13 Nov
2024
13 Nov
'24
12:12 a.m.
About that IBM-written and Sequent-written code in Linux: It wasn't just
algorithms, it was actual code. Not snippets, but dozens, even hundreds of
lines of code put into Linux from AIX and Dynix. This was different from
the copyright case, where there was disagreement about whether copying
occurred. (Again, the code I'm talking about from AIX and Sequent was NOT
AT&T code. It was entirely authored by IBM and Sequent.)
Whether IBM did anything unethical, illegal, or inappropriate is another
matter entirely. I think the lawyers argued back and forth for years, over
a decade even, about this, but it was way out of my scope.
Marc
On Mon, Nov 11, 2024 at 7:34 PM Kevin Bowling <kevin.bowling(a)kev009.com>
wrote:
On Mon, Nov 11, 2024 at 6:55 PM Kevin Bowling
<kevin.bowling(a)kev009.com>
wrote:
--
*My new email address is mrochkind(a)gmail.com <mrochkind(a)gmail.com>*
On Mon, Nov 4, 2024 at 8:14 PM Marc Rochkind <mrochkind(a)gmail.com>
wrote:
>
> Just to repeat, because of a bunch of confused posts here: The breach
of
contract case was not about System V code in Linux. It was about
non-AT&T code from System V derivatives (e.g., AIX, Dynix) into Linux. (The
copyright case was completely different.) You may wonder why non-AT&T code
from a System V derivative into LInux should be a legal issue. To find the
answer you have to read the contract. If it sounds bonkers, then we can
agree that the contract was bonkers.
Marc,
I want to thank you for disclosing your experience. My own
understanding of all this was basically whatever groklaw said and now
that all the dust is settled it's easier to hear and consider what
else was happening.
Dynix was a BSD 4.2 derivative which would make a lot of the
surrounding discussion in this thread appropriate. Although with
It was hard to re-find an authority for the BSD root but thanks to
this list [1] which is a neat paper all around. It also seems like by
the time it was called Dynix/ptx it was fully enmeshed with SysV not
unlike SunOS->Solaris. Maintained a concept called "universes" which
altered the ABI and commands to suit BSD or SysV environments.
There's a great dump of information here
https://www.krsaborio.net/unix-scalability/index.html but it doesn't
go back far enough for Sequent.
[1]
https://archive.org/details/1985-proceedings-summer-portland/page/254/mode/…
commercial OS there is no telling what kind of
mixing went on, for
instance Chalie has variously described the BSD and SysV mixing going
on in AIX on this list and elsewhere.
It is pretty clear that RCU in Linux was a direct teleport of the
algorithms developed at Sequent but maybe the code underwent some
intentional churn as Grog mentions of the JFS work (for the record the
JFS in Linux is more affined to OS/2, the JFS1 and JFS2 in AIX is a
little different than both).
I wonder if at some point SCO scored an "own-goal" on both cases in
essentially the same way that USL did where during discovery you find
out that some legally dubious things happened in both directions. It
seems like they probably could have executed some kind of shakedown or
at least a favorable situation with IBM had the stakes been lower, but
the cases were both very wide reaching and burnt off whatever kinetic
value was there into lawyer heat.
Regards,
Kevin
> I don't know how strong the copyright case was. I didn't work on it.
>
> Marc
>
> On Mon, Nov 4, 2024 at 7:13 PM Warner Losh <imp(a)bsdimp.com> wrote:
>>
>>
>>
>> On Mon, Nov 4, 2024, 6:54 PM Larry McVoy <lm(a)mcvoy.com> wrote:
>>>
>>> On Mon, Nov 04, 2024 at 06:35:30PM -0700, Warner Losh wrote:
>>> > On Mon, Nov 4, 2024 at 6:09???PM Larry McVoy <lm(a)mcvoy.com>
wrote:
>>> >
>>> > > The thing I never got a reasonable answer to was I found code in
BSD that
>>> > > was identical to code
going back to at least V7. Find bmap() in
the UFS
>>> > > code and then find the
same in V7. I might be wrong about V7,
might be
>>> > > 32V, might be V6. I
don't think it matters, it's the same in
all of them.
>>> >
>>> >
>>> > bmap() is the code that maps a logical block to a phsyical block,
>>> > > I'm quite familiar with it because I rewrote it to
bmap_write()
and
>>> > > bmap_read() as part of
making UFS do extents:
>>> > >
>>> > > http://mcvoy.com/lm/papers/SunOS.ufs_clustering.pdf
>>> > >
>>> > > When all the lawsuits were going on, since I knew that code
really well,
>>> > > I went off and looked and
the BSD code at that time had bit for
bit
>>> > > identical bmap()
implementations.
>>> > >
>>> > > I never understood why BSD could claim they rewrote everything
when they
>>> > > clearly had not rewritten
that.
>>> > >
>>> > > I've raised this question before and I just went and looked,
bmap() has
>>> > > changed. I'm pretty
sure I have Kirk's BSD source releases, if
I do,
>>> > > I'm 100% sure I can
back up what I'm saying. Not sure I care
enough to
>>> > > do so, it's all water
under the bridge at this point.
>>> > >
>>> >
>>> > The short answer is that ffs_bmap.c was one of the 70 files that
had
>>> > a AT&T copyright notice
added to it as part of the AT&T vs Regents
suit.
>>> > By the time 4.4BSD had been
released, the file had been
substantially
>>> > rewritten, but some traces of
original AT&T code remained.
>>>
>>> Yeah, this is completely a false claim. It was identical. At least
>>> in 4.3 BSD, I can imagine that 4.4 changed it because I was pointing
>>> this out around then.
>>
>>
>> 4.3bsd wasn't claimed to be a rewrite. 4.4bsd definitely was very
different. I checked before I posted. So what i said is not false. I
literally had the code up side by side 20 minutes ago. It is definitely
different though clearly related and derived a bit. That function is
absolutely not 100% copied.
>>
>>> For the record, I'm a BSD guy, my OS was SunOS 4.x, it was a bug
fixed
>>> BSD. If there ever was a guy that
wanted this to be true, it's me.
>>> It's not true, BSD ripped off Bell Labs code, that's a fact.
>>
>>
>> Except not in 4.4. 4.3 never was claimed to be a rewrite. You needed
a
AT&T license, prior to the ancient Unix license to get that. So there was
no claim to originality prior to 4.4. I didn't look at net/2 though.
>>
>> I'll check after dinner for 4.3bsd and 4.2bsd, but since FFS/UFS is
on disk different than v7fs I don't expect it to be identical.
>>
>> Warner
>
>
>
> --
> My new email address is mrochkind(a)gmail.com