aufs5.7-20200622.patch

+	IMustLock(dir);
+
+	name = &dentry->d_name;
+	path.dentry = vfsub_lookup_one_len(name->name, parent, name->len);
+	if (IS_ERR(path.dentry)) {
+		file = (void *)path.dentry;
+		pr_err("%pd lookup err %ld\n", dentry, PTR_ERR(path.dentry));
+		goto out;
+	}
+
+	/* no need to mnt_want_write() since we call dentry_open() later */
+	err = vfs_create(dir, path.dentry, 0666, NULL);
+	if (unlikely(err)) {
+		file = ERR_PTR(err);
+		pr_err("%pd create err %d\n", dentry, err);
+		goto out_dput;
+	}
+
+	path.mnt = base->mnt;
+	file = vfsub_dentry_open(&path,
+				 O_RDWR | O_CREAT | O_EXCL | O_LARGEFILE
+				 /* | __FMODE_NONOTIFY */);
+	if (IS_ERR(file)) {
+		pr_err("%pd open err %ld\n", dentry, PTR_ERR(file));
+		goto out_dput;
+	}
+
+	delegated = NULL;
+	err = vfsub_unlink(dir, &file->f_path, &delegated, /*force*/0);
+	au_xino_unlock_dir(&ldir);
+	do_unlock = 0;
+	if (unlikely(err == -EWOULDBLOCK)) {
+		pr_warn("cannot retry for NFSv4 delegation"
+			" for an internal unlink\n");
+		iput(delegated);
+	}
+	if (unlikely(err)) {
+		pr_err("%pd unlink err %d\n", dentry, err);
+		goto out_fput;
+	}
+
+	if (copy_src) {
+		/* no one can touch copy_src xino */
+		err = au_copy_file(file, copy_src, vfsub_f_size_read(copy_src));
+		if (unlikely(err)) {
+			pr_err("%pd copy err %d\n", dentry, err);
+			goto out_fput;
+		}
+	}
+	goto out_dput; /* success */
+
+out_fput:
+	fput(file);
+	file = ERR_PTR(err);
+out_dput:
+	dput(path.dentry);
+out:
+	if (do_unlock)
+		au_xino_unlock_dir(&ldir);
+	return file;
+}
+
+struct file *au_xino_file1(struct au_xino *xi)
+{
+	struct file *file;
+	unsigned int u, nfile;
+
+	file = NULL;
+	nfile = xi->xi_nfile;
+	for (u = 0; u < nfile; u++) {
+		file = xi->xi_file[u];
+		if (file)
+			break;
+	}
+
+	return file;
+}
+
+static int au_xino_file_set(struct au_xino *xi, int idx, struct file *file)
+{
+	int err;
+	struct file *f;
+	void *p;
+
+	if (file)
+		get_file(file);
+
+	err = 0;
+	f = NULL;
+	if (idx < xi->xi_nfile) {
+		f = xi->xi_file[idx];
+		if (f)
+			fput(f);
+	} else {
+		p = au_kzrealloc(xi->xi_file,
+				 sizeof(*xi->xi_file) * xi->xi_nfile,
+				 sizeof(*xi->xi_file) * (idx + 1),
+				 GFP_NOFS, /*may_shrink*/0);
+		if (p) {
+			MtxMustLock(&xi->xi_mtx);
+			xi->xi_file = p;
+			xi->xi_nfile = idx + 1;
+		} else {
+			err = -ENOMEM;
+			if (file)
+				fput(file);
+			goto out;
+		}
+	}
+	xi->xi_file[idx] = file;
+
+out:
+	return err;
+}
+
+/*
+ * if @xinew->xi is not set, then create new xigen file.
