gatos-devel Mailing List for GATOS

Hi,
Wednesday, 30 April, 2003, 21:01:59, Bobby Weinmann wrote:
> Has anyone gotten capture to work for Mach64 PCI card.  I have an AIW Rage
> Pro card and have had only system lockups.  I already mentioned this
> problem in previous posts, but I was curious if anyone has gotten one of
> these cards working.
Yes, I have capture working with my Rage Pro (AGP version) rock stable.
However, some other people were not able to get it working so far. The
reason is still unknown unfortunately.
Perhaps try my experimental version from http://zhubr.tamb.ru/km-cvs-hackB.tgz
(It might appear more reliable)
-- 
Best regards,
 Nikolai Zhubr


> Bobby Weinmann




> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel

Lourens

> Well, you could use that output to record copyrighted works (a DVD 
> perhaps) to video right? The TV-Out capable driver would enable you 
> to do that if it didn't turn on Macrovision. And if it didn't, it 
> would be a device enabling you to circumvent digital copyright 
> protection, and thus be prohibited under the DMCA. So what if we 
> had an open source driver that did turn on Macrovision? Well, it's 
> open source right? So anyone can modify it (probably trivially, ie. 
> commenting out a line of code) to not turn on Macrovision. Then 
> doesn't that still make it a device that enables someone (by 
> slightly modifying and then using) to circumvent digital copyright 
> protection?

I'm based in the UK and I'm not sure to what extent the DMCA applies. Do 
you think that it would be OK for USA citizens to use a driver just 
because it was written in an obscure country?

In the case you describe, isn't it more likely that someone is breaking 
the DMCA by using software to decode the DVD contents? That is digital, 
whereas the TV out is analogue. The TV out is being used but it's use is 
incidental to the offence. It's like saying the TV out is responsible if 
you bludgeon someone to death with the card (presumably after buying it 
and finding it's not supported!).

If we wrote a driver in good faith that turns on a generic out-of-range 
sync signal during the flyback period (to fool the AGC on a video 
recorder), would we be at risk of prosecution? How would the author of 
that driver be any more to blame than ATI, who made the hardware? Sure 
someone could modify it to remove the signal, but would that implicate 
the author of the driver?

I bought a TV-out card so I could output the display to a TV screen, and 
that is all. Most of the time I won't be watching movies on the screen, 
and the only use I can think of for recording to VHS is to send a photo 
album to Granny. I could probably do this with existing drivers (vesa), 
or with a simple dongle and some crafty modelines that turn the SVGA 
output into RGB+composite sync (http://www.sput.nl/hardware/tv-x.html) - 
no tv-out hardware needed.

I wonder if any corporations are interested in prosecuting someone who 
makes TV-out work. If you have a copyrighted, decrypted work on your 
hard disk, you can illegally p2p share it, you could illegally burn it 
onto CDs or DVDs. Those sorts of activities could be a big threat to the 
copyright holder. You could spool it out to VHS in real-time but I'm 
guessing that such activities would not be a large threat.

Macrovision is probably (I'm guessing about the business model) used 
under license. So it's not likely that an open source driver would be 
able to implement it - unless you had permission from them to include 
the signal that prevents video recording.

> The DMCA and similar laws are unclear about this, and there isn't 
> much case law yet. Would you risk going to jail over TV-out? Or pay 
> for a lawyer to defend yourself even if eventually you were found 
> innocent?

I can sympathise with USA residents who are concerned over it, but at 
the same time I find it hard to believe that it could happen in the UK. 
The legal system here might not be perfect but they usually seem to be 
reasonable and fair. To summarise:

* there is already support for TV-out using the vesa drivers
* there is already hardware for TV-out
* TV-out is analogue, not digital
* TV-out is not primarily designed to circumvent copy protection
* TV-out has substantial legitimate uses
* TV-out is not marketed for use in circumvention
* there are legitimate uses for viewing the TV-out signal on a TV
* there are legitimate uses for recording the TV-out signal
* who is going to prosecute someone who writes a driver?

Incidentally, how does this work on other OSs? Is the Macrovision signal 
turned on and off by the DVD decoder software? Is the signal always 
present? When I've got a spare moment I'm going to try recording from 
the TV-out in vesa mode and see if it works...

