Stuart Kreitman wrote:
> Garrett:
>
> If you indeed manage to get ATI's attention on these matters, and keep
> your foot in the door so that we can pursue programming specs for all
> of their current products, we would be incredibly impressed and
> indebted, but what connection do you have that can go where no Sun VP
> or Xorg developer (including a contractor doing ddx work under hire to
> ATI and under strict NDA) has gone before?
I can't entirely say, but lets just say my connection got a fully
working radeon driver with BIOS parser (that's key new information, not
ever released before) for NetBSD released last month. Take a look at
src/sys/dev/pci/radeonfb*.[ch] in the public NetBSD tree. I wrote that
code, using ATI NDA information, and got ATI to review it internally and
approve it for public open source release.
I think the same channel would work for releasing the preexisting Sun
driver.
I do _not_ think I will be able to get programming specs for _new_
hardware (or 3D specs for old hardware, beyond what I already have), out
of them, however.
In that regard, ATI seem to be getting more and more closed-lipped. But
pfb and and m64 drivers are 2D only, and represent no substantially new
technology that would enable one of their competitors.
-- Garrett
>
> Stuart Kreitman
> Sun X server group
>
> Garrett D'Amore wrote:
>> Linda Fellingham wrote:
>>> Garrett,
>>>
>>> I don't intend to "pass the buck" - it is just a fact that Sun OEM'd
>>> both the hardware and software for pgx32 from TechSource and we never
>>> had source for it.
>>>
>>> Similarly, we cannot release source for our drivers for the 3DLabs and
>>> ATI-based cards because we have signed NDA's with those companies not
>>> to release their proprietary information. The binaries, though, are
>>> part of OpenSolaris.
>>>
>>> For the old cards, it is possible that these vendors would allow us to
>>> release source but legal resources would be required to do this. Maybe
>>> the Software organization is willing to pick this challenge up?
>>
>> I'm almost certain that this would be possible. Heck, I got permission
>> to release Radeon sources from ATI including some stuff that is not, to
>> my knowledge, in _any_ other operating system.
>>
>> I'd be very surprised if 3DLabs had much desire to keep the drivers
>> close sourced. They aren't making any more products and want to sell
>> the inventory that they already have. If open sourcing would help them
>> move that inventory, I think they'd do it in a heart beat without regard
>> to whether it might give a competitor information about their
>> implementation. (Mostly because, as I said, they're exiting the space,
>> and don't really care what their former competitors do or don't do.)
>>
>> Heck, I'll go one better, Linda.
>>
>> I have some older m64 and recent pfb sources thru my contacts at Sun. I
>> have contacts at ATI _and_ at 3DLabs.
>>
>> I'm willing to state publically, that I'm willing to approach both ATI
>> and 3DLabs, to get permission to release the current Sun source code.
>> If Sun (your group) will agree to release Sun's interest in the drivers
>> to the open source (under CDDL), I'm fairly certain that I can get ATI
>> and 3DLabs to agree to the release.
>>
>> In otherwords, if you can work out _Sun's_ interest in the drivers, I
>> will drive the other side.
>>
>> -- Garrett
>>
>>> Linda
>>>
>>> Garrett D'Amore wrote:
>>>> Linda Fellingham wrote:
>>>>> PGX32 drivers belong to TechSource. Sun does not have source.
>>>> Ah, pass that buck! :-) The drivers were part of the OpenSolaris
>>>> release, and Sun has shipped (iirc) both drivers and hardware based on
>>>> PGX32 in the past. Perhaps Sun can engage TechSource?
>>>>
>>>> Meanwhile, I'm starting another debug. Its starting to look very much
>>>> like dtgreet is implicated. Perhaps it uses a different mechanism to
>>>> blank the screen than the normal screen savers do?
>>>>
>>>> -- Garrett
>>>>
>>>>> Linda
>>>>>
>>>>> Garrett D'Amore wrote:
>>>>>> I've noticed that when the screensaver activates on my pgx32
>>>>>> (running at
>>>>>> 1280x1024, but I don't think this is terribly relevant), my screen
>>>>>> goes
>>>>>> dark _and stays dark_. I.e. it never comes back with a mouse
>>>>>> move or
>>>>>> keyboard activity.
>>>>>>
>>>>>> This can be seen just at dtlogin, with no other graphics activity.
>>>>>>
>>>>>> Killing/restarting X seems to bring the display back. (Killing X
>>>>>> will
>>>>>> bring back a reasonable console.)
>>>>>>
>>>>>> This is with b50. I _believe_ that s10 06/06 did not suffer from
>>>>>> this. I can reload and retest if necessary.
>>>>>>
>>>>>> I don't have pgx32 source, so I can't help with debug much.
>>>>>>
>>>>>> Can someone take a look at it? Probably this should be filed as a
>>>>>> bug,
>>>>>> if not already known. If someone from the X group at Sun agrees,
>>>>>> let me
>>>>>> know and I'll file one.
>>>>>>
>>>>
>>>
>>
>>
--
Garrett D'Amore, Principal Software Engineer
Tadpole Computer / Computing Technologies Division,
General Dynamics C4 Systems
http://www.tadpolecomputer.com/
Phone: 951 325-2134 Fax: 951 325-2191
