Received-SPF: pass (google.com: best guess record for domain of linux-ext4-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Mon, 6 May 2019 08:57:22 -0700
From:   Eric Biggers <ebiggers@kernel.org>
To:     fstests@vger.kernel.org
Cc:     linux-fscrypt@vger.kernel.org, linux-ext4@vger.kernel.org,
        linux-f2fs-devel@lists.sourceforge.net
Subject: Re: [RFC PATCH 0/7] xfstests: verify fscrypt-encrypted contents and
 filenames
Message-ID: <20190506155721.GB661@sol.localdomain>
References: <20190426204153.101861-1-ebiggers@kernel.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20190426204153.101861-1-ebiggers@kernel.org>
User-Agent: Mutt/1.11.4 (2019-03-13)
Sender: linux-ext4-owner@vger.kernel.org
Precedence: bulk

On Fri, Apr 26, 2019 at 01:41:46PM -0700, Eric Biggers wrote:
> Hello,
> 
> This series adds xfstests which verify that encrypted contents and
> filenames on ext4 and f2fs are actually correct, i.e. that the
> encryption uses the correct algorithms, keys, IVs, and padding amounts.
> The new tests work by creating encrypted files, unmounting the
> filesystem, reading the ciphertext from disk using dd and debugfs or
> dump.f2fs, and then comparing it against ciphertext computed
> independently by a new test program that implements the same algorithms.
> 
> These tests are important because:
> 
> - The whole point of file encryption is that the files are actually
>   encrypted correctly on-disk.  Except for generic/399, current xfstests
>   only tests the filesystem semantics, not the actual encryption.
>   generic/399 only tests for incompressibility of encrypted file
>   contents using one particular encryption setting, which isn't much.
> 
> - fscrypt now supports 4 main combinations of encryption settings,
>   rather than 1 as it did originally.  This may be doubled to 8 soon
>   (https://patchwork.kernel.org/patch/10908153/).  We should test all
>   settings.  And without tests, even if the initial implementation is
>   correct, breakage in one specific setting could go undetected.
> 
> - Though Linux's crypto API has self-tests, these only test the
>   algorithms themselves, not how they are used, e.g. by fscrypt.
> 
> Patch 1 is a cleanup patch.  Patches 2-4 add the common helpers for
> ciphertext verification tests.  Patches 5-7 add the actual tests.
> 
> These tests require e2fsprogs and f2fs-tools patches I recently sent out
> to fix printing encrypted filenames.  So, this series might not be
> suitable for merging into mainline xfstests until those patches are
> applied.  Regardless, comments are appreciated.  The needed patches are:
> 
> 	debugfs: avoid ambiguity when printing filenames (https://marc.info/?l=linux-ext4&m=155596495624232&w=2)
> 	f2fs-tools: improve filename printing (https://sourceforge.net/p/linux-f2fs/mailman/message/36648641/)
> 
> This series can also be retrieved from git at
> https://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/xfstests-dev.git
> branch "ciphertext-verification".
> 
> I also have patches on top of this series which verify the ciphertext
> produced from v2 encryption policies, which are proposed by my kernel
> patch series "fscrypt: key management improvements"
> (https://patchwork.kernel.org/cover/10908107/).  v2 encryption policies
> will use a different key derivation function, and thus their ciphertext
> will be different.  These additional patches can be found at branch
> "fscrypt-key-mgmt-improvements" of my git repo above.  But I've arranged
> things such that this shorter series can potentially be applied earlier,
> to test what's in the kernel now.
> 
> Eric Biggers (7):
>   common/encrypt: introduce helpers for set_encpolicy and get_encpolicy
>   fscrypt-crypt-util: add utility for reproducing fscrypt encrypted data
>   common/encrypt: support requiring other encryption settings
>   common/encrypt: add helper for ciphertext verification tests
>   generic: verify ciphertext of v1 encryption policies with AES-256
>   generic: verify ciphertext of v1 encryption policies with AES-128
>   generic: verify ciphertext of v1 encryption policies with Adiantum
> 
>  .gitignore               |    1 +
>  common/encrypt           |  482 ++++++++++-
>  src/Makefile             |    3 +-
>  src/fscrypt-crypt-util.c | 1645 ++++++++++++++++++++++++++++++++++++++
>  tests/ext4/024           |    3 +-
>  tests/generic/395        |   28 +-
>  tests/generic/395.out    |    2 +-
>  tests/generic/396        |   15 +-
>  tests/generic/397        |    3 +-
>  tests/generic/398        |    5 +-
>  tests/generic/399        |    3 +-
>  tests/generic/419        |    3 +-
>  tests/generic/421        |    3 +-
>  tests/generic/429        |    3 +-
>  tests/generic/435        |    3 +-
>  tests/generic/440        |    5 +-
>  tests/generic/700        |   41 +
>  tests/generic/700.out    |    5 +
>  tests/generic/701        |   41 +
>  tests/generic/701.out    |    5 +
>  tests/generic/702        |   43 +
>  tests/generic/702.out    |   10 +
>  tests/generic/group      |    3 +
>  23 files changed, 2308 insertions(+), 47 deletions(-)
>  create mode 100644 src/fscrypt-crypt-util.c
>  create mode 100755 tests/generic/700
>  create mode 100644 tests/generic/700.out
>  create mode 100755 tests/generic/701
>  create mode 100644 tests/generic/701.out
>  create mode 100755 tests/generic/702
>  create mode 100644 tests/generic/702.out
> 
> -- 
> 2.21.0.593.g511ec345e18-goog
> 

Any comments on this?

FYI, the e2fsprogs patch that these tests need was applied.

I'm still waiting for the f2fs-tools patch.

- Eric