]> git.baikalelectronics.ru Git - kernel.git/commitdiff
staging: exfat: remove symlink feature.
authorTetsuhiro Kohada <Kohada.Tetsuhiro@dc.MitsubishiElectric.co.jp>
Wed, 19 Feb 2020 05:57:27 +0000 (14:57 +0900)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sun, 23 Feb 2020 18:17:04 +0000 (19:17 +0100)
Remove symlink feature completely.

Becouse
-Uses reserved areas(defined in the Microsoft exfat specification), causing future incompatibilities.
-Not described in Microsoft exfat specifications or SD standards.
-For REMOVABLE media, causes incompatibility with other implementations.
-Not supported by other major exfat drivers.
-Not implemented symlink feature in linux FAT/VFAT.

Remove this feature completely because of serious media compatibility issues.
(Can't enable even with CONFIG)

If you have any questions about this patch, please let me know.

Reviewed-by: Takahiro Mori <Mori.Takahiro@ab.MitsubishiElectric.co.jp>
Acked-by: Valdis Kletnieks <valdis.kletnieks@vt.edu>
Signed-off-by: Tetsuhiro Kohada <Kohada.Tetsuhiro@dc.MitsubishiElectric.co.jp>
Link: https://lore.kernel.org/r/20200219055727.12867-1-Kohada.Tetsuhiro@dc.MitsubishiElectric.co.jp
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/staging/exfat/exfat_super.c

index b398114c26042ffdc384cc702ae67dc32baad649..c7bc07e91c450a755189900295db1abbc9448d32 100644 (file)
@@ -660,375 +660,6 @@ out:
        return ret;
 }
 
