gatos-devel Mailing List for GATOS

Hi,
Patrick Gleichmann wrote:
> Hi,
> I have two ATI Radeon 7000 VE cards in my pc.
> When using the patch from "http://megahurts.dk/rune/tv_output.html"  
> with the 6.8.99 CVS Xorg, then a 'Option "TVOutput "PAL"' only  enables 
> the TV-output for the first (primary) card, but not the  second one.
> 
> Any hint?
> Same result - of course - with tvo_set. Just as a test, I've added a  
> 'printf("%d", ScreenCount(dy))' to tvo_set.c right and it gives '1'.
> 

Does they both work without xinerama?

Could you tell me your setup: what is connected to which card.
And please attach Xorg log and config files.


Rune Petersen

Hi,
I have two ATI Radeon 7000 VE cards in my pc.
When using the patch from "http://megahurts.dk/rune/tv_output.html"  
with the 6.8.99 CVS Xorg, then a 'Option "TVOutput "PAL"' only  
enables the TV-output for the first (primary) card, but not the  
second one.

Any hint?
Same result - of course - with tvo_set. Just as a test, I've added a  
'printf("%d", ScreenCount(dy))' to tvo_set.c right and it gives '1'.


Thanks,
Patrick

Hi Rune

I happen to have run into some serious problem and its driving me nuts.
Here goes:
I have a PCI card, an old peice which I'd like to put to use with my
desktop, to watch some movies. (I'm sure you might have guessed by
now, where this one is headed )

OK. putting it and using the ati driver, gives me proper stuff right
away on a monitor. However, I try to attach my new TV ,  all I get is
garbeled screen.

This, and a lot of xorg.conf settings, and many prayers + forums
scanning activity later I've  decided to ask you as to what could be
wrong with the setup.

BTW, I use gentoo 2.6.10 , xorg 6.8.2 .... without any patches (binary
ones that came with the distro CDs)

Now after almost giving up the hope, yesterday I popped in a knoppix
3.6 CD, and started the system ... still garbeled screen in X. After a
careful inspection of the boot options in the CD , I pass vsync=3D60
hsync=3D30 to grub. Guess what the display works (yea it did some
probing before starting, and although I'm not certain about this, it
might be using vesa driver. (I don't mind that as long as I can watch
movies clearly).

Well the other diff. was that knoppix 3.6 uses XFree86 - 4.x.x.
So ..... I am in serious soup and I'll be grateful if you could point
me to a way to use TV out o my card on my gentoo system.

cheerz
+coolant

PS: Xorg --configure idetified my card with string "3D Rage LT Pro"
(If that gives you any better picture)., and it has 8mb ram.

there was a misunderstanding here, i'm using xorg cvs
[...]
X Window System Version 6.8.99.900 (6.9.0 RC 0)
[...]
I recompiled the whole thing on my machine
i don't know if ati4.4 is still up to date...

.

besides tuner still not working, capture is not stable
tryng to record i got kernel Oops!

on console

Message from syslogd@akira at Tue Oct 18 23:48:32 2005 ...
akira kernel: Oops: 0000 [1] PREEMPT

Message from syslogd@akira at Tue Oct 18 23:48:32 2005 ...
akira kernel: CR2: 0000000000000000

