gatos-devel Mailing List for GATOS

On Sat, 31 May 2003, Cam wrote:

> Hi,
>
> Having read all the replies from developers here I think it's more
> likely just lack of time and enthusiasm that prevents support. It's just
> a shame because proper TV-out support would enable a whole range of
> multimedia projects and encourage development further down the line.
> There may be people who can do the necessary work but the barriers to
> entry are fairly large - you have to be comfortable with compiling X,
> working with hardware, familiar with drivers, etc.

Which is the reason for the devel branch - it was meant to let people who
want to work with TV-out get something working.

If you can get devel branch to work for you - you are in position to start
messing with TV-out, provided, of course, you are sure of your legal
position.

                        best

                           Vladimir Dergachev

Hi,

For a while I was convinced that there was a conspiracy on the part of 
the open source developers, to prevent tv-out from being supported. I 
thought it would be relatively easy for someone who is already familiar 
with the drivers to make TV-out work. I figured that it was just down to 
restricting the modes that the display uses to compatible ones, and 
maybe setting up a scaling unit or something like that.

Having read all the replies from developers here I think it's more 
likely just lack of time and enthusiasm that prevents support. It's just 
a shame because proper TV-out support would enable a whole range of 
multimedia projects and encourage development further down the line. 
There may be people who can do the necessary work but the barriers to 
entry are fairly large - you have to be comfortable with compiling X, 
working with hardware, familiar with drivers, etc.

I bought a card with TV-out for a multimedia project, before it was 
abundantly clear that it was not supported. I figured that the general 
level of development on the drivers and the availability of several 
different drivers would mean that support (in general) would not be a 
problem. In fact TV-out does work under some circumstances so there were 
a few misleading, encouraging pages on the web. Now instead of going on 
to develop an interesting multimedia PC, I'm facing tweaking drivers 
with not much chance of success, or possibly buying a different video 
card and chalking it up to experience.

-Cam

PS. This isn't a complaint to the GATOS developers, I can sympathise 
with not having enough time to devote to open source stuff.

Well, the DMCA has no affect in Canada thus ATI's building is just 10 =
mins
from me. Move development to Canada :)

Shawn.


-----Original Message-----
From: gat...@li...
[mailto:gat...@li...] On Behalf Of Anthony J.
Ciani
Sent: Thursday, May 29, 2003 11:15 PM
To: gat...@li...
Subject: [GATOS]TVout / Legal Issues


Hello All,

	This message is at the same time informative, argumentative, and
moot.  To review the basic problems touched on here please see
http://sourceforge.net/mailarchive/forum.php?thread_id=3D1868811&forum_id=
=3D5014

To start, Macrovision contractual obligations are the main reasons ATI
doesn't release the specs for the TVout feature of its older cards.  =
Just as
an informative note, the reason Macrovision wants this information
restricted is NOT because they fear people will turn off the Macrovision =
and
copy movies, but because people might turn on the Macrovision and =
'protect'
video which has not bought a Macrovision license.  The Macrovision =
license
'usually' only pertains to the addition of Macrovision to the output =
signal,
not the absence of Macrovision on output which should include it.

On the above note, I'm a little surprised that you can't get ATI to meet
half-way on the TVout support.  Either they could provide a =
static-linked
binary which would allow for switching of the display and control of the =
TV
image, or they could provide TVout specs without information on enabling
Macrovision, which should be perfectly allowable under their contract =
with
Macrovision.  For the most part, TV out already works, if you turn on =
the
machine with the TV connected and use VESA calls, or use atitvout (VESA
calls).  The only things lacking are convenience, and the ability to =
adjust
the image.

**This sounds weird, but it MAY even be possible to use the windows =
binaries
to configure the video device.  Unfortunately, I don't know that much =
about
using MS libraries in *NIX programs.

As far as the DMCA goes, it doesn't even apply to computer display =
devices,
unless said device contains hardware specifically designed for playing
protected material (DXR3/Hollywood...).  In that case, it only applies =
to
the DVD playback function of the hardware.  Beyond that, it only applies
when playing protected material through the device which is designed to =
play
the protected media in a protected manner.  To the best of my knowledge, =
the
older ATI video cards do not included DVD=20
decryption/decoding ability, nor does the GATOS driver provide this.
Finally, the DMCA actually makes anything which can at any point remove =
copy
protection, illegal.  This means that GATOS would be illegal with or =
without
TVout, which is quite ludicrous, and the Supreme Court already ruled on =
that
back in 1984, when it ruled against the MPAA, and allowed the marketing =
of a
device called the Video Cassette Recorder.

However, long standing examples are always best to dissuade fears.  =
First,
there are many, many TVout projects for various cards already available =
for
Linux (NVidia, SiS, Matrox, i810), and they have all been ignored by the
MPAA (which could still sue, but would get thrown out on lack of merit).
There are still VGA to TV devices legally available for purchase, and =
none
of them contain Macrovision.  You can legally buy rather inexpensive =
video
editing devices which remove Macrovision.  The Supreme Court made these
legal back in the 1984 Betamax ruling, and subsequent rulings.  The =
TVout
feature of ATI cards can be used on Windows platforms to illegally copy
protected material to video tape.  Hmm, why hasn't ATI stopped =
supporting
TVout under Windoze?  Refer to the second paragraph for the answer.


To summarize:  ATI can give specs for TVout and multi-head display
as-long-as it does not include information to enable Macrovision.  If =
they
don't, it's because of laziness, not legality.  Reverse engineering of =
the
TVout is legally allowable, as long as you do not enable Macrovision on
unlicensed output.



PS:  Just a note.  The Macrovision and CSS protection schemes have been
fairly well proven useless.  Because of this, fewer and fewer studios =
are
licensing them for use on their DVDs.  In fact, I haven't seen anything
Macrovision protected for about 2 years now.

--=20
------------------------------------------------------------
              Anthony Ciani (ac...@ui...)
           Computational Condensed Matter Physics
   Department of Physics, University of Illinois, Chicago
              http://ciani.phy.uic.edu/~tony
------------------------------------------------------------



-------------------------------------------------------
This SF.net email is sponsored by: eBay
Get office equipment for less on eBay!
http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
_______________________________________________
Gatos-devel mailing list
Gat...@li...
https://lists.sourceforge.net/lists/listinfo/gatos-devel

On Fri, 30 May 2003, Anthony J. Ciani wrote:

> The tvout.c and i2c.c were part of GATOS way back around 0.0.5, and
> have Steaphan Greene's name on them.  From looking at the source, I'd
> almost have to say they were working from specs.  It looks as though all
> of the registers are listed that have anything to do with formatting the
> image, plus some.  It may even be complete enough to make a command line
> utility to switch on the TVout (like atitvout) and adjust the image.
>
> I think I'll also show it to Lennart Poettering, who was (until just
> recently) maintaining atitvout.  His used VESA calls, but tvout.c
> addresses the registers directly.
>

Well, the sourcecode is up on the webpage (as a cvs tarball), feel free to
experiment.

