS_INO_NODE_SZ;
if (!last_reference)
ilen += ubifs_inode(new_inode)->data_len;
} else
ilen = 0;
aligned_dlen1 = ALIGN(dlen1, 8);
aligned_dlen2 = ALIGN(dlen2, 8);
len = aligned_dlen1 + aligned_dlen2 + ALIGN(ilen, 8) + ALIGN(plen, 8);
if (old_dir != new_dir)
len += plen;
dent = kmalloc(len, GFP_NOFS);
if (!dent)
return -ENOMEM;
/* Make new dent */
dent->ch.node_type = UBIFS_DENT_NODE;
dent_key_init_flash(c, &dent->key, new_dir->i_ino, &new_dentry->d_name);
dent->inum = cpu_to_le64(old_inode->i_ino);
dent->type = get_dent_type(old_inode->i_mode);
dent->nlen = cpu_to_le16(new_dentry->d_name.len);
memcpy(dent->name, new_dentry->d_name.name, new_dentry->d_name.len);
dent->name[new_dentry->d_name.len] = '\0';
zero_dent_node_unused(dent);
ubifs_prep_grp_node(c, dent, dlen1, 0);
dent2 = (void *)dent + aligned_dlen1;
/* Make deletion dent */
dent2->ch.node_type = UBIFS_DENT_NODE;
dent_key_init_flash(c, &dent2->key, old_dir->i_ino,
&old_dentry->d_name);
dent2->inum = cpu_to_le64(0);
dent2->type = DT_UNKNOWN;
dent2->nlen = cpu_to_le16(old_dentry->d_name.len);
memcpy(dent2->name, old_dentry->d_name.name, old_dentry->d_name.len);
dent2->name[old_dentry->d_name.len] = '\0';
zero_dent_node_unused(dent2);
ubifs_prep_grp_node(c, dent2, dlen2, 0);
p = (void *)dent2 + aligned_dlen2;
if (new_inode) {
pack_inode(c, p, new_inode, 0, last_reference);
p += ALIGN(ilen, 8);
}
if (old_dir == new_dir)
pack_inode(c, p, old_dir, 1, 0);
else {
pack_inode(c, p, old_dir, 0, 0);
p += ALIGN(plen, 8);
pack_inode(c, p, new_dir, 1, 0);
}
err = make_reservation(c, BASEHD, len);
if (err)
goto out_free;
if (last_reference) {
err = ubifs_add_orphan(c, new_inode->i_ino);
if (err) {
release_head(c, BASEHD);
goto out_finish;
}
}
err = write_head(c, BASEHD, dent, len, &lnum, &offs, sync);
if (!sync && !err) {
struct ubifs_wbuf *wbuf = &c->jheads[BASEHD].wbuf;
ubifs_wbuf_add_ino_nolock(wbuf, new_dir->i_ino);
ubifs_wbuf_add_ino_nolock(wbuf, old_dir->i_ino);
}
release_head(c, BASEHD);
if (err)
goto out_ro;
if (new_inode)
ubifs_wbuf_add_ino_nolock(&c->jheads[BASEHD].wbuf,
new_inode->i_ino);
dent_key_init(c, &key, new_dir->i_ino, &new_dentry->d_name);
err = ubifs_tnc_add_nm(c, &key, lnum, offs, dlen1, &new_dentry->d_name);
if (err)
goto out_ro;
err = ubifs_add_dirt(c, lnum, dlen2);
if (err)
goto out_ro;
dent_key_init(c, &key, old_dir->i_ino, &old_dentry->d_name);
err = ubifs_tnc_remove_nm(c, &key, &old_dentry->d_name);
if (err)
goto out_ro;
offs += aligned_dlen1 + aligned_dlen2;
if (new_inode) {
ino_key_init(c, &key, new_inode->i_ino);
err = ubifs_tnc_add(c, &key, lnum, offs, ilen);
if (err)
goto out_ro;
offs += ALIGN(ilen, 8);
}
ino_key_init(c, &key, old_dir->i_ino);
err = ubifs_tnc_add(c, &key, lnum, offs, plen);
if (err)
goto out_ro;
if (old_dir != new_dir) {
offs += ALIGN(plen, 8);
ino_key_init(c, &key, new_dir->i_ino);
err = ubifs_tnc_add(c, &key, lnum, offs, plen);
if (err)
goto out_ro;
}
finish_reservation(c);
kfree(dent);
return 0;
out_ro:
ubifs_ro_mode(c, err);
if (last_reference)
ubifs_delete_orphan(c, new_inode->i_ino);
out_finish:
finish_reservation(c);
out_free:
kfree(dent);
return err;
}
/**
* ubifs_wbuf_sync_nolock - synchronize write-buffer.
* @wbuf: write-buffer to synchronize
*
* This function synchronizes write-buffer @buf and returns zero in case of
* success or a negative error code in case of failure.
*
* Note, although write-buffers are of @c->max_write_size, this function does
* not necessarily writes all @c->max_write_size bytes to the flash. Instead,
* if the write-buffer is only partially filled with data, only the used part
* of the write-buffer (aligned on @c->min_io_size boundary) is synchronized.
* This way we waste less space.
*/
int ubifs_wbuf_sync_nolock(struct ubifs_wbuf *wbuf)
{
struct ubifs_info *c = wbuf->c;
int err, dirt, sync_len;
cancel_wbuf_timer_nolock(wbuf);
if (!wbuf->used || wbuf->lnum == -1)
/* Write-buffer is empty or not seeked */
return 0;
dbg_io("LEB %d:%d, %d bytes, jhead %s",
wbuf->lnum, wbuf->offs, wbuf->used, dbg_jhead(wbuf->jhead));
ubifs_assert(!(wbuf->avail & 7));
ubifs_assert(wbuf->offs + wbuf->size <= c->leb_size);
ubifs_assert(wbuf->size >= c->min_io_size);
ubifs_assert(wbuf->size <= c->max_write_size);
ubifs_assert(wbuf->size % c->min_io_size == 0);
ubifs_assert(!c->ro_media && !c->ro_mount);
if (c->leb_size - wbuf->offs >= c->max_write_size)
ubifs_assert(!((wbuf->offs + wbuf->size) % c->max_write_size));
if (c->ro_error)
return -EROFS;
/*
* Do not write whole write buffer but write only the minimum necessary
* amount of min. I/O units.
*/
sync_len = ALIGN(wbuf->used, c->min_io_size);
dirt = sync_len - wbuf->used;
if (dirt)
ubifs_pad(c, wbuf->buf + wbuf->used, dirt);
#if defined(FEATURE_UBIFS_PERF_INDEX)
if(wbuf->jhead == DATAHD)
err = ubifs_leb_write_log(c, wbuf->lnum, wbuf->buf, wbuf->offs, sync_len);
else
#endif
err = ubifs_leb_write(c, wbuf->lnum, wbuf->buf, wbuf->offs, sync_len);
if (err)
return err;
wbuf->w_count += sync_len; //MTK
spin_lock(&wbuf->lock);
wbuf->offs += sync_len;
/*
* Now @wbuf->offs is not necessarily aligned to @c->max_write_size.
* But our goal is to optimize writes and make sure we write in
* @c->max_write_size chunks and to @c->max_write_size-aligned offset.
* Thus, if @wbuf->offs is not aligned to @c->max_write_size now, make
* sure that @wbuf->offs + @wbuf->size is aligned to
* @c->max_write_size. This way we make sure that after next
* write-buffer flush we are again at the optimal offset (aligned to
* @c->max_write_size).
*/
if (c->leb_size - wbuf->offs < c->max_write_size)
wbuf->size = c->leb_size - wbuf->offs;
else if (wbuf->offs & (c->max_write_size - 1))
wbuf->size = ALIGN(wbuf->offs, c->max_write_size) - wbuf->offs;
else
wbuf->size = c->max_write_size;
wbuf->avail = wbuf->size;
wbuf->used = 0;
wbuf->next_ino = 0;
spin_unlock(&wbuf->lock);
if (wbuf->sync_callback)
err = wbuf->sync_callback(c, wbuf->lnum,
c->leb_size - wbuf->offs, dirt);
return err;
}
/**
* ubifs_wbuf_write_nolock - write data to flash via write-buffer.
* @wbuf: write-buffer
* @buf: node to write
* @len: node length
*
* This function writes data to flash via write-buffer @wbuf. This means that
* the last piece of the node won't reach the flash media immediately if it
* does not take whole max. write unit (@c->max_write_size). Instead, the node
* will sit in RAM until the write-buffer is synchronized (e.g., by timer, or
* because more data are appended to the write-buffer).
*
* This function returns zero in case of success and a negative error code in
* case of failure. If the node cannot be written because there is no more
* space in this logical eraseblock, %-ENOSPC is returned.
*/
int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len)
{
struct ubifs_info *c = wbuf->c;
int err, written, n, aligned_len = ALIGN(len, 8);
dbg_io("%d bytes (%s) to jhead %s wbuf at LEB %d:%d", len,
dbg_ntype(((struct ubifs_ch *)buf)->node_type),
dbg_jhead(wbuf->jhead), wbuf->lnum, wbuf->offs + wbuf->used);
ubifs_assert(len > 0 && wbuf->lnum >= 0 && wbuf->lnum < c->leb_cnt);
ubifs_assert(wbuf->offs >= 0 && wbuf->offs % c->min_io_size == 0);
ubifs_assert(!(wbuf->offs & 7) && wbuf->offs <= c->leb_size);
ubifs_assert(wbuf->avail > 0 && wbuf->avail <= wbuf->size);
ubifs_assert(wbuf->size >= c->min_io_size);
ubifs_assert(wbuf->size <= c->max_write_size);
ubifs_assert(wbuf->size % c->min_io_size == 0);
ubifs_assert(mutex_is_locked(&wbuf->io_mutex));
ubifs_assert(!c->ro_media && !c->ro_mount);
ubifs_assert(!c->space_fixup);
if (c->leb_size - wbuf->offs >= c->max_write_size)
ubifs_assert(!((wbuf->offs + wbuf->size) % c->max_write_size));
if (c->leb_size - wbuf->offs - wbuf->used < aligned_len) {
err = -ENOSPC;
goto out;
}
cancel_wbuf_timer_nolock(wbuf);
if (c->ro_error)
return -EROFS;
if (aligned_len <= wbuf->avail) {
/*
* The node is not very large and fits entirely within
* write-buffer.
