Re: [PATCH] ext4: Add support for SFITRIM, an ioctl for secure FITRIM.

["Lukáš Czerner" <lczerner@redhat.com>]

On Tue, 17 Jun 2014, Dave Chinner wrote:

> Date: Tue, 17 Jun 2014 12:49:53 +1000
> From: Dave Chinner <david@fromorbit.com>
> To: Theodore Ts'o <tytso@mit.edu>
> Cc: JP Abgrall <jpa@google.com>, Eric Sandeen <sandeen@redhat.com>,
>     linux-ext4@vger.kernel.org, Geremy Condra <gcondra@google.com>,
>     "linux-fsdevel@vger.kernel.org" <linux-fsdevel@vger.kernel.org>
> Subject: Re: [PATCH] ext4: Add support for SFITRIM,
>     an ioctl for secure FITRIM.
> 
> On Fri, Jun 13, 2014 at 07:41:34PM -0400, Theodore Ts'o wrote:
> > On Fri, Jun 13, 2014 at 12:44:34PM -0700, JP Abgrall wrote:
> > > The per-file secure discard seems to be the way to go, as there are
> > > only a few places in Android where this needs to be turned on.
> > > The  current idletime-fstrim would  switch from FITRIM to SFITRIM to
> > > reduce the leftovers.
> > 
> > OK, how about this?  The following patch is in the Google data center
> > kernel, but I never got around to get it upstream (oops, was on my
> > todo list, but it never happened).
> > 
> > If you want to adopt this for usptream, and add support for
> > BLKSECDISCARD as well as BLKDISCARD, then you could for each file that
> > you want to do the per-file secure discard, you would just have to
> > open the file, call the BLKSECDISCARD ioctl, and then delete the file.
> > 
> > Cheers,
> > 
> > 					- Ted
> > 
> > commit 16ff6352b123aa134417793d636f05cd4e240eaa
> > Author: Theodore Ts'o <tytso@google.com>
> > Date:   Fri Dec 20 12:48:26 2013 -0500
> > 
> >     ext4: add support for the BLKDISCARD ioctl
> >     
> >     The blkdicard ioctl previously only worked on block devices.  Allow
> >     this ioctl to work on ext4 files.
> >     
> >     This commit is intended to be sent upstream.
> 
> Not in that form - it's an ugly API hack.
> 
> This is really just an extension of hole punching (if the blocks in
> the file are being removed) or zeroing (if the blocks are being
> retained by the file). Either way, fallocate() is the interface
> used for per-file block level manipulations, and either of these
> operations could issue a discard (secure or not) during the
> punch/zero operation....

I definitely agree with Dave here it is an ugly API hack. Fallocate
seems much more suitable for this.

New flag FALLOC_FL_ISSUE_DISCARD which would work with
FALLOC_FL_PUNCH_HOLE, FALLOC_FL_ZERO_RANGE and possibly
FALLOC_FL_COLLAPSE_RANGE might actually be useful.

-Lukas

> 
> Cheers,
> 
> Dave.
>