+ */
+struct file *au_xi_new(struct super_block *sb, struct au_xi_new *xinew)
+{
+	struct file *file;
+	int err;
+
+	SiMustAnyLock(sb);
+
+	file = au_xino_create2(sb, xinew->base, xinew->copy_src);
+	if (IS_ERR(file)) {
+		err = PTR_ERR(file);
+		pr_err("%s[%d], err %d\n",
+		       xinew->xi ? "xino" : "xigen",
+		       xinew->idx, err);
+		goto out;
+	}
+
+	if (xinew->xi)
+		err = au_xino_file_set(xinew->xi, xinew->idx, file);
+	else {
+		BUG();
+		/* todo: make xigen file an array */
+		/* err = au_xigen_file_set(sb, xinew->idx, file); */
+	}
+	fput(file);
+	if (unlikely(err))
+		file = ERR_PTR(err);
+
+out:
+	return file;
+}
+
+/* ---------------------------------------------------------------------- */
+
+/*
+ * truncate xino files
+ */
+static int au_xino_do_trunc(struct super_block *sb, aufs_bindex_t bindex,
+			    int idx, struct kstatfs *st)
+{
+	int err;
+	blkcnt_t blocks;
+	struct file *file, *new_xino;
+	struct au_xi_new xinew = {
+		.idx = idx
+	};
+
+	err = 0;
+	xinew.xi = au_sbr(sb, bindex)->br_xino;
+	file = au_xino_file(xinew.xi, idx);
+	if (!file)
+		goto out;
+
+	xinew.base = &file->f_path;
+	err = vfs_statfs(xinew.base, st);
+	if (unlikely(err)) {
+		AuErr1("statfs err %d, ignored\n", err);
+		err = 0;
+		goto out;
+	}
+
+	blocks = file_inode(file)->i_blocks;
+	pr_info("begin truncating xino(b%d-%d), ib%llu, %llu/%llu free blks\n",
+		bindex, idx, (u64)blocks, st->f_bfree, st->f_blocks);
+
+	xinew.copy_src = file;
+	new_xino = au_xi_new(sb, &xinew);
+	if (IS_ERR(new_xino)) {
+		err = PTR_ERR(new_xino);
+		pr_err("xino(b%d-%d), err %d, ignored\n", bindex, idx, err);
+		goto out;
+	}
+
+	err = vfs_statfs(&new_xino->f_path, st);
+	if (!err)
+		pr_info("end truncating xino(b%d-%d), ib%llu, %llu/%llu free blks\n",
+			bindex, idx, (u64)file_inode(new_xino)->i_blocks,
+			st->f_bfree, st->f_blocks);
+	else {
+		AuErr1("statfs err %d, ignored\n", err);
+		err = 0;
+	}
+
+out:
+	return err;
+}
+
+int au_xino_trunc(struct super_block *sb, aufs_bindex_t bindex, int idx_begin)
+{
+	int err, i;
+	unsigned long jiffy;
+	aufs_bindex_t bbot;
+	struct kstatfs *st;
+	struct au_branch *br;
+	struct au_xino *xi;
+
+	err = -ENOMEM;
+	st = kmalloc(sizeof(*st), GFP_NOFS);
+	if (unlikely(!st))
+		goto out;
+
+	err = -EINVAL;
+	bbot = au_sbbot(sb);
+	if (unlikely(bindex < 0 || bbot < bindex))
+		goto out_st;
+
+	err = 0;
+	jiffy = jiffies;
+	br = au_sbr(sb, bindex);
+	xi = br->br_xino;
+	for (i = idx_begin; !err && i < xi->xi_nfile; i++)
+		err = au_xino_do_trunc(sb, bindex, i, st);
+	if (!err)
+		au_sbi(sb)->si_xino_jiffy = jiffy;
+
+out_st:
+	au_kfree_rcu(st);
+out:
+	return err;
+}
+
+struct xino_do_trunc_args {
+	struct super_block *sb;
+	struct au_branch *br;
+	int idx;
+};
+
+static void xino_do_trunc(void *_args)
+{
+	struct xino_do_trunc_args *args = _args;
+	struct super_block *sb;
+	struct au_branch *br;
+	struct inode *dir;
+	int err, idx;
+	aufs_bindex_t bindex;
+
+	err = 0;
+	sb = args->sb;
+	dir = d_inode(sb->s_root);
+	br = args->br;
+	idx = args->idx;
+
+	si_noflush_write_lock(sb);
+	ii_read_lock_parent(dir);
+	bindex = au_br_index(sb, br->br_id);
+	err = au_xino_trunc(sb, bindex, idx);
+	ii_read_unlock(dir);
+	if (unlikely(err))
+		pr_warn("err b%d, (%d)\n", bindex, err);
+	atomic_dec(&br->br_xino->xi_truncating);
+	au_lcnt_dec(&br->br_count);
+	si_write_unlock(sb);
+	au_nwt_done(&au_sbi(sb)->si_nowait);
+	au_kfree_rcu(args);
+}
+
+/*
+ * returns the index in the xi_file array whose corresponding file is necessary
+ * to truncate, or -1 which means no need to truncate.