                        best

                          Vladimir Dergachev

>
> As far as suits, any suit against such an innocuous project would have to
> be harassment.  The end result; a counter suit and money, after another
> suit to make them pay what the judge ordered them to in the counter
> suit.  But yes, it is a pain in both cheeks, which is the reason for the
> counter suit.
>
> The SCO suit is also interesting for one reason that probably hasn't
> appeared on SlashDot.  Remember the Federal anti-trust suit against
> Microsoft?  Microsoft and the Justice Department kept negotiating more and
> more lucrative "punishments" for Microsoft, even though the judge was
> willing to put Microsoft through the blender.  The attourney representing
> the government in that case was David Boies, the same attourney
> representing SCO.  Microsoft has made a huge investment into SCO, so the
> SCO suit against IBM is really just Microsoft harasment against
> Linux.  But putting the two facts together yeilds this realization: The
> lawyer that was supposed to have played hardball against Microsoft for the
> government, went soft on Microsoft and now virtually works for them.
>
> On Fri, 30 May 2003, Vladimir Dergachev wrote:
> >
> > On Fri, 30 May 2003, Anthony J. Ciani wrote:
> >
> > > Hello again,
> > >
> > > 	I have learned to pay little attention to SlashDot.  The SCO vs
> > > IBM case hasn't even gone in front of a judge yet.  BTW, the SCO suit does
> > > have legal merit, and IBM could be in for a hurt if it did use SCO
> > > technologies in unlicensed products (i.e. Linux).  Of course, IBM didn't,
> > > but that will have to be shown in court.  In the case of TVout, there is
> > > no legal basis, thus no merit, and no lawsuits.
> >
> > You missed my point. What would *you* do if you were presented with a
> > lawsuit ? I am pretty sure you will spend a fair amount of time studying
> > the details and perhaps some being upset. And you will probably have to
> > hire a lawyer. Whether or *NOT* the lawsuit has legal basis.
> >
> > In my case, there is really no point in risking this, while I do not have
> > time for TV-out in the first place.
> >
> > Here is a short list of outstanding issues with *current* drivers:
> >
> >    * debug and verify that km from CVS works with mach64 and rage128
> >      cards
> >
> >    * debug gcc 3.2.x issues. Find out whether it is really necessary
> >      to have two binary sets of modules.
> >
> >    * merge code with XFree86 CVS when the latter is closer to release.
> >      this would be a lot of work because of progress in merging ati.2
> >      into mach64 driver
> >
> >    * Rage Theatre 200 support (if ATI does create and release the docs -
> >      it is taking them a while). This is a *very* big job, as Rage Theatre
> >      200 includes audio functions, which might require major redesign.
> >
> > Things that would be nice to have:
> >
> >    * second Xv port that uses 3d textures (playback only)
> >
> >    * complete km_api so that km 0.99.x can be started
> >
> >    * nicer capture interface for AVview
> >
> > Furthermore, ATI sells component adaptor that allows connection to HDTV
> > sets. AFAIK, this does not use Rage Theatre, but rather DVI port.
> > I would imagine this could be made working by carefully studying current
> > code - and having proper hardware, of course.
> >
> > Unlike regular TV-out, component adaptor is a technology that is not
> > likely to go away after a few years.
> >
> > Lastly, all this needs to be done in spare time, which is always hard to
> > come by, especially in large chunks.
> >
> > >
> > > 	I'm a little confused about the state of Tvout support in
> > > GATOS.  There are two objects included in GATOS (tvout.o and
> > > i2cmpp.o) which control the ImpacTV chip for TVout display.  The ImpacTV
> > > chip is (more-or-less) exactly the same as the ImpacTV2, except that it
> > > doesn't include Macrovision or support for 16:9 modes.  Did ATI give GATOS
> > > the specs for the ImpacTV chip, or were these reverse engineered?
> >
> > No idea. Which code are you referring to ? xatitv ? This has been a good
> > while ago. I do not believe tvout.c was ever finished.
> >
> > >
> > > 	Before completely nixing the idea of TVout, might I suggest
> > > contacting ATI again.  This time be rather specific that you do not want
> > > specifications for the Macrovision feature, but only stripped down specs
> > > on theory of operation, activation, positioning, and scaling.
> >
> > If you would like to work on TV-out go ahead and try. If nothing else this
> > would be a very good toy project to learn about multimedia.
> >
> >                     best
> >
> >                        Vladimir Dergachev
> >
> > >
> > > On Fri, 30 May 2003, Vladimir Dergachev wrote:
> > > > Anthony,
> > > >
> > > >    Yeah, you do have some good points.
> > > >
> > > >    You missed one though: the lawsuit does not have to have solid legal
> > > > basis under it, just be complicated enough so that the court does not
> > > > dismiss it right away (see, for example, slashdot notes about SCO).
> > > >
> > > >    So, personally, I consider that I do not want to pursue TV-out, as it
> > > > *may* have legal issues, has fairly low resolution (800x600 at best),
> > > > and understanding of TV-out will in no way help toward making future HDTV
> > > > drivers, if these will become possible.
> > > >
> > > >    You are, of course, free to form your own opinion.
> > > >
> > > >                              best
> > > >
> > > >                                 Vladimir Dergachev
> > > >
> > > > On Thu, 29 May 2003, Anthony J. Ciani wrote:
> > > >
> > > > > Hello All,
> > > > >
> > > > > 	This message is at the same time informative, argumentative, and
> > > > > moot.  To review the basic problems touched on here please see
> > > > > http://sourceforge.net/mailarchive/forum.php?thread_id=1868811&forum_id=5014
> > > > >
> > > > > To start, Macrovision contractual obligations are the main reasons ATI
> > > > > doesn't release the specs for the TVout feature of its older
> > > > > cards.  Just as an informative note, the reason Macrovision wants this
> > > > > information restricted is NOT because they fear people will turn off the
> > > > > Macrovision and copy movies, but because people might turn on the
> > > > > Macrovision and 'protect' video which has not bought a Macrovision
> > > > > license.  The Macrovision license 'usually' only pertains to the addition
> > > > > of Macrovision to the output signal, not the absence of Macrovision on
> > > > > output which should include it.
> > > > >
> > > > > On the above note, I'm a little surprised that you can't get ATI to meet
> > > > > half-way on the TVout support.  Either they could provide a
> > > > > static-linked binary which would allow for switching of the display and
> > > > > control of the TV image, or they could provide TVout specs without
> > > > > information on enabling Macrovision, which should be perfectly allowable
> > > > > under their contract with Macrovision.  For the most part, TV out already
> > > > > works, if you turn on the machine with the TV connected and use VESA
> > > > > calls, or use atitvout (VESA calls).  The only things lacking are
> > > > > convenience, and the ability to adjust the image.
> > > > >
> > > > > **This sounds weird, but it MAY even be possible to use the windows
> > > > > binaries to configure the video device.  Unfortunately, I don't know that
> > > > > much about using MS libraries in *NIX programs.
> > > > >
> > > > > As far as the DMCA goes, it doesn't even apply to computer display
> > > > > devices, unless said device contains hardware specifically designed for
> > > > > playing protected material (DXR3/Hollywood...).  In that case, it only
> > > > > applies to the DVD playback function of the hardware.  Beyond that, it
> > > > > only applies when playing protected material through the device which is
> > > > > designed to play the protected media in a protected manner.  To the best
> > > > > of my knowledge, the older ATI video cards do not included DVD
> > > > > decryption/decoding ability, nor does the GATOS driver provide
> > > > > this.  Finally, the DMCA actually makes anything which can at any point
> > > > > remove copy protection, illegal.  This means that GATOS would be illegal
> > > > > with or without TVout, which is quite ludicrous, and the Supreme Court
> > > > > already ruled on that back in 1984, when it ruled against the MPAA,
> > > > > and allowed the marketing of a device called the Video Cassette Recorder.
> > > > >
> > > > > However, long standing examples are always best to dissuade fears.  First,
> > > > > there are many, many TVout projects for various cards already available
> > > > > for Linux (NVidia, SiS, Matrox, i810), and they have all been ignored by
> > > > > the MPAA (which could still sue, but would get thrown out on lack of
> > > > > merit).  There are still VGA to TV devices legally available for purchase,
> > > > > and none of them contain Macrovision.  You can legally buy rather
> > > > > inexpensive video editing devices which remove Macrovision.  The Supreme
> > > > > Court made these legal back in the 1984 Betamax ruling, and subsequent
> > > > > rulings.  The TVout feature of ATI cards can be used on Windows platforms
> > > > > to illegally copy protected material to video tape.  Hmm, why hasn't ATI
> > > > > stopped supporting TVout under Windoze?  Refer to the second paragraph for
> > > > > the answer.
> > > > >
> > > > >
> > > > > To summarize:  ATI can give specs for TVout and multi-head display
> > > > > as-long-as it does not include information to enable Macrovision.  If they
> > > > > don't, it's because of laziness, not legality.  Reverse engineering of
> > > > > the TVout is legally allowable, as long as you do not enable Macrovision
> > > > > on unlicensed output.
> > > > >
> > > > >
> > > > >
> > > > > PS:  Just a note.  The Macrovision and CSS protection schemes have been
> > > > > fairly well proven useless.  Because of this, fewer and fewer studios are
> > > > > licensing them for use on their DVDs.  In fact, I haven't seen anything
> > > > > Macrovision protected for about 2 years now.
> > > > >
> > > > > --
> > > > > ------------------------------------------------------------
> > > > >               Anthony Ciani (ac...@ui...)
> > > > >            Computational Condensed Matter Physics
> > > > >    Department of Physics, University of Illinois, Chicago
> > > > >               http://ciani.phy.uic.edu/~tony
> > > > > ------------------------------------------------------------
> > > > >
> > > > >
> > > > >
> > > > > -------------------------------------------------------
> > > > > This SF.net email is sponsored by: eBay
> > > > > Get office equipment for less on eBay!
> > > > > http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> > > > > _______________________________________________
> > > > > Gatos-devel mailing list
> > > > > Gat...@li...
> > > > > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> > > > >
> > > >
> > >
> > > --
> > > ------------------------------------------------------------
> > >               Anthony Ciani (ac...@ui...)
> > >            Computational Condensed Matter Physics
> > >    Department of Physics, University of Illinois, Chicago
> > >               http://ciani.phy.uic.edu/~tony
> > > ------------------------------------------------------------
> > >
> > >
> > >
> > > -------------------------------------------------------
> > > This SF.net email is sponsored by: eBay
> > > Get office equipment for less on eBay!
> > > http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> > > _______________________________________________
> > > Gatos-devel mailing list
> > > Gat...@li...
> > > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> > >
> >
>
> --
> ------------------------------------------------------------
>               Anthony Ciani (ac...@ui...)
>            Computational Condensed Matter Physics
>    Department of Physics, University of Illinois, Chicago
>               http://ciani.phy.uic.edu/~tony
> ------------------------------------------------------------
>

The tvout.c and i2c.c were part of GATOS way back around 0.0.5, and 
have Steaphan Greene's name on them.  From looking at the source, I'd
almost have to say they were working from specs.  It looks as though all
of the registers are listed that have anything to do with formatting the
image, plus some.  It may even be complete enough to make a command line
utility to switch on the TVout (like atitvout) and adjust the image.

I think I'll also show it to Lennart Poettering, who was (until just
recently) maintaining atitvout.  His used VESA calls, but tvout.c
addresses the registers directly.


As far as suits, any suit against such an innocuous project would have to
be harassment.  The end result; a counter suit and money, after another
suit to make them pay what the judge ordered them to in the counter
suit.  But yes, it is a pain in both cheeks, which is the reason for the
counter suit.

The SCO suit is also interesting for one reason that probably hasn't
appeared on SlashDot.  Remember the Federal anti-trust suit against
Microsoft?  Microsoft and the Justice Department kept negotiating more and
more lucrative "punishments" for Microsoft, even though the judge was
willing to put Microsoft through the blender.  The attourney representing
the government in that case was David Boies, the same attourney 
representing SCO.  Microsoft has made a huge investment into SCO, so the
SCO suit against IBM is really just Microsoft harasment against
Linux.  But putting the two facts together yeilds this realization: The
lawyer that was supposed to have played hardball against Microsoft for the
government, went soft on Microsoft and now virtually works for them.

