]> err.no Git - linux-2.6/commit
[PATCH] hfs_fill_super returns success even if no root inode
authorEric Sandeen <sandeen@redhat.com>
Thu, 16 Nov 2006 09:19:22 +0000 (01:19 -0800)
committerLinus Torvalds <torvalds@woody.osdl.org>
Thu, 16 Nov 2006 19:43:38 +0000 (11:43 -0800)
commitd6ddf55440833fd9404138026af246c51ebeef22
tree095984ba56a4f6eb14379f77bc6b0fbd49cf7f6a
parent4c1b6d18bf2fdeb5ac725126c6928aaa98c8e22f
[PATCH] hfs_fill_super returns success even if no root inode

http://kernelfun.blogspot.com/2006/11/mokb-14-11-2006-linux-26x-selinux.html

mount that image...
fs: filesystem was not cleanly unmounted, running fsck.hfs is recommended.  mounting read-only.
hfs: get root inode failed.
BUG: unable to handle kernel NULL pointer dereference at virtual address 00000018
 printing eip
...
EIP is at superblock_doinit+0x21/0x767
...
 [] selinux_sb_kern_mount+0xc/0x4b
 [] vfs_kern_mount+0x99/0xf6
 [] do_kern_mount+0x2d/0x3e
 [] do_mount+0x5fa/0x66d
 [] sys_mount+0x77/0xae
 [] syscall_call+0x7/0xb
DWARF2 unwinder stuck at syscall_call+0x7/0xb

hfs_fill_super() returns success even if
  root_inode = hfs_iget(sb, &fd.search_key->cat, &rec);
or
  sb->s_root = d_alloc_root(root_inode);

fails.  This superblock finds its way to superblock_doinit() which does:

        struct dentry *root = sb->s_root;
        struct inode *inode = root->d_inode;

and boom.  Need to make sure the error cases return an error, I think.

[akpm@osdl.org: return -ENOMEM on oom]
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Cc: Roman Zippel <zippel@linux-m68k.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
fs/hfs/super.c