+ */
+static int xino_trunc_test(struct super_block *sb, struct au_branch *br)
+{
+	int err;
+	unsigned int u;
+	struct kstatfs st;
+	struct au_sbinfo *sbinfo;
+	struct au_xino *xi;
+	struct file *file;
+
+	/* todo: si_xino_expire and the ratio should be customizable */
+	sbinfo = au_sbi(sb);
+	if (time_before(jiffies,
+			sbinfo->si_xino_jiffy + sbinfo->si_xino_expire))
+		return -1;
+
+	/* truncation border */
+	xi = br->br_xino;
+	for (u = 0; u < xi->xi_nfile; u++) {
+		file = au_xino_file(xi, u);
+		if (!file)
+			continue;
+
+		err = vfs_statfs(&file->f_path, &st);
+		if (unlikely(err)) {
+			AuErr1("statfs err %d, ignored\n", err);
+			return -1;
+		}
+		if (div64_u64(st.f_bfree * 100, st.f_blocks)
+		    >= AUFS_XINO_DEF_TRUNC)
+			return u;
+	}
+
+	return -1;
+}
+
+static void xino_try_trunc(struct super_block *sb, struct au_branch *br)
+{
+	int idx;
+	struct xino_do_trunc_args *args;
+	int wkq_err;
+
+	idx = xino_trunc_test(sb, br);
+	if (idx < 0)
+		return;
+
+	if (atomic_inc_return(&br->br_xino->xi_truncating) > 1)
+		goto out;
+
+	/* lock and kfree() will be called in trunc_xino() */
+	args = kmalloc(sizeof(*args), GFP_NOFS);
+	if (unlikely(!args)) {
+		AuErr1("no memory\n");
+		goto out;
+	}
+
+	au_lcnt_inc(&br->br_count);
+	args->sb = sb;
+	args->br = br;
+	args->idx = idx;
+	wkq_err = au_wkq_nowait(xino_do_trunc, args, sb, /*flags*/0);
+	if (!wkq_err)
+		return; /* success */
+
+	pr_err("wkq %d\n", wkq_err);
+	au_lcnt_dec(&br->br_count);
+	au_kfree_rcu(args);
+
+out:
+	atomic_dec(&br->br_xino->xi_truncating);
+}
+
+/* ---------------------------------------------------------------------- */
+
+struct au_xi_calc {
+	int idx;
+	loff_t pos;
+};
+
+static void au_xi_calc(struct super_block *sb, ino_t h_ino,
+		       struct au_xi_calc *calc)
+{
+	loff_t maxent;
+
+	maxent = au_xi_maxent(sb);
+	calc->idx = div64_u64_rem(h_ino, maxent, &calc->pos);
+	calc->pos *= sizeof(ino_t);
+}
+
+static int au_xino_do_new_async(struct super_block *sb, struct au_branch *br,
+				struct au_xi_calc *calc)
+{
+	int err;
+	struct file *file;
+	struct au_xino *xi = br->br_xino;
+	struct au_xi_new xinew = {
+		.xi = xi
+	};
+
+	SiMustAnyLock(sb);
+
+	err = 0;
+	if (!xi)
+		goto out;
+
+	mutex_lock(&xi->xi_mtx);
+	file = au_xino_file(xi, calc->idx);
+	if (file)
+		goto out_mtx;
+
+	file = au_xino_file(xi, /*idx*/-1);
+	AuDebugOn(!file);
+	xinew.idx = calc->idx;
+	xinew.base = &file->f_path;