-static int ffsReadFile(struct inode *inode, struct file_id_t *fid, void *buffer,
-                      u64 count, u64 *rcount)
-{
-       s32 offset, sec_offset, clu_offset;
-       u32 clu;
-       int ret = 0;
-       sector_t LogSector;
-       u64 oneblkread, read_bytes;
-       struct buffer_head *tmp_bh = NULL;
-       struct super_block *sb = inode->i_sb;
-       struct fs_info_t *p_fs = &(EXFAT_SB(sb)->fs_info);
-       struct bd_info_t *p_bd = &(EXFAT_SB(sb)->bd_info);
-
-       /* check the validity of the given file id */
-       if (!fid)
-               return -EINVAL;
-
-       /* check the validity of pointer parameters */
-       if (!buffer)
-               return -EINVAL;
-
-       /* acquire the lock for file system critical section */
-       mutex_lock(&p_fs->v_mutex);
-
-       /* check if the given file ID is opened */
-       if (fid->type != TYPE_FILE) {
-               ret = -EPERM;
-               goto out;
-       }
-
-       if (fid->rwoffset > fid->size)
-               fid->rwoffset = fid->size;
-
-       if (count > (fid->size - fid->rwoffset))
-               count = fid->size - fid->rwoffset;
-
-       if (count == 0) {
-               if (rcount)
-                       *rcount = 0;
-               ret = 0;
-               goto out;
-       }
-
-       read_bytes = 0;
-
-       while (count > 0) {
-               clu_offset = (s32)(fid->rwoffset >> p_fs->cluster_size_bits);
-               clu = fid->start_clu;
-
-               if (fid->flags == 0x03) {
-                       clu += clu_offset;
-               } else {
-                       /* hint information */
-                       if ((clu_offset > 0) && (fid->hint_last_off > 0) &&
-                           (clu_offset >= fid->hint_last_off)) {
-                               clu_offset -= fid->hint_last_off;
-                               clu = fid->hint_last_clu;
-                       }
-
-                       while (clu_offset > 0) {
-                               /* clu = exfat_fat_read(sb, clu); */
-                               if (exfat_fat_read(sb, clu, &clu) == -1) {
-                                       ret = -EIO;
-                                       goto out;
-                               }
-
-                               clu_offset--;
-                       }
-               }
-
-               /* hint information */
-               fid->hint_last_off = (s32)(fid->rwoffset >>
-                                          p_fs->cluster_size_bits);
-               fid->hint_last_clu = clu;
-
-               /* byte offset in cluster */
-               offset = (s32)(fid->rwoffset & (p_fs->cluster_size - 1));
-
-               /* sector offset in cluster */
-               sec_offset = offset >> p_bd->sector_size_bits;
-
-               /* byte offset in sector */
-               offset &= p_bd->sector_size_mask;
-
-               LogSector = START_SECTOR(clu) + sec_offset;
-
-               oneblkread = (u64)(p_bd->sector_size - offset);
-               if (oneblkread > count)
-                       oneblkread = count;
-
-               if ((offset == 0) && (oneblkread == p_bd->sector_size)) {
-                       if (sector_read(sb, LogSector, &tmp_bh, 1) !=
-                           0)
-                               goto err_out;
-                       memcpy((char *)buffer + read_bytes,
-                              (char *)tmp_bh->b_data, (s32)oneblkread);
-               } else {
-                       if (sector_read(sb, LogSector, &tmp_bh, 1) !=
-                           0)
-                               goto err_out;
-                       memcpy((char *)buffer + read_bytes,
-                              (char *)tmp_bh->b_data + offset,
-                              (s32)oneblkread);
-               }
-               count -= oneblkread;
-               read_bytes += oneblkread;
-               fid->rwoffset += oneblkread;
-       }
-       brelse(tmp_bh);
-
-/* How did this ever work and not leak a brlse()?? */
-err_out:
-       /* set the size of read bytes */
-       if (rcount)
-               *rcount = read_bytes;
-
-       if (p_fs->dev_ejected)
-               ret = -EIO;
-
-out:
-       /* release the lock for file system critical section */
-       mutex_unlock(&p_fs->v_mutex);
-
-       return ret;
-}
-
-static int ffsWriteFile(struct inode *inode, struct file_id_t *fid,
-                       void *buffer, u64 count, u64 *wcount)
-{
-       bool modified = false;
-       s32 offset, sec_offset, clu_offset;
-       s32 num_clusters, num_alloc, num_alloced = (s32)~0;
-       int ret = 0;
-       u32 clu, last_clu;
-       sector_t LogSector;
-       u64 oneblkwrite, write_bytes;
-       struct chain_t new_clu;
-       struct timestamp_t tm;
-       struct dentry_t *ep, *ep2;
-       struct entry_set_cache_t *es = NULL;
-       struct buffer_head *tmp_bh = NULL;
-       struct super_block *sb = inode->i_sb;
-       struct fs_info_t *p_fs = &(EXFAT_SB(sb)->fs_info);
-       struct bd_info_t *p_bd = &(EXFAT_SB(sb)->bd_info);
-
-       /* check the validity of the given file id */
-       if (!fid)
-               return -EINVAL;
-
-       /* check the validity of pointer parameters */
-       if (!buffer)
-               return -EINVAL;
-
-       /* acquire the lock for file system critical section */
-       mutex_lock(&p_fs->v_mutex);
-
-       /* check if the given file ID is opened */
-       if (fid->type != TYPE_FILE) {
-               ret = -EPERM;
-               goto out;
-       }
-
-       if (fid->rwoffset > fid->size)
-               fid->rwoffset = fid->size;
-
-       if (count == 0) {
-               if (wcount)
-                       *wcount = 0;
-               ret = 0;
-               goto out;
-       }
-
-       fs_set_vol_flags(sb, VOL_DIRTY);
-
-       if (fid->size == 0)
-               num_clusters = 0;
-       else
-               num_clusters = (s32)((fid->size - 1) >>
-                                    p_fs->cluster_size_bits) + 1;
-
-       write_bytes = 0;
-
-       while (count > 0) {
-               clu_offset = (s32)(fid->rwoffset >> p_fs->cluster_size_bits);
-               clu = fid->start_clu;
-               last_clu = fid->start_clu;
-
-               if (fid->flags == 0x03) {
-                       if ((clu_offset > 0) && (clu != CLUSTER_32(~0))) {
-                               last_clu += clu_offset - 1;
-
-                               if (clu_offset == num_clusters)
-                                       clu = CLUSTER_32(~0);
-                               else
-                                       clu += clu_offset;
-                       }
-               } else {
-                       /* hint information */
-                       if ((clu_offset > 0) && (fid->hint_last_off > 0) &&
-                           (clu_offset >= fid->hint_last_off)) {
-                               clu_offset -= fid->hint_last_off;
