gatos-devel Mailing List for GATOS

Curtis Magyar wrote:

> I'm going to take a look at your patches, and hopefully we can come to
> some kind of common patch between the two of us.  I'm also using
> Gentoo, and I have both a Radeon 7200 (R100), and a Rage128 All In
> Wonder.  I'm trying to get my PCI Mach64 card back from a friend for a
> while so we can finally get that driver back on track too.
> 
> My patch is here for the time being:
> http://curtman.mine.nu/~curtis/ati.2.patch
> or compressed:
>  http://curtman.mine.nu/~curtis/ati.2.patch.bz2
> 
Well, my there:
ftp://ftp.penguin.cz/pub/users/utx/gatos/

> If you want to take a look.  I took a different approach to patching,
> in that I started with a diff of the Gatos 4.4.0 code -vs- XFree86
> 4.4.0's ati driver, then hand merged each change into Xorg's 6.8.1 ATI
> driver.  The R128 code I haven't tested at all though, I'll be doing
> that later tonight with your patch along for reference.

I have first hand edited patch, then applied and then manually applied
rejects. I have had opened patch in mc editor and I was trying to think
hard about all red (deleted) lines. In most cases, I was able to have
only green (added) patches or those, where deleted lines were obvious.
But not succeeded for all (see below). Sometimes I really don't know.

> It sure is great to see more people are interested in Gatos these days!

I see the best will be integrating to xorg-x11. Xorg people are much
more open for third party patches than XFree86. In this case, this
porting will be the last one. If it can happen, patch should be as clean
as possible.


I have some notes there:

Please ignore radeon_video.c in my patch (better ignore whole file,
contains many uncertain chunks).

Solve better (#ifdef, not -) unloading BIOS.


Code places, which should be verified:

Why this change?
-    if (!xf86SetDepthBpp(pScrn, 0, 0, 0, (Support24bppFb
+    if (!xf86SetDepthBpp(pScrn, 8, 8, 8, (Support24bppFb

We are incompatible here with Xorg in number of steps. Best should be
"as much as card supports or has any reason".
-   {XvSettable | XvGettable, 0, (1 << 24) - 1, "XV_COLORKEY"},
-   {XvSettable | XvGettable, -64, 63, "XV_BRIGHTNESS"},
-   {XvSettable | XvGettable, 0, 31, "XV_SATURATION"},
-   {XvSettable | XvGettable, 0, 1, "XV_DOUBLE_BUFFER"}
And probably related OUTREG lines below in this file.

My mistake. I have added duplicate code. Should be fixed:
     pPriv->brightness = 0;
-    pPriv->saturation = 16;
+    pPriv->saturation = 0;
+    pPriv->contrast = 0;
+    pPriv->hue = 0;

Verify offset computation and OUTREGs in R128DisplayVideo422().

What code is up to date and better: *Helper or R128*?


Probably more testing is needed. I still have problems with R128:
- Lost interrupt after gdm logout (not sure, whether it occurs in
6.8.1). Not GATOS relates.
- Cursor crippling for both sw (sw cursor cripples xv color mask; not
yet verified in 6.8.1) and hw (hw cursor mode {for transparent cursors}
is crippled by GL; something was improved here, in 6.8.1 the look is
reverted after GL app quits). Not GATOS related.
- Transparency support is slooow. Much slower than software
pseudo-transparency. Looks like unaccelerated X.



________________________________________________________________________
Stanislav Brabec
http://www.penguin.cz/~utx, ICQ 116020046

Hallo.

I spent this evening by making my AIW R128 tuner working again with
xorg-X11 (6.8.0+Gentoo pathes, probably equal to 6.8.1).

When I have looked at the list, I have found, that I was not only one
doing it.

Well. I have stopped merging rejects and there is my result.

xorg-x11-6.8.0-gatos.patch:
Patch for xorg-x11 to make it Gatos (after adding multimedia drivers to
the directory).

Because Curtis Magyar probably correctly ported Radeon part, patch
should be complete. I kept radeon_video.c_see_FIXME.rej and
stopped my work there (search for FIXME string in reject file).

