Re: [dm-devel] Re: dm-ioband + bio-cgroup benchmarks

[KAMEZAWA Hiroyuki <kamezawa hiroyu jp fujitsu com>]

On Fri, 19 Sep 2008 22:18:01 -0700
Balbir Singh <balbir linux vnet ibm com> wrote:

> Kamezawa-San, I am not dead against it, but I would provide a knob/control point
> for system administrator to decide if movement is important for applications,
> then let them do so (like force_empty).
> 
make sense.


> > Anyway what's next for me is
> >  1. fix current discussion to remove page->page_cgroup pointer.
> >  2. reduce locks.
> 
> Are you planning on reposting these. I've been trying other approaches at my end
> 
I'll post in next Monday. It's obvious that I should do more tests/fixes...
About performance, I'll give it up at some reasonable point.

> 1. Use radix tree per-node per-zone
> 2. Use radix trees only for 32 bit systems
> 3. Depend on CONFIG_HAVE_MEMORY_PRESENT and build a sparse data structure and
> use pre-allocation
> 
> I've posted (1) and I'll take a look at your patches as well
> 
My patch has (many) bugs. Severals are fixed but there will be still ;)
SwapCache beats me again because it easily reuse uncharged pages...

BTW why do you like radix-tree ? It's not very good for our purpose.
FLATMEM support for small system will be easy work.

> >  3. support swap and swap-cache.
> > 
> > I think algorithm for (1), (2) is now getting smart.
> > 
> 
> Yes, it is getting better
> 

Thanks,
-Kame