-                               clu = fid->hint_last_clu;
-                       }
-
-                       while ((clu_offset > 0) && (clu != CLUSTER_32(~0))) {
-                               last_clu = clu;
-                               /* clu = exfat_fat_read(sb, clu); */
-                               if (exfat_fat_read(sb, clu, &clu) == -1) {
-                                       ret = -EIO;
-                                       goto out;
-                               }
-                               clu_offset--;
-                       }
-               }
-
-               if (clu == CLUSTER_32(~0)) {
-                       num_alloc = (s32)((count - 1) >>
-                                         p_fs->cluster_size_bits) + 1;
-                       new_clu.dir = (last_clu == CLUSTER_32(~0)) ?
-                                       CLUSTER_32(~0) : last_clu + 1;
-                       new_clu.size = 0;
-                       new_clu.flags = fid->flags;
-
-                       /* (1) allocate a chain of clusters */
-                       num_alloced = exfat_alloc_cluster(sb,
-                                                         num_alloc,
-                                                         &new_clu);
-                       if (num_alloced == 0)
-                               break;
-                       if (num_alloced < 0) {
-                               ret = num_alloced;
-                               goto out;
-                       }
-
-                       /* (2) append to the FAT chain */
-                       if (last_clu == CLUSTER_32(~0)) {
-                               if (new_clu.flags == 0x01)
-                                       fid->flags = 0x01;
-                               fid->start_clu = new_clu.dir;
-                               modified = true;
-                       } else {
-                               if (new_clu.flags != fid->flags) {
-                                       exfat_chain_cont_cluster(sb,
-                                                                fid->start_clu,
-                                                                num_clusters);
-                                       fid->flags = 0x01;
-                                       modified = true;
-                               }
-                               if (new_clu.flags == 0x01)
-                                       exfat_fat_write(sb, last_clu, new_clu.dir);
-                       }
-
-                       num_clusters += num_alloced;
-                       clu = new_clu.dir;
-               }
-
-               /* hint information */
-               fid->hint_last_off = (s32)(fid->rwoffset >>
-                                          p_fs->cluster_size_bits);
-               fid->hint_last_clu = clu;
-
-               /* byte offset in cluster   */
-               offset = (s32)(fid->rwoffset & (p_fs->cluster_size - 1));
-
-               /* sector offset in cluster */
-               sec_offset = offset >> p_bd->sector_size_bits;
-
-               /* byte offset in sector    */
-               offset &= p_bd->sector_size_mask;
-
-               LogSector = START_SECTOR(clu) + sec_offset;
-
-               oneblkwrite = (u64)(p_bd->sector_size - offset);
-               if (oneblkwrite > count)
-                       oneblkwrite = count;
-
-               if ((offset == 0) && (oneblkwrite == p_bd->sector_size)) {
-                       if (sector_read(sb, LogSector, &tmp_bh, 0) !=
-                           0)
-                               goto err_out;
-                       memcpy((char *)tmp_bh->b_data,
-                              (char *)buffer + write_bytes, (s32)oneblkwrite);
-                       if (sector_write(sb, LogSector, tmp_bh, 0) !=
-                           0) {
-                               brelse(tmp_bh);
-                               goto err_out;
-                       }
-               } else {
-                       if ((offset > 0) ||
-                           ((fid->rwoffset + oneblkwrite) < fid->size)) {
-                               if (sector_read(sb, LogSector, &tmp_bh, 1) !=
-                                   0)
-                                       goto err_out;
-                       } else {
-                               if (sector_read(sb, LogSector, &tmp_bh, 0) !=
-                                   0)
-                                       goto err_out;
-                       }
-
-                       memcpy((char *)tmp_bh->b_data + offset,
-                              (char *)buffer + write_bytes, (s32)oneblkwrite);
-                       if (sector_write(sb, LogSector, tmp_bh, 0) !=
-                           0) {
-                               brelse(tmp_bh);
-                               goto err_out;
-                       }
-               }
-
-               count -= oneblkwrite;
-               write_bytes += oneblkwrite;
-               fid->rwoffset += oneblkwrite;
-
-               fid->attr |= ATTR_ARCHIVE;
-
-               if (fid->size < fid->rwoffset) {
-                       fid->size = fid->rwoffset;
-                       modified = true;
-               }
-       }
-
-       brelse(tmp_bh);
-
-       /* (3) update the direcoty entry */
-       es = get_entry_set_in_dir(sb, &fid->dir, fid->entry,
-                                 ES_ALL_ENTRIES, &ep);
-       if (!es)
-               goto err_out;
-       ep2 = ep + 1;
-
-       exfat_set_entry_time(ep, tm_current(&tm), TM_MODIFY);
-       exfat_set_entry_attr(ep, fid->attr);
-
-       if (modified) {
-               if (exfat_get_entry_flag(ep2) != fid->flags)
-                       exfat_set_entry_flag(ep2, fid->flags);
-
-               if (exfat_get_entry_size(ep2) != fid->size)
-                       exfat_set_entry_size(ep2, fid->size);
-
-               if (exfat_get_entry_clu0(ep2) != fid->start_clu)
-                       exfat_set_entry_clu0(ep2, fid->start_clu);
-       }
-
-       update_dir_checksum_with_entry_set(sb, es);
-       release_entry_set(es);
-
-#ifndef CONFIG_STAGING_EXFAT_DELAYED_SYNC
-       fs_sync(sb, true);
-       fs_set_vol_flags(sb, VOL_CLEAN);
-#endif
-
-err_out:
-       /* set the size of written bytes */
-       if (wcount)
-               *wcount = write_bytes;
-
-       if (num_alloced == 0)
-               ret = -ENOSPC;
-
-       else if (p_fs->dev_ejected)
-               ret = -EIO;
-
-out:
-       /* release the lock for file system critical section */
-       mutex_unlock(&p_fs->v_mutex);
-
-       return ret;
-}
-
 static int ffsTruncateFile(struct inode *inode, u64 old_size, u64 new_size)
 {
        s32 num_clusters;
@@ -2273,7 +1904,6 @@ static struct dentry *exfat_lookup(struct inode *dir, struct dentry *dentry,
        int err;
        struct file_id_t fid;
        loff_t i_pos;
-       u64 ret;
        mode_t i_mode;
 