from the syslog

[...]
 kernel: Processing "VIDEO_STREAM_ACTIVE"="1" int_value=1
 kernel: radeon_get_window_parameters: width=720 height=288
 kernel: radeon_get_window_parameters: width=720 height=288
 kernel: km: GUI_DMA queue (size=10) is full first=1 last=0 flag=0x80000001
 last message repeated 214 times
 kernel: radeon_get_window_parameters: width=720 height=288
 kernel: km: GUI_DMA queue (size=10) is full first=1 last=0 flag=0x80000001
 last message repeated 9 times
 kernel: Unable to handle kernel NULL pointer dereference at 
0000000000000000 RIP:
 kernel: <ffffffff88087491>{:km_api_drv:km_data_destroy_kdufpd+17}
 kernel: PGD 1a6bd067 PUD b085067 PMD 0
 kernel: Oops: 0000 [1] PREEMPT
 kernel: CPU 0
 kernel: Modules linked in: km_drv km_api_drv videodev snd_pcm_oss 
snd_mixer_oss md5 ipv6 snd_via82xx gameport snd_mpu401_uart w83627hf 
i2c_sensor i2c_isa
 kernel: Pid: 9910, comm: start_avview Not tainted 2.6.13.2
 kernel: RIP: 0010:[<ffffffff88087491>] 
<ffffffff88087491>{:km_api_drv:km_data_destroy_kdufpd+17}
 kernel: RSP: 0018:ffff810027365ee8  EFLAGS: 00010296
 kernel: RAX: ffff8100319a5000 RBX: ffffffff880969a0 RCX: 0000000000000000
 kernel: RDX: 0000000000000000 RSI: ffff81003b2c7280 RDI: 0000000000000000
 kernel: RBP: 0000000000000000 R08: ffff810027364000 R09: ffff810012080a40
 kernel: R10: 00000000ffffffff R11: 0000000000000202 R12: ffff81003ff20080
 kernel: R13: ffff81003ee32db8 R14: ffff81003ee0e128 R15: 00000000009e8460
 kernel: FS:  00002aaaadf82a40(0000) GS:ffffffff80646800(0000) 
knlGS:0000000000000000
[...]



Matthias Transier wrote:

>Hi,
>
>I am now trying to capture one of these crappy pictures my card is
>producing. But I did not manage to get km running. nebokazer and I
>have nearly the same setup, he is able to capture and I am not :(.
>When starting avview, I am getting this error message:
>
>WARNING !   Radeon memory controller is misconfigured, disabling capture
>WARNING !   upgrade your Xserver and DRM driver
>
>The differences between nebokazer's and my kernel messages are:
>irq 201 for him and irq 20 for me.
>reg_aperture: 0xffffc20000620000 vs. 0xffffc200006a0000
>Further, avview reports different ports on startup: 65 vs. 61
>
>I use kernel drm ([drm] Initialized drm 1.0.0 20040925) and Xorg 6.8.2
>with ati.4.4.0/r200_branch from cvs.
>
>Anyone any idea?
>
>Cheers,
>Matthias
>
>On 9/19/05, Matthias Transier <mat...@gm...> wrote:
>  
>
>>Hi all,
>>
>>I am having problems watching TV (PAL) with my 9800SE AIW (on amd64).
>>The initialization seems to work fine (see below), but I am not
>>getting a usable picture with AVview. The only thing that appears on
>>the screen are dark shadows (which I can turn bright but they remain
>>just shadows) of a TV picture. And the picture seems shifted.
>>Did anyone experience similar problems? Can anyone help me?
>>
>>Thanks,
>>Matthias
>>    
>>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by:
>Power Architecture Resource Center: Free content, downloads, discussions,
>and more. http://solutions.newsforge.com/ibmarch.tmpl
>_______________________________________________
>Gatos-devel mailing list
>Gat...@li...
>https://lists.sourceforge.net/lists/listinfo/gatos-devel
>
>  
>

Hi,

I am now trying to capture one of these crappy pictures my card is
producing. But I did not manage to get km running. nebokazer and I
have nearly the same setup, he is able to capture and I am not :(.
When starting avview, I am getting this error message:

WARNING !   Radeon memory controller is misconfigured, disabling capture
WARNING !   upgrade your Xserver and DRM driver

The differences between nebokazer's and my kernel messages are:
irq 201 for him and irq 20 for me.
reg_aperture: 0xffffc20000620000 vs. 0xffffc200006a0000
Further, avview reports different ports on startup: 65 vs. 61

I use kernel drm ([drm] Initialized drm 1.0.0 20040925) and Xorg 6.8.2
with ati.4.4.0/r200_branch from cvs.

Anyone any idea?

