[PATCH] hfs: NLS support

This adds NLS support to HFS.  Using the kernel options iocharset and codepage
it's possible to map the disk encoding to a local mapping.  If these options
are not used, it falls back to the old direct mapping.

Signed-off-by: Roman Zippel <zippel@linux-m68k.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

8 files changed, 176 insertions, 46 deletions

diff --git a/fs/hfs/catalog.c b/fs/hfs/catalog.c
index 65dedef..2fcd679 100644
--- a/fs/hfs/catalog.c
+++ b/fs/hfs/catalog.c
@@ -20,12 +20,12 @@
  *
  * Given the ID of the parent and the name build a search key.
  */
-void hfs_cat_build_key(btree_key *key, u32 parent, struct qstr *name)
+void hfs_cat_build_key(struct super_block *sb, btree_key *key, u32 parent, struct qstr *name)
 {
 	key->cat.reserved = 0;
 	key->cat.ParID = cpu_to_be32(parent);
 	if (name) {
-		hfs_triv2mac(&key->cat.CName, name);
+		hfs_asc2mac(sb, &key->cat.CName, name);
 		key->key_len = 6 + key->cat.CName.len;
 	} else {
 		memset(&key->cat.CName, 0, sizeof(struct hfs_name));
@@ -62,13 +62,14 @@ static int hfs_cat_build_record(hfs_cat_rec *rec, u32 cnid, struct inode *inode)
 	}
 }
 
-static int hfs_cat_build_thread(hfs_cat_rec *rec, int type,
+static int hfs_cat_build_thread(struct super_block *sb,
+				hfs_cat_rec *rec, int type,
 				u32 parentid, struct qstr *name)
 {
 	rec->type = type;
 	memset(rec->thread.reserved, 0, sizeof(rec->thread.reserved));
 	rec->thread.ParID = cpu_to_be32(parentid);
-	hfs_triv2mac(&rec->thread.CName, name);
+	hfs_asc2mac(sb, &rec->thread.CName, name);
 	return sizeof(struct hfs_cat_thread);
 }
 
@@ -93,8 +94,8 @@ int hfs_cat_create(u32 cnid, struct inode *dir, struct qstr *str, struct inode *
 	sb = dir->i_sb;
 	hfs_find_init(HFS_SB(sb)->cat_tree, &fd);
 
-	hfs_cat_build_key(fd.search_key, cnid, NULL);
-	entry_size = hfs_cat_build_thread(&entry, S_ISDIR(inode->i_mode) ?
+	hfs_cat_build_key(sb, fd.search_key, cnid, NULL);
+	entry_size = hfs_cat_build_thread(sb, &entry, S_ISDIR(inode->i_mode) ?
 			HFS_CDR_THD : HFS_CDR_FTH,
 			dir->i_ino, str);
 	err = hfs_brec_find(&fd);
@@ -107,7 +108,7 @@ int hfs_cat_create(u32 cnid, struct inode *dir, struct qstr *str, struct inode *
 	if (err)
 		goto err2;
 
-	hfs_cat_build_key(fd.search_key, dir->i_ino, str);
+	hfs_cat_build_key(sb, fd.search_key, dir->i_ino, str);
 	entry_size = hfs_cat_build_record(&entry, cnid, inode);
 	err = hfs_brec_find(&fd);
 	if (err != -ENOENT) {
@@ -127,7 +128,7 @@ int hfs_cat_create(u32 cnid, struct inode *dir, struct qstr *str, struct inode *
 	return 0;
 
 err1:
-	hfs_cat_build_key(fd.search_key, cnid, NULL);
+	hfs_cat_build_key(sb, fd.search_key, cnid, NULL);
 	if (!hfs_brec_find(&fd))
 		hfs_brec_remove(&fd);
 err2:
@@ -176,7 +177,7 @@ int hfs_cat_find_brec(struct super_block *sb, u32 cnid,
 	hfs_cat_rec rec;
 	int res, len, type;
 
-	hfs_cat_build_key(fd->search_key, cnid, NULL);
+	hfs_cat_build_key(sb, fd->search_key, cnid, NULL);
 	res = hfs_brec_read(fd, &rec, sizeof(rec));
 	if (res)
 		return res;
@@ -211,7 +212,7 @@ int hfs_cat_delete(u32 cnid, struct inode *dir, struct qstr *str)
 	sb = dir->i_sb;
 	hfs_find_init(HFS_SB(sb)->cat_tree, &fd);
 
