Re: L4 meets the Hurd at the Congress in Berlin

[Gernot Heiser]

Folks,

It seems that there are some misunderstandings here. I'll try to
clarify.

>>>>> On Mon, 31 Dec 2001 19:42:33 +0100, Moritz Schulte <address@hidden> said:
MS> Hi there,
MS> the CCC congress was fun for me, I hope for the others, too.  Not only
MS> that there were some Hurd people, but also two L4 hackers from
MS> Karlsruhe.  Neal had his really nice talk on the Hurd and after that
MS> the L4 people talked spontaneouslly a bit about L4.

MS> Jeroen and me tried to summarize the Hurd/L4 related points a bit.

MS> o Horst Wenske, an active L4 hacker who seems to be interested in the
MS>   Hurd, wants to tell some other L4 people about the Hurd and tries to
MS>   promote it a bit there.  He also suggested to make a list of Hurd/L4
MS>   related subjects he could propose for theses at his university.  See
MS>   the "Student projects" at http://www.research.ibm.com/sawmill/ as an
MS>   example of what he means.  Right now they are still doing research
MS>   on SawMill, which is dead, the Hurd is not.  That would be a big
MS>   advantage for the Hurd, since we are usually volunteers, they would
MS>   work fulltime on such projects.  Since SawMill is dead, the Hurd
MS>   would be a better system for them to try out their nice microkernel.

For one thing we certainly don't do research on a dead system. We've
got better things to do than wasting our time.

Whether or not SawMill is "dead" is more a matter of nomenclature than
anything else. The project goals have moved considerably, for example
it is no longer married to Linux source. SawMill activities have also
mostly moved from IBM to Karlsruhe. The revised focus will probably
justify a name change. Irrespective of this, the project of developing
a well-designed multi-server OS on top of L4 is certainly well alive,
and additional person power has just been allocated to it.

We also don't need to be convinced about the desirability of a
well-performing L4-based Hurd. In fact, we have been following the
discussions in l4-hurd with interest, and are evaluating whether, and
in which form, we should become actively involved in it. We have made
contact with some people active on this list with the aim of getting a
few to Karlsruhe for a few days of discussion. However, this has been
on the back burner due to more pressing issues that needed dealing
with.

In any case we see the Hurd as a potentially very interesting
application of L4. However we haven't analysed the issues thoroughly
enough yet to be able to determine how well the Hurd is suited to
be a good demonstration of L4 capabilities.

In summary, we are very interested in supporting the L4-Hurd project,
but have not yet decided how much of our own resources we can commit
to it.

MS> o He wants to convince the IDL4 hacker to release it.

No convincing necessary. IDL4 will be released when it's ready for
it.

MS> o We need to know how the new L4 API will look like, perhaps it would
MS>   make things easier for the port.  The L4 API is still evolving and
MS>   won't stop doing so in the near future.  But all L4 kernels are at
MS>   least backwards compatible.

The new API is converging. We expect to release draft specs soon.

MS> o He wants to convince the SawMill people to open the source (SawMill
MS>   is a multi-server OS running on top of L4).  Perhaps that could be
MS>   helpful since that OS is quite similar to the Hurd, he said.

"The source" is a not very well defined concept with SawMill. Also, we
don't tend to release stuff thats highly
experimental/hacked/unstable. At the moment SawMill (or whatever it
will be called) is too much of a moving target to be anywhere near a
release.



I hope this helps to clarify where Karlsruhe stands with respect to
the Hurd and related issues.

Happy New Year to you all,
Gernot
--
Gernot Heiser                   Visiting Professor
Phone:  +49-721-608-4053        Systems Architecture Group
Mobile: +49-162-861-3179        University of Karlsruhe, Germany
Fax:    +49-721-608-7664        at home at UNSW, Sydney, Australia