*/
memcpy(wbuf->buf + wbuf->used, buf, len);
if (aligned_len == wbuf->avail) {
dbg_io("flush jhead %s wbuf to LEB %d:%d",
dbg_jhead(wbuf->jhead), wbuf->lnum, wbuf->offs);
#if defined(FEATURE_UBIFS_PERF_INDEX)
if(wbuf->jhead == DATAHD)
err = ubifs_leb_write_log(c, wbuf->lnum, wbuf->buf,
wbuf->offs, wbuf->size);
else
#endif
err = ubifs_leb_write(c, wbuf->lnum, wbuf->buf,
wbuf->offs, wbuf->size);
if (err)
goto out;
wbuf->w_count += wbuf->size; //MTK
spin_lock(&wbuf->lock);
wbuf->offs += wbuf->size;
if (c->leb_size - wbuf->offs >= c->max_write_size)
wbuf->size = c->max_write_size;
else
wbuf->size = c->leb_size - wbuf->offs;
wbuf->avail = wbuf->size;
wbuf->used = 0;
wbuf->next_ino = 0;
spin_unlock(&wbuf->lock);
} else {
spin_lock(&wbuf->lock);
wbuf->avail -= aligned_len;
wbuf->used += aligned_len;
spin_unlock(&wbuf->lock);
}
goto exit;
}
written = 0;
if (wbuf->used) {
/*
* The node is large enough and does not fit entirely within
* current available space. We have to fill and flush
* write-buffer and switch to the next max. write unit.
*/
dbg_io("flush jhead %s wbuf to LEB %d:%d",
dbg_jhead(wbuf->jhead), wbuf->lnum, wbuf->offs);
memcpy(wbuf->buf + wbuf->used, buf, wbuf->avail);
#if defined(FEATURE_UBIFS_PERF_INDEX)
if(wbuf->jhead == DATAHD)
err = ubifs_leb_write_log(c, wbuf->lnum, wbuf->buf, wbuf->offs,
wbuf->size);
else
#endif
err = ubifs_leb_write(c, wbuf->lnum, wbuf->buf, wbuf->offs,
wbuf->size);
if (err)
goto out;
wbuf->w_count += wbuf->size; //MTK
wbuf->offs += wbuf->size;
len -= wbuf->avail;
aligned_len -= wbuf->avail;
written += wbuf->avail;
} else if (wbuf->offs & (c->max_write_size - 1)) {
/*
* The write-buffer offset is not aligned to
* @c->max_write_size and @wbuf->size is less than
* @c->max_write_size. Write @wbuf->size bytes to make sure the
* following writes are done in optimal @c->max_write_size
* chunks.
*/
dbg_io("write %d bytes to LEB %d:%d",
wbuf->size, wbuf->lnum, wbuf->offs);
#if defined(FEATURE_UBIFS_PERF_INDEX)
if(wbuf->jhead == DATAHD)
err = ubifs_leb_write_log(c, wbuf->lnum, wbuf->buf, wbuf->offs,
wbuf->size);
else
#endif
err = ubifs_leb_write(c, wbuf->lnum, buf, wbuf->offs,
wbuf->size);
if (err)
goto out;
wbuf->w_count += wbuf->size; //MTK
wbuf->offs += wbuf->size;
len -= wbuf->size;
aligned_len -= wbuf->size;
written += wbuf->size;
}
/*
* The remaining data may take more whole max. write units, so write the
* remains multiple to max. write unit size directly to the flash media.
* We align node length to 8-byte boundary because we anyway flash wbuf
* if the remaining space is less than 8 bytes.
*/
n = aligned_len >> c->max_write_shift;
if (n) {
n <<= c->max_write_shift;
dbg_io("write %d bytes to LEB %d:%d", n, wbuf->lnum,
wbuf->offs);
#if defined(FEATURE_UBIFS_PERF_INDEX)
if(wbuf->jhead == DATAHD)
err = ubifs_leb_write_log(c, wbuf->lnum, buf + written,
wbuf->offs, n);
else
#endif
err = ubifs_leb_write(c, wbuf->lnum, buf + written,
wbuf->offs, n);
if (err)
goto out;
wbuf->w_count += n; //MTK
wbuf->offs += n;
aligned_len -= n;
len -= n;
written += n;
}
spin_lock(&wbuf->lock);
if (aligned_len)
/*
* And now we have what's left and what does not take whole
* max. write unit, so write it to the write-buffer and we are
* done.
*/
memcpy(wbuf->buf, buf + written, len);
if (c->leb_size - wbuf->offs >= c->max_write_size)
wbuf->size = c->max_write_size;
else
wbuf->size = c->leb_size - wbuf->offs;
wbuf->avail = wbuf->size - aligned_len;
wbuf->used = aligned_len;
wbuf->next_ino = 0;
spin_unlock(&wbuf->lock);
exit:
if (wbuf->sync_callback) {
int free = c->leb_size - wbuf->offs - wbuf->used;
err = wbuf->sync_callback(c, wbuf->lnum, free, 0);
if (err)
goto out;
}
if (wbuf->used)
new_wbuf_timer_nolock(wbuf);
return 0;
out:
ubifs_err("cannot write %d bytes to LEB %d:%d, error %d",
len, wbuf->lnum, wbuf->offs, err);
dbg_dump_node(c, buf);
dbg_dump_stack();
dbg_dump_leb(c, wbuf->lnum);
return err;
}
/**
* ubifs_read_node_wbuf - read node from the media or write-buffer.
* @wbuf: wbuf to check for un-written data
* @buf: buffer to read to
* @type: node type
* @len: node length
* @lnum: logical eraseblock number
* @offs: offset within the logical eraseblock
*
* This function reads a node of known type and length, checks it and stores
* in @buf. If the node partially or fully sits in the write-buffer, this
* function takes data from the buffer, otherwise it reads the flash media.
* Returns zero in case of success, %-EUCLEAN if CRC mismatched and a negative
* error code in case of failure.
*/
int ubifs_read_node_wbuf(struct ubifs_wbuf *wbuf, void *buf, int type, int len,
int lnum, int offs)
{
const struct ubifs_info *c = wbuf->c;
int err, rlen, overlap;
struct ubifs_ch *ch = buf;
dbg_io("LEB %d:%d, %s, length %d", lnum, offs, dbg_ntype(type), len);
ubifs_assert(wbuf && lnum >= 0 && lnum < c->leb_cnt && offs >= 0);
ubifs_assert(!(offs & 7) && offs < c->leb_size);
ubifs_assert(type >= 0 && type < UBIFS_NODE_TYPES_CNT);
spin_lock(&wbuf->lock);
overlap = (lnum == wbuf->lnum && offs + len > wbuf->offs);
if (!overlap) {
/* We may safely unlock the write-buffer and read the data */
spin_unlock(&wbuf->lock);
return ubifs_read_node(c, buf, type, len, lnum, offs);
}
/* Don't read under wbuf */
rlen = wbuf->offs - offs;
if (rlen < 0)
rlen = 0;
/* Copy the rest from the write-buffer */
memcpy(buf + rlen, wbuf->buf + offs + rlen - wbuf->offs, len - rlen);
spin_unlock(&wbuf->lock);
if (rlen > 0) {
/* Read everything that goes before write-buffer */
err = ubi_read(c->ubi, lnum, buf, offs, rlen);
if (err && err != -EBADMSG) {
ubifs_err("failed to read node %d from LEB %d:%d, "
"error %d", type, lnum, offs, err);
dbg_dump_stack();
return err;
}
}
if (type != ch->node_type) {
ubifs_err("bad node type (%d but expected %d)",
ch->node_type, type);
goto out;
}
err = ubifs_check_node(c, buf, lnum, offs, 0, 0);
if (err) {
ubifs_err("expected node type %d", type);
return err;
}
rlen = le32_to_cpu(ch->len);
if (rlen != len) {
ubifs_err("bad node length %d, expected %d", rlen, len);
goto out;
}
return 0;
out:
ubifs_err("bad node at LEB %d:%d", lnum, offs);
dbg_dump_node(c, buf);
dbg_dump_stack();
return -EINVAL;
}
/**
* ubifs_check_node - check node.
* @c: UBIFS file-system description object
* @buf: node to check
* @lnum: logical eraseblock number
* @offs: offset within the logical eraseblock
* @quiet: print no messages
* @must_chk_crc: indicates whether to always check the CRC
*
* This function checks node magic number and CRC checksum. This function also
* validates node length to prevent UBIFS from becoming crazy when an attacker
* feeds it a file-system image with incorrect nodes. For example, too large
* node length in the common header could cause UBIFS to read memory outside of
* allocated buffer when checking the CRC checksum.
*
* This function may skip data nodes CRC checking if @c->no_chk_data_crc is
* true, which is controlled by corresponding UBIFS mount option. However, if
* @must_chk_crc is true, then @c->no_chk_data_crc is ignored and CRC is
* checked. Similarly, if @c->mounting or @c->remounting_rw is true (we are
* mounting or re-mounting to R/W mode), @c->no_chk_data_crc is ignored and CRC
* is checked. This is because during mounting or re-mounting from R/O mode to
* R/W mode we may read journal nodes (when replying the journal or doing the
* recovery) and the journal nodes may potentially be corrupted, so checking is
* required.
*
* This function returns zero in case of success and %-EUCLEAN in case of bad
* CRC or magic.