After quick testing, r128 tuner works with my patch.

Looking at some '-' lines, I am really in doubt - is it out-of-sync X
change, is it tuner related GATOS change, tuner-related different
register programming, an additional feature, missing in xorg-x11 or
mistake? Hard to guess...

Maybe it's a time to make all appropriate changes #ifdef GATOS, and then
try to submit them for mainstream xorg. It will prevent repeated porting
and branching of GATOS in future.

If it cannot be accepted, CVS should store also original xorg source (or
at least configure cvs to not modify xorg CVS headers (cvs admin can do
it)).


________________________________________________________________________
Stanislav Brabec
http://www.penguin.cz/~utx, ICQ 116020046

Hi Curtis,

     Thank you for the patch ! Can you produce one against Xorg 6.8.1 ?
It would be easier for most people to apply..

                      thank you !

                            Vladimir Dergachev

On Mon, 27 Sep 2004, Curtis Magyar wrote:

> Since Vladimir is such a busy guy these days, and I'm so impatient...
> I thought I'd take a shot at porting Gatos to Xorg 6.8.1.  If anyone
> has some time to examine my changes please do so, and let me know.  I
> commented some stuff that I was curious about, most of that is marked
> XXX:, so please grep for those.  There's probably some inadvertent
> whitespace changes in there too that don't need to be, I initially had
> problems syncing my tab/indents to what was there already.
>
> I would highly recommend users stay away, since I have very little
> experience with Gatos, and Xorg/XFree code.  My goal was to be able to
> watch TV again, and also to help Vladimir out if possible.  With that
> in mind, understand that capture is probably very broken, it's not
> something that I use, so I didn't test it.  It does compile, and TV
> does seem to work on my Radeon 7200 AIW.
>
> This code may cause the universe to collapse into a singularity, I'm not sure.
>
> Anyway, this patch should be applied against ati.4.4.0 in CVS, not ati.2.
>
> Thanks.
>
> -- 
> Curtis Magyar
> cu...@gm...
>
> P.S.
> My card is a:
>  ATI Technologies Inc Radeon R100 QD [Radeon 7200]
>

On Fri, 24 Sep 2004, Paul Kirschner wrote:

> In the recent announcement for experimental 1 of ati-4.4.o binaries, it w=
as=20
> mentioned that mach64 "is not supported". Unfortunately, I've got an=20
> "All-in-Wonder" Mach 64 card.
>
> When will Mach64 support be added back in?

The current plans are to repeat the port of ati.2 directly into X.org=20
CVS tree. I'll do that as soon soon as suitably large chunk of=20
spare time turns up.

I'll post an announcement as soon as there is anything to test.

                         best

                           Vladimir Dergachev

>
> BTW - I appreciate your efforts. I feel somewhat orphaned in that ATI doe=
s=20
> NOT support Win2000 (yeah, that other OS) for the All-in-Wonder. So you a=
re=20
> my only hope unless I want to boot up Win98 (shudder!!!) just to watch th=
e=20
> news.
>
> thanks...
>
> _________________________________________________________________
> Don=92t just search. Find. Check out the new MSN Search!=20
> http://search.msn.click-url.com/go/onm00200636ave/direct/01/
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
> Project Admins to receive an Apple iPod Mini FREE for your judgement on
> who ports your project to Linux PPC the best. Sponsored by IBM.
> Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

In the recent announcement for experimental 1 of ati-4.4.o binaries, it was 
mentioned that mach64 "is not supported". Unfortunately, I've got an 
"All-in-Wonder" Mach 64 card.

When will Mach64 support be added back in?

BTW - I appreciate your efforts. I feel somewhat orphaned in that ATI does 
NOT support Win2000 (yeah, that other OS) for the All-in-Wonder. So you are 
my only hope unless I want to boot up Win98 (shudder!!!) just to watch the 
news.

thanks...