-	hfs_cat_build_key(fd.search_key, dir->i_ino, str);
+	hfs_cat_build_key(sb, fd.search_key, dir->i_ino, str);
 	res = hfs_brec_find(&fd);
 	if (res)
 		goto out;
@@ -239,7 +240,7 @@ int hfs_cat_delete(u32 cnid, struct inode *dir, struct qstr *str)
 	if (res)
 		goto out;
 
-	hfs_cat_build_key(fd.search_key, cnid, NULL);
+	hfs_cat_build_key(sb, fd.search_key, cnid, NULL);
 	res = hfs_brec_find(&fd);
 	if (!res) {
 		res = hfs_brec_remove(&fd);
@@ -280,7 +281,7 @@ int hfs_cat_move(u32 cnid, struct inode *src_dir, struct qstr *src_name,
 	dst_fd = src_fd;
 
 	/* find the old dir entry and read the data */
-	hfs_cat_build_key(src_fd.search_key, src_dir->i_ino, src_name);
+	hfs_cat_build_key(sb, src_fd.search_key, src_dir->i_ino, src_name);
 	err = hfs_brec_find(&src_fd);
 	if (err)
 		goto out;
@@ -289,7 +290,7 @@ int hfs_cat_move(u32 cnid, struct inode *src_dir, struct qstr *src_name,
 			    src_fd.entrylength);
 
 	/* create new dir entry with the data from the old entry */
-	hfs_cat_build_key(dst_fd.search_key, dst_dir->i_ino, dst_name);
+	hfs_cat_build_key(sb, dst_fd.search_key, dst_dir->i_ino, dst_name);
 	err = hfs_brec_find(&dst_fd);
 	if (err != -ENOENT) {
 		if (!err)
@@ -305,7 +306,7 @@ int hfs_cat_move(u32 cnid, struct inode *src_dir, struct qstr *src_name,
 	mark_inode_dirty(dst_dir);
 
 	/* finally remove the old entry */
-	hfs_cat_build_key(src_fd.search_key, src_dir->i_ino, src_name);
+	hfs_cat_build_key(sb, src_fd.search_key, src_dir->i_ino, src_name);
 	err = hfs_brec_find(&src_fd);
 	if (err)
 		goto out;
@@ -321,7 +322,7 @@ int hfs_cat_move(u32 cnid, struct inode *src_dir, struct qstr *src_name,
 		goto out;
 
 	/* remove old thread entry */
-	hfs_cat_build_key(src_fd.search_key, cnid, NULL);
+	hfs_cat_build_key(sb, src_fd.search_key, cnid, NULL);
 	err = hfs_brec_find(&src_fd);
 	if (err)
 		goto out;
@@ -330,8 +331,8 @@ int hfs_cat_move(u32 cnid, struct inode *src_dir, struct qstr *src_name,
 		goto out;
 
 	/* create new thread entry */
-	hfs_cat_build_key(dst_fd.search_key, cnid, NULL);
-	entry_size = hfs_cat_build_thread(&entry, type == HFS_CDR_FIL ? HFS_CDR_FTH : HFS_CDR_THD,
+	hfs_cat_build_key(sb, dst_fd.search_key, cnid, NULL);
+	entry_size = hfs_cat_build_thread(sb, &entry, type == HFS_CDR_FIL ? HFS_CDR_FTH : HFS_CDR_THD,
 					dst_dir->i_ino, dst_name);
 	err = hfs_brec_find(&dst_fd);
 	if (err != -ENOENT) {
diff --git a/fs/hfs/dir.c b/fs/hfs/dir.c
index c559982..e1f24be 100644
--- a/fs/hfs/dir.c
+++ b/fs/hfs/dir.c
@@ -28,7 +28,7 @@ static struct dentry *hfs_lookup(struct inode *dir, struct dentry *dentry,
 	dentry->d_op = &hfs_dentry_operations;
 
 	hfs_find_init(HFS_SB(dir->i_sb)->cat_tree, &fd);
-	hfs_cat_build_key(fd.search_key, dir->i_ino, &dentry->d_name);
+	hfs_cat_build_key(dir->i_sb, fd.search_key, dir->i_ino, &dentry->d_name);
 	res = hfs_brec_read(&fd, &rec, sizeof(rec));
 	if (res) {
 		hfs_find_exit(&fd);
@@ -56,7 +56,7 @@ static int hfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 	struct inode *inode = filp->f_dentry->d_inode;
 	struct super_block *sb = inode->i_sb;
 	int len, err;
-	char strbuf[HFS_NAMELEN + 1];
+	char strbuf[HFS_MAX_NAMELEN];
 	union hfs_cat_rec entry;
 	struct hfs_find_data fd;
 	struct hfs_readdir_data *rd;
@@ -66,7 +66,7 @@ static int hfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 		return 0;
 