+	/* xinew.copy_src = NULL; */
+	file = au_xi_new(sb, &xinew);
+	if (IS_ERR(file))
+		err = PTR_ERR(file);
+
+out_mtx:
+	mutex_unlock(&xi->xi_mtx);
+out:
+	return err;
+}
+
+struct au_xino_do_new_async_args {
+	struct super_block *sb;
+	struct au_branch *br;
+	struct au_xi_calc calc;
+	ino_t ino;
+};
+
+struct au_xi_writing {
+	struct hlist_bl_node node;
+	ino_t h_ino, ino;
+};
+
+static int au_xino_do_write(vfs_writef_t write, struct file *file,
+			    struct au_xi_calc *calc, ino_t ino);
+
+static void au_xino_call_do_new_async(void *args)
+{
+	struct au_xino_do_new_async_args *a = args;
+	struct au_branch *br;
+	struct super_block *sb;
+	struct au_sbinfo *sbi;
+	struct inode *root;
+	struct file *file;
+	struct au_xi_writing *del, *p;
+	struct hlist_bl_head *hbl;
+	struct hlist_bl_node *pos;
+	int err;
+
+	br = a->br;
+	sb = a->sb;
+	sbi = au_sbi(sb);
+	si_noflush_read_lock(sb);
+	root = d_inode(sb->s_root);
+	ii_read_lock_child(root);
+	err = au_xino_do_new_async(sb, br, &a->calc);
+	if (unlikely(err)) {
+		AuIOErr("err %d\n", err);
+		goto out;
+	}
+
+	file = au_xino_file(br->br_xino, a->calc.idx);
+	AuDebugOn(!file);
+	err = au_xino_do_write(sbi->si_xwrite, file, &a->calc, a->ino);
+	if (unlikely(err)) {
+		AuIOErr("err %d\n", err);
+		goto out;
+	}
+
+	del = NULL;
+	hbl = &br->br_xino->xi_writing;
+	hlist_bl_lock(hbl);
+	au_hbl_for_each(pos, hbl) {
+		p = container_of(pos, struct au_xi_writing, node);
+		if (p->ino == a->ino) {
+			del = p;
+			hlist_bl_del(&p->node);
+			break;
+		}
+	}
+	hlist_bl_unlock(hbl);
+	au_kfree_rcu(del);
+
+out:
+	au_lcnt_dec(&br->br_count);
+	ii_read_unlock(root);
+	si_read_unlock(sb);
+	au_nwt_done(&sbi->si_nowait);
+	au_kfree_rcu(a);
+}
+
+/*
+ * create a new xino file asynchronously
+ */
+static int au_xino_new_async(struct super_block *sb, struct au_branch *br,
+			     struct au_xi_calc *calc, ino_t ino)
+{
+	int err;
+	struct au_xino_do_new_async_args *arg;
+
+	err = -ENOMEM;
+	arg = kmalloc(sizeof(*arg), GFP_NOFS);
+	if (unlikely(!arg))
+		goto out;
+
+	arg->sb = sb;
+	arg->br = br;
+	arg->calc = *calc;
+	arg->ino = ino;
+	au_lcnt_inc(&br->br_count);
+	err = au_wkq_nowait(au_xino_call_do_new_async, arg, sb, AuWkq_NEST);
+	if (unlikely(err)) {
+		pr_err("wkq %d\n", err);
+		au_lcnt_dec(&br->br_count);
+		au_kfree_rcu(arg);
+	}
+
+out:
+	return err;
+}
+
+/*
+ * read @ino from xinofile for the specified branch{@sb, @bindex}
+ * at the position of @h_ino.
