On Thu, 16 Jul 2015 09:19:49 +0200 Michal Hocko <mhocko@kernel.org> wrote:

> On Wed 15-07-15 13:57:11, Andrew Morton wrote:
> > On Wed, 15 Jul 2015 13:14:41 +0200 Michal Hocko <mhocko@kernel.org> wrote:
> > 
> > > mem_cgroup structure is defined in mm/memcontrol.c currently which
> > > means that the code outside of this file has to use external API even
> > > for trivial access stuff.
> > > 
> > > This patch exports mm_struct with its dependencies and makes some of the
> > > exported functions inlines. This even helps to reduce the code size a bit
> > > (make defconfig + CONFIG_MEMCG=y)
> > > 
> > > text		data    bss     dec     	 hex 	filename
> > > 12355346        1823792 1089536 15268674         e8fb42 vmlinux.before
> > > 12354970        1823792 1089536 15268298         e8f9ca vmlinux.after
> > > 
> > > This is not much (370B) but better than nothing. We also save a function
> > > call in some hot paths like callers of mem_cgroup_count_vm_event which is
> > > used for accounting.
> > > 
> > > The patch doesn't introduce any functional changes.
> > > 
> > > ...
> > >
> > >  include/linux/memcontrol.h | 369 +++++++++++++++++++++++++++++++++++++++++----
> > 
> > Boy, that's a ton of new stuff into the header file.  Do we actually
> > *need* to expose all this?
> 
> I am exporting struct mem_cgroup with its dependencies + some small
> functions which allow to inline some really trivial code and helps to
> generate a better code.
> 
> > Is some other patch dependent on it? 
> 
> Without mem_cgroup visible outside of memcontrol.c we couldn't inline
> and now we can also use some fields from mem_cgroup directly and get rid
> of some really trivial access functions.
> 
> > If
> > not then perhaps we shouldn't do this - if the code was already this
> > way, I'd be attracted to a patch which was the reverse of this one!
> 
> I agree with Johannes who originally suggested to expose mem_cgroup that
> it will allow for a better code later.

Sure, but how *much* better?  Are there a significant number of
fastpath functions involved?

From a maintainability/readability point of view, this is quite a bad
patch.  It exposes a *lot* of stuff to the whole world.  We need to get
a pretty good runtime benefit from doing this to ourselves.  I don't
think that saving 376 bytes on a fatconfig build is sufficient
justification?