Cheers,
Matthias

On 9/19/05, Matthias Transier <mat...@gm...> wrote:
> Hi all,
>
> I am having problems watching TV (PAL) with my 9800SE AIW (on amd64).
> The initialization seems to work fine (see below), but I am not
> getting a usable picture with AVview. The only thing that appears on
> the screen are dark shadows (which I can turn bright but they remain
> just shadows) of a TV picture. And the picture seems shifted.
> Did anyone experience similar problems? Can anyone help me?
>
> Thanks,
> Matthias

Given my earlier problems with my Mac Mini and trying to get TV out 
working off of one of the CVS snapshots, I've been trying to today with 
Xorg CVS HEAD instead, to see if I'd have any more luck with that. I 
managed to update Rune's tv_out patchset to work with the Xorg CVS 
(patch available at 
http://tevp.net/projects/gatos-tv_out-20051015.patch) but have now got 
an Xserver that crashes somewhere in the middle of theater_out. I 
switched on RADEON_DEBUG and got the following messages at the end of my 
Xorg log.

(**) RADEON(0): RADEONScreenInit 98000000 0
(**) RADEON(0): Allocating VIP interface(**) RADEON(0): WaitForIdle 
(entering): 64 entries, stat=0x00000140
(**) RADEON(0): TVOutput opt = PAL
(II) Loading sub module "theater_out"
(II) LoadModule: "theater_out"
(II) Loading /opt/xorg-6.8.99-dev/lib/modules/drivers/theater_out_drv.so
(II) Module theater_out: vendor="X.Org Foundation"
         compiled for 4.3.99.902, module version = 1.0.0
         ABI class: X.Org Video Driver, version 0.8
(II) RADEON(0): Embedded Rage Theater detected
(**) RADEON(0): RADEONSave
(**) RADEON(0): RADEONSaveMode(0x1025e178)
(**) RADEON(0): Read: 0x0000000c 0x00030059 0x00000000
(**) RADEON(0): Read: rd=12, fd=89, pd=3
(**) RADEON(0): Read VCLK_ECP_CNTL = 00000003
(**) RADEON(0): RADEONSaveMode returns 0x1025e178

This was from running "Xorg" from an ssh'ed command line, and gave me a 
server that can't be killed with Ctrl-C. Opening another ssh connection, 
and doing a kill -9 appears to be the only thing that'll kill it. Tried 
running it under gdb, and still no luck (program just plain dies, no 
stack trace, no core, nothing). If anyone here knows any other 
suggestions for how to debug this, I'd love to hear them.

Thanks,

Tom Parker

Hasse Hagen Johansen wrote:
>>>>>>"Rune" == Rune Petersen <ru...@me...> writes:
> 
> 
>     Rune> Hasse Hagen Johansen wrote:
>     >> Hi I'm in the middle of playing around with mythtv. I hva trie
>     >> to get the tv-out working on my old Radeon Vivo card, but have
>     >> had no success.
>     Rune> VIVO as in TV-in & TV-out? The TV-in modules interferes with
>     Rune> the TV-out module. Try removing these modules:
> 
>     Rune> 	theatre200_drv.so theatre_detect_drv.so theatre_drv.so
> 
>     Rune> from /usr/X11R6/lib/modules/multimedia
> 
> Hi Rune and thanks
> 
> By ViVo I meant video-in/video-out(I also think the ati product was
> called ViVo)
> 
> It helped a lot to remove the tv-input modules, so now it works :-)
> The quality is not so good though. I probably have to get a tv with
> svhs input. I heard that should be better?

It is and RGB is even better, but don't expect a miracle.

It's basicly a question of available bandwidth for the signals. (number 
of wires used).


Rune Petersen

Rune Petersen wrote:
> Alon Bar-Lev wrote:

>> How to report it?
>> What is the problem that I should report (except of the
>> message...)
> This should explain it:
> 
> http://xorg.freedesktop.org/wiki/#head-c3b9a4709623ada3f539d241da32a58ffffff7d1 

I know how to report... I just need to know what to 
report... I don't want to write a completely silly post...

>>> P.S.
>>> Can you get tv-out to work with ATI's fglrx driver?
>>>
> This page lists the 4 version of the ATI driver and what version of 
> Xfree & Xorg are supported:
> 
> https://support.ati.com/ics/support/default.asp?deptID=894&task=knowledge&folderID=27 

OK... does not work with 6.8.99... I thought you know a 
pre-release version.

