Re: Am I missing something?

[Marcus Brinkmann]

Hi Eric,

the reason you are only finding old info is that in the last months I
put other priorities up-front, and I was and probably still am the
main driving force behind the Hurd to L4 port (there are other people,
but their involvement, although substantial and even crucial, dates
back even further).

That doesn't mean the project is dead, it's stalled.  But note that
even if I were hacking on it 20 hours a day, it would be months from
being in a state of usability that the Hurd on Mach is.

We are painfully aware of the age of Mach, and we are positiviely
aware of L4 and its benefits.  However, please also note that L4 is
not a magic key to success: There is a lot of work required to fill
the gap between what L4 provides and what we need in the Hurd, and the
Hurd itself needs some design fixes.  I have detailed some of this in
some design documents, including the one in the doc directory of the
hurd-l4 repository.

It's a lot of hard work, and it requires intimate knowledge of
operating system design, the Hurd, and Mach (yes, it's required, and
if only to understand the Hurd source as it is right now).

Note that the Hurd is not tied to Mach conceptionally, but it's true
that all the code has to change.  I am not sure if that can be avoided
much.  Abstraction layers at that level of the system tend to kill
performance - all it takes is a function call, pointer indirection or
memory page access too much at the wrong place.

In any case, you should be delighted to hear that _what_ is there in
the hurd-l4 directory does indeed work, although it doesn't do much,
certainly hardly anything visible.

Discussion happens on address@hidden (IIRC).

Thanks,
Marcus