On Fri, 30 May 2003, Vladimir Dergachev wrote:
> 
> On Fri, 30 May 2003, Anthony J. Ciani wrote:
> 
> > Hello again,
> >
> > 	I have learned to pay little attention to SlashDot.  The SCO vs
> > IBM case hasn't even gone in front of a judge yet.  BTW, the SCO suit does
> > have legal merit, and IBM could be in for a hurt if it did use SCO
> > technologies in unlicensed products (i.e. Linux).  Of course, IBM didn't,
> > but that will have to be shown in court.  In the case of TVout, there is
> > no legal basis, thus no merit, and no lawsuits.
> 
> You missed my point. What would *you* do if you were presented with a
> lawsuit ? I am pretty sure you will spend a fair amount of time studying
> the details and perhaps some being upset. And you will probably have to
> hire a lawyer. Whether or *NOT* the lawsuit has legal basis.
> 
> In my case, there is really no point in risking this, while I do not have
> time for TV-out in the first place.
> 
> Here is a short list of outstanding issues with *current* drivers:
> 
>    * debug and verify that km from CVS works with mach64 and rage128
>      cards
> 
>    * debug gcc 3.2.x issues. Find out whether it is really necessary
>      to have two binary sets of modules.
> 
>    * merge code with XFree86 CVS when the latter is closer to release.
>      this would be a lot of work because of progress in merging ati.2
>      into mach64 driver
> 
>    * Rage Theatre 200 support (if ATI does create and release the docs -
>      it is taking them a while). This is a *very* big job, as Rage Theatre
>      200 includes audio functions, which might require major redesign.
> 
> Things that would be nice to have:
> 
>    * second Xv port that uses 3d textures (playback only)
> 
>    * complete km_api so that km 0.99.x can be started
> 
>    * nicer capture interface for AVview
> 
> Furthermore, ATI sells component adaptor that allows connection to HDTV
> sets. AFAIK, this does not use Rage Theatre, but rather DVI port.
> I would imagine this could be made working by carefully studying current
> code - and having proper hardware, of course.
> 
> Unlike regular TV-out, component adaptor is a technology that is not
> likely to go away after a few years.
> 
> Lastly, all this needs to be done in spare time, which is always hard to
> come by, especially in large chunks.
> 
> >
> > 	I'm a little confused about the state of Tvout support in
> > GATOS.  There are two objects included in GATOS (tvout.o and
> > i2cmpp.o) which control the ImpacTV chip for TVout display.  The ImpacTV
> > chip is (more-or-less) exactly the same as the ImpacTV2, except that it
> > doesn't include Macrovision or support for 16:9 modes.  Did ATI give GATOS
> > the specs for the ImpacTV chip, or were these reverse engineered?
> 
> No idea. Which code are you referring to ? xatitv ? This has been a good
> while ago. I do not believe tvout.c was ever finished.
> 
> >
> > 	Before completely nixing the idea of TVout, might I suggest
> > contacting ATI again.  This time be rather specific that you do not want
> > specifications for the Macrovision feature, but only stripped down specs
> > on theory of operation, activation, positioning, and scaling.
> 
> If you would like to work on TV-out go ahead and try. If nothing else this
> would be a very good toy project to learn about multimedia.
> 
>                     best
> 
>                        Vladimir Dergachev
> 
> >
> > On Fri, 30 May 2003, Vladimir Dergachev wrote:
> > > Anthony,
> > >
> > >    Yeah, you do have some good points.
> > >
> > >    You missed one though: the lawsuit does not have to have solid legal
> > > basis under it, just be complicated enough so that the court does not
> > > dismiss it right away (see, for example, slashdot notes about SCO).
> > >
> > >    So, personally, I consider that I do not want to pursue TV-out, as it
> > > *may* have legal issues, has fairly low resolution (800x600 at best),
> > > and understanding of TV-out will in no way help toward making future HDTV
> > > drivers, if these will become possible.
> > >
> > >    You are, of course, free to form your own opinion.
> > >
> > >                              best
> > >
> > >                                 Vladimir Dergachev
> > >
> > > On Thu, 29 May 2003, Anthony J. Ciani wrote:
> > >
> > > > Hello All,
> > > >
> > > > 	This message is at the same time informative, argumentative, and
> > > > moot.  To review the basic problems touched on here please see
> > > > http://sourceforge.net/mailarchive/forum.php?thread_id=1868811&forum_id=5014
> > > >
> > > > To start, Macrovision contractual obligations are the main reasons ATI
> > > > doesn't release the specs for the TVout feature of its older
> > > > cards.  Just as an informative note, the reason Macrovision wants this
> > > > information restricted is NOT because they fear people will turn off the
> > > > Macrovision and copy movies, but because people might turn on the
> > > > Macrovision and 'protect' video which has not bought a Macrovision
> > > > license.  The Macrovision license 'usually' only pertains to the addition
> > > > of Macrovision to the output signal, not the absence of Macrovision on
> > > > output which should include it.
> > > >
> > > > On the above note, I'm a little surprised that you can't get ATI to meet
> > > > half-way on the TVout support.  Either they could provide a
> > > > static-linked binary which would allow for switching of the display and
> > > > control of the TV image, or they could provide TVout specs without
> > > > information on enabling Macrovision, which should be perfectly allowable
> > > > under their contract with Macrovision.  For the most part, TV out already
> > > > works, if you turn on the machine with the TV connected and use VESA
> > > > calls, or use atitvout (VESA calls).  The only things lacking are
> > > > convenience, and the ability to adjust the image.
> > > >
> > > > **This sounds weird, but it MAY even be possible to use the windows
> > > > binaries to configure the video device.  Unfortunately, I don't know that
> > > > much about using MS libraries in *NIX programs.
> > > >
> > > > As far as the DMCA goes, it doesn't even apply to computer display
> > > > devices, unless said device contains hardware specifically designed for
> > > > playing protected material (DXR3/Hollywood...).  In that case, it only
> > > > applies to the DVD playback function of the hardware.  Beyond that, it
> > > > only applies when playing protected material through the device which is
> > > > designed to play the protected media in a protected manner.  To the best
> > > > of my knowledge, the older ATI video cards do not included DVD
> > > > decryption/decoding ability, nor does the GATOS driver provide
> > > > this.  Finally, the DMCA actually makes anything which can at any point
> > > > remove copy protection, illegal.  This means that GATOS would be illegal
> > > > with or without TVout, which is quite ludicrous, and the Supreme Court
> > > > already ruled on that back in 1984, when it ruled against the MPAA,
> > > > and allowed the marketing of a device called the Video Cassette Recorder.
> > > >
> > > > However, long standing examples are always best to dissuade fears.  First,
> > > > there are many, many TVout projects for various cards already available
> > > > for Linux (NVidia, SiS, Matrox, i810), and they have all been ignored by
> > > > the MPAA (which could still sue, but would get thrown out on lack of
> > > > merit).  There are still VGA to TV devices legally available for purchase,
> > > > and none of them contain Macrovision.  You can legally buy rather
> > > > inexpensive video editing devices which remove Macrovision.  The Supreme
> > > > Court made these legal back in the 1984 Betamax ruling, and subsequent
> > > > rulings.  The TVout feature of ATI cards can be used on Windows platforms
> > > > to illegally copy protected material to video tape.  Hmm, why hasn't ATI
> > > > stopped supporting TVout under Windoze?  Refer to the second paragraph for
> > > > the answer.
> > > >
> > > >
> > > > To summarize:  ATI can give specs for TVout and multi-head display
> > > > as-long-as it does not include information to enable Macrovision.  If they
> > > > don't, it's because of laziness, not legality.  Reverse engineering of
> > > > the TVout is legally allowable, as long as you do not enable Macrovision
> > > > on unlicensed output.
> > > >
> > > >
> > > >
> > > > PS:  Just a note.  The Macrovision and CSS protection schemes have been
> > > > fairly well proven useless.  Because of this, fewer and fewer studios are
> > > > licensing them for use on their DVDs.  In fact, I haven't seen anything
> > > > Macrovision protected for about 2 years now.
> > > >
> > > > --
> > > > ------------------------------------------------------------
> > > >               Anthony Ciani (ac...@ui...)
> > > >            Computational Condensed Matter Physics
> > > >    Department of Physics, University of Illinois, Chicago
> > > >               http://ciani.phy.uic.edu/~tony
> > > > ------------------------------------------------------------
> > > >
> > > >
> > > >
> > > > -------------------------------------------------------
> > > > This SF.net email is sponsored by: eBay
> > > > Get office equipment for less on eBay!
> > > > http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> > > > _______________________________________________
> > > > Gatos-devel mailing list
> > > > Gat...@li...
> > > > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> > > >
> > >
> >
> > --
> > ------------------------------------------------------------
> >               Anthony Ciani (ac...@ui...)
> >            Computational Condensed Matter Physics
> >    Department of Physics, University of Illinois, Chicago
> >               http://ciani.phy.uic.edu/~tony
> > ------------------------------------------------------------
> >
> >
> >
> > -------------------------------------------------------
> > This SF.net email is sponsored by: eBay
> > Get office equipment for less on eBay!
> > http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> > _______________________________________________
> > Gatos-devel mailing list
> > Gat...@li...
> > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> >
> 

-- 
------------------------------------------------------------
              Anthony Ciani (ac...@ui...)
           Computational Condensed Matter Physics
   Department of Physics, University of Illinois, Chicago
              http://ciani.phy.uic.edu/~tony
------------------------------------------------------------

Hmmm, is it pretty much hopeless using the vesa driver then? Perhaps I
could run both X servers somehow to enable capturing? I experimented
with this but could not get the vesa server to load (on a different vt
and :1) after the ATI driver one was loaded (complained about no legal
screens and i had to reboot to get vesa to work again).  
I may try Luca Hall's method instead if not...

Thanks for the quick reply,
Paul

On Fri, 2003-05-30 at 11:31, Vladimir Dergachev wrote:

> No, km won't be of much use without Xserver, ati.2 and AVview or xawtv.
> 
>                        best
> 
>                            Vladimir Dergachev
> 
> On Fri, 30 May 2003, Paul Gratz wrote:
> 
> > Hi,
> >    I'm new to this group but I've done my best to read back through the
> > archives and I could not find a good answer to my question. Please
> > forgive me if I err :-).
> >
> >    I own a 7500 AIW and I have used it in a multimedia machine I have
> > built. I bought it some time ago and would like to not buy another card
> > for this machine.  After much reading here and elsewhere I came to the
> > conclusion that I need to use the "vesa" X driver to get the TV out to
> > work on the card (since this is a multimedia machine this is essential,
> > I don't even have a monitor on it right now). After some testing I have
> > that working acceptably well with "freevo" (very cool software).
> >
> >   Anyways my problem is that I would like to use the tv-tuner part of
> > the card as well. To this end I downloaded the KM package from the GATOS
> > project and was able to compile and install it on my system. I already
> > have v4l on the system so that is not the problem. Infact after
> > installing the modules I see that it makes /dev/v4l/video0 and vbi0 (I'm
> > using the proc filesystem). Here is the dmesg output on loading the
> > modules:
> >
> > Linux video capture interface: v1.00
> > Kmultimedia API module version alpha-3.0 loaded
> > Kmultimedia module version alpha-3.0 loaded
> > Page size is 4096 sizeof(bm_list_descriptor)=16 sizeof(KM_STRUCT)=952
> > km: using irq 11
> > Register aperture is 0xe0020000 0x00010000
> > kms variables: reg_aperture=0xd08a5000
> > km: DMA_GUI_STATUS=0x00000002 entries=2
> > sizeof(kmfl_template)=624 sizeof(KM_FIELD)=48
> > Device ATI Technologies Inc Radeon 7500 QW 01:00.0 (0x1002:0x5157)
> > corresponds to /dev/video0
> > kms variables: reg_aperture=0xd08a5000
> >
> >
> > When I try to use them however, I get the following from mplayer:
> >
> > [snip unrelated junk]
> > Playing TV
> > Cache fill:  0.00% (0 bytes)    TV detected! ;-)
> > Selected driver: v4l
> >  name: Video 4 Linux input
> >  author: Alex Beregszaszi <al...@na...>
> >  comment: under development
> > unable to open '/dev/video0': No data available
> > Falling back on trying to parse playlist TV...
> > ============ Sorry, this file format is not recognized/supported
> > =============
> > === If this file is an AVI, ASF or MPEG stream, please contact the
> > author! ===
> >
> > I also tried avview but I got some error that was kind of similar IIRC.
> >
> > Interestingly I get the following in "dmesg" after this attempt:
> >
> > Purging transfer queue
> > WARNING !   Radeon memory controller is misconfigured, disabling capture
> > WARNING !   upgrade your Xserver and DRM driver
> > km: no data is available until AVview or xawtv is started
> >
> >
> > I even tried downloading the drm modules to see if that would help (it
> > did not).  Is there any way I can make this work, or does anyone have
> > any suggestions?
> >
> > Thanks!
> > Paul
> >
> >
> >
> >
> >
> > -------------------------------------------------------
> > This SF.net email is sponsored by: eBay
> > Get office equipment for less on eBay!
> > http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> > _______________________________________________
> > Gatos-devel mailing list
> > Gat...@li...
> > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> >
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: eBay
> Get office equipment for less on eBay!
> http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel

On Fri, 30 May 2003, Anthony J. Ciani wrote:

> Hello again,
>
> 	I have learned to pay little attention to SlashDot.  The SCO vs
> IBM case hasn't even gone in front of a judge yet.  BTW, the SCO suit does
> have legal merit, and IBM could be in for a hurt if it did use SCO
> technologies in unlicensed products (i.e. Linux).  Of course, IBM didn't,
> but that will have to be shown in court.  In the case of TVout, there is
> no legal basis, thus no merit, and no lawsuits.

You missed my point. What would *you* do if you were presented with a
lawsuit ? I am pretty sure you will spend a fair amount of time studying
the details and perhaps some being upset. And you will probably have to
hire a lawyer. Whether or *NOT* the lawsuit has legal basis.