-Cam

Lourens Veen wrote:

>Well, you could use that output to record copyrighted works (a DVD 
>perhaps) to video right? The TV-Out capable driver would enable you 
>to do that if it didn't turn on Macrovision. And if it didn't, it 
>would be a device enabling you to circumvent digital copyright 
>protection, and thus be prohibited under the DMCA. So what if we 
>had an open source driver that did turn on Macrovision? Well, it's 
>open source right? So anyone can modify it (probably trivially, ie. 
>commenting out a line of code) to not turn on Macrovision. Then 
>doesn't that still make it a device that enables someone (by 
>slightly modifying and then using) to circumvent digital copyright 
>protection?
>
>The DMCA and similar laws are unclear about this, and there isn't 
>much case law yet. Would you risk going to jail over TV-out? Or pay 
>for a lawyer to defend yourself even if eventually you were found 
>innocent?
>
>It's likely to be illegal, and very likely to cost a lot of money 
>should legal action come into play, and the current developers 
>would rather not take the risk. End of story, unless someone in a 
>country not yet governed by big media companies wants to write the 
>code.
>
>Lourens
>  
>
In response to my inability to get TV-out working under Linux, I have 
decided to 're-imberse' my consumer satisfaction levels by copying all 
DVDs I can get my hands on. And I can get my hands on a lot. Of couse 
sometimes my friends come around and steal the copies from me 
temporarily, but if you can't trust your friends...

For me, being able to watch a DVD over and over again long after I have 
returned the original makes up for not being able to watch it once on my 
TV. For this re-imbersement action, I use a number of open source tools 
which are developed in the US and hosted in the US. They allow 
point-and-click copying of any DVD to basically any format you can name. 
I prefer DivX ;) 5 / Ogg Vorbis.

In light of the fact that this is all possible and happening, I can't 
see the the relevence of the argument against TV-out support for my 
Radeon on the grounds that I may tend to copy copyrighted material. 
What's that saying? Closing the gate after the horse has bolted? Ah 
well. Good old mplayer / DirectFB is it I'm afraid.

Has anyone gotten capture to work for Mach64 PCI card.  I have an AIW Rage
Pro card and have had only system lockups.  I already mentioned this
problem in previous posts, but I was curious if anyone has gotten one of
these cards working.

Bobby Weinmann

On Wed 30 April 2003 17:35, micron wrote:
> I wanted to compile ati.2 so I downladed the source from cvs.
> I type: xmkmf /usr/X11R6 but it returns:
>
> imake -I/usr/X11R6/config/cf -DTOPDIR=3D/usr/X11R6 -DCURDIR=3D.
> Imakefile.c:35: Imake.tmpl: No such file or directory
> imake: Exit code 1.
>   Stop.
>
> What can I do?

Are you sure you downloaded and compiled the XFree86 sources? See=20
instructions at http://gatos.sourceforge.net/ati.2.php

Lourens
--=20
GPG public key: http://home.student.utwente.nl/l.e.veen/lourens.key

On Wed 30 April 2003 14:53, Cam wrote:
> Hi,
>
> > I think the best chance for this is for someone in a country
> > without DMCA or DMCA-like laws
>
> I bought a card with a TV-out which is actually composite video
> out (ie. not RF modulated). I'm personally not interested in
> recording the output to video but I would have thought that it
> was a legitimate use in almost all cases. There is a lot of
> confusion about legal issues here.

Well, you could use that output to record copyrighted works (a DVD=20
perhaps) to video right? The TV-Out capable driver would enable you=20
to do that if it didn't turn on Macrovision. And if it didn't, it=20
would be a device enabling you to circumvent digital copyright=20
protection, and thus be prohibited under the DMCA. So what if we=20
had an open source driver that did turn on Macrovision? Well, it's=20
open source right? So anyone can modify it (probably trivially, ie.=20
commenting out a line of code) to not turn on Macrovision. Then=20
doesn't that still make it a device that enables someone (by=20
slightly modifying and then using) to circumvent digital copyright=20
protection?

The DMCA and similar laws are unclear about this, and there isn't=20
much case law yet. Would you risk going to jail over TV-out? Or pay=20
for a lawyer to defend yourself even if eventually you were found=20
innocent?

It's likely to be illegal, and very likely to cost a lot of money=20
should legal action come into play, and the current developers=20
would rather not take the risk. End of story, unless someone in a=20
country not yet governed by big media companies wants to write the=20
code.

Lourens
--=20
GPG public key: http://home.student.utwente.nl/l.e.veen/lourens.key

I wanted to compile ati.2 so I downladed the source from cvs.
I type: xmkmf /usr/X11R6 but it returns:

imake -I/usr/X11R6/config/cf -DTOPDIR=/usr/X11R6 -DCURDIR=.
Imakefile.c:35: Imake.tmpl: No such file or directory
imake: Exit code 1.
  Stop.

What can I do?

Thanks in advance
	micron

Hi,

there is already some support for TV out in linux - using the vesa 
driver. There is also support for the acceleration, in ati.2. So I would 
guess that reverse engineering would start by looking at register dumps 
under various vesa settings, ignoring the differences that relate to 
parts of the hardware that are already supported by ati.2. I would hope 
that TV-out is a relatively simple feature...

> I think the best chance for this is for someone in a country without 
> DMCA or DMCA-like laws

I bought a card with a TV-out which is actually composite video out (ie. 
not RF modulated). I'm personally not interested in recording the output 
to video but I would have thought that it was a legitimate use in almost 
all cases. There is a lot of confusion about legal issues here.

