Re: [PATCH v5 7/9] fscrypt: add inline encryption support

From: Eric Biggers <ebiggers@kernel.org>
To: Christoph Hellwig <hch@infradead.org>
Cc: Satya Tangirala <satyat@google.com>,
	linux-scsi@vger.kernel.org, Kim Boojin <boojin.kim@samsung.com>,
	Kuohong Wang <kuohong.wang@mediatek.com>,
	Barani Muthukumaran <bmuthuku@qti.qualcomm.com>,
	linux-f2fs-devel@lists.sourceforge.net,
	linux-block@vger.kernel.org, linux-fscrypt@vger.kernel.org,
	linux-fsdevel@vger.kernel.org
Subject: Re: [PATCH v5 7/9] fscrypt: add inline encryption support
Date: Mon, 4 Nov 2019 19:12:22 -0800	[thread overview]
Message-ID: <20191105031222.GE692@sol.localdomain> (raw)
In-Reply-To: <20191031212103.GA6244@infradead.org>

On Thu, Oct 31, 2019 at 02:21:03PM -0700, Christoph Hellwig wrote:
> > > 
> > > Btw, I'm not happy about the 8-byte IV assumptions everywhere here.
> > > That really should be a parameter, not hardcoded.
> > 
> > To be clear, the 8-byte IV assumption doesn't really come from fs/crypto/, but
> > rather in what the blk-crypto API provides.  If blk-crypto were to provide
> > longer IV support, fs/crypto/ would pretty easily be able to make use of it.
> 
> That's what I meant - we hardcode the value in fscrypt.  Instead we need
> to expose the size from blk-crypt and check for it.
> 
> > 
> > (And if IVs >= 24 bytes were supported and we added AES-128-CBC-ESSIV and
> > Adiantum support to blk-crypto.c, then inline encryption would be able to do
> > everything that the existing filesystem-layer contents encryption can do.)
> > 
> > Do you have anything in mind for how to make the API support longer IVs in a
> > clean way?  Are you thinking of something like:
> > 
> > 	#define BLK_CRYPTO_MAX_DUN_SIZE	24
> > 
> > 	u8 dun[BLK_CRYPTO_MAX_DUN_SIZE];
> > 	int dun_size;
> > 
> > We do have to perform arithmetic operations on it, so a byte array would make it
> > ugly and slower, but it should be possible...
> 
> Well, we could make it an array of u64s, which means we can do all the
> useful arithmetics on components on one of them.  But I see the point,
> this adds significant complexity for no real short term gain, and we
> should probably postponed it until needed.  Maybe just document the
> assumptions a little better.

Just in case it's not obvious to anyone, I should also mention that being
limited to specifying a 64-bit DUN doesn't prevent hardware that accepts a
longer IV (e.g. 128 bits) from being used.  It would just be a matter of
zero-padding the IV in the driver rather than in hardware.

The actual limitation we're talking about here is in the range of IVs that can
be specified.  A 64-bit DUN only allows the first 64 bits of the IV to be
nonzero.  That works for fscrypt in all cases except DIRECT_KEY policies, and it
would work for dm-crypt using the usual dm-crypt IV generator (plain64).

But for inline encryption to be compatible with DIRECT_KEY fscrypt policies or
with certain other dm-crypt IV generators, we'd need the ability to specify more
IV bits.

- Eric