Re: [Ext4 Secure Delete 0/7 v4] Ext4 secure delete

[Allison Henderson <achender@linux.vnet.ibm.com>]

On 10/07/2011 10:07 AM, Allison Henderson wrote:
> On 10/07/2011 08:21 AM, Andreas Dilger wrote:
>> On 2011-10-07, at 1:10 AM, Allison Henderson<achender@linux.vnet.ibm.com>   wrote:
>>> Sorry for the delay in getting this next version out.
>>> I had some tasks to take care of, and now I'm picking up my
>>> secure delete work again. I'm still not quite done yet,
>>> but a lot has changed and I wanted to update people so that
>>> we have an idea of where its going.  Currently the patch
>>> deals with data blocks, meta blocks, directory entries,
>>> journal blocks, and also provides an option for secure
>>> deleting with random data instead of just zeros.
>>> I'm also planning on adding some more patches to
>>> deal with inodes and also a mount option that turns
>>> on secure delete by default.  Im still not quite done
>>> debugging, but Im just sending it out early to get
>>> some more eyes on it. Feed back appreciated!  :)
>>>
>>> v3->v4
>>> Added a new file attribute flag EXT4_SECRM_RANDOM_FL
>>> This flag causes the secure delete operations to over write
>>> blocks with random data instead of zeros.
>>
>> Since inode flags are in short supply, and I suspect users that want this want it for all files, this should probably be a superblock flag?
>>
> That is a really good point.  The first thing that comes to mind though would be the fact that it is a lot slower when the random flag is on especially for really big files.  So that would be one case where I could imagine a user might want the ability to set different options per file.  But since the flags are a limited resource, I can see where we may not want to spend it so quickly.  I will see if maybe there is some way I can optimize it, but I would like to see more folks weigh in on this topic too.
>

I've had another suggestion for this come up, so I wanted to put it out 
here to see what people think.  Instead of a new flag, we could use a 
flag that is mutually exclusive like the undelete flag, since a file 
that is secure deleted, should not be undeletable.  Would this be 
something that people would be interested in doing?


>>> New function ext4_secure_delete_lblks added to walk
>>> data blocks and secure delete them before any blocks
>>> are removed.
>>>
>>> Meta blocks are secure deleted before they are
>>> released
>>>
>>> New function added to identify holes in ind files.
>>> Used by ext4_secure_delete_lblks to skip over holes
>>> during secure delete.
>>>
>>> Added another list in the journal structure to track
>>> journal blocks so that they can be secure deleted later.
>>>
>>> Added new ext4_secure_delete_jblks that secure deletes
>>> journal blocks that were used to journal the specified
>>> logical blocks
>>>
>>> Allison Henderson (7):
>>>    ext4: Secure Delete: Add new EXT4_SECRM_RANDOM_FL flag
>>>    ext4: Secure Delete: Add ext4_ind_hole_lookup function
>>>    ext4: Secure Delete: Add secure delete functions
>>>    ext4: Secure Delete: Secure delete file data
>>>    ext4: Secure Delete: Secure delete directory entry
>>>    ext4: Secure Delete: Secure delete meta data blocks
>>>    ext4/jbd2: Secure Delete: Secure delete journal blocks
>>>
>>> fs/ext4/ext4.h         |   28 +++-
>>> fs/ext4/ext4_extents.h |    2 +
>>> fs/ext4/extents.c      |   21 +++-
>>> fs/ext4/indirect.c     |    2 +-
>>> fs/ext4/inode.c        |  391 ++++++++++++++++++++++++++++++++++++++++++++++++
>>> fs/ext4/mballoc.c      |    8 +
>>> fs/ext4/namei.c        |   64 +++++++-
>>> fs/jbd2/commit.c       |    6 +
>>> fs/jbd2/journal.c      |  112 ++++++++++++++
>>> include/linux/jbd2.h   |   21 +++
>>> 10 files changed, 642 insertions(+), 13 deletions(-)
>>>
>>> --
>>> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
>>> the body of a message to majordomo@vger.kernel.org
>>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html