 	hfs_find_init(HFS_SB(sb)->cat_tree, &fd);
-	hfs_cat_build_key(fd.search_key, inode->i_ino, NULL);
+	hfs_cat_build_key(sb, fd.search_key, inode->i_ino, NULL);
 	err = hfs_brec_find(&fd);
 	if (err)
 		goto out;
@@ -111,7 +111,7 @@ static int hfs_readdir(struct file *filp, void *dirent, filldir_t filldir)
 		}
 		hfs_bnode_read(fd.bnode, &entry, fd.entryoffset, fd.entrylength);
 		type = entry.type;
-		len = hfs_mac2triv(strbuf, &fd.key->cat.CName);
+		len = hfs_mac2asc(sb, strbuf, &fd.key->cat.CName);
 		if (type == HFS_CDR_DIR) {
 			if (fd.entrylength < sizeof(struct hfs_cat_dir)) {
 				printk("HFS: small dir entry\n");
@@ -307,7 +307,8 @@ static int hfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 			   old_dir, &old_dentry->d_name,
 			   new_dir, &new_dentry->d_name);
 	if (!res)
-		hfs_cat_build_key((btree_key *)&HFS_I(old_dentry->d_inode)->cat_key,
+		hfs_cat_build_key(old_dir->i_sb,
+				  (btree_key *)&HFS_I(old_dentry->d_inode)->cat_key,
 				  new_dir->i_ino, &new_dentry->d_name);
 	return res;
 }
diff --git a/fs/hfs/hfs.h b/fs/hfs/hfs.h
index df6b33a..88099ab 100644
--- a/fs/hfs/hfs.h
+++ b/fs/hfs/hfs.h
@@ -25,6 +25,7 @@
 #define HFS_SECTOR_SIZE		512    /* size of an HFS sector */
 #define HFS_SECTOR_SIZE_BITS	9      /* log_2(HFS_SECTOR_SIZE) */
 #define HFS_NAMELEN		31     /* maximum length of an HFS filename */
+#define HFS_MAX_NAMELEN		128
 #define HFS_MAX_VALENCE		32767U
 
 /* Meanings of the drAtrb field of the MDB,
diff --git a/fs/hfs/hfs_fs.h b/fs/hfs/hfs_fs.h
index 0dc8ef8..aae019a 100644
--- a/fs/hfs/hfs_fs.h
+++ b/fs/hfs/hfs_fs.h
@@ -141,6 +141,8 @@ struct hfs_sb_info {
 
 	int session, part;
 
+	struct nls_table *nls_io, *nls_disk;
+
 	struct semaphore bitmap_lock;
 
 	unsigned long flags;
@@ -168,7 +170,7 @@ extern int hfs_cat_create(u32, struct inode *, struct qstr *, struct inode *);
 extern int hfs_cat_delete(u32, struct inode *, struct qstr *);
 extern int hfs_cat_move(u32, struct inode *, struct qstr *,
 			struct inode *, struct qstr *);
-extern void hfs_cat_build_key(btree_key *, u32, struct qstr *);
+extern void hfs_cat_build_key(struct super_block *, btree_key *, u32, struct qstr *);
 
 /* dir.c */
 extern struct file_operations hfs_dir_operations;
@@ -222,8 +224,8 @@ extern int hfs_strcmp(const unsigned char *, unsigned int,
 extern int hfs_compare_dentry(struct dentry *, struct qstr *, struct qstr *);
 
 /* trans.c */
-extern void hfs_triv2mac(struct hfs_name *, struct qstr *);
-extern int hfs_mac2triv(char *, const struct hfs_name *);
+extern void hfs_asc2mac(struct super_block *, struct hfs_name *, struct qstr *);
+extern int hfs_mac2asc(struct super_block *, char *, const struct hfs_name *);
 
 extern struct timezone sys_tz;
 
diff --git a/fs/hfs/inode.c b/fs/hfs/inode.c
index 7519123..f1570b9 100644
--- a/fs/hfs/inode.c
+++ b/fs/hfs/inode.c
@@ -160,7 +160,7 @@ struct inode *hfs_new_inode(struct inode *dir, struct qstr *name, int mode)
 
 	init_MUTEX(&HFS_I(inode)->extents_lock);
 	INIT_LIST_HEAD(&HFS_I(inode)->open_dir_list);
-	hfs_cat_build_key((btree_key *)&HFS_I(inode)->cat_key, dir->i_ino, name);
+	hfs_cat_build_key(sb, (btree_key *)&HFS_I(inode)->cat_key, dir->i_ino, name);
 	inode->i_ino = HFS_SB(sb)->next_id++;
 	inode->i_mode = mode;
 	inode->i_uid = current->fsuid;
diff --git a/fs/hfs/mdb.c b/fs/hfs/mdb.c
index 217e32f..0a473f7 100644
--- a/fs/hfs/mdb.c
+++ b/fs/hfs/mdb.c
@@ -10,6 +10,7 @@
 