In my case, there is really no point in risking this, while I do not have
time for TV-out in the first place.

Here is a short list of outstanding issues with *current* drivers:

   * debug and verify that km from CVS works with mach64 and rage128
     cards

   * debug gcc 3.2.x issues. Find out whether it is really necessary
     to have two binary sets of modules.

   * merge code with XFree86 CVS when the latter is closer to release.
     this would be a lot of work because of progress in merging ati.2
     into mach64 driver

   * Rage Theatre 200 support (if ATI does create and release the docs -
     it is taking them a while). This is a *very* big job, as Rage Theatre
     200 includes audio functions, which might require major redesign.

Things that would be nice to have:

   * second Xv port that uses 3d textures (playback only)

   * complete km_api so that km 0.99.x can be started

   * nicer capture interface for AVview

Furthermore, ATI sells component adaptor that allows connection to HDTV
sets. AFAIK, this does not use Rage Theatre, but rather DVI port.
I would imagine this could be made working by carefully studying current
code - and having proper hardware, of course.

Unlike regular TV-out, component adaptor is a technology that is not
likely to go away after a few years.

Lastly, all this needs to be done in spare time, which is always hard to
come by, especially in large chunks.

>
> 	I'm a little confused about the state of Tvout support in
> GATOS.  There are two objects included in GATOS (tvout.o and
> i2cmpp.o) which control the ImpacTV chip for TVout display.  The ImpacTV
> chip is (more-or-less) exactly the same as the ImpacTV2, except that it
> doesn't include Macrovision or support for 16:9 modes.  Did ATI give GATOS
> the specs for the ImpacTV chip, or were these reverse engineered?

No idea. Which code are you referring to ? xatitv ? This has been a good
while ago. I do not believe tvout.c was ever finished.

>
> 	Before completely nixing the idea of TVout, might I suggest
> contacting ATI again.  This time be rather specific that you do not want
> specifications for the Macrovision feature, but only stripped down specs
> on theory of operation, activation, positioning, and scaling.

If you would like to work on TV-out go ahead and try. If nothing else this
would be a very good toy project to learn about multimedia.

                    best

                       Vladimir Dergachev

>
> On Fri, 30 May 2003, Vladimir Dergachev wrote:
> > Anthony,
> >
> >    Yeah, you do have some good points.
> >
> >    You missed one though: the lawsuit does not have to have solid legal
> > basis under it, just be complicated enough so that the court does not
> > dismiss it right away (see, for example, slashdot notes about SCO).
> >
> >    So, personally, I consider that I do not want to pursue TV-out, as it
> > *may* have legal issues, has fairly low resolution (800x600 at best),
> > and understanding of TV-out will in no way help toward making future HDTV
> > drivers, if these will become possible.
> >
> >    You are, of course, free to form your own opinion.
> >
> >                              best
> >
> >                                 Vladimir Dergachev
> >
> > On Thu, 29 May 2003, Anthony J. Ciani wrote:
> >
> > > Hello All,
> > >
> > > 	This message is at the same time informative, argumentative, and
> > > moot.  To review the basic problems touched on here please see
> > > http://sourceforge.net/mailarchive/forum.php?thread_id=1868811&forum_id=5014
> > >
> > > To start, Macrovision contractual obligations are the main reasons ATI
> > > doesn't release the specs for the TVout feature of its older
> > > cards.  Just as an informative note, the reason Macrovision wants this
> > > information restricted is NOT because they fear people will turn off the
> > > Macrovision and copy movies, but because people might turn on the
> > > Macrovision and 'protect' video which has not bought a Macrovision
> > > license.  The Macrovision license 'usually' only pertains to the addition
> > > of Macrovision to the output signal, not the absence of Macrovision on
> > > output which should include it.
> > >
> > > On the above note, I'm a little surprised that you can't get ATI to meet
> > > half-way on the TVout support.  Either they could provide a
> > > static-linked binary which would allow for switching of the display and
> > > control of the TV image, or they could provide TVout specs without
> > > information on enabling Macrovision, which should be perfectly allowable
> > > under their contract with Macrovision.  For the most part, TV out already
> > > works, if you turn on the machine with the TV connected and use VESA
> > > calls, or use atitvout (VESA calls).  The only things lacking are
> > > convenience, and the ability to adjust the image.
> > >
> > > **This sounds weird, but it MAY even be possible to use the windows
> > > binaries to configure the video device.  Unfortunately, I don't know that
> > > much about using MS libraries in *NIX programs.
> > >
> > > As far as the DMCA goes, it doesn't even apply to computer display
> > > devices, unless said device contains hardware specifically designed for
> > > playing protected material (DXR3/Hollywood...).  In that case, it only
> > > applies to the DVD playback function of the hardware.  Beyond that, it
> > > only applies when playing protected material through the device which is
> > > designed to play the protected media in a protected manner.  To the best
> > > of my knowledge, the older ATI video cards do not included DVD
> > > decryption/decoding ability, nor does the GATOS driver provide
> > > this.  Finally, the DMCA actually makes anything which can at any point
> > > remove copy protection, illegal.  This means that GATOS would be illegal
> > > with or without TVout, which is quite ludicrous, and the Supreme Court
> > > already ruled on that back in 1984, when it ruled against the MPAA,
> > > and allowed the marketing of a device called the Video Cassette Recorder.
> > >
> > > However, long standing examples are always best to dissuade fears.  First,
> > > there are many, many TVout projects for various cards already available
> > > for Linux (NVidia, SiS, Matrox, i810), and they have all been ignored by
> > > the MPAA (which could still sue, but would get thrown out on lack of
> > > merit).  There are still VGA to TV devices legally available for purchase,
> > > and none of them contain Macrovision.  You can legally buy rather
> > > inexpensive video editing devices which remove Macrovision.  The Supreme
> > > Court made these legal back in the 1984 Betamax ruling, and subsequent
> > > rulings.  The TVout feature of ATI cards can be used on Windows platforms
> > > to illegally copy protected material to video tape.  Hmm, why hasn't ATI
> > > stopped supporting TVout under Windoze?  Refer to the second paragraph for
> > > the answer.
> > >
> > >
> > > To summarize:  ATI can give specs for TVout and multi-head display
> > > as-long-as it does not include information to enable Macrovision.  If they
> > > don't, it's because of laziness, not legality.  Reverse engineering of
> > > the TVout is legally allowable, as long as you do not enable Macrovision
> > > on unlicensed output.
> > >
> > >
> > >
> > > PS:  Just a note.  The Macrovision and CSS protection schemes have been
> > > fairly well proven useless.  Because of this, fewer and fewer studios are
> > > licensing them for use on their DVDs.  In fact, I haven't seen anything
> > > Macrovision protected for about 2 years now.
> > >
> > > --
> > > ------------------------------------------------------------
> > >               Anthony Ciani (ac...@ui...)
> > >            Computational Condensed Matter Physics
> > >    Department of Physics, University of Illinois, Chicago
> > >               http://ciani.phy.uic.edu/~tony
> > > ------------------------------------------------------------
> > >
> > >
> > >
> > > -------------------------------------------------------
> > > This SF.net email is sponsored by: eBay
> > > Get office equipment for less on eBay!
> > > http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> > > _______________________________________________
> > > Gatos-devel mailing list
> > > Gat...@li...
> > > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> > >
> >
>
> --
> ------------------------------------------------------------
>               Anthony Ciani (ac...@ui...)
>            Computational Condensed Matter Physics
>    Department of Physics, University of Illinois, Chicago
>               http://ciani.phy.uic.edu/~tony
> ------------------------------------------------------------
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: eBay
> Get office equipment for less on eBay!
> http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

How I got TV-OUT working in Linux ( fbsd also )
These directions are only for ATI AIW 7500.

[FreeBSD directions are the same execpt for the kernel ofcourse.]

Since it took me almost three days ( 18hs per) to get this working
ive decided to write this so others dont have to go through the pain
(or at least not at much) that I did. This will let you use Tv-Out
at 800x600@60Hz, although I also have it working at 640x480 but its way
to huge. This is a bit getto I have to warn you ,gentle reader, as after
you have the output going to the tv you cant a) switch video modes and
b) exit X. Doing either a or b will give your monitor the ability to lose
sync, aka not turn off but the green light stay on and flash fast as hell,
while the monitor remains black. I havent figured out any way to resync the
monitor but if anyone else has plaese let me know. Enough talk, be warned. I
never seen much to fsck up linux to the point of hard reboot
but these will, youll prob have to reboot once or twice ;) to get this
going.

1) First read this http://www.reades.com/radeon.html ( < kudos ), most all
of my linux info comes from here and you can apply most all to fbsd. This is
the main doc to read. Im mainly going to restate whats here and add stuff to
the XF86Config section.

Get Linux 2.4.20 it from http://kernel.org/

cd /usr/src
wget http://www.kernel.org/pub/linux/kernel/v2.4/linux-2.4.20.tar.gz
tar -xvzf linux-2.4.20.tar.gz
rm linux
ln -s /usr/src/linux-2.4.20 linux
cd linux
make xconfig <- if X is running
 else make menuconfig
Read the "Kernel Configuation" of http://www.reades.com/radeon.html
and follow all the directions there.
Number 6) finished kernel config; for lilo run /sbin/lilo, never used grub.
MAKE A BOOT DISK, and reboot.

Get XFree86 4.2.0
cd /usr/local
mkdir x420
cd x420
ftp://ftp.xfree86.org/pub/XFree86/4.2.0/source/
get X420src-1.tgz X420src-2.tgz X420src-3.tgz
tar -xvzf /usr/local/X420src-(1 2 and 3).tgz
cd xc
make World
leave; then come back and hour later and say omfg this takes a long time.
make install
make install.man
Follow instructions on http://www.reades.com/radeon.html and BACKUP clean
dir.
reboot to make sure it went ok

Edit the x config file
Follow instructions on http://www.reades.com/radeon.html for "X Config file"
_EXECPT_ dont put in the dri stuff aka Load "dri" and Section "DRI". We'll
put that
in later. Putting that in totally locked my sys and when I moved the mouse,
this must
be a hidden feature, my sys would instantly reboot! wtf, dont know. SO leave
that out
as to not enable hidden features.

DRM Kernel
Again follow instructions on http://www.reades.com/radeon.html "DRM Kernel"
( execpt in 6 its 2.4.20 not 17 )
and then the "ATI.2" below it.
 #5 xmkmf is in /usr/X11R6/bin. export PATH=$PATH:/usr/X11R6/bin
 #5 xmkmf /usr/local/x420/xc

Plug in the cord that connects composite or tv-out into the card.

reboot again

Easy part is over this is where the _FUN_ begins!

If you run startx now youll notcie it fails with 
No Valid Mode Found DFP/LCD
Screen(s) found, but none have a usable configuration. 