-Cam



-- 
ca...@me...                                                 <--

Please don't send HTML mail to mailinglists...

On Wed 30 April 2003 11:03, jf wrote:
<snip>
>
> If I have well understood, the problem is that ati doesn't want
> to give any doc about tv out because they have a contract with a
> Macrovison and no developer want/can to reverse Windows drivers
> because it may constitue a DMCA violation, isn't it ?

That's correct yes.

> This morning when I woke up I had the following idea (surely
> someone already had it ;)) why not using directly the windows
> drivers ? As avifile does with video codec ? I know video codec
> are really simple peace of code but I'm sure the function that
> activate the tv out is rather simple too

Well, Microsoft has a DDK (Driver Development Kit) up for download=20
on their site, which might make it possible to figure out the=20
driver interfaces Windows uses. However, the Windows kernel is=20
vastly different from the Linux kernel and/or XFree86. Doing this=20
would probably mean emulating Windows to a large extent to be able=20
to give the Windows driver an environment it can work in. Ofcourse,=20
we could try to figure out what part of the Windows driver does=20
TV-Out, but that would be reverse engineering.

I think the best chance for this is for someone in a country without=20
DMCA or DMCA-like laws (ie not the US, I'm not too sure on Europe,=20
I think the EUCD is now effective but it's a directive, not a law.=20
The individual countries still have to pass laws implementing the=20
EUCD before it's really in effect in terms of getting arrested for=20
violations) to boot Windows, fiddle with settings in the Windows=20
driver and make register dumps of the various settings, and then=20
use this information to patch the drivers. Or you would have to do=20
it in secret and post the resulting driver to Freenet anonymously,=20
but ofcourse that would make you a criminal.

Then again, I bought a DVD recently and played it on my Linux=20
machine using libdvdcss. If the EUCD were implemented in Holland=20
right now (I don't think it is yet) I'd be a criminal already. I=20
own the DVD, I own the computer I used to play it, and I have=20
permission to use the software I used from its authors for every=20
program involved. And yet I'm a criminal? These laws are crazy. But=20
unless the high-level politicians get a clue (not very likely=20
indeed) we're stuck with them...

Lourens
--=20
GPG public key: http://home.student.utwente.nl/l.e.veen/lourens.key

jf wrote:

> Hi,
> It is still a pity that tv out doesn't work on my AIW radeon 7500, 
> I've tried many things including atitvout which doesn't work for me... 
> I've tried the clonedisplay option of the "recent" X 4.3 (thought I'm 
> not sure I well understood the documentation) which give nothing 
> new... I always read carefully  threads with "tv out" in the subject, 
> with hope that some day I could read "Now tv out is supported !" but 
> it never comes....
>
> If I have well understood, the problem is that ati doesn't want to 
> give any doc about tv out because they have a contract with a 
> Macrovison and no developer want/can to reverse Windows drivers 
> because it may constitue a DMCA violation, isn't it ?
>
> This morning when I woke up I had the following idea (surely someone 
> already had it ) why not using directly the windows drivers ? As 
> avifile does with video codec ? I know video codec are really simple 
> peace of code but I'm sure the function that activate the tv out is 
> rather simple too
>
> Please tell me if I'm wrong, if anybody has ever work on it
>
> JF
> PS : congratulations to all people that work on gatos for they really 
> good job 

I share your pain.
I _had_ tv-out working under very special conditions with xfree86-4.2.1. 
The steps to get it running:

 - Edit /etc/X11/XF86Config and set the hsync / vrefresh rates to 
something relatively low, that the TV can handle. I can't remember 
what's a decent value.
 - Switch X to 800x600. It's OK to run X at whatever you want, and use 
CTRL-ALT-PLUS to change the resolution to 800x600 while X is running.
 - Run atitvout. I used to use: atitvout -r auto (I think). You may have 
to try ALL options. For me, when atitvout was 'successful' I would get a 
blank monitor screen. This initially made me think that it was 
unsuccessful, and it was a minor miracle that I realised that it had 
almost worked. See below...
 - Run mplayer / xine directly after atitvout. They do *something* 
magical, and the TV _and_ montior come alive. I then have to use the 
mouse to centre the display properly.

When the video clip is finished, the display gets corrupted, and can be 
restored by using CTRL-ALT-MINUS. This will re-activate the monitor, and 
corrupt the TV display (mine goes green).