*/
int ubifs_check_node(const struct ubifs_info *c, const void *buf, int lnum,
int offs, int quiet, int must_chk_crc)
{
int err = -EINVAL, type, node_len;
uint32_t crc, node_crc, magic;
const struct ubifs_ch *ch = buf;
ubifs_assert(lnum >= 0 && lnum < c->leb_cnt && offs >= 0);
ubifs_assert(!(offs & 7) && offs < c->leb_size);
magic = le32_to_cpu(ch->magic);
if (magic != UBIFS_NODE_MAGIC) {
if (!quiet)
ubifs_err("bad magic %#08x, expected %#08x",
magic, UBIFS_NODE_MAGIC);
err = -EUCLEAN;
goto out;
}
type = ch->node_type;
if (type < 0 || type >= UBIFS_NODE_TYPES_CNT) {
if (!quiet)
ubifs_err("bad node type %d", type);
goto out;
}
node_len = le32_to_cpu(ch->len);
if (node_len + offs > c->leb_size)
goto out_len;
if (c->ranges[type].max_len == 0) {
if (node_len != c->ranges[type].len)
goto out_len;
} else if (node_len < c->ranges[type].min_len ||
node_len > c->ranges[type].max_len)
goto out_len;
if (!must_chk_crc && type == UBIFS_DATA_NODE && !c->mounting &&
!c->remounting_rw && c->no_chk_data_crc)
return 0;
crc = crc32(UBIFS_CRC32_INIT, buf + 8, node_len - 8);
node_crc = le32_to_cpu(ch->crc);
if (crc != node_crc) {
if (!quiet)
ubifs_err("bad CRC: calculated %#08x, read %#08x",
crc, node_crc);
err = -EUCLEAN;
goto out;
}
return 0;
out_len:
if (!quiet)
ubifs_err("bad node length %d", node_len);
out:
if (!quiet) {
ubifs_err("bad node at LEB %d:%d", lnum, offs);
dbg_dump_node(c, buf);
dbg_dump_stack();
}
return err;
}
static int ubifs_fill_super(struct super_block *sb, void *data, int silent)
{
struct ubi_volume_desc *ubi = sb->s_fs_info;
struct ubifs_info *c;
struct inode *root;
int err;
c = kzalloc(sizeof(struct ubifs_info), GFP_KERNEL);
if (!c)
return -ENOMEM;
spin_lock_init(&c->cnt_lock);
spin_lock_init(&c->cs_lock);
spin_lock_init(&c->buds_lock);
spin_lock_init(&c->space_lock);
spin_lock_init(&c->orphan_lock);
init_rwsem(&c->commit_sem);
mutex_init(&c->lp_mutex);
mutex_init(&c->tnc_mutex);
mutex_init(&c->log_mutex);
mutex_init(&c->mst_mutex);
mutex_init(&c->umount_mutex);
init_waitqueue_head(&c->cmt_wq);
c->buds = RB_ROOT;
c->old_idx = RB_ROOT;
c->size_tree = RB_ROOT;
c->orph_tree = RB_ROOT;
INIT_LIST_HEAD(&c->infos_list);
INIT_LIST_HEAD(&c->idx_gc);
INIT_LIST_HEAD(&c->replay_list);
INIT_LIST_HEAD(&c->replay_buds);
INIT_LIST_HEAD(&c->uncat_list);
INIT_LIST_HEAD(&c->empty_list);
INIT_LIST_HEAD(&c->freeable_list);
INIT_LIST_HEAD(&c->frdi_idx_list);
INIT_LIST_HEAD(&c->unclean_leb_list);
INIT_LIST_HEAD(&c->old_buds);
INIT_LIST_HEAD(&c->orph_list);
INIT_LIST_HEAD(&c->orph_new);
c->highest_inum = UBIFS_FIRST_INO;
c->lhead_lnum = c->ltail_lnum = UBIFS_LOG_LNUM;
ubi_get_volume_info(ubi, &c->vi);
ubi_get_device_info(c->vi.ubi_num, &c->di);
/* Re-open the UBI device in read-write mode */
c->ubi = ubi_open_volume(c->vi.ubi_num, c->vi.vol_id, UBI_READONLY);
if (IS_ERR(c->ubi)) {
err = PTR_ERR(c->ubi);
goto out_free;
}
c->vfs_sb = sb;
sb->s_fs_info = c;
sb->s_magic = UBIFS_SUPER_MAGIC;
sb->s_blocksize = UBIFS_BLOCK_SIZE;
sb->s_blocksize_bits = UBIFS_BLOCK_SHIFT;
sb->s_dev = c->vi.cdev;
sb->s_maxbytes = c->max_inode_sz = key_max_inode_size(c);
if (c->max_inode_sz > MAX_LFS_FILESIZE)
sb->s_maxbytes = c->max_inode_sz = MAX_LFS_FILESIZE;
if (c->rw_incompat) {
ubifs_err("the file-system is not R/W-compatible");
ubifs_msg("on-flash format version is w%d/r%d, but software "
"only supports up to version w%d/r%d", c->fmt_version,
c->ro_compat_version, UBIFS_FORMAT_VERSION,
UBIFS_RO_COMPAT_VERSION);
return -EROFS;
}
mutex_lock(&c->umount_mutex);
err = mount_ubifs(c);
if (err) {
ubifs_assert(err < 0);
goto out_unlock;
}
/* Read the root inode */
root = ubifs_iget(sb, UBIFS_ROOT_INO);
if (IS_ERR(root)) {
err = PTR_ERR(root);
goto out_umount;
}
sb->s_root = NULL;
mutex_unlock(&c->umount_mutex);
return 0;
out_umount:
ubifs_umount(c);
out_unlock:
mutex_unlock(&c->umount_mutex);
ubi_close_volume(c->ubi);
out_free:
kfree(c);
return err;
}
static int ubifs_link(struct dentry *old_dentry, struct inode *dir,
struct dentry *dentry)
{
struct ubifs_info *c = dir->i_sb->s_fs_info;
struct inode *inode = d_inode(old_dentry);
struct ubifs_inode *ui = ubifs_inode(inode);
struct ubifs_inode *dir_ui = ubifs_inode(dir);
int err, sz_change = CALC_DENT_SIZE(dentry->d_name.len);
struct ubifs_budget_req req = { .new_dent = 1, .dirtied_ino = 2,
.dirtied_ino_d = ALIGN(ui->data_len, 8) };
/*
* Budget request settings: new direntry, changing the target inode,
* changing the parent inode.
*/
dbg_gen("dent '%pd' to ino %lu (nlink %d) in dir ino %lu",
dentry, inode->i_ino,
inode->i_nlink, dir->i_ino);
ubifs_assert(mutex_is_locked(&dir->i_mutex));
ubifs_assert(mutex_is_locked(&inode->i_mutex));
err = dbg_check_synced_i_size(c, inode);
if (err)
return err;
err = ubifs_budget_space(c, &req);
if (err)
return err;
lock_2_inodes(dir, inode);
inc_nlink(inode);
ihold(inode);
inode->i_ctime = ubifs_current_time(inode);
dir->i_size += sz_change;
dir_ui->ui_size = dir->i_size;
dir->i_mtime = dir->i_ctime = inode->i_ctime;
err = ubifs_jnl_update(c, dir, &dentry->d_name, inode, 0, 0);
if (err)
goto out_cancel;
unlock_2_inodes(dir, inode);
ubifs_release_budget(c, &req);
d_instantiate(dentry, inode);
return 0;
out_cancel:
dir->i_size -= sz_change;
dir_ui->ui_size = dir->i_size;
drop_nlink(inode);
unlock_2_inodes(dir, inode);
ubifs_release_budget(c, &req);
iput(inode);
return err;
}
static int ubifs_unlink(struct inode *dir, struct dentry *dentry)
{
struct ubifs_info *c = dir->i_sb->s_fs_info;
struct inode *inode = d_inode(dentry);
struct ubifs_inode *dir_ui = ubifs_inode(dir);
int sz_change = CALC_DENT_SIZE(dentry->d_name.len);
int err, budgeted = 1;
struct ubifs_budget_req req = { .mod_dent = 1, .dirtied_ino = 2 };
unsigned int saved_nlink = inode->i_nlink;
/*
* Budget request settings: deletion direntry, deletion inode (+1 for
* @dirtied_ino), changing the parent directory inode. If budgeting
* fails, go ahead anyway because we have extra space reserved for
* deletions.
*/
dbg_gen("dent '%pd' from ino %lu (nlink %d) in dir ino %lu",
dentry, inode->i_ino,
inode->i_nlink, dir->i_ino);
ubifs_assert(mutex_is_locked(&dir->i_mutex));
ubifs_assert(mutex_is_locked(&inode->i_mutex));
err = dbg_check_synced_i_size(c, inode);
if (err)
return err;
err = ubifs_budget_space(c, &req);
if (err) {
if (err != -ENOSPC)
return err;
budgeted = 0;
}
lock_2_inodes(dir, inode);
inode->i_ctime = ubifs_current_time(dir);
drop_nlink(inode);
dir->i_size -= sz_change;
dir_ui->ui_size = dir->i_size;
dir->i_mtime = dir->i_ctime = inode->i_ctime;
err = ubifs_jnl_update(c, dir, &dentry->d_name, inode, 1, 0);
if (err)
goto out_cancel;
unlock_2_inodes(dir, inode);
if (budgeted)
ubifs_release_budget(c, &req);
else {
/* We've deleted something - clean the "no space" flags */
c->bi.nospace = c->bi.nospace_rp = 0;
smp_wmb();
}
return 0;
out_cancel:
dir->i_size += sz_change;
dir_ui->ui_size = dir->i_size;
set_nlink(inode, saved_nlink);
unlock_2_inodes(dir, inode);
if (budgeted)
ubifs_release_budget(c, &req);
return err;
}
/**
* check_dir_empty - check if a directory is empty or not.
* @c: UBIFS file-system description object
* @dir: VFS inode object of the directory to check
*
* This function checks if directory @dir is empty. Returns zero if the
* directory is empty, %-ENOTEMPTY if it is not, and other negative error codes
* in case of of errors.
*/
static int check_dir_empty(struct ubifs_info *c, struct inode *dir)
{
struct qstr nm = { .name = NULL };
struct ubifs_dent_node *dent;
union ubifs_key key;
int err;
lowest_dent_key(c, &key, dir->i_ino);
dent = ubifs_tnc_next_ent(c, &key, &nm);
if (IS_ERR(dent)) {
err = PTR_ERR(dent);
if (err == -ENOENT)
err = 0;
} else {
kfree(dent);
err = -ENOTEMPTY;
}
return err;
}
static int ubifs_rmdir(struct inode *dir, struct dentry *dentry)
{
struct ubifs_info *c = dir->i_sb->s_fs_info;
struct inode *inode = d_inode(dentry);
int sz_change = CALC_DENT_SIZE(dentry->d_name.len);
int err, budgeted = 1;
struct ubifs_inode *dir_ui = ubifs_inode(dir);
struct ubifs_budget_req req = { .mod_dent = 1, .dirtied_ino = 2 };
/*
* Budget request settings: deletion direntry, deletion inode and
* changing the parent inode. If budgeting fails, go ahead anyway
* because we have extra space reserved for deletions.