+ */
+int au_xino_read(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino,
+		 ino_t *ino)
+{
+	int err;
+	ssize_t sz;
+	struct au_xi_calc calc;
+	struct au_sbinfo *sbinfo;
+	struct file *file;
+	struct au_xino *xi;
+	struct hlist_bl_head *hbl;
+	struct hlist_bl_node *pos;
+	struct au_xi_writing *p;
+
+	*ino = 0;
+	if (!au_opt_test(au_mntflags(sb), XINO))
+		return 0; /* no xino */
+
+	err = 0;
+	au_xi_calc(sb, h_ino, &calc);
+	xi = au_sbr(sb, bindex)->br_xino;
+	file = au_xino_file(xi, calc.idx);
+	if (!file) {
+		hbl = &xi->xi_writing;
+		hlist_bl_lock(hbl);
+		au_hbl_for_each(pos, hbl) {
+			p = container_of(pos, struct au_xi_writing, node);
+			if (p->h_ino == h_ino) {
+				AuDbg("hi%llu, i%llu, found\n",
+				      (u64)p->h_ino, (u64)p->ino);
+				*ino = p->ino;
+				break;
+			}
+		}
+		hlist_bl_unlock(hbl);
+		return 0;
+	} else if (vfsub_f_size_read(file) < calc.pos + sizeof(*ino))
+		return 0; /* no xino */
+
+	sbinfo = au_sbi(sb);
+	sz = xino_fread(sbinfo->si_xread, file, ino, sizeof(*ino), &calc.pos);
+	if (sz == sizeof(*ino))
+		return 0; /* success */
+
+	err = sz;
+	if (unlikely(sz >= 0)) {
+		err = -EIO;
+		AuIOErr("xino read error (%zd)\n", sz);
+	}
+	return err;
+}
+
+static int au_xino_do_write(vfs_writef_t write, struct file *file,
+			    struct au_xi_calc *calc, ino_t ino)
+{
+	ssize_t sz;
+
+	sz = xino_fwrite(write, file, &ino, sizeof(ino), &calc->pos);
+	if (sz == sizeof(ino))
+		return 0; /* success */
+
+	AuIOErr("write failed (%zd)\n", sz);
+	return -EIO;
+}
+
+/*
+ * write @ino to the xinofile for the specified branch{@sb, @bindex}
+ * at the position of @h_ino.
+ * even if @ino is zero, it is written to the xinofile and means no entry.
+ * if the size of the xino file on a specific filesystem exceeds the watermark,
+ * try truncating it.
+ */
+int au_xino_write(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino,
+		  ino_t ino)
+{
+	int err;
+	unsigned int mnt_flags;
+	struct au_xi_calc calc;
+	struct file *file;
+	struct au_branch *br;
+	struct au_xino *xi;
+	struct au_xi_writing *p;
+
+	SiMustAnyLock(sb);
+
+	mnt_flags = au_mntflags(sb);
+	if (!au_opt_test(mnt_flags, XINO))
+		return 0;
+
+	au_xi_calc(sb, h_ino, &calc);
+	br = au_sbr(sb, bindex);
+	xi = br->br_xino;
+	file = au_xino_file(xi, calc.idx);
+	if (!file) {
+		/* store the inum pair into the list */
+		p = kmalloc(sizeof(*p), GFP_NOFS | __GFP_NOFAIL);
+		p->h_ino = h_ino;
+		p->ino = ino;
+		au_hbl_add(&p->node, &xi->xi_writing);
+
+		/* create and write a new xino file asynchronously */
+		err = au_xino_new_async(sb, br, &calc, ino);
+		if (!err)
+			return 0; /* success */
+		goto out;
+	}
+
+	err = au_xino_do_write(au_sbi(sb)->si_xwrite, file, &calc, ino);
+	if (!err) {
+		br = au_sbr(sb, bindex);
+		if (au_opt_test(mnt_flags, TRUNC_XINO)
+		    && au_test_fs_trunc_xino(au_br_sb(br)))
+			xino_try_trunc(sb, br);
+		return 0; /* success */
+	}
+
+out:
+	AuIOErr("write failed (%d)\n", err);
+	return -EIO;
+}
+
+static ssize_t xino_fread_wkq(vfs_readf_t func, struct file *file, void *buf,
+			      size_t size, loff_t *pos);
+
+/* todo: unnecessary to support mmap_sem since kernel-space? */
