[Developer] Fwd: Greeting.

[Hal Ashburner]

JoonHyun Bae wrote:
> Hi, all.
Hi JoonHyun!
>
> I am lucky enough to be the second guy posting this list.
> Yeah!, again :-)
>
> I hope OK Labs to be a champion in embedded virtualization market.
> Because OK Labs has the open source micor-kernel and I can see the source
> code. (bravo!)
>
> First of all, my question is, what is the license of OKL4?
> Does it has another license model againt L4:Pistachio-embedded from
> NICTA?
>
I'm not a lawyer and obviously if licensing is important to you, you
should definitely speak to a lawyer take their advice and ignore what
I'm saying here. Having said that my understanding is that the license
is the same. It's still a BSD like license that now says OKL instead of
NICTA. If you make a billion dollars using our software, the sales that
use our software have to say "Contains OKL software" or something like
this. It is all very standard and normal open source licensing.  Just to
repeat, if you have any doubts about the license, you should see your
lawyer.
> Does the source code base is branching or is the same as the NICTA code?
The NICTA code has now become the OKL code. It's the same code base with
further work done it. The OKL code base is the one that will continue to
be developed.
> I'm at a glance of VirtualLogix (http://www.virtuallogix.com/) .
> Does anybody knows the VirtualLogix?
> I guess it is hypervisor or para-virtualization. isn't it?
> Would it be a competitor of OKL4?
> I want to compare the diff. between L4 and virtuallogix, but I can't
> get the
> VirtualLogix's Code. :-(.
These guys are competitors. We haven't seen their source either, it's
not available, it's closed source software. Having said that I'm not
sure how keen we would be to see it closely even if it were open. Our
understanding from their public website is that their virtualization
technology runs guest Operating Systems in Kernel mode which we think
limits the value of virtualization considerably.
>
> And the last question is, why does it be so slow in PC99 emulation with
> skyeye on my desktop?
Skyeye is the arm simulator. The slowness is a limitation of that simulator.
An alternative is qemu for i386 (pc99) emulation, which is considerably
faster. If you install qemu you can try it with:
$ ./tools/build.py machine=ia32_pc99 project=iguana wombat=True
toolprefix=i686-unknown-linux-gnu- simulate

Which requires that you have the recommended toolchain installed and on
your path.
You can get it here:
http://www.ertos.nicta.com.au/downloads/i686-gcc-3.3.4-glibc-2.3.3-2006-06-02.tar.gz

> I am expecting that the running speed should be fast enough.
> But with a user's point of view. it seems so slow.
> How could I explain it?
It _is_ fast on real hardware. In fact for some of the LMBench linux
benchmarks, virtualised wombat achieves faster numbers than
unvirtualized (bare metal) linux.
I think the best explanation is that skyeye, whilst being a fairly
accurate simulator is slow compared to real hardware.

All the best,
Hal Ashburner