*/
dbg_gen("directory '%pd', ino %lu in dir ino %lu", dentry,
inode->i_ino, dir->i_ino);
ubifs_assert(mutex_is_locked(&dir->i_mutex));
ubifs_assert(mutex_is_locked(&inode->i_mutex));
err = check_dir_empty(c, d_inode(dentry));
if (err)
return err;
err = ubifs_budget_space(c, &req);
if (err) {
if (err != -ENOSPC)
return err;
budgeted = 0;
}
lock_2_inodes(dir, inode);
inode->i_ctime = ubifs_current_time(dir);
clear_nlink(inode);
drop_nlink(dir);
dir->i_size -= sz_change;
dir_ui->ui_size = dir->i_size;
dir->i_mtime = dir->i_ctime = inode->i_ctime;
err = ubifs_jnl_update(c, dir, &dentry->d_name, inode, 1, 0);
if (err)
goto out_cancel;
unlock_2_inodes(dir, inode);
if (budgeted)
ubifs_release_budget(c, &req);
else {
/* We've deleted something - clean the "no space" flags */
c->bi.nospace = c->bi.nospace_rp = 0;
smp_wmb();
}
return 0;
out_cancel:
dir->i_size += sz_change;
dir_ui->ui_size = dir->i_size;
inc_nlink(dir);
set_nlink(inode, 2);
unlock_2_inodes(dir, inode);
if (budgeted)
ubifs_release_budget(c, &req);
return err;
}
static int ubifs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode)
{
struct inode *inode;
struct ubifs_inode *dir_ui = ubifs_inode(dir);
struct ubifs_info *c = dir->i_sb->s_fs_info;
int err, sz_change = CALC_DENT_SIZE(dentry->d_name.len);
struct ubifs_budget_req req = { .new_ino = 1, .new_dent = 1 };
/*
* Budget request settings: new inode, new direntry and changing parent
* directory inode.
*/
dbg_gen("dent '%pd', mode %#hx in dir ino %lu",
dentry, mode, dir->i_ino);
err = ubifs_budget_space(c, &req);
if (err)
return err;
inode = ubifs_new_inode(c, dir, S_IFDIR | mode);
if (IS_ERR(inode)) {
err = PTR_ERR(inode);
goto out_budg;
}
err = ubifs_init_security(dir, inode, &dentry->d_name);
if (err)
goto out_inode;
mutex_lock(&dir_ui->ui_mutex);
insert_inode_hash(inode);
inc_nlink(inode);
inc_nlink(dir);
dir->i_size += sz_change;
dir_ui->ui_size = dir->i_size;
dir->i_mtime = dir->i_ctime = inode->i_ctime;
err = ubifs_jnl_update(c, dir, &dentry->d_name, inode, 0, 0);
if (err) {
ubifs_err(c, "cannot create directory, error %d", err);
goto out_cancel;
}
mutex_unlock(&dir_ui->ui_mutex);
ubifs_release_budget(c, &req);
d_instantiate(dentry, inode);
return 0;
out_cancel:
dir->i_size -= sz_change;
dir_ui->ui_size = dir->i_size;
drop_nlink(dir);
mutex_unlock(&dir_ui->ui_mutex);
out_inode:
make_bad_inode(inode);
iput(inode);
out_budg:
ubifs_release_budget(c, &req);
return err;
}
static int ubifs_mknod(struct inode *dir, struct dentry *dentry,
umode_t mode, dev_t rdev)
{
struct inode *inode;
struct ubifs_inode *ui;
struct ubifs_inode *dir_ui = ubifs_inode(dir);
struct ubifs_info *c = dir->i_sb->s_fs_info;
union ubifs_dev_desc *dev = NULL;
int sz_change = CALC_DENT_SIZE(dentry->d_name.len);
int err, devlen = 0;
struct ubifs_budget_req req = { .new_ino = 1, .new_dent = 1,
.new_ino_d = ALIGN(devlen, 8),
.dirtied_ino = 1 };
/*
* Budget request settings: new inode, new direntry and changing parent
* directory inode.
*/
dbg_gen("dent '%pd' in dir ino %lu", dentry, dir->i_ino);
if (!new_valid_dev(rdev))
return -EINVAL;
if (S_ISBLK(mode) || S_ISCHR(mode)) {
dev = kmalloc(sizeof(union ubifs_dev_desc), GFP_NOFS);
if (!dev)
return -ENOMEM;
devlen = ubifs_encode_dev(dev, rdev);
}
err = ubifs_budget_space(c, &req);
if (err) {
kfree(dev);
return err;
}
inode = ubifs_new_inode(c, dir, mode);
if (IS_ERR(inode)) {
kfree(dev);
err = PTR_ERR(inode);
goto out_budg;
}
init_special_inode(inode, inode->i_mode, rdev);
inode->i_size = ubifs_inode(inode)->ui_size = devlen;
ui = ubifs_inode(inode);
ui->data = dev;
ui->data_len = devlen;
err = ubifs_init_security(dir, inode, &dentry->d_name);
if (err)
goto out_inode;
mutex_lock(&dir_ui->ui_mutex);
dir->i_size += sz_change;
dir_ui->ui_size = dir->i_size;
dir->i_mtime = dir->i_ctime = inode->i_ctime;
err = ubifs_jnl_update(c, dir, &dentry->d_name, inode, 0, 0);
if (err)
goto out_cancel;
mutex_unlock(&dir_ui->ui_mutex);
ubifs_release_budget(c, &req);
insert_inode_hash(inode);
d_instantiate(dentry, inode);
return 0;
out_cancel:
dir->i_size -= sz_change;
dir_ui->ui_size = dir->i_size;
mutex_unlock(&dir_ui->ui_mutex);
out_inode:
make_bad_inode(inode);
iput(inode);
out_budg:
ubifs_release_budget(c, &req);
return err;
}
static int ubifs_symlink(struct inode *dir, struct dentry *dentry,
const char *symname)
{
struct inode *inode;
struct ubifs_inode *ui;
struct ubifs_inode *dir_ui = ubifs_inode(dir);
struct ubifs_info *c = dir->i_sb->s_fs_info;
int err, len = strlen(symname);
int sz_change = CALC_DENT_SIZE(dentry->d_name.len);
struct ubifs_budget_req req = { .new_ino = 1, .new_dent = 1,
.new_ino_d = ALIGN(len, 8),
.dirtied_ino = 1 };
/*
* Budget request settings: new inode, new direntry and changing parent
* directory inode.
*/
dbg_gen("dent '%pd', target '%s' in dir ino %lu", dentry,
symname, dir->i_ino);
if (len > UBIFS_MAX_INO_DATA)
return -ENAMETOOLONG;
err = ubifs_budget_space(c, &req);
if (err)
return err;
inode = ubifs_new_inode(c, dir, S_IFLNK | S_IRWXUGO);
if (IS_ERR(inode)) {
err = PTR_ERR(inode);
goto out_budg;
}
ui = ubifs_inode(inode);
ui->data = kmalloc(len + 1, GFP_NOFS);
if (!ui->data) {
err = -ENOMEM;
goto out_inode;
}
memcpy(ui->data, symname, len);
((char *)ui->data)[len] = '\0';
inode->i_link = ui->data;
/*
* The terminating zero byte is not written to the flash media and it
* is put just to make later in-memory string processing simpler. Thus,
* data length is @len, not @len + %1.
*/
ui->data_len = len;
inode->i_size = ubifs_inode(inode)->ui_size = len;
err = ubifs_init_security(dir, inode, &dentry->d_name);
if (err)
goto out_inode;
mutex_lock(&dir_ui->ui_mutex);
dir->i_size += sz_change;
dir_ui->ui_size = dir->i_size;
dir->i_mtime = dir->i_ctime = inode->i_ctime;
err = ubifs_jnl_update(c, dir, &dentry->d_name, inode, 0, 0);
if (err)
goto out_cancel;
mutex_unlock(&dir_ui->ui_mutex);
ubifs_release_budget(c, &req);
insert_inode_hash(inode);
d_instantiate(dentry, inode);
return 0;
out_cancel:
dir->i_size -= sz_change;
dir_ui->ui_size = dir->i_size;
mutex_unlock(&dir_ui->ui_mutex);
out_inode:
make_bad_inode(inode);
iput(inode);
out_budg:
ubifs_release_budget(c, &req);
return err;
}
/**
* lock_3_inodes - a wrapper for locking three UBIFS inodes.
* @inode1: first inode
* @inode2: second inode
* @inode3: third inode
*
* This function is used for 'ubifs_rename()' and @inode1 may be the same as
* @inode2 whereas @inode3 may be %NULL.
*
* We do not implement any tricks to guarantee strict lock ordering, because
* VFS has already done it for us on the @i_mutex. So this is just a simple
* wrapper function.
*/
static void lock_3_inodes(struct inode *inode1, struct inode *inode2,
struct inode *inode3)
{
mutex_lock_nested(&ubifs_inode(inode1)->ui_mutex, WB_MUTEX_1);
if (inode2 != inode1)
mutex_lock_nested(&ubifs_inode(inode2)->ui_mutex, WB_MUTEX_2);
if (inode3)
mutex_lock_nested(&ubifs_inode(inode3)->ui_mutex, WB_MUTEX_3);
}
/**
* unlock_3_inodes - a wrapper for unlocking three UBIFS inodes for rename.
