Follow up on command line auditing

[William Roberts]

On Mon, Dec 2, 2013 at 9:18 AM, Richard Guy Briggs <rgb at redhat.com> wrote:
> On Mon, Dec 02, 2013 at 08:20:10AM -0800, William Roberts wrote:
>> On Mon, Dec 2, 2013 at 8:07 AM, Richard Guy Briggs <rgb at redhat.com> wrote:
>> > On Mon, Dec 02, 2013 at 07:42:20AM -0800, William Roberts wrote:
>> >> Changelog since last post:
>> >> * Rebase on latest master
>> >>
>> >> [PATCH] audit: Audit proc cmdline value
>> >
>> > Hi Bill,
>> >
>> > I wasn't expecting that you would squash everything down into one patch.
>> > I think it should be at least two.  I'm comfortable with the changes in
>> > the audit subsystem.  Could those be one patch?  As for the changes to
>> > proc (including base and util) those might be better as a seperate
>> > patch.
>>
>> Richard,
>> Ok so what do you think the best way forward is? I don't want to duplicate
>> code from proc/base.c. I would need to export proc_pid_cmdline()
>> in the first patch or re-implement it in the audit subsystem, followed
>> by a patch
>> to merge the functionality. What would you prefer?
>
> I would split them into 3 patches:
>
> 1) implement the length and copy funcitons:
>  include/linux/mm.h |    7 +++++
>  mm/util.c          |   48 ++++++++++++++++++++++++++++++
>
> 2) use them in the proc call:
>  fs/proc/base.c     |   35 +++++++---------------
>
> 3) use them in audit:
>  kernel/audit.h     |    1 +
>  kernel/auditsc.c   |   82 ++++++++++++++++++++++++++++++++++++++++++++++++++++
>
>
> Does this split make sense?  Combining 1 and 2 might be acceptable to
> those subsystem maintainers...

You read my mind here after I sent this, this is exactly what I was thinking.

When I am done do I publish this to kernel mainline, here, or elsewhere?

Bill

<snip>