dtrace discuss - Mar 2007 - Licensing for Dtrace headers?

Hello;

I have been following the dtrace port for FreeBSD but unfortunately there have
not been new commits in a while. I recall there were problems with the Dtrace
header licensing that makes it impossible to ship a kernel that supports dtrace.
The problem is only with some headers; the main dtrace code is in a loadable
kernel module which would have to be loaded if one wants to use dtrace at all.

It looks like dtrace won''t be available in FreeBSD 7.0, is there any
chance the licensing situation will be solved in the near future?

best regards,

Pedro.
 
 
This message posted from opensolaris.org

>Hello;
>I have been following the dtrace port for FreeBSD but unfortunately
>there have not been new commit s in a while. I recall there were
>problems with the Dtrace header licensing that makes it impossibl e to
>ship a kernel that supports dtrace. The problem is only with some
>headers; the main dtrace cod e is in a loadable kernel module which
>would have to be loaded if one wants to use dtrace at all.
>It looks like dtrace won''t be available in FreeBSD 7.0, is there
any
>chance the licensing situatio n will be solved in the near future?
Why is this our problem?

We cannot easily solve "we don''t accept licenses other than
X" issues;
those are not our issues, but other people''s issues.

Casper

In response to Dik Casper;

I somehow had the impression SUN wanted to see dtrace in other Operating
Systems, but you are absolutely right, It''s certainly not your problem.

I also hereby apologize, this forum is not really meant to discuss boring
licensing issues.

best regards,

    Pedro.
 
 
This message posted from opensolaris.org

Pedro Giffuni <giffunip at yahoo.com> wrote:
> In response to Dik Casper;
>
> I somehow had the impression SUN wanted to see dtrace in other Operating
Systems, but you are absolutely right, It''s certainly not your problem.
I had a discussion with the NetBSD folks about a year
ago and my impression was that they only like to see it implemented
in a way that does not make dtrace a default basic feature.

The same thing should be true for FreeBSD.

J?rg

-- 
 EMail:joerg at schily.isdn.cs.tu-berlin.de (home) J?rg Schilling D-13353 Berlin
       js at cs.tu-berlin.de                (uni)  
       schilling at fokus.fraunhofer.de     (work) Blog:
http://schily.blogspot.com/
 URL:  http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily

>In response to Dik Casper;
>
>I somehow had the impression SUN wanted to see dtrace in other
>Operating Systems, but you are abso lutely right, It''s certainly
not
>your problem.
>I also hereby apologize, this forum is not really meant to discuss
>boring licensing issues.
It''s certainly something for license-discuss and not dtrace-discuss.

Casper

On 3/31/07, Pedro Giffuni <giffunip at yahoo.com>
wrote:
> I somehow had the impression SUN wanted to see dtrace in other Operating
> Systems
That''s my impression too...

BTW... from what I learned in the SCO vs IBM case, the "#define"s in
header files are not subject to copyright protection.

In fact, in BrandZ (an OpenSolaris project to run Linux user-sparce
programs in Zones), all the "#define"s for syscall numbers are
recreated in a Solaris header file.

Pedro, can you let me know exact which file(s) is causing problem??

Rayson





 but you are absolutely right, It''s certainly not your
problem.
>
> I also hereby apologize, this forum is not really meant to discuss boring
licensing issues.
>
> best regards,
>
>    Pedro.
>
>
> This message posted from opensolaris.org
> _______________________________________________
> dtrace-discuss mailing list
> dtrace-discuss at opensolaris.org
>

On 31/03/07, Pedro Giffuni <giffunip at yahoo.com>
wrote:
> In response to Dik Casper;
>
> I somehow had the impression SUN wanted to see dtrace in other Operating
Systems, but you are absolutely right, It''s certainly not your problem.
>
They do. It is up to others whether or not to use it under the terms
it is available. As others said, I suggest license-discuss.

-- 
"Less is only more where more is no good." --Frank Lloyd Wright

Shawn Walker, Software and Systems Analyst
binarycrusader at gmail.com - http://binarycrusader.blogspot.com/

On Sat, Mar 31, 2007 at 05:31:01PM +0200, Casper.Dik at sun.com
wrote:
> >It looks like dtrace won''t be available in FreeBSD 7.0, is
there any
> >chance the licensing situatio n will be solved in the near future?
> 
> Why is this our problem?
Casper, with all due respect, this _is_ our problem. This is the discussion
list for the broadly defined DTrace community and that includes users who
may be using DTrace on other platforms. The DTrace community wants to
encourage adoption and proliferation of DTrace as the gold standard of
dynamic tracing, and we want to be aware of any impediments to that end.

Adam

-- 
Adam Leventhal, Solaris Kernel Development       http://blogs.sun.com/ahl

I for one do not think this is a boring topic, and would appreciate any
information on the status of this issue, and wether its likely to be resolved in
the future.

Im sorry if there is a more apropriate forum for this.
 
 
This message posted from opensolaris.org

On 08/04/07, Shell Woolvic <ktheroot at gmail.com>
wrote:
> I for one do not think this is a boring topic, and would appreciate any
information on the status of this issue, and wether its likely to be resolved in
the future.
>
> Im sorry if there is a more apropriate forum for this.
I think the main problem is that no one has stated exactly what the
specific issue and why it is an issue. I''ve seen some claims that
using the includes would somehow affect programs using the headers,
but I''ve seen nothing to substantiate that other than guesses.

The real thing is to figure out if the license would *actually* affect
things that do an #include on the DTrace headers, and then see if a
license change is needed for them.

What I''ve read from the various SCO Cases and the like is that simple
"#define" and API declarations have no effect.

Perhaps the easiest approach here is not to change *which* license is
used, but to clarify the interpretation of the CDDL in this scenario,
or amend it as the FSF has done theirs.

I think most people would agree that it is rather silly to expect
doing an #include of DTrace headers to have any affect at all. I don''t
think that was the intention when the license was chosen.

-- 
"Less is only more where more is no good." --Frank Lloyd Wright

Shawn Walker, Software and Systems Analyst
binarycrusader at gmail.com - http://binarycrusader.blogspot.com/