_________________________________________________________________
Don’t just search. Find. Check out the new MSN Search! 
http://search.msn.click-url.com/go/onm00200636ave/direct/01/

Hurray!!!

	Thanks for this release.=20

Now I can go and install Fedora Core 2 to my box.

Regards,
Ivo.

On Wed, 2004-09-22 at 09:06 +0200, Lourens Veen wrote:
> A new set of binary Gatos drivers for X.org 6.7.0 is now (finally, sorr=
y!)=20
> available. These binaries are compiled against X.org 6.7.0, but may als=
o work=20
> with XFree86 4.4.0. There has only been some very limited testing, so t=
his=20
> release is as experimental as the other Gatos releases.
>=20
> This release is an unpatched version of the ati.4.4.0 CVS branch, and a=
s such=20
> it does _not_ include support for TV-out. Mach64 cards are also unsuppo=
rted.=20
> Watch the gatos-devel list if you want TV-out. I hope to release a patc=
hed=20
> binary later today or tomorrow as well, but it probably won't be on the=
=20
> official SourceForge page.
>=20
> You can get the files from=20
> https://sourceforge.net/project/showfiles.php?group_id=3D12629&package_=
id=3D126120&release_id=3D269650
>=20
> Installation instructions are included. The procedure is basically the =
same as=20
> for the other releases, but do check the README.
>=20
> Please send comments and test results to gat...@li....=
net.
>=20
> Cheers,
>=20
> Lourens
>=20
>=20
> -------------------------------------------------------
> This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
> Project Admins to receive an Apple iPod Mini FREE for your judgement on
> who ports your project to Linux PPC the best. Sponsored by IBM.
> Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
--=20
Ivo S=C3=A4rak <iv...@ve...>

A new set of binary Gatos drivers for X.org 6.7.0 is now (finally, sorry!) 
available. These binaries are compiled against X.org 6.7.0, but may also work 
with XFree86 4.4.0. There has only been some very limited testing, so this 
release is as experimental as the other Gatos releases.

This release is an unpatched version of the ati.4.4.0 CVS branch, and as such 
it does _not_ include support for TV-out. Mach64 cards are also unsupported. 
Watch the gatos-devel list if you want TV-out. I hope to release a patched 
binary later today or tomorrow as well, but it probably won't be on the 
official SourceForge page.

You can get the files from 
https://sourceforge.net/project/showfiles.php?group_id=12629&package_id=126120&release_id=269650

Installation instructions are included. The procedure is basically the same as 
for the other releases, but do check the README.

Please send comments and test results to gat...@li....

Cheers,

Lourens

On Tue, 21 Sep 2004 08:52:03 +0200, G=E9rard Delafond <ge...@de...>=
 wrote:
> Is GATOS included in X.org R 6.8, as I read somewhere ? Completely or
> partially ?

As far as I know, there is absolutely no Gatos code in Xorg 6.8.0, and
the plan is to begin merging it in Xorg CVS for possible inclusion in
the *next* release.  From what I hear Gatos is also incompatible with
6.8.0 at this point, and that is unlikely to change unless someone
other than Vladimir champions the cause.  He's busy with the R300
project and the merging into Xorg CVS.  Those should be of higher
priority IMHO.  Maybe as the merger happens a backport would become
possible, but I wouldn't hold my breath.

See:
  http://article.gmane.org/gmane.comp.video.gatos.devel/126

--=20
Curtis Magyar
cu...@gm...

Hi all (and especially Vladimir),

I've been working on the ati.4.4.0 branch for a bit, with the goal of finally
getting some Xorg 6.7.0 binaries released. Unfortunately, I messed up the file
upload to SourceForge, and it won't give me another try for 24 hours.

In the mean time I'm putting together an even more experimental than plain
ati.4.4.0 CVS, and I wanted to include support for the ForceCRT option as posted
to the list by Scott Alfter in
http://sourceforge.net/mailarchive/message.php?msg_id=6821588

As I tried to port the patch, I noticed that things have changed quite a bit,
and I think that this option is now no longer necessary, since you can use
MonitorLayout now. Is this correct?