 #include <linux/cdrom.h>
 #include <linux/genhd.h>
+#include <linux/nls.h>
 
 #include "hfs_fs.h"
 #include "btree.h"
@@ -343,6 +344,11 @@ void hfs_mdb_put(struct super_block *sb)
 	brelse(HFS_SB(sb)->mdb_bh);
 	brelse(HFS_SB(sb)->alt_mdb_bh);
 
+	if (HFS_SB(sb)->nls_io)
+		unload_nls(HFS_SB(sb)->nls_io);
+	if (HFS_SB(sb)->nls_disk)
+		unload_nls(HFS_SB(sb)->nls_disk);
+
 	kfree(HFS_SB(sb));
 	sb->s_fs_info = NULL;
 }
diff --git a/fs/hfs/super.c b/fs/hfs/super.c
index fd78414..c5074ae 100644
--- a/fs/hfs/super.c
+++ b/fs/hfs/super.c
@@ -17,6 +17,7 @@
 #include <linux/blkdev.h>
 #include <linux/mount.h>
 #include <linux/init.h>
+#include <linux/nls.h>
 #include <linux/parser.h>
 #include <linux/seq_file.h>
 #include <linux/vfs.h>
@@ -130,6 +131,10 @@ static int hfs_show_options(struct seq_file *seq, struct vfsmount *mnt)
 		seq_printf(seq, ",part=%u", sbi->part);
 	if (sbi->session >= 0)
 		seq_printf(seq, ",session=%u", sbi->session);
+	if (sbi->nls_disk)
+		seq_printf(seq, ",codepage=%s", sbi->nls_disk->charset);
+	if (sbi->nls_io)
+		seq_printf(seq, ",iocharset=%s", sbi->nls_io->charset);
 	if (sbi->s_quiet)
 		seq_printf(seq, ",quiet");
 	return 0;
@@ -163,6 +168,7 @@ static struct super_operations hfs_super_operations = {
 enum {
 	opt_uid, opt_gid, opt_umask, opt_file_umask, opt_dir_umask,
 	opt_part, opt_session, opt_type, opt_creator, opt_quiet,
+	opt_codepage, opt_iocharset,
 	opt_err
 };
 
@@ -177,6 +183,8 @@ static match_table_t tokens = {
 	{ opt_type, "type=%s" },
 	{ opt_creator, "creator=%s" },
 	{ opt_quiet, "quiet" },
+	{ opt_codepage, "codepage=%s" },
+	{ opt_iocharset, "iocharset=%s" },
 	{ opt_err, NULL }
 };
 
@@ -282,11 +290,46 @@ static int parse_options(char *options, struct hfs_sb_info *hsb)
 		case opt_quiet:
 			hsb->s_quiet = 1;
 			break;
+		case opt_codepage:
+			if (hsb->nls_disk) {
+				printk("HFS+-fs: unable to change codepage\n");
+				return 0;
+			}
+			p = match_strdup(&args[0]);
+			hsb->nls_disk = load_nls(p);
+			if (!hsb->nls_disk) {
+				printk("HFS+-fs: unable to load codepage \"%s\"\n", p);
+				kfree(p);
+				return 0;
+			}
+			kfree(p);
+			break;
+		case opt_iocharset:
+			if (hsb->nls_io) {
+				printk("HFS: unable to change iocharset\n");
+				return 0;
+			}
+			p = match_strdup(&args[0]);
+			hsb->nls_io = load_nls(p);
+			if (!hsb->nls_io) {
+				printk("HFS: unable to load iocharset \"%s\"\n", p);
+				kfree(p);
+				return 0;
+			}
+			kfree(p);
+			break;
 		default:
 			return 0;
 		}
 	}
 
+	if (hsb->nls_disk && !hsb->nls_io) {
+		hsb->nls_io = load_nls_default();
+		if (!hsb->nls_io) {
+			printk("HFS: unable to load default iocharset\n");
+			return 0;
+		}
+	}
 	hsb->s_dir_umask &= 0777;
 	hsb->s_file_umask &= 0577;
 
diff --git a/fs/hfs/trans.c b/fs/hfs/trans.c
index fb9720a..e673a88 100644
--- a/fs/hfs/trans.c
+++ b/fs/hfs/trans.c
@@ -9,12 +9,15 @@
  * with ':' vs. '/' as the path-element separator.
  */
 