I have not been able to get TV-out going under X with atitvout without 
using mplayer / xine directly after atitvout.
I used a script:

atitvout -r auto
for I in *
do
    mplayer -vo xvidix -ao alsa9 -double $I
done

atitvout no longer works for me under xfree86-4.3.0

I have tried the vesa driver. It works kinda OK, but you don't get xv 
acceleration, so the movie performance is fairly shit. Also I like my 
3D. I've also tried using a Geforce 440MX with tv-out, and I can assure 
you that you're not missing out on much from the nvidia camp. The 
drivers are horribly unstable. I locked my box hard about 8 times in one 
night by just starting X, finding that TV-out had not been activated for 
whatever reason, stopping X, editing my /etc/X11/XF86Config, and 
restarting X. When I *finally* got tv-out working (and it took a while), 
I discovered that even though my monitor was running at 800x600, the TV 
was only displaying the top 640x480 of the screen. All the locking 
eventually led to my system not getting past the 'Init booting...' stage 
just after the kernel fires up, and I had to do a complete re-install. 
Also, while the nvidia drivers support the XRandr extension normally, 
when TV-out is activated, the display is locked at the inital 
resolution. So there's no chance of running at 1024x768 and dropping 
back to 800x600 for a movie ... you've got to run at 800x600 the whole 
time! That's just not enough for me.  So anyway, I got pissed off enough 
with the nvidia setup to chuck my Geforce back on the growing pile of 
cards next to my computer and put my Radeon back in. I'm now content 
with having to reboot and run mplayer from the f**king console with the 
DirectFB driver (as the vesa driver now segfaults). I'm not particularly 
impressed with this setup. A lot of DVDs have those scene selection 
things which are quite handy and cool, but mplayer doesn't support them 
and this looks set in stone. So maybe xine will be up a DirectFB driver?

I dunno. I've said it before, and I'll say it again. I paid for a 
f**king card with tv-out support, and screw anyone who tries to stop me 
from using it. While nvidia's tv-out support isn't exactly something to 
be proud of, at least it exists. The same should be true for the DRI / 
Gatos drivers (the part about existing, not the reference to quality). I 
will be VERY careful when buying my next video card. But unfortunately 
it looks like I'll be heading back to nvidia. The DRI drivers won't run 
Tribes 2, Unreal Tournament 2003 or Neverwinter Nights, and the Gatos 
drivers won't to TV-out, and my card seems to be not detected at all by 
the km module, so I can't do video capture. But I can run glxgears with 
the new TCL drivers at 1000 fps ... cool! All discussions about blame 
and patents and US legal bullshit aside, as an end user of video cards, 
it is quite difficult to talk fondly of my experience with open source 
video drivers. It really pains me to say that, and I appreciate all the 
time open source developers put into the software that I use. In most 
cases I can say I'm extremely happy with the open source software that I 
use. But video drivers have always been a thorn in my paw, and I don't 
see that changing - especially with the 'sue / bomb / screw the world' 
attitude that we're getting from the US. Think I'm going overboard? Then 
ask one of the Gatos developers why they won't implement tv-out. Still 
not convinced? Then ask one of the DRI developers why they won't 
implement S3 Texture Compression.

As George Michael says, "I wanna kick back mumma, what's this world 
coming to lately?"

i've installed experimental-9 on my slackbox
and i found DRI works all right for me (it
used to hang the X server formerly) even while
watching TV (pal/italy), so everything sees to
work, BUT only at first cause if i close the
TV app (avvview or xawtv) it seems that something
is not closed properly, i get this behaviour:
- strange dotted stripes here and there when X does redrawing
- xv seems to work no more (colormap mess'd up etc.)
- if I reopen avview X refresh is all right again
- if I reopen xawtv I get no image whatsoever
- if I play any movie with mplayer and xv driver
   everything is ok

I can send some screenshots of the stripes if
can be of any help

by the way, avview does not stop the tuner
on exit so I get television channel audio
all the time :(

thanx

nebok

P.S.
the card is a ati AIW-rage128pro and
the uP is an Athlon-TBird



On 2003.04.19 23:04 Vladimir Dergachev wrote:
> 
> 2 new binary sets of ati.2 drivers have been posted.
> 
>   * experimental-9 contains a possible fix for r128 DRI problems
> 
>   * experimental-5 has been reposted as it appears to be the only
>     one that has correctly working radeon driver with PAL/SECAM due
>     to recent VBI changes. (it would still be nice to have correct
>     VBI constants..)
> 
>                         best
> 
>                           Vladimir Dergachev
> 
> 
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> Gatos-announce mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-announce
> 

I read Adam Richter's port attempt e-mails from January on the list, but I couldn't access his ftp site, so I don't know how far he got.

This is based on km-0.2.1, not CVS.  And I removed the stuff that required km_api. I wanted to start with the simplest thing possible.

What works for me:

1) Convert km_v4l to use new file_operations struct and use file_operations.private_data to store KM_STRUCT.

2) Use kmalloc(., GFP_KERNEL) for dma_table, and use pci_alloc_consistent for frame_buffer memory (ripped off the idea from the via82xx DMA stuff in the ALSA driver).

3) use virt_to_phys for kvirt_to_pa (necessary?)