TIA,

Lourens (from a different address than normally, since the spamfilter blocks my
mail otherwise)

I'm using Mandrake 10 (download edition) Kernel
2.6.3-4mdksmp and XFree86-4.3-30mdk

I have ati.2-4.3.20030708-1mdk and atitvout installed

The TV out works, although only in text mode (80x25).
I get simultaneous display on the TV and the monitor
from the time the computer is booted up right thru to
the point where I type 'startx' (my default runlevel
is 3).This means that the card/driver is working. Its
only when the X is started that the display from the
TV (and the monitor) is lost.

My guess is that its some frequency (Modeline) issue?
But my expertise is well below adjusting modelines in
XF86Config-4. The modelines that I think related to
this problem
----
# TV fullscreen mode or DVD fullscreen output.
# 768x576 @ 79 Hz, 50 kHz hsync
  ModeLine "768x576"     50.00  768  832  846 1000  
576  590  595  630
----
[root@omni root]# atitvout detect
CRT is attached.
TV is attached via Composite.
[root@omni root]# atitvout vbe   
VBE Version: 2.0
VBE OEM Identification: ATI RAGE128
[root@omni root]# atitvout active
CRT is active.
Autoswitch is active

 --- Prash <get...@gm...> wrote: 
> AFAIK - there are only two options to get TVout
> workiing on ATI -  a)
> Gatos b) ATI's closed source driver.
> 
> a)
> What distro are you on? Gatos works only with XFree.
> For TVOut to work
> you need ati.2.
> Ati.4.4.0 does not support Tvout yet.
> 
> 
> b)
> ATI's closed source driver also work with XFree and
> 2.4 kernel only.
> See http://www.ati.com
> 
> 
> On Tue, 21 Sep 2004 04:43:51 -0400 (EDT), Ashley
> Moore
> <lin...@ya...> wrote:
> > Hi,
> > I just obtained a second computer and installed
> Mdk 10
> > community download, currently running kernel
> > 2.6.3-4mdksmp. I plan to use it mainly as my test
> > platform and to watch DVD's. The h/w config is:
> Dual
> > AMD Athlon 1.5Ghz / ATI Rage128 (A-I-W Pro) /
> Sound
> > Blaster Live / ancient 14" Fujikama PVGA-1024A
> > monitor. All the h/w seems to be configured right,
> the
> > display card is configured as an 'ATI Rage 128
> TVout
> > (it was detected as an 'ATI Rage 128). I have
> > installed xine with the correct drivers and am
> able to
> > watch DVD's on the monitor perfectly.
> > 
> > The display appears simultaneously on the monitor
> and
> > the TV when its connected to the tv out (on
> reboot).
> > On issuing 'startx' the display vanishes from the
> TV
> > and a garbled image on the monitor as if the
> frequency
> > is incorrect. i.e I can see the output on the TV
> from
> > the boot up BIOS screen (and the lilo graphic
> screen)
> > right upto the point when I start X. (default
> runlevel
> > is 3)
> > 
> > I could watch DVD's on the TV under WindowsXP
> using
> > the same display card on my main computer (a dual
> > boot, which has been relocated. I swapped out the
> > display card to the second PC).
> > 
> > I looked for similar problems with google without
> any
> > luck. Am I missing something? Any help, pointers
> would
> > be appreciated. I'd hate to bring back the
> WindowsXP
> > PC just to watch DVD's on TV
> > __________________
> > Linux User #229125
> > 
> >
>
-------------------------------------------------------
> > This SF.Net email is sponsored by: YOU BE THE
> JUDGE. Be one of 170
> > Project Admins to receive an Apple iPod Mini FREE
> for your judgement on
> > who ports your project to Linux PPC the best.
> Sponsored by IBM.
> > Deadline: Sept. 24. Go here:
> http://sf.net/ppc_contest.php
> > _______________________________________________
> > Gatos-devel mailing list
> > Gat...@li...
> >
>
https://lists.sourceforge.net/lists/listinfo/gatos-devel
> >
>  

Hi !