If this fails as it did for me you have to get modeline's but try this first
and for the love of god pray it works.

cd /etc/X11
cp XF86Config-4 XF86Config
pico XF86Config

In the Module Section have ( among other things )
Load "dbe"
Load "type1"
Load "freetype"
Load "glx"
Load "record"
Load "dri" <- can put it in now
Load "xtrap"
Load "speedo" 

Add a section under it:
Section "DRI"
   Mode 0666
EndSection

Go down to Monitor section
I have HorizSync at 30-50 which is safe for most all monitors
if your sure of it out it in else put in 30-50, which is very close to
the default anyway.
Now the VertRefresh has to be 60 for tv-out so
  VertRefresh 60

Open a new window and type lspci and find the card. Youll see a number like
00:01.0 or something similar, mine was 01:00.0 but it will most likely be
different.

Go to the section device for your card. Fill in what you dont have:
Section "Device"
   Identifier "ATI All-in-Wonder"
   Driver "ati"
   Option "CrtScreen" "true"
   # fill in number PCI:number of above, replace . with :  on mine its
   # i got 01:00.0 so mine is
   BusID "PCI:01:00:00"
   Option "AGPMode" "4" # if this dosent work try "2"
EndSection

Now go to the screen section you should have
have the default depth be 16 bits per pixel:
  DefaultDepth 16
int the subsection for 16 have
  SubSection "Display"
     Depth 16
     Modes "800x600" # and any others you want, but start with 800x600
     ViewPort 0 0 # if you dont want a virtual screen which is annoyning as
hell
   EndSection

Ok your now ready to test, startx>&x_err.log
If you can see your screen your done.
If not ( most likely ) you have to do some other things.
If you can see some colors and lines your getting there but still have
more things to do. First make sure that your in 800x600, run xvidtune, that
will tell you what mode your in. If your not in it goto it. by either 
ctrl+alt+ [+\-] on num pad or next in xvidtune. In 800x600 look again.
Next try cycling through all the modes and see if any work or come close.
Still dosent work? Try this go to Monitor Section. Under VertRefresh 60
put these lines ( assumes NTSC mode tv [america] ). These worked on 3 tvs.

ModeLine "800x600NTSC" 38.76 800 812 814 880 600 646 649 735
Modeline "640x480NTSC" 28.19 640 656 658 784 480 520 525 600

Now go back to SubSection "Display" and comment out Modes.
Type a new Modes with out modelines
 Modes "800x600NTSC" "640x480NTSC"
Try starting x again.

Last effort
Run xf86cfg, go to the modeline option. click the dropdown list for all the
modelines that are 800x600, 640x480, and 1024x768 that your monitor can
handle and click add for each one. write out the file somewhere like
/tmp/x_blah. cut all the modelines from that file and paste it where we put
the other two modelines ( erasing them ). Now go back to the SubSection
"Display" comment out the last Modes and add a new one. For each of the
Modelines you just pasted you need the corresponding name here, so for
example if you had ModeLine "800x600@60" # # # ... in the modes you would
have Modes "800x600@60" . . . This is what I had to do but in the end it
worked. I hope this helps out, I needed to do something while i copy my 28
gigs of music from backup.
 One more thing FSCK DMCA!!!

Because the hardware is not set back to the original state.


On 30-May-03 at 13:22, Hall, Luca (ha...@bn...) wrote:
> 
>  Ive searched the lists and only find ppl complaining about this but
> I havent found any fix yet. The problem is after I have tv-out working and
> I want to switch to another mode or close X my monitor wont resync - just
> goes black. Anyone have any fix or can tell me why this happens.
> 
> luca
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: eBay
> Get office equipment for less on eBay!
> http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel

 Ive searched the lists and only find ppl complaining about this but
I havent found any fix yet. The problem is after I have tv-out working and I
want to switch to another mode or close X my monitor wont resync - just goes
black. Anyone have any fix or can tell me why this happens.

luca

No, km won't be of much use without Xserver, ati.2 and AVview or xawtv.

                       best

                           Vladimir Dergachev

On Fri, 30 May 2003, Paul Gratz wrote:

> Hi,
>    I'm new to this group but I've done my best to read back through the
> archives and I could not find a good answer to my question. Please
> forgive me if I err :-).
>
>    I own a 7500 AIW and I have used it in a multimedia machine I have
> built. I bought it some time ago and would like to not buy another card
> for this machine.  After much reading here and elsewhere I came to the
> conclusion that I need to use the "vesa" X driver to get the TV out to
> work on the card (since this is a multimedia machine this is essential,
> I don't even have a monitor on it right now). After some testing I have
> that working acceptably well with "freevo" (very cool software).
>
>   Anyways my problem is that I would like to use the tv-tuner part of
> the card as well. To this end I downloaded the KM package from the GATOS
> project and was able to compile and install it on my system. I already
> have v4l on the system so that is not the problem. Infact after
> installing the modules I see that it makes /dev/v4l/video0 and vbi0 (I'm
> using the proc filesystem). Here is the dmesg output on loading the
> modules:
>
> Linux video capture interface: v1.00
> Kmultimedia API module version alpha-3.0 loaded
> Kmultimedia module version alpha-3.0 loaded
> Page size is 4096 sizeof(bm_list_descriptor)=16 sizeof(KM_STRUCT)=952
> km: using irq 11
> Register aperture is 0xe0020000 0x00010000
> kms variables: reg_aperture=0xd08a5000
> km: DMA_GUI_STATUS=0x00000002 entries=2
> sizeof(kmfl_template)=624 sizeof(KM_FIELD)=48
> Device ATI Technologies Inc Radeon 7500 QW 01:00.0 (0x1002:0x5157)
> corresponds to /dev/video0
> kms variables: reg_aperture=0xd08a5000
>
>
> When I try to use them however, I get the following from mplayer:
>
> [snip unrelated junk]
> Playing TV
> Cache fill:  0.00% (0 bytes)    TV detected! ;-)
> Selected driver: v4l
>  name: Video 4 Linux input
>  author: Alex Beregszaszi <al...@na...>
>  comment: under development
> unable to open '/dev/video0': No data available
> Falling back on trying to parse playlist TV...
> ============ Sorry, this file format is not recognized/supported
> =============
> === If this file is an AVI, ASF or MPEG stream, please contact the
> author! ===
>
> I also tried avview but I got some error that was kind of similar IIRC.
>
> Interestingly I get the following in "dmesg" after this attempt:
>
> Purging transfer queue
> WARNING !   Radeon memory controller is misconfigured, disabling capture
> WARNING !   upgrade your Xserver and DRM driver
> km: no data is available until AVview or xawtv is started
>
>
> I even tried downloading the drm modules to see if that would help (it
> did not).  Is there any way I can make this work, or does anyone have
> any suggestions?
>
> Thanks!
> Paul
>
>
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: eBay
> Get office equipment for less on eBay!
> http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

Hello again,

	I have learned to pay little attention to SlashDot.  The SCO vs
IBM case hasn't even gone in front of a judge yet.  BTW, the SCO suit does
have legal merit, and IBM could be in for a hurt if it did use SCO
technologies in unlicensed products (i.e. Linux).  Of course, IBM didn't,
but that will have to be shown in court.  In the case of TVout, there is
no legal basis, thus no merit, and no lawsuits.

	I'm a little confused about the state of Tvout support in
GATOS.  There are two objects included in GATOS (tvout.o and 
i2cmpp.o) which control the ImpacTV chip for TVout display.  The ImpacTV
chip is (more-or-less) exactly the same as the ImpacTV2, except that it
doesn't include Macrovision or support for 16:9 modes.  Did ATI give GATOS
the specs for the ImpacTV chip, or were these reverse engineered?

	Before completely nixing the idea of TVout, might I suggest
contacting ATI again.  This time be rather specific that you do not want
specifications for the Macrovision feature, but only stripped down specs
on theory of operation, activation, positioning, and scaling.

On Fri, 30 May 2003, Vladimir Dergachev wrote:
> Anthony,
> 
>    Yeah, you do have some good points.
> 
>    You missed one though: the lawsuit does not have to have solid legal
> basis under it, just be complicated enough so that the court does not
> dismiss it right away (see, for example, slashdot notes about SCO).
> 
>    So, personally, I consider that I do not want to pursue TV-out, as it
> *may* have legal issues, has fairly low resolution (800x600 at best),
> and understanding of TV-out will in no way help toward making future HDTV
> drivers, if these will become possible.
> 
>    You are, of course, free to form your own opinion.
> 
>                              best
> 
>                                 Vladimir Dergachev
> 
> On Thu, 29 May 2003, Anthony J. Ciani wrote:
> 
> > Hello All,
> >
> > 	This message is at the same time informative, argumentative, and
> > moot.  To review the basic problems touched on here please see
> > http://sourceforge.net/mailarchive/forum.php?thread_id=1868811&forum_id=5014
> >
> > To start, Macrovision contractual obligations are the main reasons ATI
> > doesn't release the specs for the TVout feature of its older
> > cards.  Just as an informative note, the reason Macrovision wants this
> > information restricted is NOT because they fear people will turn off the
> > Macrovision and copy movies, but because people might turn on the
> > Macrovision and 'protect' video which has not bought a Macrovision
> > license.  The Macrovision license 'usually' only pertains to the addition
> > of Macrovision to the output signal, not the absence of Macrovision on
> > output which should include it.
> >
> > On the above note, I'm a little surprised that you can't get ATI to meet
> > half-way on the TVout support.  Either they could provide a
> > static-linked binary which would allow for switching of the display and
> > control of the TV image, or they could provide TVout specs without
> > information on enabling Macrovision, which should be perfectly allowable
> > under their contract with Macrovision.  For the most part, TV out already
> > works, if you turn on the machine with the TV connected and use VESA
> > calls, or use atitvout (VESA calls).  The only things lacking are
> > convenience, and the ability to adjust the image.
> >
> > **This sounds weird, but it MAY even be possible to use the windows
> > binaries to configure the video device.  Unfortunately, I don't know that
> > much about using MS libraries in *NIX programs.
> >
> > As far as the DMCA goes, it doesn't even apply to computer display
> > devices, unless said device contains hardware specifically designed for
> > playing protected material (DXR3/Hollywood...).  In that case, it only
> > applies to the DVD playback function of the hardware.  Beyond that, it
> > only applies when playing protected material through the device which is
> > designed to play the protected media in a protected manner.  To the best
> > of my knowledge, the older ATI video cards do not included DVD
> > decryption/decoding ability, nor does the GATOS driver provide
> > this.  Finally, the DMCA actually makes anything which can at any point
> > remove copy protection, illegal.  This means that GATOS would be illegal
> > with or without TVout, which is quite ludicrous, and the Supreme Court
> > already ruled on that back in 1984, when it ruled against the MPAA,
> > and allowed the marketing of a device called the Video Cassette Recorder.
> >
> > However, long standing examples are always best to dissuade fears.  First,
> > there are many, many TVout projects for various cards already available
> > for Linux (NVidia, SiS, Matrox, i810), and they have all been ignored by
> > the MPAA (which could still sue, but would get thrown out on lack of
> > merit).  There are still VGA to TV devices legally available for purchase,
> > and none of them contain Macrovision.  You can legally buy rather
> > inexpensive video editing devices which remove Macrovision.  The Supreme
> > Court made these legal back in the 1984 Betamax ruling, and subsequent
> > rulings.  The TVout feature of ATI cards can be used on Windows platforms
> > to illegally copy protected material to video tape.  Hmm, why hasn't ATI
> > stopped supporting TVout under Windoze?  Refer to the second paragraph for
> > the answer.
> >
> >
> > To summarize:  ATI can give specs for TVout and multi-head display
> > as-long-as it does not include information to enable Macrovision.  If they
> > don't, it's because of laziness, not legality.  Reverse engineering of
> > the TVout is legally allowable, as long as you do not enable Macrovision
> > on unlicensed output.
> >
> >
> >
> > PS:  Just a note.  The Macrovision and CSS protection schemes have been
> > fairly well proven useless.  Because of this, fewer and fewer studios are
> > licensing them for use on their DVDs.  In fact, I haven't seen anything
> > Macrovision protected for about 2 years now.
> >
> > --
> > ------------------------------------------------------------
> >               Anthony Ciani (ac...@ui...)
> >            Computational Condensed Matter Physics
> >    Department of Physics, University of Illinois, Chicago
> >               http://ciani.phy.uic.edu/~tony
> > ------------------------------------------------------------
> >
> >
> >
> > -------------------------------------------------------
> > This SF.net email is sponsored by: eBay
> > Get office equipment for less on eBay!
> > http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> > _______________________________________________
> > Gatos-devel mailing list
> > Gat...@li...
> > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> >
> 

