From: Eric Biggers Subject: Re: [PATCH] MAINTAINERS: fscrypto: recommend linux-fsdevel for fscrypto patches Date: Mon, 24 Oct 2016 15:58:26 -0700 Message-ID: <20161024225826.GC83082@google.com> References: <1477340955-145575-1-git-send-email-ebiggers@google.com> <8426b631-dc03-b45a-ec58-fc3ae5cdf0c9@nod.at> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-fsdevel@vger.kernel.org, linux-ext4@vger.kernel.org, linux-f2fs-devel@lists.sourceforge.net, linux-kernel@vger.kernel.org, tytso@mit.edu, jaegeuk@kernel.org To: Richard Weinberger Return-path: Content-Disposition: inline In-Reply-To: <8426b631-dc03-b45a-ec58-fc3ae5cdf0c9@nod.at> Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-ext4.vger.kernel.org On Mon, Oct 24, 2016 at 10:44:41PM +0200, Richard Weinberger wrote: > That's a very good idea! > Do we also have designated git tree for fscrypto? > As soon UBIFS file encryption is mainline I'll have to monitor changes > to fscrypto very closely because changes in fscrypto could introduce > changes in the on-flash format of UBIFS. > i.e. a subtle change in filename encryption... :-) There isn't yet a dedicated fscrypto tree. Ted has been taking in most of my fscrypto patches through the ext4 tree. This was discussed briefly before, but as designed filesystems do not need to be aware of the format of the xattr (fscrypt_context) or the format of the encrypted data per se. Thus, new modes/formats can be introduced without requiring changes to filesystem-specific code. The point that was brought up is whether some filesystems would additionally like to track in their superblock which encryption modes and flags are used and either deny mounts or warn at mount-time if some are unrecognized. So it should be decided soon if UBIFS is going to do that or not. Eric