+ssize_t xino_fread(vfs_readf_t func, struct file *file, void *kbuf, size_t size,
+		   loff_t *pos)
+{
+	ssize_t err;
+	mm_segment_t oldfs;
+	union {
+		void *k;
+		char __user *u;
+	} buf;
+	int i;
+	const int prevent_endless = 10;
+
+	i = 0;
+	buf.k = kbuf;
+	oldfs = get_fs();
+	set_fs(KERNEL_DS);
+	do {
+		err = func(file, buf.u, size, pos);
+		if (err == -EINTR
+		    && !au_wkq_test()
+		    && fatal_signal_pending(current)) {
+			set_fs(oldfs);
+			err = xino_fread_wkq(func, file, kbuf, size, pos);
+			BUG_ON(err == -EINTR);
+			oldfs = get_fs();
+			set_fs(KERNEL_DS);
+		}
+	} while (i++ < prevent_endless
+		 && (err == -EAGAIN || err == -EINTR));
+	set_fs(oldfs);
+
+#if 0 /* reserved for future use */
+	if (err > 0)
+		fsnotify_access(file->f_path.dentry);
+#endif
+
+	return err;
+}
+
+struct xino_fread_args {
+	ssize_t *errp;
+	vfs_readf_t func;
+	struct file *file;
+	void *buf;
+	size_t size;
+	loff_t *pos;
+};
+
+static void call_xino_fread(void *args)
+{
+	struct xino_fread_args *a = args;
+	*a->errp = xino_fread(a->func, a->file, a->buf, a->size, a->pos);
+}
+
+static ssize_t xino_fread_wkq(vfs_readf_t func, struct file *file, void *buf,
+			      size_t size, loff_t *pos)
+{
+	ssize_t err;
+	int wkq_err;
+	struct xino_fread_args args = {
+		.errp	= &err,
+		.func	= func,
+		.file	= file,
+		.buf	= buf,
+		.size	= size,
+		.pos	= pos
+	};
+
+	wkq_err = au_wkq_wait(call_xino_fread, &args);
+	if (unlikely(wkq_err))
+		err = wkq_err;
+
+	return err;
+}
+
+static ssize_t xino_fwrite_wkq(vfs_writef_t func, struct file *file, void *buf,
+			       size_t size, loff_t *pos);
+
+static ssize_t do_xino_fwrite(vfs_writef_t func, struct file *file, void *kbuf,
+			      size_t size, loff_t *pos)
+{
+	ssize_t err;
+	mm_segment_t oldfs;
+	union {
+		void *k;
+		const char __user *u;
+	} buf;
+	int i;
+	const int prevent_endless = 10;
+
+	i = 0;
+	buf.k = kbuf;
+	oldfs = get_fs();
+	set_fs(KERNEL_DS);
+	do {
+		err = func(file, buf.u, size, pos);
+		if (err == -EINTR
+		    && !au_wkq_test()
+		    && fatal_signal_pending(current)) {
+			set_fs(oldfs);
+			err = xino_fwrite_wkq(func, file, kbuf, size, pos);
+			BUG_ON(err == -EINTR);
+			oldfs = get_fs();
+			set_fs(KERNEL_DS);
+		}
+	} while (i++ < prevent_endless
+		 && (err == -EAGAIN || err == -EINTR));
+	set_fs(oldfs);
+
+#if 0 /* reserved for future use */
+	if (err > 0)
+		fsnotify_modify(file->f_path.dentry);
+#endif
+
+	return err;
+}
+
+struct do_xino_fwrite_args {
+	ssize_t *errp;
+	vfs_writef_t func;
+	struct file *file;
+	void *buf;
+	size_t size;
+	loff_t *pos;
+};
+
+static void call_do_xino_fwrite(void *args)
+{
+	struct do_xino_fwrite_args *a = args;
+	*a->errp = do_xino_fwrite(a->func, a->file, a->buf, a->size, a->pos);
+}
+
+static ssize_t xino_fwrite_wkq(vfs_writef_t func, struct file *file, void *buf,
+			       size_t size, loff_t *pos)
+{
+	ssize_t err;
+	int wkq_err;
+	struct do_xino_fwrite_args args = {
+		.errp	= &err,
+		.func	= func,
+		.file	= file,
+		.buf	= buf,
+		.size	= size,
+		.pos	= pos
+	};
+
+	/*
+	 * it breaks RLIMIT_FSIZE and normal user's limit,
+	 * users should care about quota and real 'filesystem full.'