Attached is a log file with RADEON_DEBUG... Maybe it will 
help...

Best Regards,
Alon Bar-Lev.

>>>>> "Rune" == Rune Petersen <ru...@me...> writes:

    Rune> Hasse Hagen Johansen wrote:
    >> Hi I'm in the middle of playing around with mythtv. I hva trie
    >> to get the tv-out working on my old Radeon Vivo card, but have
    >> had no success.
    Rune> VIVO as in TV-in & TV-out? The TV-in modules interferes with
    Rune> the TV-out module. Try removing these modules:

    Rune> 	theatre200_drv.so theatre_detect_drv.so theatre_drv.so

    Rune> from /usr/X11R6/lib/modules/multimedia

Hi Rune and thanks

By ViVo I meant video-in/video-out(I also think the ati product was
called ViVo)

It helped a lot to remove the tv-input modules, so now it works :-)
The quality is not so good though. I probably have to get a tv with
svhs input. I heard that should be better?

Thanks for the help

/Hasse

Alon Bar-Lev wrote:
> Hello,
> 
> Rune Petersen wrote:
> 
>  > Can you log in (SSH) and kill X? and if you can, do
>  > you get the keyboard back?
> 
> Hmm... I don't have another computer... but the laptop power
> buttons are not working as well... (different interface,
> different daemon), I will try to have another computer.
if you are able to get a second PC to connect to the laptop also try 
starting Xorg via the ssh connection like this:
	gdb Xorg

then type run and <enter>.
> 
> 
>>
>>  > Requesting insufficient memory window!: start: 0x3c00 end: 0x3cff 
>> size 0xc0120100
>>  > (EE) Cannot find a replacement memory range
>>
>> This may be unrelated, but it looks bad, please report it to Xorg 
>> (bugzilla or mailing-list).
This should explain it:

http://xorg.freedesktop.org/wiki/#head-c3b9a4709623ada3f539d241da32a58ffffff7d1

> 
> How to report it?
> What is the problem that I should report (except of the
> message...)
> 
>>
>> A mobile... you may have a hard time getting tv-out working. there 
>> seams to be a problem with using LCD-monitor and getting tv-out to work.
>> If you are willing to do the "leg" work I'll help all I can
> 
> 
> Thank you for your patient!!!
> I want my TV out working... (And my modem... :( )
> I will be glad to receive any help.
> 
>>
>> the main problem is that timing used are from the LCD monitor not the 
>> modeline...
>> something like this:
>> Option     "IgnoreEDID" "true
>> Option     "MonitorLayout"    "NONE,AUTO"
>>
> Tried.
> Now the modeline is OK now!!!!
> But it still hung...
> Attached is the new log.
> 
>>
>>
>> P.S.
>> Can you get tv-out to work with ATI's fglrx driver?
>>
> 
> Oh... I thought of that after I migrated X... But ATI
> drivers do not work with the 6.8.99 version of X... or I
> don't know how to run them... Which driver version should work?

This page lists the 4 version of the ATI driver and what version of 
Xfree & Xorg are supported:

https://support.ati.com/ics/support/default.asp?deptID=894&task=knowledge&folderID=27


Rune Petersen

Hasse Hagen Johansen wrote:
> Hi
> 
> I'm in the middle of playing around with mythtv. I hva trie to get the
> tv-out working on my old Radeon Vivo card, but have had no success.
VIVO as in TV-in & TV-out? The TV-in modules interferes with the TV-out 
module. Try removing these modules:

	theatre200_drv.so
	theatre_detect_drv.so
	theatre_drv.so

from /usr/X11R6/lib/modules/multimedia

> 
> lspci output is: VGA compatible controller: ATI Technologies Inc Radeon R100 QD [Radeon 7200]
> 
> I got the latest snapshot from Xorg and patched it with the Tv-out
> pathces from http://megahurts.dk/rune/tv_output.html and compiled it,
> and then put Option "TVOutput" "PAL" in the xorg.conf file. My normal
> monitor turns off and the tv switches from the console to black
> screen(earlier it has also been white)
> 
> I know that my TV doesn't accept a S-VHS signal. I have an euro scart
> which convert it to composite(I believe that is how it works)
Nope SCART is wired for composite, S-VIDEO/S-VHS, and RGB; no conversion 
needed.
> 
> My console is displyed fine, but when starting Xorg it just turns
> black
> 
> Any suggestions would be welcomed, also if it not worth the trouble to
> get it working(the quality on the console isn't that good...but that
> could be because the tv doesn't accept S-VHS)
Then your TV will support composite. If you try S-VHS on that is not 
supported you will get a bad image, but still an image.