-- 
------------------------------------------------------------
              Anthony Ciani (ac...@ui...)
           Computational Condensed Matter Physics
   Department of Physics, University of Illinois, Chicago
              http://ciani.phy.uic.edu/~tony
------------------------------------------------------------

On Fri, 30 May 2003, Nikolai Zhubr wrote:

> Hi,
> Friday, 30 May, 2003, 4:24:25, Vladimir Dergachev wrote:
> > On Fri, 30 May 2003, Nikolai Zhubr wrote:
>
> >> Hi Vladimir,
> >> Thursday, 22 May, 2003, 23:02:00, you wrote:
> >> [...]
> >> > I would be very thankful if someone pointed out to me how one can find out
> >> > register aperture address in Windows (i.e. scan PCI bus).
> >> What about pcilib (the one from pciutils)? Wouldn't it make you happy?
>
> > I used it for Linux version of hw_script, but I have no idea how to
> > compile it under Windows (especially so it works under Win 2000) and I
> > don't know whether it will not induce conflicts when scanning PCI bus.
> Well, it compiles with no big problem with mingw32, however still it
> doesn't work properly for me (yet?). That is, it reports proper Manuf.
> ID and device ID and IRQ number but the rest of info is messed up. I
> think ideally it would be better to use some dedicated windows api
> instead of accessing ports directly, though I see no reason why just
> simply reading out pci regs could hurt it.
> I'll try to investigate this further then.
> And yes, for win2000 some more stuff will be necessary in order that
> it worked.

At some point I thought that Windows should store such information in the
registry (which is accessible from Tcl/Tk), however I was not able to find
the proper place.

                       best

                            Vladimir Dergachev

> --
> Best regards,
>  Nikolai Zhubr
>
> >                           best
>
> >                              Vladimir Dergachev
>
> >> --
> >> Best regards,
> >>  Nikolai Zhubr
> >>
> >>
> >> >                       best
> >>
> >> >                          Vladimir Dergachev
> >>
> >>
> >>
> >>
> >> -------------------------------------------------------
> >> This SF.net email is sponsored by: eBay
> >> Get office equipment for less on eBay!
> >> http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> >> _______________________________________________
> >> Gatos-devel mailing list
> >> Gat...@li...
> >> https://lists.sourceforge.net/lists/listinfo/gatos-devel
> >>
>
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: eBay
> Get office equipment for less on eBay!
> http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

Hi, 
   I'm new to this group but I've done my best to read back through the
archives and I could not find a good answer to my question. Please
forgive me if I err :-). 

   I own a 7500 AIW and I have used it in a multimedia machine I have
built. I bought it some time ago and would like to not buy another card
for this machine.  After much reading here and elsewhere I came to the
conclusion that I need to use the "vesa" X driver to get the TV out to
work on the card (since this is a multimedia machine this is essential,
I don't even have a monitor on it right now). After some testing I have
that working acceptably well with "freevo" (very cool software).

  Anyways my problem is that I would like to use the tv-tuner part of
the card as well. To this end I downloaded the KM package from the GATOS
project and was able to compile and install it on my system. I already
have v4l on the system so that is not the problem. Infact after
installing the modules I see that it makes /dev/v4l/video0 and vbi0 (I'm
using the proc filesystem). Here is the dmesg output on loading the
modules:

Linux video capture interface: v1.00
Kmultimedia API module version alpha-3.0 loaded
Kmultimedia module version alpha-3.0 loaded
Page size is 4096 sizeof(bm_list_descriptor)=16 sizeof(KM_STRUCT)=952
km: using irq 11
Register aperture is 0xe0020000 0x00010000
kms variables: reg_aperture=0xd08a5000
km: DMA_GUI_STATUS=0x00000002 entries=2
sizeof(kmfl_template)=624 sizeof(KM_FIELD)=48
Device ATI Technologies Inc Radeon 7500 QW 01:00.0 (0x1002:0x5157)
corresponds to /dev/video0
kms variables: reg_aperture=0xd08a5000


When I try to use them however, I get the following from mplayer:

[snip unrelated junk]
Playing TV
Cache fill:  0.00% (0 bytes)    TV detected! ;-)
Selected driver: v4l
 name: Video 4 Linux input
 author: Alex Beregszaszi <al...@na...>
 comment: under development
unable to open '/dev/video0': No data available
Falling back on trying to parse playlist TV...
============ Sorry, this file format is not recognized/supported
=============
=== If this file is an AVI, ASF or MPEG stream, please contact the
author! ===

I also tried avview but I got some error that was kind of similar IIRC.

Interestingly I get the following in "dmesg" after this attempt:

Purging transfer queue
WARNING !   Radeon memory controller is misconfigured, disabling capture
WARNING !   upgrade your Xserver and DRM driver
km: no data is available until AVview or xawtv is started


I even tried downloading the drm modules to see if that would help (it
did not).  Is there any way I can make this work, or does anyone have
any suggestions?

Thanks!
Paul

Hi,
Friday, 30 May, 2003, 4:24:25, Vladimir Dergachev wrote:
> On Fri, 30 May 2003, Nikolai Zhubr wrote:

>> Hi Vladimir,
>> Thursday, 22 May, 2003, 23:02:00, you wrote:
>> [...]
>> > I would be very thankful if someone pointed out to me how one can find out
>> > register aperture address in Windows (i.e. scan PCI bus).
>> What about pcilib (the one from pciutils)? Wouldn't it make you happy?

> I used it for Linux version of hw_script, but I have no idea how to
> compile it under Windows (especially so it works under Win 2000) and I
> don't know whether it will not induce conflicts when scanning PCI bus.
Well, it compiles with no big problem with mingw32, however still it
doesn't work properly for me (yet?). That is, it reports proper Manuf.
ID and device ID and IRQ number but the rest of info is messed up. I
think ideally it would be better to use some dedicated windows api
instead of accessing ports directly, though I see no reason why just
simply reading out pci regs could hurt it.
I'll try to investigate this further then.
And yes, for win2000 some more stuff will be necessary in order that
it worked.
-- 
Best regards,
 Nikolai Zhubr

>                           best

>                              Vladimir Dergachev

>> --
>> Best regards,
>>  Nikolai Zhubr
>>
>>
>> >                       best
>>
>> >                          Vladimir Dergachev
>>
>>
>>
>>
>> -------------------------------------------------------
>> This SF.net email is sponsored by: eBay
>> Get office equipment for less on eBay!
>> http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
>> _______________________________________________
>> Gatos-devel mailing list
>> Gat...@li...
>> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>>

On Thu, 29 May 2003, Mark Cuss wrote:

> Vladimir,
>
> Thanks for finding the problem for me - I was barking up the wrong tree
> entirely...
>
> Do you know if this bug is fixed in the current XFree CVS or if a patch
> is available elsewhere?

No. It looks like more investigation is needed. But you could try (after
reading the thread) to tweak XFree86 source yourself..

See thread "Athlon related mystery" (but, no, it is not entirely Athlon
related) on de...@xf... mailing list.