+	 */
+	wkq_err = au_wkq_wait(call_do_xino_fwrite, &args);
+	if (unlikely(wkq_err))
+		err = wkq_err;
+
+	return err;
+}
+
+ssize_t xino_fwrite(vfs_writef_t func, struct file *file, void *buf,
+		    size_t size, loff_t *pos)
+{
+	ssize_t err;
+
+	if (rlimit(RLIMIT_FSIZE) == RLIM_INFINITY) {
+		lockdep_off();
+		err = do_xino_fwrite(func, file, buf, size, pos);
+		lockdep_on();
+	} else {
+		lockdep_off();
+		err = xino_fwrite_wkq(func, file, buf, size, pos);
+		lockdep_on();
+	}
+
+	return err;
+}
+
+/* ---------------------------------------------------------------------- */
+
+/*
+ * inode number bitmap
+ */
+static const int page_bits = (int)PAGE_SIZE * BITS_PER_BYTE;
+static ino_t xib_calc_ino(unsigned long pindex, int bit)
+{
+	ino_t ino;
+
+	AuDebugOn(bit < 0 || page_bits <= bit);
+	ino = AUFS_FIRST_INO + pindex * page_bits + bit;
+	return ino;
+}
+
+static void xib_calc_bit(ino_t ino, unsigned long *pindex, int *bit)
+{
+	AuDebugOn(ino < AUFS_FIRST_INO);
+	ino -= AUFS_FIRST_INO;
+	*pindex = ino / page_bits;
+	*bit = ino % page_bits;
+}
+
+static int xib_pindex(struct super_block *sb, unsigned long pindex)
+{
+	int err;
+	loff_t pos;
+	ssize_t sz;
+	struct au_sbinfo *sbinfo;
+	struct file *xib;
+	unsigned long *p;
+
+	sbinfo = au_sbi(sb);
+	MtxMustLock(&sbinfo->si_xib_mtx);
+	AuDebugOn(pindex > ULONG_MAX / PAGE_SIZE
+		  || !au_opt_test(sbinfo->si_mntflags, XINO));
+
+	if (pindex == sbinfo->si_xib_last_pindex)
+		return 0;
+
+	xib = sbinfo->si_xib;
+	p = sbinfo->si_xib_buf;
+	pos = sbinfo->si_xib_last_pindex;
+	pos *= PAGE_SIZE;
+	sz = xino_fwrite(sbinfo->si_xwrite, xib, p, PAGE_SIZE, &pos);
+	if (unlikely(sz != PAGE_SIZE))
+		goto out;
+
+	pos = pindex;
+	pos *= PAGE_SIZE;
+	if (vfsub_f_size_read(xib) >= pos + PAGE_SIZE)
+		sz = xino_fread(sbinfo->si_xread, xib, p, PAGE_SIZE, &pos);
+	else {
+		memset(p, 0, PAGE_SIZE);
+		sz = xino_fwrite(sbinfo->si_xwrite, xib, p, PAGE_SIZE, &pos);
+	}
+	if (sz == PAGE_SIZE) {
+		sbinfo->si_xib_last_pindex = pindex;
+		return 0; /* success */
+	}
+
+out:
+	AuIOErr1("write failed (%zd)\n", sz);
+	err = sz;
+	if (sz >= 0)
+		err = -EIO;
+	return err;
+}
+
+static void au_xib_clear_bit(struct inode *inode)
+{
+	int err, bit;
+	unsigned long pindex;
+	struct super_block *sb;
+	struct au_sbinfo *sbinfo;
+
+	AuDebugOn(inode->i_nlink);
+
+	sb = inode->i_sb;
+	xib_calc_bit(inode->i_ino, &pindex, &bit);
