All of lore.kernel.org
 help / color / mirror / Atom feed
* Question on ext4 directory hashes in combination with file name encryption
@ 2016-09-30 14:09 Richard Weinberger
  2016-09-30 20:20 ` Andreas Dilger
  2016-09-30 22:05 ` Theodore Ts'o
  0 siblings, 2 replies; 3+ messages in thread
From: Richard Weinberger @ 2016-09-30 14:09 UTC (permalink / raw)
  To: linux-ext4, linux-fsdevel, linux-kernel
  Cc: Theodore Ts'o, Eric Biggers, David Gstir

Hi,

if I read the ext4 code correctly, you pass encrypted filenames to ext4fs_dirhash().
These filenames are not encoded and therefore binary gibberish.
Isn't this a problem for the ext4 hash functions? My fear is that these hashes are optimized
for ASCII strings and produce more collisions when binary data is used as input.

Thanks,
//richard

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Question on ext4 directory hashes in combination with file name encryption
  2016-09-30 14:09 Question on ext4 directory hashes in combination with file name encryption Richard Weinberger
@ 2016-09-30 20:20 ` Andreas Dilger
  2016-09-30 22:05 ` Theodore Ts'o
  1 sibling, 0 replies; 3+ messages in thread
From: Andreas Dilger @ 2016-09-30 20:20 UTC (permalink / raw)
  To: Richard Weinberger
  Cc: Ext4 Developers List, linux-fsdevel, linux-kernel,
	Theodore Ts'o, Eric Biggers, David Gstir

[-- Attachment #1: Type: text/plain, Size: 663 bytes --]

On Sep 30, 2016, at 8:09 AM, Richard Weinberger <richard@nod.at> wrote:
> 
> Hi,
> 
> if I read the ext4 code correctly, you pass encrypted filenames to ext4fs_dirhash().
> These filenames are not encoded and therefore binary gibberish.
> Isn't this a problem for the ext4 hash functions? My fear is that these hashes are optimized
> for ASCII strings and produce more collisions when binary data is used as input.

The default hash function (half-md4) is an (old) crypto hash and works
fine with binary data.  Some of the other hash functions are less strong,
but I don't think anyone changes the hash function for ext4.

Cheers, Andreas






[-- Attachment #2: Message signed with OpenPGP using GPGMail --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Question on ext4 directory hashes in combination with file name encryption
  2016-09-30 14:09 Question on ext4 directory hashes in combination with file name encryption Richard Weinberger
  2016-09-30 20:20 ` Andreas Dilger
@ 2016-09-30 22:05 ` Theodore Ts'o
  1 sibling, 0 replies; 3+ messages in thread
From: Theodore Ts'o @ 2016-09-30 22:05 UTC (permalink / raw)
  To: Richard Weinberger
  Cc: linux-ext4, linux-fsdevel, linux-kernel, Eric Biggers, David Gstir

On Fri, Sep 30, 2016 at 04:09:09PM +0200, Richard Weinberger wrote:
> 
> if I read the ext4 code correctly, you pass encrypted filenames to ext4fs_dirhash().
> These filenames are not encoded and therefore binary gibberish.

That's correct.

> Isn't this a problem for the ext4 hash functions? My fear is that these hashes are optimized
> for ASCII strings and produce more collisions when binary data is used as input.

I'm not particularly worried.  In general, while the converse is true;
that hashes that assume that they are used for binary data might not
work as well for ASCII strings, the ext4 hash functions should work
fine for binary data.

					- Ted

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2016-10-01  0:44 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-09-30 14:09 Question on ext4 directory hashes in combination with file name encryption Richard Weinberger
2016-09-30 20:20 ` Andreas Dilger
2016-09-30 22:05 ` Theodore Ts'o

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.