If you still can get it to work, please send me your Xorg config an logfile.


Rune Petersen

Clarity on these issues is emerging. When I found that I could not even 
compile my own Motif C code in this partition for want of many missing 
include files, I discovered that even though the openmotif-devel package 
was listed as installed in the rpm database, there were tons of missing 
files, among them all the include files. So after reinstalling the 
openmotif and openmotif-devel packages, motv compiled fine in the xawtv 
suite with just a --enable-motif ./configure option.

I agree, motv is a nice looking and easy to use viewing tool. I get good 
320x240 jpeg screen captures through your module with the halfwidth=1 
option, about all that is realistic with my ancient 4MB card.

On the matter of the freezing and crashing I saw with xawtv and this 
driver module, I checked with the glxinfo tool and direct rendering is 
indeed disabled. However, I'm starting to think that the OS installation 
may have been at fault. If half the files in the openmotif rpm were 
missing, what else is corrupted? I think either the initial install in 
this test partition was messed up, or perhaps when I replaced the 
original X.org by compiling from the CVS (in an attempt to get ati.2 
drivers that would drive an AIW classic) that things got messed up. I do 
see other problems with the X screensaver as well.

I played around with various windows and other applications for about a 
half an hour while viewing with your v4l driver and motv, and no crashes 
this time. So I'm going to reformat that partition and install the 
latest Scientific Linux version, recompile your module, and really give 
it some reliability testing. If all goes well this will be my ticket for 
(finally) getting to upgrade to a 2.6 kernel for my main partition.

Thanks for your help-
John

Eric Sellers wrote:
> 
> Still not sure why you are having this problem...
> Do you have direct rendering enabled?
> (it should not be enabled)
> 
> glxinfo | grep "direct rendering"
> 
> direct rendering: No
> 
 > My mistake, yes it uses openmotif.
 >
 > Here is how it was configured on my gentoo system.
 > ./configure --prefix=/usr --host=i686-pc-linux-gnu 
--mandir=/usr/share/man
 > --infodir=/usr/share/info --datadir=/usr/share --sysconfdir=/etc
 > --localstatedir=/var/lib --build=i686-pc-linux-gnu --with-x
 > --enable-xfree-ext --enable-xvideo --enable-dv --enable-mmx 
--enable-motif
 > --enable-quicktime --enable-alsa --enable-lirc --enable-gl --enable-zvbi
 > --disable-aa
 >
 > Eric

-- 
--------------------------------------------------------
John Fattaruso                     voice: 214-567-3654
High Speed Communications          fax: 214-480-3555
Texas Instruments                  e-mail: jo...@ti...
PO Box 660199   MS 8675            Office: FL E-3189
Dallas, TX       75266             TI PC Drop: PFLE

No trees were destroyed in the sending of this message;
however, a significant number of electrons were terribly
inconvenienced.
--------------------------------------------------------

Hello,

Rune Petersen wrote:

  > Can you log in (SSH) and kill X? and if you can, do
  > you get the keyboard back?

Hmm... I don't have another computer... but the laptop power
buttons are not working as well... (different interface,
different daemon), I will try to have another computer.


>
>  > Requesting insufficient memory window!: start: 0x3c00 end: 0x3cff 
> size 0xc0120100
>  > (EE) Cannot find a replacement memory range
> 
> This may be unrelated, but it looks bad, please report it to Xorg 
> (bugzilla or mailing-list).

How to report it?
What is the problem that I should report (except of the
message...)