* @inode1: first inode
* @inode2: second inode
* @inode3: third inode
*/
static void unlock_3_inodes(struct inode *inode1, struct inode *inode2,
struct inode *inode3)
{
if (inode3)
mutex_unlock(&ubifs_inode(inode3)->ui_mutex);
if (inode1 != inode2)
mutex_unlock(&ubifs_inode(inode2)->ui_mutex);
mutex_unlock(&ubifs_inode(inode1)->ui_mutex);
}
static int ubifs_rename(struct inode *old_dir, struct dentry *old_dentry,
struct inode *new_dir, struct dentry *new_dentry)
{
struct ubifs_info *c = old_dir->i_sb->s_fs_info;
struct inode *old_inode = d_inode(old_dentry);
struct inode *new_inode = d_inode(new_dentry);
struct ubifs_inode *old_inode_ui = ubifs_inode(old_inode);
int err, release, sync = 0, move = (new_dir != old_dir);
int is_dir = S_ISDIR(old_inode->i_mode);
int unlink = !!new_inode;
int new_sz = CALC_DENT_SIZE(new_dentry->d_name.len);
int old_sz = CALC_DENT_SIZE(old_dentry->d_name.len);
struct ubifs_budget_req req = { .new_dent = 1, .mod_dent = 1,
.dirtied_ino = 3 };
struct ubifs_budget_req ino_req = { .dirtied_ino = 1,
.dirtied_ino_d = ALIGN(old_inode_ui->data_len, 8) };
struct timespec time;
unsigned int uninitialized_var(saved_nlink);
/*
* Budget request settings: deletion direntry, new direntry, removing
* the old inode, and changing old and new parent directory inodes.
*
* However, this operation also marks the target inode as dirty and
* does not write it, so we allocate budget for the target inode
* separately.
*/
dbg_gen("dent '%pd' ino %lu in dir ino %lu to dent '%pd' in dir ino %lu",
old_dentry, old_inode->i_ino, old_dir->i_ino,
new_dentry, new_dir->i_ino);
ubifs_assert(mutex_is_locked(&old_dir->i_mutex));
ubifs_assert(mutex_is_locked(&new_dir->i_mutex));
if (unlink)
ubifs_assert(mutex_is_locked(&new_inode->i_mutex));
if (unlink && is_dir) {
err = check_dir_empty(c, new_inode);
if (err)
return err;
}
err = ubifs_budget_space(c, &req);
if (err)
return err;
err = ubifs_budget_space(c, &ino_req);
if (err) {
ubifs_release_budget(c, &req);
return err;
}
lock_3_inodes(old_dir, new_dir, new_inode);
/*
* Like most other Unix systems, set the @i_ctime for inodes on a
* rename.
*/
time = ubifs_current_time(old_dir);
old_inode->i_ctime = time;
/* We must adjust parent link count when renaming directories */
if (is_dir) {
if (move) {
/*
* @old_dir loses a link because we are moving
* @old_inode to a different directory.
*/
drop_nlink(old_dir);
/*
* @new_dir only gains a link if we are not also
* overwriting an existing directory.
*/
if (!unlink)
inc_nlink(new_dir);
} else {
/*
* @old_inode is not moving to a different directory,
* but @old_dir still loses a link if we are
* overwriting an existing directory.
*/
if (unlink)
drop_nlink(old_dir);
}
}
old_dir->i_size -= old_sz;
ubifs_inode(old_dir)->ui_size = old_dir->i_size;
old_dir->i_mtime = old_dir->i_ctime = time;
new_dir->i_mtime = new_dir->i_ctime = time;
/*
* And finally, if we unlinked a direntry which happened to have the
* same name as the moved direntry, we have to decrement @i_nlink of
* the unlinked inode and change its ctime.
*/
if (unlink) {
/*
* Directories cannot have hard-links, so if this is a
* directory, just clear @i_nlink.
*/
saved_nlink = new_inode->i_nlink;
if (is_dir)
clear_nlink(new_inode);
else
drop_nlink(new_inode);
new_inode->i_ctime = time;
} else {
new_dir->i_size += new_sz;
ubifs_inode(new_dir)->ui_size = new_dir->i_size;
}
/*
* Do not ask 'ubifs_jnl_rename()' to flush write-buffer if @old_inode
* is dirty, because this will be done later on at the end of
* 'ubifs_rename()'.
*/
if (IS_SYNC(old_inode)) {
sync = IS_DIRSYNC(old_dir) || IS_DIRSYNC(new_dir);
if (unlink && IS_SYNC(new_inode))
sync = 1;
}
err = ubifs_jnl_rename(c, old_dir, old_dentry, new_dir, new_dentry,
sync);
if (err)
goto out_cancel;
unlock_3_inodes(old_dir, new_dir, new_inode);
ubifs_release_budget(c, &req);
mutex_lock(&old_inode_ui->ui_mutex);
release = old_inode_ui->dirty;
mark_inode_dirty_sync(old_inode);
mutex_unlock(&old_inode_ui->ui_mutex);
if (release)
ubifs_release_budget(c, &ino_req);
if (IS_SYNC(old_inode))
err = old_inode->i_sb->s_op->write_inode(old_inode, NULL);
return err;
out_cancel:
if (unlink) {
set_nlink(new_inode, saved_nlink);
} else {
new_dir->i_size -= new_sz;
ubifs_inode(new_dir)->ui_size = new_dir->i_size;
}
old_dir->i_size += old_sz;
ubifs_inode(old_dir)->ui_size = old_dir->i_size;
if (is_dir) {
if (move) {
inc_nlink(old_dir);
if (!unlink)
drop_nlink(new_dir);
} else {
if (unlink)
inc_nlink(old_dir);
}
}
unlock_3_inodes(old_dir, new_dir, new_inode);
ubifs_release_budget(c, &ino_req);
ubifs_release_budget(c, &req);
return err;
}
int ubifs_getattr(struct vfsmount *mnt, struct dentry *dentry,
struct kstat *stat)
{
loff_t size;
struct inode *inode = d_inode(dentry);
struct ubifs_inode *ui = ubifs_inode(inode);
mutex_lock(&ui->ui_mutex);
generic_fillattr(inode, stat);
stat->blksize = UBIFS_BLOCK_SIZE;
stat->size = ui->ui_size;
/*
* Unfortunately, the 'stat()' system call was designed for block
* device based file systems, and it is not appropriate for UBIFS,
* because UBIFS does not have notion of "block". For example, it is
* difficult to tell how many block a directory takes - it actually
* takes less than 300 bytes, but we have to round it to block size,
* which introduces large mistake. This makes utilities like 'du' to
* report completely senseless numbers. This is the reason why UBIFS
* goes the same way as JFFS2 - it reports zero blocks for everything
* but regular files, which makes more sense than reporting completely
* wrong sizes.
*/
if (S_ISREG(inode->i_mode)) {
size = ui->xattr_size;
size += stat->size;
size = ALIGN(size, UBIFS_BLOCK_SIZE);
/*
* Note, user-space expects 512-byte blocks count irrespectively
* of what was reported in @stat->size.
*/
stat->blocks = size >> 9;
} else
stat->blocks = 0;
mutex_unlock(&ui->ui_mutex);
return 0;
}
const struct inode_operations ubifs_dir_inode_operations = {
.lookup = ubifs_lookup,
.create = ubifs_create,
.link = ubifs_link,
.symlink = ubifs_symlink,
.unlink = ubifs_unlink,
.mkdir = ubifs_mkdir,
.rmdir = ubifs_rmdir,
.mknod = ubifs_mknod,
.rename = ubifs_rename,
.setattr = ubifs_setattr,
.getattr = ubifs_getattr,
.setxattr = ubifs_setxattr,
.getxattr = ubifs_getxattr,
.listxattr = ubifs_listxattr,
.removexattr = ubifs_removexattr,
};
const struct file_operations ubifs_dir_operations = {
.llseek = generic_file_llseek,
.release = ubifs_dir_release,
.read = generic_read_dir,
.iterate = ubifs_readdir,
.fsync = ubifs_fsync,
.unlocked_ioctl = ubifs_ioctl,
#ifdef CONFIG_COMPAT
.compat_ioctl = ubifs_compat_ioctl,
#endif
};
/**
* ubifs_wbuf_write_nolock - write data to flash via write-buffer.
* @wbuf: write-buffer
* @buf: node to write
* @len: node length
*
* This function writes data to flash via write-buffer @wbuf. This means that
* the last piece of the node won't reach the flash media immediately if it
* does not take whole minimal I/O unit. Instead, the node will sit in RAM
* until the write-buffer is synchronized (e.g., by timer).
*
* This function returns zero in case of success and a negative error code in
* case of failure. If the node cannot be written because there is no more
* space in this logical eraseblock, %-ENOSPC is returned.
*/
int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len)
{
struct ubifs_info *c = wbuf->c;
int err, written, n, aligned_len = ALIGN(len, 8), offs;
dbg_io("%d bytes (%s) to wbuf at LEB %d:%d", len,
dbg_ntype(((struct ubifs_ch *)buf)->node_type), wbuf->lnum,
wbuf->offs + wbuf->used);
ubifs_assert(len > 0 && wbuf->lnum >= 0 && wbuf->lnum < c->leb_cnt);
ubifs_assert(wbuf->offs >= 0 && wbuf->offs % c->min_io_size == 0);
ubifs_assert(!(wbuf->offs & 7) && wbuf->offs <= c->leb_size);
ubifs_assert(wbuf->avail > 0 && wbuf->avail <= c->min_io_size);
ubifs_assert(mutex_is_locked(&wbuf->io_mutex));
if (c->leb_size - wbuf->offs - wbuf->used < aligned_len) {
err = -ENOSPC;
goto out;
}
cancel_wbuf_timer_nolock(wbuf);
if (c->ro_media)
return -EROFS;
if (aligned_len <= wbuf->avail) {
/*
* The node is not very large and fits entirely within
* write-buffer.
*/
memcpy(wbuf->buf + wbuf->used, buf, len);
if (aligned_len == wbuf->avail) {
dbg_io("flush wbuf to LEB %d:%d", wbuf->lnum,
wbuf->offs);
err = ubi_leb_write(c->ubi, wbuf->lnum, wbuf->buf,
wbuf->offs, c->min_io_size,
wbuf->dtype);
if (err)
goto out;
spin_lock(&wbuf->lock);
wbuf->offs += c->min_io_size;
wbuf->avail = c->min_io_size;
wbuf->used = 0;
wbuf->next_ino = 0;
spin_unlock(&wbuf->lock);
} else {
spin_lock(&wbuf->lock);
wbuf->avail -= aligned_len;
wbuf->used += aligned_len;
spin_unlock(&wbuf->lock);
}
goto exit;
}
/*
* The node is large enough and does not fit entirely within current
* minimal I/O unit. We have to fill and flush write-buffer and switch
* to the next min. I/O unit.