I am a new subscriber on the list.

I live in France, I have been a Linux user for years, but I have very few 
skill in programming.

I have a laptop described here :
http://www.delafond.org/maxdata/

At the beginning, I used XFree 3.x with UTAH-GLX, and it worked fine.

Then XFree 4.x came, and I had to use the unified DRI-xvideo patch from 
retinalburn :
http://www.retinalburn.net/linux/index.html

With this patch, I could get TV-out and DRI working. At last, I could play 
tuxracer on the TV :-).

Now X.org R 6.8 is here, and I thought it included all the patches, but after 
trying, my card is still not accelerated.

My question is : why is this patch not included in Gatos or X.org project ? Is 
it difficult to do it ? Is the quality bad ? Is there a problem with the 
license ?

It would give full support for 6 cards out of the box, which would be very 
cool !

For the moment, I did not try to aply the patch : not sure if it applies to 
the actual X.org and Gatos.

Is GATOS included in X.org R 6.8, as I read somewhere ? Completely or 
partially ?

Thanks in advance for reading and for your answers.

regards

-- 
Gerard

Hi,
I just obtained a second computer and installed Mdk 10
community download, currently running kernel
2.6.3-4mdksmp. I plan to use it mainly as my test
platform and to watch DVD's. The h/w config is: Dual
AMD Athlon 1.5Ghz / ATI Rage128 (A-I-W Pro) / Sound
Blaster Live / ancient 14" Fujikama PVGA-1024A
monitor. All the h/w seems to be configured right, the
display card is configured as an 'ATI Rage 128 TVout
(it was detected as an 'ATI Rage 128). I have
installed xine with the correct drivers and am able to
watch DVD's on the monitor perfectly.

The display appears simultaneously on the monitor and
the TV when its connected to the tv out (on reboot).
On issuing 'startx' the display vanishes from the TV
and a garbled image on the monitor as if the frequency
is incorrect. i.e I can see the output on the TV from
the boot up BIOS screen (and the lilo graphic screen)
right upto the point when I start X. (default runlevel
is 3)

I could watch DVD's on the TV under WindowsXP using
the same display card on my main computer (a dual
boot, which has been relocated. I swapped out the
display card to the second PC).

I looked for similar problems with google without any
luck. Am I missing something? Any help, pointers would
be appreciated. I'd hate to bring back the WindowsXP
PC just to watch DVD's on TV
__________________
Linux User #229125

Hi Bill,

    If you want to play with 3d acceleration features of your card take a 
look at http://r300.sf.net

                          best

                            Vladimir Dergachev

On Tue, 21 Sep 2004, Bill Sharer wrote:

> lspci -vv for a 9800 Pro AIW
>
> 01:00.1 Display controller: ATI Technologies Inc Radeon R350 [Radeon 9800] 
> (Secondary)
>       Subsystem: ATI Technologies Inc: Unknown device 4f73
>       Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- 
> Stepping+ SERR- FastB2B-
>       Status: Cap+ 66Mhz+ UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- 
> <TAbort- <MAbort- >SERR- <PERR-
>       Latency: 64 (2000ns min), cache line size 40
>       Region 0: Memory at d0000000 (32-bit, prefetchable) [size=128M]
>       Region 1: Memory at fcc00000 (32-bit, non-prefetchable) [size=64K]
>       Capabilities: [50] Power Management version 2
>               Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA 
> PME(D0-,D1-,D2-,D3hot-,D3cold-)
>               Status: D0 PME-Enable- DSel=0 DScale=0 PME-
>
>
> I'll do an Xfree86 log if anyone is interested but the succint summary is 
> thus:
>
> (WW) RADEON(0): Direct rendering not yet supported on Radeon 9500/9700 and 
> newer  cards
>
> Bill
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
> Project Admins to receive an Apple iPod Mini FREE for your judgement on
> who ports your project to Linux PPC the best. Sponsored by IBM.
> Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