        __lock_super(sb);
@@ -2295,18 +1925,6 @@ static struct dentry *exfat_lookup(struct inode *dir, struct dentry *dentry,
        }
 
        i_mode = inode->i_mode;
-       if (S_ISLNK(i_mode) && !EXFAT_I(inode)->target) {
-               EXFAT_I(inode)->target = kmalloc(i_size_read(inode) + 1,
-                                                GFP_KERNEL);
-               if (!EXFAT_I(inode)->target) {
-                       err = -ENOMEM;
-                       goto error;
-               }
-               ffsReadFile(dir, &fid, EXFAT_I(inode)->target,
-                           i_size_read(inode), &ret);
-               *(EXFAT_I(inode)->target + i_size_read(inode)) = '\0';
-       }
-
        alias = d_find_alias(inode);
        if (alias && !exfat_d_anon_disconn(alias)) {
                BUG_ON(d_unhashed(alias));
@@ -2398,73 +2016,6 @@ out:
        return err;
 }
 
-static int exfat_symlink(struct inode *dir, struct dentry *dentry,
-                        const char *target)
-{
-       struct super_block *sb = dir->i_sb;
-       struct timespec64 curtime;
-       struct inode *inode;
-       struct file_id_t fid;
-       loff_t i_pos;
-       int err;
-       u64 len = (u64)strlen(target);
-       u64 ret;
-
-       __lock_super(sb);
-
-       pr_debug("%s entered\n", __func__);
-
-       err = ffsCreateFile(dir, (u8 *)dentry->d_name.name, FM_SYMLINK, &fid);
-       if (err)
-               goto out;
-
-
-       err = ffsWriteFile(dir, &fid, (char *)target, len, &ret);
-
-       if (err) {
-               ffsRemoveFile(dir, &fid);
-               goto out;
-       }
-
-       INC_IVERSION(dir);
-       curtime = current_time(dir);
-       dir->i_ctime = curtime;
-       dir->i_mtime = curtime;
-       dir->i_atime = curtime;
-       if (IS_DIRSYNC(dir))
-               (void)exfat_sync_inode(dir);
-       else
-               mark_inode_dirty(dir);
-
-       i_pos = ((loff_t)fid.dir.dir << 32) | (fid.entry & 0xffffffff);
-
-       inode = exfat_build_inode(sb, &fid, i_pos);
-       if (IS_ERR(inode)) {
-               err = PTR_ERR(inode);
-               goto out;
-       }
-       INC_IVERSION(inode);
-       curtime = current_time(inode);
-       inode->i_mtime = curtime;
-       inode->i_atime = curtime;
-       inode->i_ctime = curtime;
-       /* timestamp is already written, so mark_inode_dirty() is unneeded. */
-
-       EXFAT_I(inode)->target = kmemdup(target, len + 1, GFP_KERNEL);
-       if (!EXFAT_I(inode)->target) {
-               err = -ENOMEM;
-               goto out;
-       }
-
-       dentry->d_time = GET_IVERSION(dentry->d_parent->d_inode);
-       d_instantiate(dentry, inode);
-
-out:
-       __unlock_super(sb);
-       pr_debug("%s exited\n", __func__);
-       return err;
-}
-
 static int exfat_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode)
 {
        struct super_block *sb = dir->i_sb;
@@ -2838,7 +2389,6 @@ static const struct inode_operations exfat_dir_inode_operations = {
        .create        = exfat_create,
        .lookup        = exfat_lookup,
        .unlink        = exfat_unlink,
-       .symlink       = exfat_symlink,
        .mkdir         = exfat_mkdir,
        .rmdir         = exfat_rmdir,
        .rename        = exfat_rename,