> 
> A mobile... you may have a hard time getting tv-out working. there seams 
> to be a problem with using LCD-monitor and getting tv-out to work.
> If you are willing to do the "leg" work I'll help all I can

Thank you for your patient!!!
I want my TV out working... (And my modem... :( )
I will be glad to receive any help.

> 
> the main problem is that timing used are from the LCD monitor not the 
> modeline...
> something like this:
> Option     "IgnoreEDID" "true
> Option     "MonitorLayout"    "NONE,AUTO"
> 
Tried.
Now the modeline is OK now!!!!
But it still hung...
Attached is the new log.

> 
> 
> P.S.
> Can you get tv-out to work with ATI's fglrx driver?
> 

Oh... I thought of that after I migrated X... But ATI
drivers do not work with the 6.8.99 version of X... or I
don't know how to run them... Which driver version should work?

Hmmm... The message it too large again... Trying again.

Best Regards,
Alon Bar-Lev

Hello,

Rune Petersen wrote:

 > Can you log in (SSH) and kill X? and if you can, do
 > you get the keyboard back?

Hmm... I don't have another computer... but the laptop power 
buttons are not working as well... (different interface, 
different daemon), I will try to have another computer.


>
>  > Requesting insufficient memory window!: start: 0x3c00 end: 0x3cff 
> size 0xc0120100
>  > (EE) Cannot find a replacement memory range
> 
> This may be unrelated, but it looks bad, please report it to Xorg 
> (bugzilla or mailing-list).

How to report it?
What is the problem that I should report (except of the 
message...)

> 
> A mobile... you may have a hard time getting tv-out working. there seams 
> to be a problem with using LCD-monitor and getting tv-out to work.
> If you are willing to do the "leg" work I'll help all I can

Thank you for your patient!!!
I want my TV out working... (And my modem... :( )
I will be glad to receive any help.

> 
> the main problem is that timing used are from the LCD monitor not the 
> modeline...
> something like this:
> Option     "IgnoreEDID" "true
> Option     "MonitorLayout"    "NONE,AUTO"
> 
Tried.
Now the modeline is OK now!!!!
But it still hung...
Attached is the new log.

> 
> 
> P.S.
> Can you get tv-out to work with ATI's fglrx driver?
> 

Oh... I thought of that after I migrated X... But ATI 
drivers do not work with the 6.8.99 version of X... or I 
don't know how to run them... Which driver version should work?

Best Regards,
Alon Bar-Lev

Alon Bar-Lev wrote:
> Rune Petersen wrote:
> 
>> Alon Bar-Lev wrote:
>>
>>> Rune Petersen wrote:
>>>
>>>>
>>>> TV-in is merged, not TV-out... main reason for it not being merged 
>>>> is that the code is GPL..
>>>>
>>>
>>> Hello Rune,
>>>
>>> Thank you for your help!
>>>
>>> After migrating to X Window System Version 6.99.99.900 (7.0.0 RC 0) I 
>>> tried to apply the patch of TV Output from 
>>> http://megahurts.dk/rune/tv_output.html, I use modular X so I had to 
>>> change the Makefile.am (The Imakefile was invalid in the patch).
>>
>>
>> Please the monolithic source, it should be more stable.
>>
> 
> OK...
> Back to X Window System Version 6.8.99.15...
> [Took me a whole day to recompile my system to 6.9.99... And
> then back a whole day to revert... :-(]
> 
> Now... After I am using your modeline the computer just
> hangs... It is just like I specify AGPMode > 1... The screen
> is black and the keyboard is dead.
Can you log in (SSH) and kill X? and if you can, do you get the keyboard 
back?
> 
> Just to remind... I have:
> 0000:01:00.0 VGA compatible controller: ATI Technologies Inc
> RV350 [Mobility Radeon 9600 M10]
> 
> Attached is Xorg.0.log and my tv.conf.
> 
> Is there any way to make it produce more trace so it will
> help in solving this issue?
set RADEON_DEBUG to 1 in radeon.h

 > Requesting insufficient memory window!: start: 0x3000 end: 0x30ff 
size 0xc0120100
 > Requesting insufficient memory window!: start: 0x3400 end: 0x34ff 
size 0xc0120100
 > Requesting insufficient memory window!: start: 0x3800 end: 0x38ff 
size 0xc0120100
 > Requesting insufficient memory window!: start: 0x3c00 end: 0x3cff 
size 0xc0120100
 > (EE) Cannot find a replacement memory range

This may be unrelated, but it looks bad, please report it to Xorg 
(bugzilla or mailing-list).


