[dcmf] Lockless ROMIO driver
Rob Ross
rross at mcs.anl.gov
Wed Feb 13 11:53:31 CST 2008
Hi Joe,
Ok. I appreciate your concern, and hopefully we can come up with a
streamlined process to minimize overhead.
I'm going to explain my understanding of the patch, then I'll ask how
we should proceed in the future.
My understanding is that RobL's patch includes one new file, adio/
common/ad_write_nolock.c, with a Univ. of Chicago copyright applied
(as per all our other code). All the other files that he touches
already exist and existed in the ROMIO version that you obtained from
us, so those files already have the Univ. of Chicago copyright in them
(and thus there is no copyright line in the diff -- it's the same!).
So post-patch the copyrights are all *correct* (I believe), but I
think you are asking something different?
I'm a little concerned about the second point, but before I flip
out :), how has IBM dealt with this in the past?
Thanks,
Rob
On Feb 13, 2008, at 11:21 AM, Joseph Ratterman wrote:
>
> Rob,
>
> I think these will be handled pretty quickly. As I understand it,
> the problem is on two fronts:
> • The patch lacks an explicit copyright notice, so we don't know
> that it can be used.
> • We don't have the written guarantee that the patch was written by
> R Latham and therefore legal.
> I certainly don't intend to imply that either of these are actually
> a problem, but it is important that we ensure that we follow the
> best practices. IBM is especially concerned about these sorts of
> issues, and follow the rules from the beginning.
>
> All that being said, our current changes are moving quite a bit
> faster than can be seen externally. I just uploaded another tarball
> to the wiki, but that's not really a good way to do all this. We
> are really hoping that a git server will be available soon.
>
>
> Thanks,
> Joe Ratterman
> jratt at us.ibm.com
>
>
>
>
> Rob Ross <rross at mcs.anl.gov>
> 02/13/08 11:05 AM
>
> To
> Joseph Ratterman/Rochester/IBM at IBMUS
> cc
> Rob Latham <robl at mcs.anl.gov>, Judith W Hjortness/Rochester/
> IBM at IBMUS, Todd Inglett/Rochester/IBM at IBMUS, dcmf at lists.anl-external.org
> Subject
> Re: [dcmf] Lockless ROMIO driver
>
>
>
>
>
> Hi Joe,
>
> I agree that in general we must be careful that patches are provided
> under a consistent license. In the case of Argonne patches to this
> software, we are providing patches under the same license and
> copyright under which the code was originally written. Can we fast-
> track these?
>
> Thanks,
>
> Rob
>
> On Feb 13, 2008, at 9:28 AM, Joseph Ratterman wrote:
>
> >
> > This one is mostly for external contributors.
> >
> > Before we can begin to take and use any patches, we need to have a
> > contributor agreement signed by the author. We want to ensure that
> > the DCMF community follows the best practices for avoiding
> copyright/
> > licensing issues, so this is important to all people using the DCMF
> > source code. We plan to get the final agreement information out on
> > the mail list and wiki later today--we are currently wrapping up a
> > few loose ends.
> >
> > After we get this going, we will have to keep a list of
> > signatories. It seems that there are a few ways to record these,
> > and I would like people to comment.
> > + Keep an internal list at IBM (or the future maintainer). This
> > ensures privacy.
> > + Keep a list on the Wiki. Since changes are public and pages can
> > be protected. This is more open, and could also be included with
> > the next:
> > + Maintain a "developers" file inside the git repository. This will
> > be more accessible when we get a git server, and it is even more
> > secure than the wiki list.
> >
> > Thanks,
> > Joe Ratterman
> > jratt at us.ibm.com
> >
> >
> >
> >
> >
> > Rob Latham <robl at mcs.anl.gov>
> > Sent by: dcmf-bounces at lists.anl-external.org
> > 02/12/08 02:22 PM
> >
> > To
> > dcmf at lists.anl-external.org
> > cc
> > Subject
> > [dcmf] Lockless ROMIO driver
> >
> >
> >
> >
> >
> > These patches re-introduce ad_ufs to ROMIO, but do so in a lock-free
> > way. This
> > is my first git experience, so sorry for the messy subject lines.
> >
> >
> > _______________________________________________
> > dcmf mailing list
> > dcmf at lists.anl-external.org
> > http://lists.anl-external.org/cgi-bin/mailman/listinfo/dcmf
> > http://dcmf.anl-external.org/wiki
> >
> > _______________________________________________
> > dcmf mailing list
> > dcmf at lists.anl-external.org
> > http://lists.anl-external.org/cgi-bin/mailman/listinfo/dcmf
> > http://dcmf.anl-external.org/wiki
>
>
More information about the dcmf
mailing list