linux-ext4 - [PATCH 0/6] fscrypt: fixes for presentation of long encrypted filenames

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20170424170013.85175-1-ebiggers3@gmail.com>
Date:   Mon, 24 Apr 2017 10:00:07 -0700
From:   Eric Biggers <ebiggers3@...il.com>
To:     linux-fscrypt@...r.kernel.org
Cc:     "Theodore Y . Ts'o" <tytso@....edu>,
        Jaegeuk Kim <jaegeuk@...nel.org>,
        linux-f2fs-devel@...ts.sourceforge.net, linux-ext4@...r.kernel.org,
        linux-mtd@...ts.infradead.org,
        Gwendal Grignou <gwendal@...omium.org>, hashimoto@...omium.org,
        kinaba@...omium.org, Eric Biggers <ebiggers@...gle.com>
Subject: [PATCH 0/6] fscrypt: fixes for presentation of long encrypted filenames

From: Eric Biggers <ebiggers@...gle.com>

This series fixes the bugs that have been identified with how filesystems handle
presenting long encrypted filenames without the key.

Patch 1 is Jaegeuk's fix to make f2fs start checking the ciphertext portion of
the digested names.  I made one change to this patch which is that to determine
whether we should use the hash from the fscrypt_name structure rather than
compute the hash, we should check for 'disk_name.name' being NULL rather than
'hash' being nonzero, since 0 is a valid hash value.

Patch 2 fixes the bug found on Chrome OS where the wrong part of the ciphertext
was included in the digested names, causing collisions and undeletable files.

Patches 3-6 clean things up to be less insane and confusing, e.g. by introducing
a shared function for name matching and a struct to represent a digested name.

Patches 1-2 will need to be backported and I think they should be merged into
4.12 through the fscrypt tree.  The other patches are nice to have but it's not
a big deal if they need to wait for next cycle.

This patch series leaves out UBIFS; it can be changed to use the common matching
function once available, if desired.

Eric Biggers (5):
  fscrypt: avoid collisions when presenting long encrypted filenames
  fscrypt: introduce helper function for filename matching
  ext4: switch to using fscrypt_match_name()
  f2fs: switch to using fscrypt_match_name()
  ext4: clean up ext4_match() and callers

Jaegeuk Kim (1):
  f2fs: check entire encrypted bigname when finding a dentry

 fs/crypto/fname.c               |  90 +++++++++++++++++++++++++++--------
 fs/crypto/fscrypt_private.h     |   2 -
 fs/ext4/namei.c                 | 103 ++++++++++++----------------------------
 fs/f2fs/dir.c                   |  25 ++--------
 fs/f2fs/f2fs.h                  |   3 +-
 fs/f2fs/hash.c                  |   7 ++-
 fs/f2fs/inline.c                |   4 +-
 include/linux/fscrypt_notsupp.h |   9 ++++
 include/linux/fscrypt_supp.h    |  78 ++++++++++++++++++++++++++++++
 9 files changed, 202 insertions(+), 119 deletions(-)

-- 
2.12.2.816.g2cccc81164-goog