 > II) RADEON(0): Primary:
 >  Monitor   -- LVDS
 >  Connector -- VGA
 >  DAC Type  -- Primary
 >  TMDS Type -- NONE
 >  DDC Type  -- VGA_DDC

A mobile... you may have a hard time getting tv-out working. there seams 
to be a problem with using LCD-monitor and getting tv-out to work.
If you are willing to do the "leg" work I'll help all I can

the main problem is that timing used are from the LCD monitor not the 
modeline...
something like this:
Option     "IgnoreEDID" "true
Option     "MonitorLayout"	"NONE,AUTO"


Rune Petersen


P.S.
Can you get tv-out to work with ATI's fglrx driver?

well, probably the man page is a little outdated
cause your version is the one actually working
at least on my system

i wonder if it would from a 32 bit client to a 64 bit server
but i'm going too far ;)

well, just for the records
it works even from a 64 to a 32 box :)
bye

Rune Petersen wrote:
> Alon Bar-Lev wrote:
> 
>> Rune Petersen wrote:
>>
>>>
>>> TV-in is merged, not TV-out... main reason for it not being merged is 
>>> that the code is GPL..
>>>
>>
>> Hello Rune,
>>
>> Thank you for your help!
>>
>> After migrating to X Window System Version 6.99.99.900 (7.0.0 RC 0) I 
>> tried to apply the patch of TV Output from 
>> http://megahurts.dk/rune/tv_output.html, I use modular X so I had to 
>> change the Makefile.am (The Imakefile was invalid in the patch).
> 
> Please the monolithic source, it should be more stable.
> 

OK...
Back to X Window System Version 6.8.99.15...
[Took me a whole day to recompile my system to 6.9.99... And
then back a whole day to revert... :-(]

Now... After I am using your modeline the computer just
hangs... It is just like I specify AGPMode > 1... The screen
is black and the keyboard is dead.

Just to remind... I have:
0000:01:00.0 VGA compatible controller: ATI Technologies Inc
RV350 [Mobility Radeon 9600 M10]

Attached is Xorg.0.log and my tv.conf.

Is there any way to make it produce more trace so it will
help in solving this issue?

Hmm... Forgot the attachemnts...
Hmm... Now I hope they are not too large...

Best Regards,
Alon Bar-Lev.

Rune Petersen wrote:
> Alon Bar-Lev wrote:
> 
>> Rune Petersen wrote:
>>
>>>
>>> TV-in is merged, not TV-out... main reason for it not being merged is 
>>> that the code is GPL..
>>>
>>
>> Hello Rune,
>>
>> Thank you for your help!
>>
>> After migrating to X Window System Version 6.99.99.900 (7.0.0 RC 0) I 
>> tried to apply the patch of TV Output from 
>> http://megahurts.dk/rune/tv_output.html, I use modular X so I had to 
>> change the Makefile.am (The Imakefile was invalid in the patch).
> 
> Please the monolithic source, it should be more stable.
> 

OK...
Back to X Window System Version 6.8.99.15...
[Took me a whole day to recompile my system to 6.9.99... And
then back a whole day to revert... :-(]

Now... After I am using your modeline the computer just
hangs... It is just like I specify AGPMode > 1... The screen
is black and the keyboard is dead.

Just to remind... I have:
0000:01:00.0 VGA compatible controller: ATI Technologies Inc
RV350 [Mobility Radeon 9600 M10]

Attached is Xorg.0.log and my tv.conf.

Is there any way to make it produce more trace so it will
help in solving this issue?

Hmm... forgot the attachemnts...

Best Regards,
Alon Bar-Lev.