lspci -vv for a 9800 Pro AIW

01:00.1 Display controller: ATI Technologies Inc Radeon R350 [Radeon 
9800] (Secondary)
        Subsystem: ATI Technologies Inc: Unknown device 4f73
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- 
ParErr- Stepping+ SERR- FastB2B-
        Status: Cap+ 66Mhz+ UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- 
<TAbort- <MAbort- >SERR- <PERR-
        Latency: 64 (2000ns min), cache line size 40
        Region 0: Memory at d0000000 (32-bit, prefetchable) [size=128M]
        Region 1: Memory at fcc00000 (32-bit, non-prefetchable) [size=64K]
        Capabilities: [50] Power Management version 2
                Flags: PMEClk- DSI- D1+ D2+ AuxCurrent=0mA 
PME(D0-,D1-,D2-,D3hot-,D3cold-)
                Status: D0 PME-Enable- DSel=0 DScale=0 PME-


I'll do an Xfree86 log if anyone is interested but the succint summary 
is thus:

(WW) RADEON(0): Direct rendering not yet supported on Radeon 9500/9700 
and newer  cards

Bill

Gang

Just put together a new AMD64 3200+ on a socket 754 mb (ASUS K8V SE).  I 
dropped a 9800 Pro AIW in alongside a Hauppage PVR-350.   I'll be 
playing with ivtv and mythtv on the PVR-350 while I wait for something I 
can test for you guys.  I remember when I picked up one of the first 
Radeon AIW's that it took a fair while for the Linux side to catch up 
with support.

I'll be running Mandrake 10 and playing with Gentoo for the first time 
on the Linux side.

Bill

Hi,

I have noticed that my new Radeon 9200 card has an
S-Video cable for TV output. However, my monitor's
vertical refresh rate is marked as 55-75 Hz, and I
understand that not being able to run at 50 Hz for PAL
is a Bad Thing...

Is there any way that the TV Output driver can switch
my monitor OFF, so that DVI output and TV output are
mutually exclusive? My monitor *does* support DPMS,
and so can be shut down programatically
(theoretically, anyway). This wouldn't be too bad a
solution; I basically would like to be able to watch
DVDs on my TV screen.

Any thoughts, please?
Thanks,
Chris



	
	
		
___________________________________________________________ALL-NEW Yahoo! Messenger - all new features - even more fun!  http://uk.messenger.yahoo.com

thanks, now it works, but:

        (WW) RADEON(0): Option "TVOutput" is not used

using cvs w/  tv output

@800x600 --> tvo_set: Invalid attributes in XV adaptor

There was a post several weeks ago about the ati.4.4.0 binaries for xorg 
6.7.0 being available soon. I don't see them on sourceforge. Are they 
available some place?

thanks...

_________________________________________________________________
Don’t just search. Find. Check out the new MSN Search! 
http://search.msn.click-url.com/go/onm00200636ave/direct/01/

>> sent to the card).
>
> I guess I should have specified, that I don't have any hardware newer
> than a Radeon 7200, so there's not much that I can do outside of
> dissasembly to help out.
>> From what I understand, it should be perfectly legal for me to
> dissasemble the driver, and answer questions that you guys have about
> its inner workings.  Especially since I didn't have to agree to any
> license in order to download it, and I don't have any reason to
> install it which would require me to agree to one.
>

Still a bit fishy - we better keep our noses clean.

I don't know about other developers, but I personally, don't want to 
know what is inside that driver - spoils the fun of writing one's own GL 
driver.

I do want to know how to access the hardware, but for that disassembly is 
not required.

                        best

                           Vladimir Dergachev

>
> -- 
> Curtis Magyar
> cu...@gm...
>

On Sat, 18 Sep 2004 13:52:09 -0400 (EDT), Vladimir Dergachev
<vo...@mi...> wrote:
> Actually, reverse engineering does not imply disassembly :))
> 
> In fact, you can join r300.sf.net which already has a demo program that
> uses R300/RV350 GPUs (paints textured rectangles for example) and there
> was no disassebly involved in any of this - just some deductive skills
> and knowledge of existing Xorg and DRI drivers.
> 
> There is still much to do - we don't know what many registers are for and
> it requires a lot of experimentation to find out their purpose.
> 
> If you have a working driver installed (say fglrx) you might enjoy playing
> with glxtest as well (it is able to capture streams of command packets
> sent to the card).