+#include <linux/types.h>
+#include <linux/nls.h>
+
 #include "hfs_fs.h"
 
 /*================ Global functions ================*/
 
 /*
- * hfs_mac2triv()
+ * hfs_mac2asc()
  *
  * Given a 'Pascal String' (a string preceded by a length byte) in
  * the Macintosh character set produce the corresponding filename using
@@ -27,23 +30,58 @@
  * by ':' which never appears in HFS filenames.	 All other characters
  * are passed unchanged from input to output.
  */
-int hfs_mac2triv(char *out, const struct hfs_name *in)
+int hfs_mac2asc(struct super_block *sb, char *out, const struct hfs_name *in)
 {
-	const char *p;
-	char c;
-	int i, len;
+	struct nls_table *nls_disk = HFS_SB(sb)->nls_disk;
+	struct nls_table *nls_io = HFS_SB(sb)->nls_io;
+	const char *src;
+	char *dst;
+	int srclen, dstlen, size;
+
+	src = in->name;
+	srclen = in->len;
+	dst = out;
+	dstlen = HFS_MAX_NAMELEN;
+	if (nls_io) {
+		wchar_t ch;
 
-	len = in->len;
-	p = in->name;
-	for (i = 0; i < len; i++) {
-		c = *p++;
-		*out++ = c == '/' ? ':' : c;
+		while (srclen > 0) {
+			if (nls_disk) {
+				size = nls_disk->char2uni(src, srclen, &ch);
+				if (size <= 0) {
+					ch = '?';
+					size = 1;
+				}
+				src += size;
+				srclen -= size;
+			} else {
+				ch = *src++;
+				srclen--;
+			}
+			if (ch == '/')
+				ch = ':';
+			size = nls_io->uni2char(ch, dst, dstlen);
+			if (size < 0) {
+				if (size == -ENAMETOOLONG)
+					goto out;
+				*dst = '?';
+				size = 1;
+			}
+			dst += size;
+			dstlen -= size;
+		}
+	} else {
+		char ch;
+
+		while (--srclen >= 0)
+			*dst++ = (ch = *src++) == '/' ? ':' : ch;
 	}
-	return i;
+out:
+	return dst - out;
 }
 
 /*
- * hfs_triv2mac()
+ * hfs_asc2mac()
  *
  * Given an ASCII string (not null-terminated) and its length,
  * generate the corresponding filename in the Macintosh character set
@@ -54,19 +92,57 @@ int hfs_mac2triv(char *out, const struct hfs_name *in)
  * This routine is a inverse to hfs_mac2triv().
  * A ':' is replaced by a '/'.
  */
-void hfs_triv2mac(struct hfs_name *out, struct qstr *in)
+void hfs_asc2mac(struct super_block *sb, struct hfs_name *out, struct qstr *in)
 {
+	struct nls_table *nls_disk = HFS_SB(sb)->nls_disk;
+	struct nls_table *nls_io = HFS_SB(sb)->nls_io;
 	const char *src;
-	char *dst, c;
-	int i, len;
+	char *dst;
+	int srclen, dstlen, size;
 
-	out->len = len = min((unsigned int)HFS_NAMELEN, in->len);
 	src = in->name;
+	srclen = in->len;
 	dst = out->name;
-	for (i = 0; i < len; i++) {
-		c = *src++;
-		*dst++ = c == ':' ? '/' : c;
+	dstlen = HFS_NAMELEN;
+	if (nls_io) {
+		wchar_t ch;
+
+		while (srclen > 0) {
+			size = nls_io->char2uni(src, srclen, &ch);
+			if (size < 0) {
+				ch = '?';
+				size = 1;
+			}
+			src += size;
+			srclen -= size;
+			if (ch == ':')
+				ch = '/';
+			if (nls_disk) {
+				size = nls_disk->uni2char(ch, dst, dstlen);
+				if (size < 0) {
+					if (size == -ENAMETOOLONG)
+						goto out;
+					*dst = '?';
+					size = 1;
+				}
+				dst += size;
+				dstlen -= size;
+			} else {
+				*dst++ = ch > 0xff ? '?' : ch;
+				dstlen--;
+			}
+		}
+	} else {
+		char ch;
+
+		if (dstlen > srclen)
+			dstlen = srclen;
+		while (--dstlen >= 0)
+			*dst++ = (ch = *src++) == ':' ? '/' : ch;
 	}
-	for (; i < HFS_NAMELEN; i++)
+out:
+	out->len = dst - (char *)out->name;
+	dstlen = HFS_NAMELEN - out->len;
+	while (--dstlen >= 0)
 		*dst++ = 0;
 }