4) I also played around a little with the irq handler logic.  (necessary?) 

5) Use kbuild-style Makefile.

This is XF 4.2.1 with (ati.2 driver for XF4.2).  

2.5.67-bk9 seems to want to OOPS on umount at reboot on my box (well after km has been rmmod'ed), so there may be something I'm missing, or it may be the kernel.

Fred

Hi,
Tuesday, 29 April, 2003, 9:12:36, Matt Mannikka wrote:
>   I have tried GATOS before with limited success.  The ati.2 binaries
> have helped out tremendously (glxgears was in the teens, now is at 300 
> fps).  However, I have not been able to get video capture to work.  I 
> would like to get my system working with freevo, but I need the 
> video4linux functionality.
AFAIK freevo is not quite compatible with gatos drivers right now (Or,
gatos drivers are not compatible with freevo, whatever)
Anyway. Timeshifting is _not_ possible on all-in-wonders yet. However,
just plain video recording should work quite reliable i suppose.

> I have a dual PIII 1 Ghz machine dual with Red Hat 8.0.  I am using 
> stock Red Hat Xfree 4.2, kernel 2.4.18-27.8.0smp, ffmpeg 0.4.6, and 
> xawtv 3.74.  I compiled modules from drm-kernel-1.2.0-16.tar.gz and 
> km-0.2.1.tgz.  I copied the modules from km to the 
> /lib/modules/2.4.18-27.8.0smp/misc/ directory.

> What I think is supposed to happen is I should be able to do...

>    1. xawtv
>    2. /sbin/modprobe videodev
>    3. /sbin/insmod km_api
>    4. /sbin/insmod km_drv
>    5. /sbin/insmod km_v4l
no you shouldn't do that, this one is kinda internal (temporary)
object.

>    6. ffmpeg /tmp/temp.mpg

> ...and get video captured in the temp.mpg file.  It doesn't quite happen 
> that way.  On the fifth step I get warnings about the kernel being 
> tainted, but it still loads.  When I try the last step I get:
>     /dev/video: Is a directory
>     Could not find video grab device
Remove 'video' directory from /dev and create a symlink video->video0
(all within /dev).

> I had compile problems with avview, but would prefer to use xawtv.  Can 
> anyone offer some help?
Probably Vladimir will help you about avview when he is back.
I don't use it, sorry.
-- 
Best regards,
 Nikolai Zhubr

> Matt

Bob McElrath wrote:

>I have a Radeon 7200 AIW 32MB, lspci reports it as an "R100".
>
>When trying to run glxgears with the latest drivers (both drm-kernel
>with the trivial fix for rh9 and ati.2), the machine hangs hard.
>Slightly older versions (like the one included with rh9) have the same
>problem.  TV viewing with xawtv seems to work fine.
>  
>
I have noticed this as a new symptom of 4.3.0 w ATI.2.  When I have DRI 
disabled ( runnning xinerama ) any GL app, including xscreensaver, the 
system hardlocks ( no IP response ).  This is excepted by glxgears which 
seems to be the only app that detects DRI is disabled and runs in 
software.  Vladimir, is there a chance that the Mesa libraries have 
changed just slightly, so that the GATOS patched versions break DRI 
detection in some cases??

Its nice to have both heads up and running finally.  Without xinerama, 
TV, XV, and DRI all work on both cards without a hitch.  Thanks!  
However, its a little bit annoying to have to remember to disable my 3D 
screensaver, when xinerama is enable. (most of the time).

Christopher Crawford

>I have tried futzing around with the AGP settings, and turning it off by
>not loading the agpgart module and using the "ForcePCIMode" option.
>When I do that it just crashes hard when X is started.
>
>Any suggestions on this?
>
>I'm using:
>
>    ATI-4.3.0-9.i386.tar.gz
>    drm-kernel-1.100.0-9.tar.gz
>    XFree86 4.3  (redhat 9)
>
>Everything else is straight rh9.
>
>Cheers,
>Bob McElrath [Univ. of Wisconsin at Madison, Department of Physics]
>
>    "You measure democracy by the freedom it gives its dissidents, not the
>    freedom it gives its assimilated conformists." -- Abbie Hoffman
>
>  
>

  I have tried GATOS before with limited success.  The ati.2 binaries 
have helped out tremendously (glxgears was in the teens, now is at 300 
fps).  However, I have not been able to get video capture to work.  I 
would like to get my system working with freevo, but I need the 
video4linux functionality.

I have a dual PIII 1 Ghz machine dual with Red Hat 8.0.  I am using 
stock Red Hat Xfree 4.2, kernel 2.4.18-27.8.0smp, ffmpeg 0.4.6, and 
xawtv 3.74.  I compiled modules from drm-kernel-1.2.0-16.tar.gz and 
km-0.2.1.tgz.  I copied the modules from km to the 
/lib/modules/2.4.18-27.8.0smp/misc/ directory.