I guess I should have specified, that I don't have any hardware newer
than a Radeon 7200, so there's not much that I can do outside of
dissasembly to help out.
From what I understand, it should be perfectly legal for me to
dissasemble the driver, and answer questions that you guys have about
its inner workings.  Especially since I didn't have to agree to any
license in order to download it, and I don't have any reason to
install it which would require me to agree to one.


-- 
Curtis Magyar
cu...@gm...

On Sat, 18 Sep 2004, Curtis Magyar wrote:

> On Fri, 17 Sep 2004 21:51:32 -0400 (EDT), Vladimir Dergachev
> <vo...@mi...> wrote:
>> No GLX support. And no documentation available to any developers that I
>> know.
>
> Hello again.  I noticed your mail on the xorg list stating that it
> would be of some use to you to have people who were not under NDA who
> could possibly help reverse engineer drivers.  I'm curious how this
> would work, as I'm not under NDA, and I am fairly capable when it
> comes to groking other people's code.  However, assembly isn't my
> forte by any means.  How much direction would be required do you
> think?  As I have a licensed copy of IDA pro that a previous employer
> bought for me, and I also obviously have access to the GNU
> dissasembler tools.  Are you just looking for people to poke through
> the dissasembled code and say this register looks like its for this,
> and that one is for that?

Actually, reverse engineering does not imply disassembly :))

In fact, you can join r300.sf.net which already has a demo program that 
uses R300/RV350 GPUs (paints textured rectangles for example) and there 
was no disassebly involved in any of this - just some deductive skills
and knowledge of existing Xorg and DRI drivers.

There is still much to do - we don't know what many registers are for and 
it requires a lot of experimentation to find out their purpose.

If you have a working driver installed (say fglrx) you might enjoy playing 
with glxtest as well (it is able to capture streams of command packets 
sent to the card).

                         best

                             Vladimir Dergachev

>
> I can certainly give it a shot, if you can give me a goal to search
> for - I'll do my best to find it.  :)
>
> -- 
> Curtis Magyar
> cu...@gm...
>

On Fri, 17 Sep 2004 21:51:32 -0400 (EDT), Vladimir Dergachev
<vo...@mi...> wrote:
> No GLX support. And no documentation available to any developers that I
> know.

Hello again.  I noticed your mail on the xorg list stating that it
would be of some use to you to have people who were not under NDA who
could possibly help reverse engineer drivers.  I'm curious how this
would work, as I'm not under NDA, and I am fairly capable when it
comes to groking other people's code.  However, assembly isn't my
forte by any means.  How much direction would be required do you
think?  As I have a licensed copy of IDA pro that a previous employer
bought for me, and I also obviously have access to the GNU
dissasembler tools.  Are you just looking for people to poke through
the dissasembled code and say this register looks like its for this,
and that one is for that?

I can certainly give it a shot, if you can give me a goal to search
for - I'll do my best to find it.  :)

-- 
Curtis Magyar
cu...@gm...

On Sat, 18 Sep 2004, Dousak May wrote:

> Oh...Is there any possibility to make the old gatos (ibtk+gatos) work
> under X 4.3 so I wouldn'd have to search for another card?

Actually it stopped working starting with XFree86 4.0.0. The reason is 
that before it was possible to force X to think it has less framebuffer 
memory then is actually available and use the remaining piece for xatitv.

You might be able to hack the current driver (of any X version) to achieve 
the same result.

Also, newer X drivers know about overlay (which is used to display TV 
picture) and reset it occasionally - also something to be bypassed.