*/
dbg_io("flush wbuf to LEB %d:%d", wbuf->lnum, wbuf->offs);
memcpy(wbuf->buf + wbuf->used, buf, wbuf->avail);
err = ubi_leb_write(c->ubi, wbuf->lnum, wbuf->buf, wbuf->offs,
c->min_io_size, wbuf->dtype);
if (err)
goto out;
offs = wbuf->offs + c->min_io_size;
len -= wbuf->avail;
aligned_len -= wbuf->avail;
written = wbuf->avail;
/*
* The remaining data may take more whole min. I/O units, so write the
* remains multiple to min. I/O unit size directly to the flash media.
* We align node length to 8-byte boundary because we anyway flash wbuf
* if the remaining space is less than 8 bytes.
*/
n = aligned_len >> c->min_io_shift;
if (n) {
n <<= c->min_io_shift;
dbg_io("write %d bytes to LEB %d:%d", n, wbuf->lnum, offs);
err = ubi_leb_write(c->ubi, wbuf->lnum, buf + written, offs, n,
wbuf->dtype);
if (err)
goto out;
offs += n;
aligned_len -= n;
len -= n;
written += n;
}
spin_lock(&wbuf->lock);
if (aligned_len)
/*
* And now we have what's left and what does not take whole
* min. I/O unit, so write it to the write-buffer and we are
* done.
*/
memcpy(wbuf->buf, buf + written, len);
wbuf->offs = offs;
wbuf->used = aligned_len;
wbuf->avail = c->min_io_size - aligned_len;
wbuf->next_ino = 0;
spin_unlock(&wbuf->lock);
exit:
if (wbuf->sync_callback) {
int free = c->leb_size - wbuf->offs - wbuf->used;
err = wbuf->sync_callback(c, wbuf->lnum, free, 0);
if (err)
goto out;
}
if (wbuf->used)
new_wbuf_timer_nolock(wbuf);
return 0;
out:
ubifs_err("cannot write %d bytes to LEB %d:%d, error %d",
len, wbuf->lnum, wbuf->offs, err);
dbg_dump_node(c, buf);
dbg_dump_stack();
dbg_dump_leb(c, wbuf->lnum);
return err;
}
/**
* reserve_space - reserve space in the journal.
* @c: UBIFS file-system description object
* @jhead: journal head number
* @len: node length
*
* This function reserves space in journal head @head. If the reservation
* succeeded, the journal head stays locked and later has to be unlocked using
* 'release_head()'. 'write_node()' and 'write_head()' functions also unlock
* it. Returns zero in case of success, %-EAGAIN if commit has to be done, and
* other negative error codes in case of other failures.
*/
static int reserve_space(struct ubifs_info *c, int jhead, int len)
{
int err = 0, err1, retries = 0, avail, lnum, offs, free, squeeze;
struct ubifs_wbuf *wbuf = &c->jheads[jhead].wbuf;
/*
* Typically, the base head has smaller nodes written to it, so it is
* better to try to allocate space at the ends of eraseblocks. This is
* what the squeeze parameter does.
*/
squeeze = (jhead == BASEHD);
again:
mutex_lock_nested(&wbuf->io_mutex, wbuf->jhead);
avail = c->leb_size - wbuf->offs - wbuf->used;
if (wbuf->lnum != -1 && avail >= len)
return 0;
/*
* Write buffer wasn't seek'ed or there is no enough space - look for an
* LEB with some empty space.
*/
lnum = ubifs_find_free_space(c, len, &free, squeeze);
if (lnum >= 0) {
/* Found an LEB, add it to the journal head */
offs = c->leb_size - free;
err = ubifs_add_bud_to_log(c, jhead, lnum, offs);
if (err)
goto out_return;
/* A new bud was successfully allocated and added to the log */
goto out;
}
err = lnum;
if (err != -ENOSPC)
goto out_unlock;
/*
* No free space, we have to run garbage collector to make
* some. But the write-buffer mutex has to be unlocked because
* GC have to sync write buffers, which may lead a deadlock.
*/
dbg_jrn("no free space jhead %d, run GC", jhead);
mutex_unlock(&wbuf->io_mutex);
lnum = ubifs_garbage_collect(c, 0);
if (lnum < 0) {
err = lnum;
if (err != -ENOSPC)
return err;
/*
* GC could not make a free LEB. But someone else may
* have allocated new bud for this journal head,
* because we dropped the 'io_mutex', so try once
* again.
*/
dbg_jrn("GC couldn't make a free LEB for jhead %d", jhead);
if (retries++ < 2) {
dbg_jrn("retry (%d)", retries);
goto again;
}
dbg_jrn("return -ENOSPC");
return err;
}
mutex_lock_nested(&wbuf->io_mutex, wbuf->jhead);
dbg_jrn("got LEB %d for jhead %d", lnum, jhead);
avail = c->leb_size - wbuf->offs - wbuf->used;
if (wbuf->lnum != -1 && avail >= len) {
/*
* Someone else has switched the journal head and we have
* enough space now. This happens when more then one process is
* trying to write to the same journal head at the same time.
*/
dbg_jrn("return LEB %d back, already have LEB %d:%d",
lnum, wbuf->lnum, wbuf->offs + wbuf->used);
err = ubifs_return_leb(c, lnum);
if (err)
goto out_unlock;
return 0;
}
err = ubifs_add_bud_to_log(c, jhead, lnum, 0);
if (err)
goto out_return;
offs = 0;
out:
err = ubifs_wbuf_seek_nolock(wbuf, lnum, offs, UBI_SHORTTERM);
if (err)
goto out_unlock;
return 0;
out_unlock:
mutex_unlock(&wbuf->io_mutex);
return err;
out_return:
/* An error occurred and the LEB has to be returned to lprops */
ubifs_assert(err < 0);
err1 = ubifs_return_leb(c, lnum);
if (err1 && err == -EAGAIN)
/*
* Return original error code 'err' only if it is not
* '-EAGAIN', which is not really an error. Otherwise, return
* the error code of 'ubifs_return_leb()'.
*/
err = err1;
mutex_unlock(&wbuf->io_mutex);
return err;
}
/**
* ubifs_budget_space - ensure there is enough space to complete an operation.
* @c: UBIFS file-system description object
* @req: budget request
*
* This function allocates budget for an operation. It uses pessimistic
* approximation of how much flash space the operation needs. The goal of this
* function is to make sure UBIFS always has flash space to flush all dirty
* pages, dirty inodes, and dirty znodes (liability). This function may force
* commit, garbage-collection or write-back. Returns zero in case of success,
* %-ENOSPC if there is no free space and other negative error codes in case of
* failures.
*/
int ubifs_budget_space(struct ubifs_info *c, struct ubifs_budget_req *req)
{
int uninitialized_var(cmt_retries), uninitialized_var(wb_retries);
int err, idx_growth, data_growth, dd_growth;
struct retries_info ri;
ubifs_assert(req->dirtied_ino <= 4);
ubifs_assert(req->dirtied_ino_d <= UBIFS_MAX_INO_DATA * 4);
data_growth = calc_data_growth(c, req);
dd_growth = calc_dd_growth(c, req);
if (!data_growth && !dd_growth)
return 0;
idx_growth = calc_idx_growth(c, req);
memset(&ri, 0, sizeof(struct retries_info));
again:
spin_lock(&c->space_lock);
ubifs_assert(c->budg_idx_growth >= 0);
ubifs_assert(c->budg_data_growth >= 0);
ubifs_assert(c->budg_dd_growth >= 0);
if (unlikely(c->nospace) && (c->nospace_rp || !can_use_rp(c))) {
dbg_budg("no space");
spin_unlock(&c->space_lock);
return -ENOSPC;
}
c->budg_idx_growth += idx_growth;
c->budg_data_growth += data_growth;
c->budg_dd_growth += dd_growth;
err = do_budget_space(c);
if (likely(!err)) {
req->idx_growth = idx_growth;
req->data_growth = data_growth;
req->dd_growth = dd_growth;
spin_unlock(&c->space_lock);
return 0;
}
/* Restore the old values */
c->budg_idx_growth -= idx_growth;
c->budg_data_growth -= data_growth;
c->budg_dd_growth -= dd_growth;
spin_unlock(&c->space_lock);
if (req->fast) {
dbg_budg("no space for fast budgeting");
return err;
}
err = make_free_space(c, &ri);
if (err == -EAGAIN) {
dbg_budg("try again");
cond_resched();
goto again;
} else if (err == -ENOSPC) {
dbg_budg("FS is full, -ENOSPC");
c->nospace = 1;
if (can_use_rp(c) || c->rp_size == 0)
c->nospace_rp = 1;
smp_wmb();
} else
ubifs_err("cannot budget space, error %d", err);
return err;
}
/**
* make_reservation - reserve journal space.
* @c: UBIFS file-system description object
* @jhead: journal head
* @len: how many bytes to reserve
*
* This function makes space reservation in journal head @jhead. The function
* takes the commit lock and locks the journal head, and the caller has to
* unlock the head and finish the reservation with 'finish_reservation()'.
* Returns zero in case of success and a negative error code in case of
* failure.
*
* Note, the journal head may be unlocked as soon as the data is written, while
* the commit lock has to be released after the data has been added to the
* TNC.
*/
static int make_reservation(struct ubifs_info *c, int jhead, int len)
{
int err, cmt_retries = 0, nospc_retries = 0;
ubifs_assert(len <= c->dark_wm);
again:
down_read(&c->commit_sem);
err = reserve_space(c, jhead, len);
if (!err)
return 0;
up_read(&c->commit_sem);
if (err == -ENOSPC) {
/*
* GC could not make any progress. We should try to commit
* once because it could make some dirty space and GC would
* make progress, so make the error -EAGAIN so that the below
* will commit and re-try.
