Re: [Freeipmi-users] SUNBMC failure

[Al Chu]

Hey Dave,

On Mon, 2010-05-03 at 12:52 -0700, Dave Love wrote:
> Al Chu <address@hidden> writes:
> 
> > GPL and the CDDL (the Sun license) have incompatibilities with each
> > other (see http://*www.*gnu.org/licenses/license-list.html), which is why
> > I did it this way.  However, it appears the incompatibilities may exist
> > only w/ linking and not with code.  So presumably I could copy the code
> > in.  But I'm not sure.  I've been trying to figure this out, but as you
> > can expect, licenses are hard to understand :-)  I'll ask around, see if
> > I can tweak this or not.
> 
> Sure if it was CDDL code, you couldn't distribute it to combine with
> GPL'ed code unless the copyright holders on all the GPL code allowed an
> exception.  However, as I understand it, a header with no code isn't a
> problem.  Anyway, what I pointed to is BSD-ish, not CDDL, which is why I
> think there's no problem -- you'd just need to add the copyright notice
> at most.  That version is trivially different to the one in OpenSolaris.

Ahh!  That was a subtlety I missed before.  There is the bmc_intf.h from
OpenSolaris and the bmc_intf.h from ipmitool.  I had always assumed
ipmitool copied it from OpenSolaris and thus both were CDDL.  But that
is not the case.  Ipmitool's bmc_intf.h appears to be "forked" from an
earlier version, so I could freely use that version.

Thanks for the catch.  I'll look into it a bit more deeply.  Assuming no
snags, I can copy that file in.  I'll make a beta, and if you have the
chance, could you try out the beta for me?

Al

> > Hmmm.  I think the reason it fails like that on Linux is b/c /dev/bmc
> > doesn't exist.  This is in contrast to Solaris where /dev/bmc does
> > exist, but something doesn't allow the code to work with it (thus
> > "device not found" would not be a good error code).  Looking through the
> > error code list, IPMI_ERR_DEVICE_NOT_SUPPORTED is probably the better
> > one.  Sound like a good error code to change to?
> 
> Right, that sounds right -- definitely better than an internal error.
> If you can just distribute the header, it's moot anyway.
> 
> For what it's worth, here's the meat of the licence on bmc_intf.h from
> ipmitool:
> 
>  * Redistribution and use in source and binary forms, with or without
>  * modification, are permitted provided that the following conditions
>  * are met:
>  * 
>  * Redistribution of source code must retain the above copyright
>  * notice, this list of conditions and the following disclaimer.
>  * 
>  * Redistribution in binary form must reproduce the above copyright
>  * notice, this list of conditions and the following disclaimer in the
>  * documentation and/or other materials provided with the distribution.
> 
-- 
Albert Chu
address@hidden
Computer Scientist
High Performance Systems Division
Lawrence Livermore National Laboratory