Re: [PATCH 3/9] bio-cgroup controller

[KAMEZAWA Hiroyuki <kamezawa.hiroyu-+CUm20s59erQFUHtdCDX3A@public.gmane.org>]

On Fri, 17 Apr 2009 17:48:54 +0900
KAMEZAWA Hiroyuki <kamezawa.hiroyu-+CUm20s59erQFUHtdCDX3A@public.gmane.org> wrote:

> On Fri, 17 Apr 2009 17:00:16 +0900
> KAMEZAWA Hiroyuki <kamezawa.hiroyu-+CUm20s59erQFUHtdCDX3A@public.gmane.org> wrote:
> 
> > On Fri, 17 Apr 2009 16:22:01 +0900 (JST)
> > Ryo Tsuruta <ryov-jCdQPDEk3idL9jVzuh4AOg@public.gmane.org> wrote:
> > 
> > > In the case where the bio-cgroup data is allocated dynamically,
> > >    - Sometimes quite a large amount of memory get marked dirty.
> > >      In this case it requires more kernel memory than that of the
> > >      current implementation.
> > >    - The operation is expansive due to memory allocations and exclusive
> > >      controls by such as spinlocks.
> > > 
> > > In the case where the bio-cgroup data is allocated by delayed allocation, 
> > >   - It makes the operation complicated and expensive, because
> > >     sometimes a bio has to be created in the context of other
> > >     processes, such as aio and swap-out operation.
> > > 
> > > I'd prefer a simple and lightweight implementation. bio-cgroup only
> > > needs 4bytes unlike memory controller. The reason why bio-cgroup chose
> > > this approach is to minimize the overhead.
> > > 
> > My point is, plz do your best to reduce memory usage here. You increase
> > size of page_cgroup just because you cannot increase size of struct page.
> > It's not be sane reason to increase size of this object.
> > It's a cheat in my point of view.
> > 
> 
> Can't this work sanely ?
> Hmm, endian is obstacle ?
> ==
> 	sturct page_cgroup {
> 		union {
> 			struct {
> 				unsigned long memcg_field:16;
> 				unsigned long blockio_field:16;
> 			} field;
> 			unsigned long flags; /* unsigned long is not 32bits */
> 		} flags;
> 	}
> ==
> 
....sorry plz ignore.
-Kame