                        best

                          Vladimir Dergachev

>
> Thanks again for the help,
>
> Mark
>
> -----Original Message-----
> From: Vladimir Dergachev <vo...@mi...>
> To: Mark Cuss <mc...@cd...>
> Cc: gat...@so...
> Date: Thu, 29 May 2003 20:28:55 -0400 (EDT)
> Subject: Re: [GATOS]YUV to RGB Conversion on Radeon Mobility 9000
>
> >
> >
> > On Thu, 29 May 2003, Mark Cuss wrote:
> >
> > > Vladimir,
> > >
> > > Thanks for the information.  I ran the tests on both machines as you
> > > suggested...  Here are the results:
> > >
> > > Laptop:  putimage500 ~= 75 per second; shmput500 ~= 96 per second
> > > Desktop: putimage500 ~= 96 per second; shmput500 ~= 143 per second
> > >
> > > So, the desktop is definitely quicker - is this the margin you
> > expected, or
> > > should it be greater than this?
> >
> > No, I think it shows the problem quite clearly. You see, on both
> > computers
> > we are limited by the speed of PCI bus - as such accesses do not use
> > AGP
> > transfers. And the PCI bus speed should be identical.
> > The actual speed (that depends on how the data is written out) degraded
> > by
> > whole 33% - and all the time that cpu waits for the data to be
> > transferred it does nothing.
> >
> >                        best
> >
> >                          Vladimir Dergachev
> >
> > >
> > > Thanks
> > >
> > > Mark
> > >
> > > ----- Original Message -----
> > > From: "Vladimir Dergachev" <vo...@mi...>
> > > To: "Mark Cuss" <mc...@cd...>
> > > Cc: <gat...@so...>
> > > Sent: Thursday, May 29, 2003 2:40 PM
> > > Subject: Re: [GATOS]YUV to RGB Conversion on Radeon Mobility 9000
> > >
> > >
> > > >
> > > > YUV->RGB overlay works on your chip, however there was recent
> > discussion
> > > > on XFree86 devel mailing list that has unearthed a bug in how
> > XFree86
> > > > 4.3.0 sets up mtrr and accesses card apertures. The bug is present
> > in
> > > > Suse and Redhat compiles of XFree86 4.3.0 and was meant to be a
> > bugfix
> > > > for some AMD systems.
> > > >
> > > > The bug shows up as dramatic decrease in bandwidth when
> > transferring data
> > > > to graphics memory - in particular when transferring YUV data. This
> > is the
> > > > reasons for the slowdown you are seeing.
> > > >
> > > > To verify that this bug does indeed take place try using x11perf to
> > > > measure perfomance of PutImage (or ShmPutImage) functions - your
> > notebook
> > > > should show a much smaller number. As PutImage or ShmPutImage
> > simply
> > > > transer graphics data to graphics memory (and do not use YUV->RGB
> > > > acceleration) they are a good indicator for the available bandwidth
> > to the
> > > > graphics aperture.
> > > >
> > > >                        best
> > > >
> > > >                           Vladimir Dergachev
> > > >
> > > > On Thu, 29 May 2003, Mark Cuss wrote:
> > > >
> > > > > Hello All
> > > > >
> > > > > I have a question regarding support for YUV to RGB conversion in
> > > hardware on the Radeon Mobility 9000 Chip.  I've been doing testing
> > with two
> > > computers:
> > > > >
> > > > > 1)  Custom Desktop System (P4 @ 2.8 GHz):
> > > > >     - Red Hat 8; stock 4.2.0 X Server; GATOS 4.2.0 experimental
> > 16
> > > drivers
> > > > >     - ATI AIW Radeon 8500 DV Video card (R200 BB)
> > > > >
> > > > > 2)  Dell Inpsiron 8500 Laptop (P4 Mobile @ 2.4 GHz)
> > > > >     - Red Hat 8; XFree 4.3.0 binaries from FTP site, GATOS 4.3.0
> > > experimental 9 drivers
> > > > >     - ATI Radeon Mobility 9000 Video chip (R250 Lf)
> > > > >
> > > > > I've been decoding MPEG-2 on both of these machines.  On machine
> > 1, I
> > > can easily decode an MPEG-2 stream at 30 fps.  During this operation,
> > the X
> > > server isn't loading up the CPU at all (using maybe 5%).
> > > > >
> > > > > When I run the same software on the laptop, the X server takes up
> > over
> > > half of the CPU.  I've had experience with a similar problem with the
> > SMI
> > > chip, and I'm wondering if the problem is the same here - the problem
> > was
> > > that the chip couldn't convert the decoded YUV image into RGB in
> > hardware,
> > > so it was falling back to a software conversion.
> > > > >
> > > > > So, my question is - do you folks know if the R250 Lf can convert
> > YUV
> > > data (specifically I420) to RGB in hardware?  If so, does the driver
> > support
> > > this functionality?  I've attached the xvinfo data from both units
> > for
> > > anyone who is interested...
> > > > >
> > > > > Any tips would be greatly appreciated.
> > > > >
> > > > > Thanks!
> > > > >
> > > > > Mark
> > > > >
> > > > > Mark Cuss, B. Sc.
> > > > > Real Time Systems Analyst
> > > > > CDL Systems Ltd
> > > > > Suite 230
> > > > > 3553 - 31 Street NW
> > > > > Calgary, AB, Canada
> > > > >
> > > > > Phone: 403 289 1733 ext 226
> > > > > Fax: 403 282 1238
> > > > > www.cdlsystems.com
> > > >
> > > >
> > > > -------------------------------------------------------
> > > > This SF.net email is sponsored by: eBay
> > > > Get office equipment for less on eBay!
> > > > http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> > > > _______________________________________________
> > > > Gatos-devel mailing list
> > > > Gat...@li...
> > > > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> > > >
> > >
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: eBay
> Get office equipment for less on eBay!
> http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

Open "setup window" and select some video format - like AVI or MPEG.

The reason it complains is that it needs to know how to mix audio and
video in the output file.

                        best

                           Vladimir Dergachev

On Wed, 28 May 2003, Ed Ruano wrote:

> Hello,
>   I have been trying to use avview for video capture and I have been
> seeing the following message come up in a dialog:/  If you want to
> record both audio and video the file format must not be "none"./  If I
> restart avview I can usually start a capture, but changing the file name
> seems to break things.  Also, when capture does work I cannot seem to
> get any sound when I play the file back using xine.  I have Alsa 0.9
> installed and working.  Previous attempts to capture using xawtv and
> ffmpeg did have sound.  Any tips would be greatly appreciated.
>
>   Video capture looks great though!
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: eBay
> Get office equipment for less on eBay!
> http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

Anthony,

   Yeah, you do have some good points.

   You missed one though: the lawsuit does not have to have solid legal
basis under it, just be complicated enough so that the court does not
dismiss it right away (see, for example, slashdot notes about SCO).

   So, personally, I consider that I do not want to pursue TV-out, as it
*may* have legal issues, has fairly low resolution (800x600 at best),
and understanding of TV-out will in no way help toward making future HDTV
drivers, if these will become possible.

   You are, of course, free to form your own opinion.

                             best

                                Vladimir Dergachev

On Thu, 29 May 2003, Anthony J. Ciani wrote:

> Hello All,
>
> 	This message is at the same time informative, argumentative, and
> moot.  To review the basic problems touched on here please see
> http://sourceforge.net/mailarchive/forum.php?thread_id=1868811&forum_id=5014
>
> To start, Macrovision contractual obligations are the main reasons ATI
> doesn't release the specs for the TVout feature of its older
> cards.  Just as an informative note, the reason Macrovision wants this
> information restricted is NOT because they fear people will turn off the
> Macrovision and copy movies, but because people might turn on the
> Macrovision and 'protect' video which has not bought a Macrovision
> license.  The Macrovision license 'usually' only pertains to the addition
> of Macrovision to the output signal, not the absence of Macrovision on
> output which should include it.
>
> On the above note, I'm a little surprised that you can't get ATI to meet
> half-way on the TVout support.  Either they could provide a
> static-linked binary which would allow for switching of the display and
> control of the TV image, or they could provide TVout specs without
> information on enabling Macrovision, which should be perfectly allowable
> under their contract with Macrovision.  For the most part, TV out already
> works, if you turn on the machine with the TV connected and use VESA
> calls, or use atitvout (VESA calls).  The only things lacking are
> convenience, and the ability to adjust the image.
>
> **This sounds weird, but it MAY even be possible to use the windows
> binaries to configure the video device.  Unfortunately, I don't know that
> much about using MS libraries in *NIX programs.
>
> As far as the DMCA goes, it doesn't even apply to computer display
> devices, unless said device contains hardware specifically designed for
> playing protected material (DXR3/Hollywood...).  In that case, it only
> applies to the DVD playback function of the hardware.  Beyond that, it
> only applies when playing protected material through the device which is
> designed to play the protected media in a protected manner.  To the best
> of my knowledge, the older ATI video cards do not included DVD
> decryption/decoding ability, nor does the GATOS driver provide
> this.  Finally, the DMCA actually makes anything which can at any point
> remove copy protection, illegal.  This means that GATOS would be illegal
> with or without TVout, which is quite ludicrous, and the Supreme Court
> already ruled on that back in 1984, when it ruled against the MPAA,
> and allowed the marketing of a device called the Video Cassette Recorder.
>
> However, long standing examples are always best to dissuade fears.  First,
> there are many, many TVout projects for various cards already available
> for Linux (NVidia, SiS, Matrox, i810), and they have all been ignored by
> the MPAA (which could still sue, but would get thrown out on lack of
> merit).  There are still VGA to TV devices legally available for purchase,
> and none of them contain Macrovision.  You can legally buy rather
> inexpensive video editing devices which remove Macrovision.  The Supreme
> Court made these legal back in the 1984 Betamax ruling, and subsequent
> rulings.  The TVout feature of ATI cards can be used on Windows platforms
> to illegally copy protected material to video tape.  Hmm, why hasn't ATI
> stopped supporting TVout under Windoze?  Refer to the second paragraph for
> the answer.
>
>
> To summarize:  ATI can give specs for TVout and multi-head display
> as-long-as it does not include information to enable Macrovision.  If they
> don't, it's because of laziness, not legality.  Reverse engineering of
> the TVout is legally allowable, as long as you do not enable Macrovision
> on unlicensed output.
>
>
>
> PS:  Just a note.  The Macrovision and CSS protection schemes have been
> fairly well proven useless.  Because of this, fewer and fewer studios are
> licensing them for use on their DVDs.  In fact, I haven't seen anything
> Macrovision protected for about 2 years now.
>
> --
> ------------------------------------------------------------
>               Anthony Ciani (ac...@ui...)
>            Computational Condensed Matter Physics
>    Department of Physics, University of Illinois, Chicago
>               http://ciani.phy.uic.edu/~tony
> ------------------------------------------------------------
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: eBay
> Get office equipment for less on eBay!
> http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

Vladimir,

Thanks for finding the problem for me - I was barking up the wrong tree 
entirely...  

Do you know if this bug is fixed in the current XFree CVS or if a patch 
is available elsewhere?

Thanks again for the help,

Mark

-----Original Message-----
From: Vladimir Dergachev <vo...@mi...>
To: Mark Cuss <mc...@cd...>
Cc: gat...@so...
Date: Thu, 29 May 2003 20:28:55 -0400 (EDT)
Subject: Re: [GATOS]YUV to RGB Conversion on Radeon Mobility 9000

