Re: mmotm 2009-04-05-21-45 uploaded (sysctl)

[ebiederm@xmission.com (Eric W. Biederman)]

Andrew Morton <akpm@linux-foundation.org> writes:

> On Mon, 06 Apr 2009 21:12:26 -0700 Randy Dunlap <randy.dunlap@oracle.com> wrote:
>
>> Randy Dunlap wrote:
>> > On Sun, 05 Apr 2009 21:50:36 -0700 akpm@linux-foundation.org wrote:
>> > 
>> >> The mm-of-the-moment snapshot 2009-04-05-21-45 has been uploaded to
>> >>
>> >>    http://userweb.kernel.org/~akpm/mmotm/
>> >>
>> >> and will soon be available at
>> >>
>> >>    git://git.zen-sources.org/zen/mmotm.git
>> >>
>> >> It contains the following patches against 2.6.29:
>> > 
>> > 
>> > kernel/sysctl.c: In function '__register_sysctl_paths':
>> > kernel/sysctl.c:2059: error: 'struct ctl_table_header' has no member named 'dep_map'
>> > 
>> > 
>> > config attached.
>> > 
>> > ---
>> > ~Randy
>> 
>> kernel/sysctl.c: s/CONFIG_DEBUG_LOCK_ALLOC/CONFIG_PROVE_LOCKING/ at line 2056.
>
> OK, thanks, sysctl-lockdep-support-for-sysctl-reference-counting.patch.
>
> I have a note here that Peter had issues with that patch anyway.  I'm
> not yet sure if it's due to be fixed or dropped or what?

Let's call it due to be fixed.

Right now I see this going one of two ways.
1) I get the locking model fixed as requested.
2) I move the whole use counting/locking up to the struct file level,
   solve this generically for sysfs, sysctl, proc, uio and whatever
   other subsystem needs this logic.

I should know by the end of the week if I can do this generically.

If I can not do this generically it make sense to fix the patch,
and I will.  It is on my todo list.

If I can do this generically it makes sense to drop the patch,
and completely rewrite all of that code, and make certain lockdep
works on my reimplementation.

Eric