Re: [PATCH] v3 RCU: the bloatwatch edition

["Paul E. McKenney" <paulmck@linux.vnet.ibm.com>]

On Thu, Apr 09, 2009 at 01:55:29AM +0900, Paul Mundt wrote:
> On Wed, Apr 08, 2009 at 06:38:38PM +0200, Ingo Molnar wrote:
> > 
> > * Paul Mundt <lethal@linux-sh.org> wrote:
> > 
> > > On Fri, Apr 03, 2009 at 12:36:05AM +0200, Ingo Molnar wrote:
> > > > * Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:
> > > > > This patch is a version of RCU designed for (!SMP && EMBEDDED) 
> > > > > provided as a proof of concept of a small-footprint RCU 
> > > > > implementation. In particular, the implementation of 
> > > > > synchronize_rcu() is extremely lightweight and high performance.  
> > > > > It passes rcutorture testing in each of the four relevant 
> > > > > configurations (combinations of NO_HZ and PREEMPT) on x86.  This 
> > > > > saves about 900 bytes compared to Classic RCU, and a couple 
> > > > > kilobytes compared to Hierarchical RCU:
> > > > 
> > > > Andrew, what do you think?
> > > > 
> > > > A worry is yet another RCU variant - we already have 3.
> > > > 
> > > > A trick we could use would be to put it into Documentation/rcu/, 
> > > > linked in via some clever Makefile magic and only usable if a 
> > > > ultra-embedded developer does a build with something like 
> > > > CONFIG_RCU_TINY=y. That way there's no real maintenance and testing 
> > > > overhead.
> > > > 
> > > > It _does_ have documentation value beyond the ~900 bytes: it's the 
> > > > simplest and smallest possible still-working UP RCU implementation 
> > > > so it would be easy to teach RCU concepts via that, gradually.
> > > 
> > > A similar argument could have been used for tiny-shmem when it was 
> > > first integrated. As this is hiding behind CONFIG_EMBEDDED, most 
> > > users are not going to run in to it, so the confusion of 1 more 
> > > RCU variant is not likely to be a problem for those that aren't 
> > > actively seeking it out.
> > > 
> > > So, personally I think it is a good idea, and I have no 
> > > reservations about default enabling it for a number of more 
> > > constrained SH platforms.
> > 
> > but at least tiny-shmem is now nicely hidden in mm/shmem.c, in an 
> > unintrusive !CONFIG_SHMEM branch. There's no CONFIG_TINY_SHMEM 
> > option anymore - it's all done in the !CONFIG_SHMEM case.
> > 
> Now it is, yes, but it was not originally, and it was still useful when
> it was split out. If we are going to tolerate multiple RCU
> implementations in the kernel, then I see no reason to not include
> tiny-RCU in the same category. Even in the case where some of the other
> RCU variants go away, tiny-RCU remains a viable option for simple
> platforms that are more concerned about memory than anything else, so
> it's always a valid alternative.
> 
> If in the future things are more consolidated and the config option goes
> away then great, but that hardly seems like a sane prerequisite for
> merging it. CONFIG_EMBEDDED handles this just fine. You don't need to
> enable it if you don't wish to, but it's certainly measurable enough to
> be useful for those of us that have no problems enabling it ;-)

>From a kernel-size viewpoint:

  788 kernel/rcuclassic.c
  190 include/linux/rcuclassic.h
  978 total

  288 kernel/rcutiny.c
   68 include/linux/rcutiny.h
  356 total

Almost a 3x decrease in lines of code.  So, Seems to me that dropping
rcuclassic (as rcutree proves itself) and taking up rcutiny instead is
a good step forward.  ;-)

							Thanx, Paul