*/
if (nospc_retries++ < 2) {
dbg_jrn("no space, retry");
err = -EAGAIN;
}
/*
* This means that the budgeting is incorrect. We always have
* to be able to write to the media, because all operations are
* budgeted. Deletions are not budgeted, though, but we reserve
* an extra LEB for them.
*/
}
if (err != -EAGAIN)
goto out;
/*
* -EAGAIN means that the journal is full or too large, or the above
* code wants to do one commit. Do this and re-try.
*/
if (cmt_retries > 128) {
/*
* This should not happen unless the journal size limitations
* are too tough.
*/
ubifs_err("stuck in space allocation");
err = -ENOSPC;
goto out;
} else if (cmt_retries > 32)
ubifs_warn("too many space allocation re-tries (%d)",
cmt_retries);
dbg_jrn("-EAGAIN, commit and retry (retried %d times)",
cmt_retries);
cmt_retries += 1;
err = ubifs_run_commit(c);
if (err)
return err;
goto again;
out:
ubifs_err("cannot reserve %d bytes in jhead %d, error %d",
len, jhead, err);
if (err == -ENOSPC) {
/* This are some budgeting problems, print useful information */
down_write(&c->commit_sem);
spin_lock(&c->space_lock);
dbg_dump_stack();
dbg_dump_budg(c);
spin_unlock(&c->space_lock);
dbg_dump_lprops(c);
cmt_retries = dbg_check_lprops(c);
up_write(&c->commit_sem);
}
return err;
}
static int ubifs_printdir(struct file *file, void *dirent)
{
int err, over = 0;
struct qstr nm;
union ubifs_key key;
struct ubifs_dent_node *dent;
struct inode *dir = file->f_path.dentry->d_inode;
struct ubifs_info *c = dir->i_sb->s_fs_info;
dbg_gen("dir ino %lu, f_pos %#llx", dir->i_ino, file->f_pos);
if (file->f_pos > UBIFS_S_KEY_HASH_MASK || file->f_pos == 2)
/*
* The directory was seek'ed to a senseless position or there
* are no more entries.
*/
return 0;
if (file->f_pos == 1) {
/* Find the first entry in TNC and save it */
lowest_dent_key(c, &key, dir->i_ino);
nm.name = NULL;
dent = ubifs_tnc_next_ent(c, &key, &nm);
if (IS_ERR(dent)) {
err = PTR_ERR(dent);
goto out;
}
file->f_pos = key_hash_flash(c, &dent->key);
file->private_data = dent;
}
dent = file->private_data;
if (!dent) {
/*
* The directory was seek'ed to and is now readdir'ed.
* Find the entry corresponding to @file->f_pos or the
* closest one.
*/
dent_key_init_hash(c, &key, dir->i_ino, file->f_pos);
nm.name = NULL;
dent = ubifs_tnc_next_ent(c, &key, &nm);
if (IS_ERR(dent)) {
err = PTR_ERR(dent);
goto out;
}
file->f_pos = key_hash_flash(c, &dent->key);
file->private_data = dent;
}
while (1) {
dbg_gen("feed '%s', ino %llu, new f_pos %#x",
dent->name, (unsigned long long)le64_to_cpu(dent->inum),
key_hash_flash(c, &dent->key));
ubifs_assert(le64_to_cpu(dent->ch.sqnum) > ubifs_inode(dir)->creat_sqnum);
nm.len = le16_to_cpu(dent->nlen);
over = filldir(c, (char *)dent->name, nm.len,
le64_to_cpu(dent->inum), dent->type);
if (over)
return 0;
/* Switch to the next entry */
key_read(c, &dent->key, &key);
nm.name = (char *)dent->name;
dent = ubifs_tnc_next_ent(c, &key, &nm);
if (IS_ERR(dent)) {
err = PTR_ERR(dent);
goto out;
}
kfree(file->private_data);
file->f_pos = key_hash_flash(c, &dent->key);
file->private_data = dent;
cond_resched();
}
out:
if (err != -ENOENT) {
ubifs_err("cannot find next direntry, error %d", err);
return err;
}
kfree(file->private_data);
file->private_data = NULL;
file->f_pos = 2;
return 0;
}
/**
* ubifs_jrn_write_data - write a data node to the journal.
* @c: UBIFS file-system description object
* @inode: inode the data node belongs to
* @key: node key
* @buf: buffer to write
* @len: data length (must not exceed %UBIFS_BLOCK_SIZE)
*
* This function writes a data node to the journal. Returns %0 if the data node
* was successfully written, and a negative error code in case of failure.
*/
int ubifs_jrn_write_data(struct ubifs_info *c, const struct inode *inode,
const union ubifs_key *key, const void *buf, int len)
{
int err, lnum, offs, compr_type, out_len;
int dlen = UBIFS_DATA_NODE_SZ + UBIFS_BLOCK_SIZE * WORST_COMPR_FACTOR;
const struct ubifs_inode *ui = ubifs_inode(inode);
struct ubifs_data_node *data;
dbg_jrn_key(c, key, "ino %lu, blk %u, len %d, key ",
key_ino(c, key), key_block(c, key), len);
ubifs_assert(len <= UBIFS_BLOCK_SIZE);
data = kmalloc(dlen, GFP_NOFS);
if (!data)
return -ENOMEM;
data->ch.node_type = UBIFS_DATA_NODE;
key_write(c, key, &data->key);
data->size = cpu_to_le32(len);
zero_data_node_unused(data);
if (!(ui->flags && UBIFS_COMPR_FL))
/* Compression is disabled for this inode */
compr_type = UBIFS_COMPR_NONE;
else
compr_type = ui->compr_type;
out_len = dlen - UBIFS_DATA_NODE_SZ;
ubifs_compress(buf, len, &data->data, &out_len, &compr_type);
ubifs_assert(out_len <= UBIFS_BLOCK_SIZE);
dlen = UBIFS_DATA_NODE_SZ + out_len;
data->compr_type = cpu_to_le16(compr_type);
err = make_reservation(c, DATAHD, dlen);
if (err)
goto out_free;
err = write_node(c, DATAHD, data, dlen, &lnum, &offs);
if (!err)
ubifs_wbuf_add_ino_nolock(&c->jheads[DATAHD].wbuf,
key_ino(c, key));
release_head(c, DATAHD);
if (err)
goto out_ro;
err = ubifs_tnc_add(c, key, lnum, offs, dlen);
if (err)
goto out_ro;
finish_reservation(c);
kfree(data);
return 0;
out_ro:
ubifs_ro_mode(c, err);
finish_reservation(c);
out_free:
kfree(data);
return err;
}
/**
* ubifs_add_bud_to_log - add a new bud to the log.
* @c: UBIFS file-system description object
* @jhead: journal head the bud belongs to
* @lnum: LEB number of the bud
* @offs: starting offset of the bud
*
* This function writes reference node for the new bud LEB @lnum it to the log,
* and adds it to the buds tress. It also makes sure that log size does not
* exceed the 'c->max_bud_bytes' limit. Returns zero in case of success,
* %-EAGAIN if commit is required, and a negative error codes in case of
* failure.
*/
int ubifs_add_bud_to_log(struct ubifs_info *c, int jhead, int lnum, int offs)
{
int err;
struct ubifs_bud *bud;
struct ubifs_ref_node *ref;
bud = kmalloc(sizeof(struct ubifs_bud), GFP_NOFS);
if (!bud)
return -ENOMEM;
ref = kzalloc(c->ref_node_alsz, GFP_NOFS);
if (!ref) {
kfree(bud);
return -ENOMEM;
}
mutex_lock(&c->log_mutex);
ubifs_assert(!c->ro_media && !c->ro_mount);
if (c->ro_error) {
err = -EROFS;
goto out_unlock;
}
/* Make sure we have enough space in the log */
if (empty_log_bytes(c) - c->ref_node_alsz < c->min_log_bytes) {
dbg_log("not enough log space - %lld, required %d",
empty_log_bytes(c), c->min_log_bytes);
ubifs_commit_required(c);
err = -EAGAIN;
goto out_unlock;
}
/*
* Make sure the amount of space in buds will not exceed the
* 'c->max_bud_bytes' limit, because we want to guarantee mount time
* limits.
*
* It is not necessary to hold @c->buds_lock when reading @c->bud_bytes
* because we are holding @c->log_mutex. All @c->bud_bytes take place
* when both @c->log_mutex and @c->bud_bytes are locked.
*/
if (c->bud_bytes + c->leb_size - offs > c->max_bud_bytes) {
dbg_log("bud bytes %lld (%lld max), require commit",
c->bud_bytes, c->max_bud_bytes);
ubifs_commit_required(c);
err = -EAGAIN;
goto out_unlock;
}
/*
* If the journal is full enough - start background commit. Note, it is
* OK to read 'c->cmt_state' without spinlock because integer reads
* are atomic in the kernel.
*/
if (c->bud_bytes >= c->bg_bud_bytes &&
c->cmt_state == COMMIT_RESTING) {
dbg_log("bud bytes %lld (%lld max), initiate BG commit",
c->bud_bytes, c->max_bud_bytes);
ubifs_request_bg_commit(c);
}
bud->lnum = lnum;
bud->start = offs;
bud->jhead = jhead;
ref->ch.node_type = UBIFS_REF_NODE;
ref->lnum = cpu_to_le32(bud->lnum);
ref->offs = cpu_to_le32(bud->start);
ref->jhead = cpu_to_le32(jhead);
if (c->lhead_offs > c->leb_size - c->ref_node_alsz) {
c->lhead_lnum = ubifs_next_log_lnum(c, c->lhead_lnum);
c->lhead_offs = 0;
}
if (c->lhead_offs == 0) {
/* Must ensure next log LEB has been unmapped */
err = ubifs_leb_unmap(c, c->lhead_lnum);
if (err)
goto out_unlock;
}
if (bud->start == 0) {
/*
* Before writing the LEB reference which refers an empty LEB
* to the log, we have to make sure it is mapped, because
* otherwise we'd risk to refer an LEB with garbage in case of
* an unclean reboot, because the target LEB might have been
* unmapped, but not yet physically erased.