+	AuDebugOn(page_bits <= bit);
+	sbinfo = au_sbi(sb);
+	mutex_lock(&sbinfo->si_xib_mtx);
+	err = xib_pindex(sb, pindex);
+	if (!err) {
+		clear_bit(bit, sbinfo->si_xib_buf);
+		sbinfo->si_xib_next_bit = bit;
+	}
+	mutex_unlock(&sbinfo->si_xib_mtx);
+}
+
+/* ---------------------------------------------------------------------- */
+
+/*
+ * truncate a xino bitmap file
+ */
+
+/* todo: slow */
+static int do_xib_restore(struct super_block *sb, struct file *file, void *page)
+{
+	int err, bit;
+	ssize_t sz;
+	unsigned long pindex;
+	loff_t pos, pend;
+	struct au_sbinfo *sbinfo;
+	vfs_readf_t func;
+	ino_t *ino;
+	unsigned long *p;
+
+	err = 0;
+	sbinfo = au_sbi(sb);
+	MtxMustLock(&sbinfo->si_xib_mtx);
+	p = sbinfo->si_xib_buf;
+	func = sbinfo->si_xread;
+	pend = vfsub_f_size_read(file);
+	pos = 0;
+	while (pos < pend) {
+		sz = xino_fread(func, file, page, PAGE_SIZE, &pos);
+		err = sz;
+		if (unlikely(sz <= 0))
+			goto out;
+
+		err = 0;
+		for (ino = page; sz > 0; ino++, sz -= sizeof(ino)) {
+			if (unlikely(*ino < AUFS_FIRST_INO))
+				continue;
+
+			xib_calc_bit(*ino, &pindex, &bit);
+			AuDebugOn(page_bits <= bit);
+			err = xib_pindex(sb, pindex);
+			if (!err)
+				set_bit(bit, p);
+			else
+				goto out;
+		}
+	}
+
+out:
+	return err;
+}
+
+static int xib_restore(struct super_block *sb)
+{
+	int err, i;
+	unsigned int nfile;
+	aufs_bindex_t bindex, bbot;
+	void *page;
+	struct au_branch *br;
+	struct au_xino *xi;
+	struct file *file;
+
+	err = -ENOMEM;
+	page = (void *)__get_free_page(GFP_NOFS);
+	if (unlikely(!page))
+		goto out;
+
+	err = 0;
+	bbot = au_sbbot(sb);
+	for (bindex = 0; !err && bindex <= bbot; bindex++)
+		if (!bindex || is_sb_shared(sb, bindex, bindex - 1) < 0) {
+			br = au_sbr(sb, bindex);
+			xi = br->br_xino;
+			nfile = xi->xi_nfile;
+			for (i = 0; i < nfile; i++) {
+				file = au_xino_file(xi, i);
+				if (file)
+					err = do_xib_restore(sb, file, page);
+			}
+		} else
+			AuDbg("skip shared b%d\n", bindex);
+	free_page((unsigned long)page);
+
+out:
+	return err;
+}
+
+int au_xib_trunc(struct super_block *sb)
+{
+	int err;
+	ssize_t sz;
+	loff_t pos;
+	struct au_sbinfo *sbinfo;
+	unsigned long *p;
+	struct file *file;
+
+	SiMustWriteLock(sb);
+
+	err = 0;
+	sbinfo = au_sbi(sb);
+	if (!au_opt_test(sbinfo->si_mntflags, XINO))
+		goto out;
+
+	file = sbinfo->si_xib;
+	if (vfsub_f_size_read(file) <= PAGE_SIZE)
+		goto out;
+