> 
> 
> On Thu, 29 May 2003, Mark Cuss wrote:
> 
> > Vladimir,
> >
> > Thanks for the information.  I ran the tests on both machines as you
> > suggested...  Here are the results:
> >
> > Laptop:  putimage500 ~= 75 per second; shmput500 ~= 96 per second
> > Desktop: putimage500 ~= 96 per second; shmput500 ~= 143 per second
> >
> > So, the desktop is definitely quicker - is this the margin you
> expected, or
> > should it be greater than this?
> 
> No, I think it shows the problem quite clearly. You see, on both
> computers
> we are limited by the speed of PCI bus - as such accesses do not use
> AGP
> transfers. And the PCI bus speed should be identical.
> The actual speed (that depends on how the data is written out) degraded
> by
> whole 33% - and all the time that cpu waits for the data to be
> transferred it does nothing.
> 
>                        best
> 
>                          Vladimir Dergachev
> 
> >
> > Thanks
> >
> > Mark
> >
> > ----- Original Message -----
> > From: "Vladimir Dergachev" <vo...@mi...>
> > To: "Mark Cuss" <mc...@cd...>
> > Cc: <gat...@so...>
> > Sent: Thursday, May 29, 2003 2:40 PM
> > Subject: Re: [GATOS]YUV to RGB Conversion on Radeon Mobility 9000
> >
> >
> > >
> > > YUV->RGB overlay works on your chip, however there was recent
> discussion
> > > on XFree86 devel mailing list that has unearthed a bug in how
> XFree86
> > > 4.3.0 sets up mtrr and accesses card apertures. The bug is present
> in
> > > Suse and Redhat compiles of XFree86 4.3.0 and was meant to be a
> bugfix
> > > for some AMD systems.
> > >
> > > The bug shows up as dramatic decrease in bandwidth when
> transferring data
> > > to graphics memory - in particular when transferring YUV data. This
> is the
> > > reasons for the slowdown you are seeing.
> > >
> > > To verify that this bug does indeed take place try using x11perf to
> > > measure perfomance of PutImage (or ShmPutImage) functions - your
> notebook
> > > should show a much smaller number. As PutImage or ShmPutImage
> simply
> > > transer graphics data to graphics memory (and do not use YUV->RGB
> > > acceleration) they are a good indicator for the available bandwidth
> to the
> > > graphics aperture.
> > >
> > >                        best
> > >
> > >                           Vladimir Dergachev
> > >
> > > On Thu, 29 May 2003, Mark Cuss wrote:
> > >
> > > > Hello All
> > > >
> > > > I have a question regarding support for YUV to RGB conversion in
> > hardware on the Radeon Mobility 9000 Chip.  I've been doing testing
> with two
> > computers:
> > > >
> > > > 1)  Custom Desktop System (P4 @ 2.8 GHz):
> > > >     - Red Hat 8; stock 4.2.0 X Server; GATOS 4.2.0 experimental
> 16
> > drivers
> > > >     - ATI AIW Radeon 8500 DV Video card (R200 BB)
> > > >
> > > > 2)  Dell Inpsiron 8500 Laptop (P4 Mobile @ 2.4 GHz)
> > > >     - Red Hat 8; XFree 4.3.0 binaries from FTP site, GATOS 4.3.0
> > experimental 9 drivers
> > > >     - ATI Radeon Mobility 9000 Video chip (R250 Lf)
> > > >
> > > > I've been decoding MPEG-2 on both of these machines.  On machine
> 1, I
> > can easily decode an MPEG-2 stream at 30 fps.  During this operation,
> the X
> > server isn't loading up the CPU at all (using maybe 5%).
> > > >
> > > > When I run the same software on the laptop, the X server takes up
> over
> > half of the CPU.  I've had experience with a similar problem with the
> SMI
> > chip, and I'm wondering if the problem is the same here - the problem
> was
> > that the chip couldn't convert the decoded YUV image into RGB in
> hardware,
> > so it was falling back to a software conversion.
> > > >
> > > > So, my question is - do you folks know if the R250 Lf can convert
> YUV
> > data (specifically I420) to RGB in hardware?  If so, does the driver
> support
> > this functionality?  I've attached the xvinfo data from both units
> for
> > anyone who is interested...
> > > >
> > > > Any tips would be greatly appreciated.
> > > >
> > > > Thanks!
> > > >
> > > > Mark
> > > >
> > > > Mark Cuss, B. Sc.
> > > > Real Time Systems Analyst
> > > > CDL Systems Ltd
> > > > Suite 230
> > > > 3553 - 31 Street NW
> > > > Calgary, AB, Canada
> > > >
> > > > Phone: 403 289 1733 ext 226
> > > > Fax: 403 282 1238
> > > > www.cdlsystems.com
> > >
> > >
> > > -------------------------------------------------------
> > > This SF.net email is sponsored by: eBay
> > > Get office equipment for less on eBay!
> > > http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> > > _______________________________________________
> > > Gatos-devel mailing list
> > > Gat...@li...
> > > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> > >
> >

On Thu, May 29, 2003 at 05:50:21PM -0800, Pranay Kumar wrote:
> Hi,
hi

I am not sure about what I will say next, because the reason Vladimir says =
is
new to me, but I have been experiencing similar problems since about 4.2 wi=
th
a r128 chip, even with my own builds, and noone yet seems to have expained =
it
until now. It was so bad that I even stopped using X on that machine.

> Does this mean that if I am calling PutImage/XvPutImage/XvShmPutImage
> family of functions I do not use AGP transfers??
The problem I was experiencing was causing same CPU load regardless of whet=
her
busmastering was used.

> My understanding of bus transfers is limited, but if I have a AGP card,
> don't I have a dedicated bus (the AGP bus) for video transfers.
Not all drivers actually support this, only r128 and nvidia last time I che=
cked.

> -Pranay
>=20
> --------Vladimir Wrote------------
> No, I think it shows the problem quite clearly. You see, on both
> computers we are limited by the speed of PCI bus - as such accesses do
> not use AGP transfers. And the PCI bus speed should be identical. The
> actual speed (that depends on how the data is written out) degraded by
> whole 33% - and all the time that cpu waits for the data to be
> transferred it does nothing.
This 33% seems exactly what I was experiencing. It was so horrible that whe=
n I
turned busmastering off, it actually doubled, and while watching DVDs X was
eating > 50% (half in user and half in system), which definitely sucked.

I remember I tried to find the bug in the code but couldn't, and other
developers kept ignoring me :-(.

>                        best
>=20
>                          Vladimir Dergachev
Bye,

Peter Surda (Shurdeek) <shu...@pa...>, ICQ 10236103, +436505=
122023

--=20
              To boldly go where I surely don't belong.

Hello,
  I have been trying to use avview for video capture and I have been 
seeing the following message come up in a dialog:/  If you want to 
record both audio and video the file format must not be "none"./  If I 
restart avview I can usually start a capture, but changing the file name 
seems to break things.  Also, when capture does work I cannot seem to 
get any sound when I play the file back using xine.  I have Alsa 0.9 
installed and working.  Previous attempts to capture using xawtv and 
ffmpeg did have sound.  Any tips would be greatly appreciated.

  Video capture looks great though!

Hello All,

	This message is at the same time informative, argumentative, and
moot.  To review the basic problems touched on here please see
http://sourceforge.net/mailarchive/forum.php?thread_id=1868811&forum_id=5014

To start, Macrovision contractual obligations are the main reasons ATI
doesn't release the specs for the TVout feature of its older
cards.  Just as an informative note, the reason Macrovision wants this
information restricted is NOT because they fear people will turn off the
Macrovision and copy movies, but because people might turn on the
Macrovision and 'protect' video which has not bought a Macrovision
license.  The Macrovision license 'usually' only pertains to the addition
of Macrovision to the output signal, not the absence of Macrovision on
output which should include it.

On the above note, I'm a little surprised that you can't get ATI to meet
half-way on the TVout support.  Either they could provide a
static-linked binary which would allow for switching of the display and
control of the TV image, or they could provide TVout specs without
information on enabling Macrovision, which should be perfectly allowable
under their contract with Macrovision.  For the most part, TV out already
works, if you turn on the machine with the TV connected and use VESA
calls, or use atitvout (VESA calls).  The only things lacking are
convenience, and the ability to adjust the image.

**This sounds weird, but it MAY even be possible to use the windows
binaries to configure the video device.  Unfortunately, I don't know that
much about using MS libraries in *NIX programs.

As far as the DMCA goes, it doesn't even apply to computer display
devices, unless said device contains hardware specifically designed for
playing protected material (DXR3/Hollywood...).  In that case, it only
applies to the DVD playback function of the hardware.  Beyond that, it
only applies when playing protected material through the device which is
designed to play the protected media in a protected manner.  To the best
of my knowledge, the older ATI video cards do not included DVD 
decryption/decoding ability, nor does the GATOS driver provide
this.  Finally, the DMCA actually makes anything which can at any point
remove copy protection, illegal.  This means that GATOS would be illegal
with or without TVout, which is quite ludicrous, and the Supreme Court
already ruled on that back in 1984, when it ruled against the MPAA,
and allowed the marketing of a device called the Video Cassette Recorder.

However, long standing examples are always best to dissuade fears.  First,
there are many, many TVout projects for various cards already available
for Linux (NVidia, SiS, Matrox, i810), and they have all been ignored by
the MPAA (which could still sue, but would get thrown out on lack of
merit).  There are still VGA to TV devices legally available for purchase,
and none of them contain Macrovision.  You can legally buy rather
inexpensive video editing devices which remove Macrovision.  The Supreme
Court made these legal back in the 1984 Betamax ruling, and subsequent
rulings.  The TVout feature of ATI cards can be used on Windows platforms
to illegally copy protected material to video tape.  Hmm, why hasn't ATI
stopped supporting TVout under Windoze?  Refer to the second paragraph for
the answer.


To summarize:  ATI can give specs for TVout and multi-head display
as-long-as it does not include information to enable Macrovision.  If they
don't, it's because of laziness, not legality.  Reverse engineering of
the TVout is legally allowable, as long as you do not enable Macrovision
on unlicensed output.



PS:  Just a note.  The Macrovision and CSS protection schemes have been
fairly well proven useless.  Because of this, fewer and fewer studios are
licensing them for use on their DVDs.  In fact, I haven't seen anything
Macrovision protected for about 2 years now.

-- 
------------------------------------------------------------
              Anthony Ciani (ac...@ui...)
           Computational Condensed Matter Physics
   Department of Physics, University of Illinois, Chicago
              http://ciani.phy.uic.edu/~tony
------------------------------------------------------------

On Thu, 29 May 2003, Pranay Kumar wrote:

> Hi,
>
> Does this mean that if I am calling PutImage/XvPutImage/XvShmPutImage
> family of functions I do not use AGP transfers?? My understanding of bus
> transfers is limited, but if I have a AGP card, don't I have a dedicated
> bus (the AGP bus) for video transfers.
>

No. AGP transfers need kernel side support (as they access hardware
memory) and thus are only used by DRI.

The reason is that AGP bus was created to allow part of main memory to be
used instead of video memory to store texture for example.

Thus AGP only accelerates transfers between hardware RAM and video card,
not between CPU and video card (i.e. register writing). Everything else
goes over standard PCI bus - this includes writes, reads, IRQ and DMA.

Now the reason for that slowdown, is that video memory is mapped into the
address space of the CPU, and the CPU has incorrect flags - it thinks it
is as volatile as register aperture and all writes are performed
syncronously, instead of being batched for PCI transfers.

                       best

                             Vladimir Dergachev

> -Pranay
>
> --------Vladimir Wrote------------
> No, I think it shows the problem quite clearly. You see, on both
> computers we are limited by the speed of PCI bus - as such accesses do
> not use AGP transfers. And the PCI bus speed should be identical. The
> actual speed (that depends on how the data is written out) degraded by
> whole 33% - and all the time that cpu waits for the data to be
> transferred it does nothing.
>
>                        best
>
>                          Vladimir Dergachev
>
>
>

Hi,

Does this mean that if I am calling PutImage/XvPutImage/XvShmPutImage
family of functions I do not use AGP transfers?? My understanding of bus
transfers is limited, but if I have a AGP card, don't I have a dedicated
bus (the AGP bus) for video transfers.

-Pranay

--------Vladimir Wrote------------
No, I think it shows the problem quite clearly. You see, on both
computers we are limited by the speed of PCI bus - as such accesses do
not use AGP transfers. And the PCI bus speed should be identical. The
actual speed (that depends on how the data is written out) degraded by
whole 33% - and all the time that cpu waits for the data to be
transferred it does nothing.

                       best

                         Vladimir Dergachev

On Fri, 30 May 2003, Nikolai Zhubr wrote:

> Hi Vladimir,
> Thursday, 22 May, 2003, 23:02:00, you wrote:
> [...]
> > I would be very thankful if someone pointed out to me how one can find out
> > register aperture address in Windows (i.e. scan PCI bus).
> What about pcilib (the one from pciutils)? Wouldn't it make you happy?

I used it for Linux version of hw_script, but I have no idea how to
compile it under Windows (especially so it works under Win 2000) and I
don't know whether it will not induce conflicts when scanning PCI bus.

                          best

                             Vladimir Dergachev

> --
> Best regards,
>  Nikolai Zhubr
>
>
> >                       best
>
> >                          Vladimir Dergachev
>
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: eBay
> Get office equipment for less on eBay!
> http://adfarm.mediaplex.com/ad/ck/711-11697-6916-5
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>