What I think is supposed to happen is I should be able to do...

   1. xawtv
   2. /sbin/modprobe videodev
   3. /sbin/insmod km_api
   4. /sbin/insmod km_drv
   5. /sbin/insmod km_v4l
   6. ffmpeg /tmp/temp.mpg

...and get video captured in the temp.mpg file.  It doesn't quite happen 
that way.  On the fifth step I get warnings about the kernel being 
tainted, but it still loads.  When I try the last step I get:
    /dev/video: Is a directory
    Could not find video grab device

I had compile problems with avview, but would prefer to use xawtv.  Can 
anyone offer some help?

Matt

I have a Radeon 7200 AIW 32MB, lspci reports it as an "R100".

When trying to run glxgears with the latest drivers (both drm-kernel
with the trivial fix for rh9 and ati.2), the machine hangs hard.
Slightly older versions (like the one included with rh9) have the same
problem.  TV viewing with xawtv seems to work fine.

I have tried futzing around with the AGP settings, and turning it off by
not loading the agpgart module and using the "ForcePCIMode" option.
When I do that it just crashes hard when X is started.

Any suggestions on this?

I'm using:

    ATI-4.3.0-9.i386.tar.gz
    drm-kernel-1.100.0-9.tar.gz
    XFree86 4.3  (redhat 9)

Everything else is straight rh9.

Cheers,
Bob McElrath [Univ. of Wisconsin at Madison, Department of Physics]

    "You measure democracy by the freedom it gives its dissidents, not the
    freedom it gives its assimilated conformists." -- Abbie Hoffman

all,

I sent this message to Leif around getting the mach64 DRI driver and tv in/capture (km) to work 
together to create a v4l device that works for tuning, capture and playback.  I ahve included Leif's 
response, as it helps illustrate the current situation and a possible solution.

I have tried several version of km, include 0.2.0 and 0.2.1, as well as the
km-hack-8 drivers from Nikolai Zhubr.  No luck as of yet.

Any hints or helps on how to get these working prior to the mach64 merge?  I am willing to do 
testing if needed.

Dave




*********  Included mail thread:


I was afraid someone was going to ask me about this eventually. ;)  The problem is, I don't have an 
All-in-Wonder card or any tuner/capture cards.
So I've only worked on the merge as far as getting XVideo PutImage working
-- enough for video overlay, DVD etc.  That's all I'm able to test myself.
It's quite possible that there could be problems when using the km kernel module and the mach64 DRI 
kernel module together.  I'd expect the tuner drivers to work, as those are just copied from the 
GATOS tree, but capture is another story.  Both km and the DRM kernel module use the card's DMA 
engine, so they need to play nice together.

My main focus is to get the mach64 DRI driver completed and merged into
XFree86.  When that happens it can get merged downstream into GATOS and
Vladimir and crew can work on getting 3D and GATOS working together.

Since Marc Aurele La France has committed XVideo PutImage support for mach64 to XFree86 CVS, I'll 
stop maintaining my patch once that code is merged into the DRI tree.  My patch was always meant as 
a temporary solution until that happened.  So you'll either have to hack on the patch yourself, or 
wait for the DRI code to get into the GATOS tree.  At any
rate, I'll try to answer questions about the DRI driver when someone
starts working on TV-in/capture + DRI for mach64.

--Leif

On Mon, 28 Apr 2003, David Fannin wrote:

 > First off, thanks for the great work on the mach64 drivers.  I have a
 > question for you for using these drivers with the rest of Gatos -
 > specifically, for using the ATI tv tuner with mplayer.  ( I know this
 > may be a gatos/mplayer question, but I checked the mailing lists for
 > both, and I was not able to find any pointers).
 >
 > Here is my setup:
 >
 > Dual P3-700 machine , 1GB ram
 > ATI AIW Pro 8mb (old, but cheap!) AGP  (Rage Pro chip)
 > Redhat 9.0
 > kernel 2.4.20 (fresh one from kernel.org)
 > XFree 4.3.0 (complied and installed from XFree86.org)
 > mach64-06 patch (complied and mach64.o copied into kernel)
 >
 > The mach64.o module works great, and I have confirmed that the Xvideo
 > and XFree86-DRI extensions are activated, and working in my Xserver.
 > I can also play videos (with mplayer) with -vo xv.  TV-out works great
 > as well. Note that I did not compile the ati.2 or drm-kernel packages
 > for gatos. My assumption is that this source code was replaced by your
 > mach64 package. Please correct me if I am wrong.
 >
 > I have problems on getting tv-in and tv-capture to work.  My goal is
 > to get the tv-in/capture to work for mplayer, since I would like to
 > use this for Freevo.
 >
 > For these, I installed and compiled the km (both the 2.1,2.0 versions)
 > modules, and installed them (with module videodev).  They appear to load ok,
 > and a /dev/video0 device is active.    However, here is the output of
 > the v4l-info command:
 > ### video4linux device info [/dev/video0] ###
 > general info
 >     VIDIOCGCAP
 >         name                    : "Km"
 >         type                    : 0x1 [CAPTURE]
 >         channels                : 1
 >         audios                  : 0
 >         maxwidth                : 640
 >         maxheight               : 240
 >         minwidth                : 640
 >         minheight               : 240
 >
 > channels
 > ioctl VIDIOCGCHAN: Invalid argument
 > tuner
 > ioctl VIDIOCGTUNER: Invalid argument
 > audio
 > ioctl VIDIOCGAUDIO: Invalid argument
 > picture
 >     VIDIOCGPICT
 >         brightness              : 32768
 >         hue                     : 32768
 >         colour                  : 32768
 >         contrast                : 0
 >         whiteness               : 32768
 >         depth                   : 16
 >         palette                 : YUV422
 >
 > buffer
 > ioctl VIDIOCGFBUF: Invalid argument
 > window
 >     VIDIOCGWIN
 >         x                       : 0
 >         y                       : 0
 >         width                   : 640
 >         height                  : 240
 >         chromakey               : 0
 >         flags                   : 4169646320
 >
 >
 > This appears to indicating that the tuner/tv-in is not active, or
 > there is an api issue (ioctl errors).
 >
 > When I run mplayer to view the tv input, I get the following error:
 > (abbreviated)
 > ....
 > Playing TV channel 4
 > TV detected! ;-)
 > Selected driver: v4l
 >  name: Video 4 Linux input
 >  author: Alex Beregszaszi <al...@na...>
 >  comment: under development
 > Selected device: Km
 >  Capabilites: capture
 >  Device type: 1
 >  Supported sizes: 640x240 => 640x240
 >  Inputs: 1
 > ioctl get channel failed: Invalid argument
 > ioctl get mbuf failed: Invalid argument
 > Falling back on trying to parse playlist TV channel 4... Trying asx...
 >
 > This appears to indicate that the v4l device (km) is failing here.
 >
 > Additionally, I have attempted to use avview for tv viewing, but I am
 > running into a bunch of errors trying to start it up.  It compiled
 > correctly (with the tcl/tk 8.4 versions and ffmpeg), but during the
 > startup, after the inital configuation window, it goes into an infinite loop, with the stout
 > spitting out "read line =" over and over, until I kill it.    The strange
 > thing with avview is that with the km 0.2.0 module, it will show the live tv
 > tuner in the tv window!   I tried different version of avview (11.10, 11.5),
 > but they all had the same problem.  If I try avview with the km 0.2.1
 > module, then my machine does a hard lockup, and I have to reset.
 >
 > One of bit of information is that xawtv works for tv-in viewing, but
 > only in the Xvideo mode.  When I try to get xawtv to work with l4v
 > mode, it does not work.
 >
 >
 > My goal is to get the AIW Pro card to work with mplayer for tv-in viewing
 > and tv capture.   mplayer does not appear to have a Xvideo capture mode, so it needs v4l to work. 
  Based on my problems, do you have any suggestions or hints
 > for I may need to do to get this work?
 >
 > Again, thanks for you help, and the great job with the mach64 drivers - they
 > worked great, and I have no issues with them.   It was, however, a major
 > PITA to have to recompile the kernel and XFree86, but thats a redhat
 > issue.
 >
 > Dave
 > Dublin, Ca, USA


-- 
Leif Delgass
http://www.retinalburn.net

Stefan Misch wrote:
> 
> Where can I get it?
> 
> I've googled for it but "libzvbi" just finds some debian sites with
> v0.1. Some people here reported to have even v2.4 but I can't find it
> and Slackware doesn't seem to have it either. At least I can't find it.
> 
> Thanks in advance :)

Visit zapping.sf.net ...

Gary

Where can I get it?

I've googled for it but "libzvbi" just finds some debian sites with 
v0.1. Some people here reported to have even v2.4 but I can't find it 
and Slackware doesn't seem to have it either. At least I can't find it.

Thanks in advance :)

Hello,
 
I installed the ati.2 drivers, as well as the DRM kernel modules.
 
Everything was running fine, then I rebooted and couldn't login to X.
 
I got the error:
 
Failed to initialize Rage Theatre, chip disabled.
 
I'm not exactly sure what information you would need to figure out this
problem, but I will provide it.
 
I'm running RedHat 8.0 with all updates, an ATI Radeon 8500LE 128 MB.
 
Thanks,
Brad Atkins
 
 

> sync
> /sbin/insmod ./km_api.o
> /sbin/insmod ./km_drv.o km_debug=0
> ./km_drv.o: unresolved symbol
> video_register_device_R535abbbd
> ./km_drv.o: unresolved symbol
> video_unregister_device_R62a6f4c8
> make: *** [test] Error 1
Do 
	modprobe videodev 
before

Hi -

First, I last tried Gatos about 18 months ago and, although I was able 
to get a picture, there were a couple of problems.  I just tried it 
again this past week and it works quite well with avview. 
Congratulations on the improvements!