*/
err = ubifs_leb_map(c, bud->lnum);
if (err)
goto out_unlock;
}
dbg_log("write ref LEB %d:%d",
c->lhead_lnum, c->lhead_offs);
err = ubifs_write_node(c, ref, UBIFS_REF_NODE_SZ, c->lhead_lnum,
c->lhead_offs);
if (err)
goto out_unlock;
c->lhead_offs += c->ref_node_alsz;
ubifs_add_bud(c, bud);
mutex_unlock(&c->log_mutex);
kfree(ref);
return 0;
out_unlock:
mutex_unlock(&c->log_mutex);
kfree(ref);
kfree(bud);
return err;
}
/**
* ubifs_read_superblock - read superblock.
* @c: UBIFS file-system description object
*
* This function finds, reads and checks the superblock. If an empty UBI volume
* is being mounted, this function creates default superblock. Returns zero in
* case of success, and a negative error code in case of failure.
*/
int ubifs_read_superblock(struct ubifs_info *c)
{
int err, sup_flags;
struct ubifs_sb_node *sup;
if (c->empty) {
err = create_default_filesystem(c);
if (err)
return err;
}
sup = ubifs_read_sb_node(c);
if (IS_ERR(sup))
return PTR_ERR(sup);
c->fmt_version = le32_to_cpu(sup->fmt_version);
c->ro_compat_version = le32_to_cpu(sup->ro_compat_version);
/*
* The software supports all previous versions but not future versions,
* due to the unavailability of time-travelling equipment.
*/
if (c->fmt_version > UBIFS_FORMAT_VERSION) {
struct super_block *sb = c->vfs_sb;
int mounting_ro = sb->s_flags & MS_RDONLY;
ubifs_assert(!c->ro_media || mounting_ro);
if (!mounting_ro ||
c->ro_compat_version > UBIFS_RO_COMPAT_VERSION) {
ubifs_err("on-flash format version is w%d/r%d, but "
"software only supports up to version "
"w%d/r%d", c->fmt_version,
c->ro_compat_version, UBIFS_FORMAT_VERSION,
UBIFS_RO_COMPAT_VERSION);
if (c->ro_compat_version <= UBIFS_RO_COMPAT_VERSION) {
ubifs_msg("only R/O mounting is possible");
err = -EROFS;
} else
err = -EINVAL;
goto out;
}
/*
* The FS is mounted R/O, and the media format is
* R/O-compatible with the UBIFS implementation, so we can
* mount.
*/
c->rw_incompat = 1;
}
if (c->fmt_version < 3) {
ubifs_err("on-flash format version %d is not supported",
c->fmt_version);
err = -EINVAL;
goto out;
}
switch (sup->key_hash) {
case UBIFS_KEY_HASH_R5:
c->key_hash = key_r5_hash;
c->key_hash_type = UBIFS_KEY_HASH_R5;
break;
case UBIFS_KEY_HASH_TEST:
c->key_hash = key_test_hash;
c->key_hash_type = UBIFS_KEY_HASH_TEST;
break;
};
c->key_fmt = sup->key_fmt;
switch (c->key_fmt) {
case UBIFS_SIMPLE_KEY_FMT:
c->key_len = UBIFS_SK_LEN;
break;
default:
ubifs_err("unsupported key format");
err = -EINVAL;
goto out;
}
c->leb_cnt = le32_to_cpu(sup->leb_cnt);
c->max_leb_cnt = le32_to_cpu(sup->max_leb_cnt);
c->max_bud_bytes = le64_to_cpu(sup->max_bud_bytes);
c->log_lebs = le32_to_cpu(sup->log_lebs);
c->lpt_lebs = le32_to_cpu(sup->lpt_lebs);
c->orph_lebs = le32_to_cpu(sup->orph_lebs);
c->jhead_cnt = le32_to_cpu(sup->jhead_cnt) + NONDATA_JHEADS_CNT;
c->fanout = le32_to_cpu(sup->fanout);
c->lsave_cnt = le32_to_cpu(sup->lsave_cnt);
c->rp_size = le64_to_cpu(sup->rp_size);
c->rp_uid = le32_to_cpu(sup->rp_uid);
c->rp_gid = le32_to_cpu(sup->rp_gid);
sup_flags = le32_to_cpu(sup->flags);
if (!c->mount_opts.override_compr)
c->default_compr = le16_to_cpu(sup->default_compr);
c->vfs_sb->s_time_gran = le32_to_cpu(sup->time_gran);
memcpy(&c->uuid, &sup->uuid, 16);
c->big_lpt = !!(sup_flags & UBIFS_FLG_BIGLPT);
/* Automatically increase file system size to the maximum size */
c->old_leb_cnt = c->leb_cnt;
if (c->leb_cnt < c->vi.size && c->leb_cnt < c->max_leb_cnt) {
c->leb_cnt = min_t(int, c->max_leb_cnt, c->vi.size);
if (c->vfs_sb->s_flags & MS_RDONLY)
dbg_mnt("Auto resizing (ro) from %d LEBs to %d LEBs",
c->old_leb_cnt, c->leb_cnt);
else {
dbg_mnt("Auto resizing (sb) from %d LEBs to %d LEBs",
c->old_leb_cnt, c->leb_cnt);
sup->leb_cnt = cpu_to_le32(c->leb_cnt);
err = ubifs_write_sb_node(c, sup);
if (err)
goto out;
c->old_leb_cnt = c->leb_cnt;
}
}
/**
* do_commit - commit the journal.
* @c: UBIFS file-system description object
*
* This function implements UBIFS commit. It has to be called with commit lock
* locked. Returns zero in case of success and a negative error code in case of
* failure.
*/
static int do_commit(struct ubifs_info *c)
{
int err, new_ltail_lnum, old_ltail_lnum, i;
struct ubifs_zbranch zroot;
struct ubifs_lp_stats lst;
dbg_cmt("start");
ubifs_assert(!c->ro_media && !c->ro_mount);
if (c->ro_error) {
err = -EROFS;
goto out_up;
}
if (nothing_to_commit(c)) {
up_write(&c->commit_sem);
err = 0;
goto out_cancel;
}
/* Sync all write buffers (necessary for recovery) */
for (i = 0; i < c->jhead_cnt; i++) {
err = ubifs_wbuf_sync(&c->jheads[i].wbuf);
if (err)
goto out_up;
}
c->cmt_no += 1;
err = ubifs_gc_start_commit(c);
if (err)
goto out_up;
err = dbg_check_lprops(c);
if (err)
goto out_up;
err = ubifs_log_start_commit(c, &new_ltail_lnum);
if (err)
goto out_up;
err = ubifs_tnc_start_commit(c, &zroot);
if (err)
goto out_up;
err = ubifs_lpt_start_commit(c);
if (err)
goto out_up;
err = ubifs_orphan_start_commit(c);
if (err)
goto out_up;
ubifs_get_lp_stats(c, &lst);
up_write(&c->commit_sem);
err = ubifs_tnc_end_commit(c);
if (err)
goto out;
err = ubifs_lpt_end_commit(c);
if (err)
goto out;
err = ubifs_orphan_end_commit(c);
if (err)
goto out;
old_ltail_lnum = c->ltail_lnum;
err = ubifs_log_end_commit(c, new_ltail_lnum);
if (err)
goto out;
err = dbg_check_old_index(c, &zroot);
if (err)
goto out;
mutex_lock(&c->mst_mutex);
c->mst_node->cmt_no = cpu_to_le64(c->cmt_no);
c->mst_node->log_lnum = cpu_to_le32(new_ltail_lnum);
c->mst_node->root_lnum = cpu_to_le32(zroot.lnum);
c->mst_node->root_offs = cpu_to_le32(zroot.offs);
c->mst_node->root_len = cpu_to_le32(zroot.len);
c->mst_node->ihead_lnum = cpu_to_le32(c->ihead_lnum);
c->mst_node->ihead_offs = cpu_to_le32(c->ihead_offs);
c->mst_node->index_size = cpu_to_le64(c->bi.old_idx_sz);
c->mst_node->lpt_lnum = cpu_to_le32(c->lpt_lnum);
c->mst_node->lpt_offs = cpu_to_le32(c->lpt_offs);
c->mst_node->nhead_lnum = cpu_to_le32(c->nhead_lnum);
c->mst_node->nhead_offs = cpu_to_le32(c->nhead_offs);
c->mst_node->ltab_lnum = cpu_to_le32(c->ltab_lnum);
c->mst_node->ltab_offs = cpu_to_le32(c->ltab_offs);
c->mst_node->lsave_lnum = cpu_to_le32(c->lsave_lnum);
c->mst_node->lsave_offs = cpu_to_le32(c->lsave_offs);
c->mst_node->lscan_lnum = cpu_to_le32(c->lscan_lnum);
c->mst_node->empty_lebs = cpu_to_le32(lst.empty_lebs);
c->mst_node->idx_lebs = cpu_to_le32(lst.idx_lebs);
c->mst_node->total_free = cpu_to_le64(lst.total_free);
c->mst_node->total_dirty = cpu_to_le64(lst.total_dirty);
c->mst_node->total_used = cpu_to_le64(lst.total_used);
c->mst_node->total_dead = cpu_to_le64(lst.total_dead);
c->mst_node->total_dark = cpu_to_le64(lst.total_dark);
if (c->no_orphs)
c->mst_node->flags |= cpu_to_le32(UBIFS_MST_NO_ORPHS);
else
c->mst_node->flags &= ~cpu_to_le32(UBIFS_MST_NO_ORPHS);
err = ubifs_write_master(c);
mutex_unlock(&c->mst_mutex);
if (err)
goto out;
err = ubifs_log_post_commit(c, old_ltail_lnum);
if (err)
goto out;
err = ubifs_gc_end_commit(c);
if (err)
goto out;
err = ubifs_lpt_post_commit(c);
if (err)
goto out;
out_cancel:
spin_lock(&c->cs_lock);
c->cmt_state = COMMIT_RESTING;
wake_up(&c->cmt_wq);
dbg_cmt("commit end");
spin_unlock(&c->cs_lock);
return 0;
out_up:
up_write(&c->commit_sem);
out:
ubifs_err("commit failed, error %d", err);
spin_lock(&c->cs_lock);
c->cmt_state = COMMIT_BROKEN;
wake_up(&c->cmt_wq);
spin_unlock(&c->cs_lock);
ubifs_ro_mode(c, err);
return err;
}
