Re: [PATCH 3/5] xfs_db: fix metadump name obfuscation for ascii-ci filesystems

["Darrick J. Wong" <djwong@kernel.org>]

On Mon, Jun 05, 2023 at 11:59:45AM -0500, Eric Sandeen wrote:
> On 6/5/23 10:36 AM, Darrick J. Wong wrote:
> > @@ -1205,9 +1264,9 @@ generate_obfuscated_name(
> >   	/* Obfuscate the name (if possible) */
> > -	hash = libxfs_da_hashname(name, namelen);
> > -	obfuscate_name(hash, namelen, name);
> > -	ASSERT(hash == libxfs_da_hashname(name, namelen));
> > +	hash = dirattr_hashname(ino != 0, name, namelen);
> > +	obfuscate_name(hash, namelen, name, ino != 0);
> > +	ASSERT(hash == dirattr_hashname(ino != 0, name, namelen));
> 
> This makes sense to me - comments above here remind us that "inode == 0"
> means we're obfuscating an xattr value, not a filename or path name, but ...
> 
> >   	/*
> >   	 * Make sure the name is not something already seen.  If we
> > @@ -1320,7 +1379,7 @@ obfuscate_path_components(
> >   			/* last (or single) component */
> >   			namelen = strnlen((char *)comp, len);
> >   			hash = libxfs_da_hashname(comp, namelen);
> > -			obfuscate_name(hash, namelen, comp);
> > +			obfuscate_name(hash, namelen, comp, false);
> >   			ASSERT(hash == libxfs_da_hashname(comp, namelen));
> >   			break;
> >   		}
> > @@ -1332,7 +1391,7 @@ obfuscate_path_components(
> >   			continue;
> >   		}
> >   		hash = libxfs_da_hashname(comp, namelen);
> > -		obfuscate_name(hash, namelen, comp);
> > +		obfuscate_name(hash, namelen, comp, false);
> >   		ASSERT(hash == libxfs_da_hashname(comp, namelen));
> >   		comp += namelen + 1;
> >   		len -= namelen + 1;
> > 
> 
> here, why is "is_dirent" false? Shouldn't a symlink path component match the
> associated dirents, and be obsucated the same way?

Name obfuscation replaces every byte except for the last five bytes with
a random printable character, and then flips bits in those last five
bytes to make the hash match.  Chances are good that calling
obfuscate_name() twice on the same name will return different results,
which means that symlink targets won't point anywhere useful after the
obfuscation.

One could make metadump remember the (input -> output) pairs instead of
regenerating the names every time, but this comes at a cost of higher
memory consumption.  I actually did this for parent pointers so that
obfuscated dumped pptrs are still verifiable by xfs_repair.

However, symlink targets aren't required to point to a valid path, so
there doesn't seem to be much reason to add that overhead.

--D

> -Eric