Avview works fine for me, but Xawtv does not work.  I am able to obtain 
sound from all channels, but in the Xawtv display window I am not able 
to obtain a picture.  The display window is black, instead of a TV 
picture.  I have tried moving the display window around, 
enlarging/reducing it, without success.  I have tried both xvideo and 
ntsc under "TV norm"

Any ideas?

I am using ATI AIW Pro 32 MB, RedHat 8.0, XF86 4.2.  Modules km_drv, 
km_api, and r128 are loaded.

Thanks,
-David Betz

Did you apply the kernel patchs? Also make sure that TOPDIR is set 
properly in the km Makefile

-David
 



"David S. Brown" <dav...@ya...> 
Sent by: gat...@li...
04/28/2003 02:46 AM

To
gat...@li...
cc

Subject
[GATOS]make test km






I get the following error when doing make test on km
on redhat 7.3.

sync
/sbin/insmod ./km_api.o
/sbin/insmod ./km_drv.o km_debug=0
./km_drv.o: unresolved symbol
video_register_device_R535abbbd
./km_drv.o: unresolved symbol
video_unregister_device_R62a6f4c8
make: *** [test] Error 1




__________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
http://search.yahoo.com


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
Gatos-devel mailing list
Gat...@li...
https://lists.sourceforge.net/lists/listinfo/gatos-devel

I get the following error when doing make test on km
on redhat 7.3.

sync
/sbin/insmod ./km_api.o
/sbin/insmod ./km_drv.o km_debug=0
./km_drv.o: unresolved symbol
video_register_device_R535abbbd
./km_drv.o: unresolved symbol
video_unregister_device_R62a6f4c8
make: *** [test] Error 1




__________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
http://search.yahoo.com

I tried with experimental 5, 8 and 9.
With 5 and 8, X crashes and refuses even to show an image.
With Experimental 9, I get the TV working but I have the strange 
patterns all the time and the bug with Mozilla.
Since I don't want to spend a day recompiling XFree to try compiling 
ati.2 from CVS source, I will give up for now; there is no solution to 
get my tuner working.

benoit wrote:

>Bonjour,
>
>At a first step if you do not need dri there is no need to worry about drm.
>Installing ati.2 drivers is enough to view TV.
>Also you may want to try ati.2 experimental 9 for Xfree 4.3.0.
>
>I have AIW 128 (not pro).
>Parts of the screen get strange pattern but only when I exit xawtv or avview.
>
>Benoit
>
>
>On Sunday 27 April 2003 08:51 pm, Eric Buist wrote:
>  
>
>>Hello,
>>Since approximately one week, I am trying to get my ATI All-In-Wonder
>>128 Pro TV Tuner working under RedHat Linux 9 with XFree86 4.3. I was
>>able to compile the DRM kernel module from GATOS by modifying drmP.h in
>>order to make the module behave as if it were on a 2.5 kernel,  like I
>>found by browsing this mailing list. I used the ATI.2 binary for XFree86
>>4.3, experimental 8. Now my TV Tuner works approximately great, but
>>there are two strange problems.
>>- Only 3/4 of the xawtv window shows a TV program, the rest is filled
>>with garbage data. The same problem produced itself with AVView 0.10.8.
>>I tried to downgrade to DRM/ATI.2 experimental 7 but my XFree server
>>locked at startup.
>>- When I listen to TV with xawtv (probably when the ATI overlay is
>>activated), Mozilla 1.2 and 1.3 starts showing random black horizontal
>>lines when I use the mouse wheel to scroll up and down. The black bars
>>disappear when the screen is refreshed and the bug continues some times
>>after xawtv is closed.
>>I tried to download kernel 2.4.20 from ftp.kernel.org and recompile it
>>without any improvement. km could be compiled against the plain
>>non-RedHat-customized kernel but my ATI was still badly behaving.
>>I tried the ATI.2/DRM CVS versions but I couldn't get ATI.2 compiled
>>from source because it requires recompiling the entire XFree tree which
>>is very long for possibly no result.
>>ATI TV Tuner worked great for me with RedHat Linux 7.3 and Mandrake
>>Linux 9. Problems started with RedHat 9, maybe something related to
>>XFree86 4.3 since I tried with a plain 2.4.20 kernel.
>>
>>
>>
>>-------------------------------------------------------
>>This sf.net email is sponsored by:ThinkGeek
>>Welcome to geek heaven.
>>http://thinkgeek.com/sf
>>_______________________________________________
>>Gatos-devel mailing list
>>Gat...@li...
>>https://lists.sourceforge.net/lists/listinfo/gatos-devel
>>    
>>