Re: [Patch] Fix Buffer overflow in hfsplus with a corrupted image

[Eric Sesterhenn <snakebyte@gmx.de>]

* Pekka Enberg (penberg@cs.helsinki.fi) wrote:
> Hi Eric,
> 
> On Sun, Aug 24, 2008 at 4:07 PM, Eric Sesterhenn <snakebyte@gmx.de> wrote:
> > when an hfsplus image gets corrupted it might happen that the catalog
> > namelength field gets b0rked. If we mount such an image
> > the memcpy() in hfsplus_cat_build_key_uni() writes more than the 255
> > that fit in the name field. Depending on the size of the overwritten
> > data, we either only get memory corruption or also trigger an oops like
> > this:
> >
> > --- linux/fs/hfsplus/catalog.c.orig     2008-08-24 14:52:03.000000000 +0200
> > +++ linux/fs/hfsplus/catalog.c  2008-08-24 14:54:15.000000000 +0200
> > @@ -168,6 +168,11 @@ int hfsplus_find_cat(struct super_block
> >                return -EIO;
> >        }
> >
> > +       if (be16_to_cpu(tmp.thread.nodeName.length) >= 127) {
> > +               printk(KERN_ERR "hfs: catalog name length corrupted\n");
> > +               return -EIO;
> > +       }
> 
> So, where does this 127 come from? I can only find reference to a
> maximum length of 255 unicode characters (16 bits per character) in
> the following technical note for HFS+ (see sections "HFS Plus Names"
> and "Catalog Thread Records"):
> 
>   http://developer.apple.com/technotes/tn/tn1150.html
> 
> Hmm?

Ah, i missed that the name array is __be16, i somehow assumed it was a
char array, and tried to account for the multiplication by 2 in 
hfsplus_cat_build_key_uni(). here is an updated fix.

Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>

--- linux/fs/hfsplus/catalog.c.orig	2008-08-24 14:52:03.000000000 +0200
+++ linux/fs/hfsplus/catalog.c	2008-08-24 14:54:15.000000000 +0200
@@ -168,6 +168,11 @@ int hfsplus_find_cat(struct super_block 
 		return -EIO;
 	}
 
+	if (be16_to_cpu(tmp.thread.nodeName.length) > 255) {
+		printk(KERN_ERR "hfs: catalog name length corrupted\n");
+		return -EIO;
+	}
+
 	hfsplus_cat_build_key_uni(fd->search_key, be32_to_cpu(tmp.thread.parentID),
 				 &tmp.thread.nodeName);
 	return hfs_brec_find(fd);