                          best

                            Vladimir Dergachev

>
> On Sat, 2004-09-18 at 03:57, Vladimir Dergachev wrote:
>
>> On Fri, 17 Sep 2004, Dousak May wrote:
>>
>>> So, years ago I used ati mach 64 + ati ISA tuner (AMC cable) without
>>> problems with the old gatos software (ibtk+xattitv). I tried to use that
>>> combination again, but on X 4.3. Gatos doesn't work anymore (xatitv just
>>> doesn't start), ati.2 + avview work neither. Xatitv just doesn't show
>>> up, but avview doesn't have any channels (neither v4l options- only
>>> "Auto" and "none")
>>> I tried compiling kernel with v4l support, then even I added bttv (of
>>> course it doesn't work) but nothing helps. Can someone tell what could
>>> be wrong from my logs?
>>>
>>> Here is what AVVIEW writes out:
>>
>>
>> Hi Dousak,
>>
>>     This is a known issue with ati.2 drivers. Apparently when I ported the
>> mach64 i2c bus access one (or two) of the modes used by "ISA tuners" were
>> screwed up. And these never worked with older mach64 cards (they were
>> reported working with Rage Pro and Rage 128 cards).
>>
>>     I have checked the code many times and just could not see what went
>> wrong. Perhaps, it is a timing issue.
>>
>>     If you are curious try to debug the issue, or, alternatively, you might
>> want to get a Rage128 card to plug it into - there are good chances it
>> would work. I don't know of anyone who tried this with Radeons, not sure
>> even whether they have the VESA port to connect the "ISA tuner" to.
>>
>>                        best
>>
>>                           Vladimir Dergachev
>>
>>
>
>
>
> BOFH excuse #305: IRQ-problems with the Un-Interruptible-Power-Supply
>

Oh...Is there any possibility to make the old gatos (ibtk+gatos) work
under X 4.3 so I wouldn'd have to search for another card?

On Sat, 2004-09-18 at 03:57, Vladimir Dergachev wrote:

> On Fri, 17 Sep 2004, Dousak May wrote:
> 
> > So, years ago I used ati mach 64 + ati ISA tuner (AMC cable) without
> > problems with the old gatos software (ibtk+xattitv). I tried to use that
> > combination again, but on X 4.3. Gatos doesn't work anymore (xatitv just
> > doesn't start), ati.2 + avview work neither. Xatitv just doesn't show
> > up, but avview doesn't have any channels (neither v4l options- only
> > "Auto" and "none")
> > I tried compiling kernel with v4l support, then even I added bttv (of
> > course it doesn't work) but nothing helps. Can someone tell what could
> > be wrong from my logs?
> >
> > Here is what AVVIEW writes out:
> 
> 
> Hi Dousak,
> 
>     This is a known issue with ati.2 drivers. Apparently when I ported the 
> mach64 i2c bus access one (or two) of the modes used by "ISA tuners" were 
> screwed up. And these never worked with older mach64 cards (they were 
> reported working with Rage Pro and Rage 128 cards).
> 
>     I have checked the code many times and just could not see what went 
> wrong. Perhaps, it is a timing issue.
> 
>     If you are curious try to debug the issue, or, alternatively, you might 
> want to get a Rage128 card to plug it into - there are good chances it 
> would work. I don't know of anyone who tried this with Radeons, not sure 
> even whether they have the VESA port to connect the "ISA tuner" to.
> 
>                        best
> 
>                           Vladimir Dergachev
> 
> 



BOFH excuse #305: IRQ-problems with the Un-Interruptible-Power-Supply

On Fri, 17 Sep 2004, ernstern wrote:

> compiled ati.2 against the xfree86 4.3.0 tree. using kernel 2.6.2 & 2.6.8.
> starting the xserver doesn't give any errors, but screen is black?
>
> any suggestions?

No idea - what kind of monitor do you have ? Try adding
Option "compositesync" "no"

to the device section.

                          best

                            Vladimir Dergachev

>
>
>
>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
> Project Admins to receive an Apple iPod Mini FREE for your judgement on
> who ports your project to Linux PPC the best. Sponsored by IBM.
> Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>