Rune Petersen wrote:
> Alon Bar-Lev wrote:
> 
>> Rune Petersen wrote:
>>
>>>
>>> TV-in is merged, not TV-out... main reason for it not being merged is 
>>> that the code is GPL..
>>>
>>
>> Hello Rune,
>>
>> Thank you for your help!
>>
>> After migrating to X Window System Version 6.99.99.900 (7.0.0 RC 0) I 
>> tried to apply the patch of TV Output from 
>> http://megahurts.dk/rune/tv_output.html, I use modular X so I had to 
>> change the Makefile.am (The Imakefile was invalid in the patch).
> 
> Please the monolithic source, it should be more stable.
> 

OK...
Back to X Window System Version 6.8.99.15...
[Took me a whole day to recompile my system to 6.9.99... And 
then back a whole day to revert... :-(]

Now... After I am using your modeline the computer just 
hangs... It is just like I specify AGPMode > 1... The screen 
is black and the keyboard is dead.

Just to remind... I have:
0000:01:00.0 VGA compatible controller: ATI Technologies Inc 
RV350 [Mobility Radeon 9600 M10]

Attached is Xorg.0.log and my tv.conf.

Is there any way to make it produce more trace so it will 
help in solving this issue?

Best Regards,
Alon Bar-Lev.

On Sun, 9 Oct 2005, nebokazer wrote:

> i wonder if the fix is right, instead of using int32
> he enlarged the buffer to use int64 like this:
>
>
> version 1.14, Sat Jan 3 18:21:19 2004 UTC 	version 1.15, Wed Feb 9 
> 16:53:34 2005 UTC
> *Line 327* 
> 	*Line 327* 
>        return TCL_ERROR; 	         return TCL_ERROR;
>        } 	         }
> Tk_PhotoGetImage(image, &pib); 	 Tk_PhotoGetImage(image, &pib);
> data=alloca(pib.width*pib.height*4+8); 
> data=alloca((pib.width*pib.height+2)*sizeof(long));
> data[0]=pib.width; 	 data[0]=pib.width;
> data[1]=pib.height; 	 data[1]=pib.height;
> for(j=0;j<pib.height;j++){ 	 for(j=0;j<pib.height;j++){
>
>
> ------------------------------------------------------------------------
>
> but later uses 32 in the format field:
>
> property=Tk_InternAtom(tkwin, "_NET_WM_ICON");
> type=Tk_InternAtom(tkwin, "CARDINAL");
> XSync(d, False);
> XChangeProperty(d, win2, property, type, 32, PropModeReplace, (unsigned char 
> *)data, pib.width*pib.height+2);
>
> i never got too far with xlib but this does not look right

I admit it looks strange, but XChangeProperty manual page 
<http://www.x.org/X11R6.8.2/doc/XChangeProperty.3.html> says:

"If the specified format is 32, the property data must be a long array."

I figured that in the worst case using long instead of int just allocates 
twice the memory it actually needs and causes no harm.

If somebody provides proof that the above quoted manual page is in error 
and it should say uint32 instead of long, I will of course make the 
change.

-- 
Antti Ajanki

now i can see the change Annji made, probably i did something wrong earlier

i wonder if the fix is right, instead of using int32
he enlarged the buffer to use int64 like this:


version 1.14, Sat Jan 3 18:21:19 2004 UTC 	version 1.15, Wed Feb 9 
16:53:34 2005 UTC
*Line 327* 

	
*Line 327* 

         return TCL_ERROR; 	         return TCL_ERROR;
         } 	         }
 Tk_PhotoGetImage(image, &pib); 	 Tk_PhotoGetImage(image, &pib);
 data=alloca(pib.width*pib.height*4+8); 
 data=alloca((pib.width*pib.height+2)*sizeof(long));
 data[0]=pib.width; 	 data[0]=pib.width;
 data[1]=pib.height; 	 data[1]=pib.height;
 for(j=0;j<pib.height;j++){ 	 for(j=0;j<pib.height;j++){


------------------------------------------------------------------------

but later uses 32 in the format field:

property=Tk_InternAtom(tkwin, "_NET_WM_ICON");
type=Tk_InternAtom(tkwin, "CARDINAL");
XSync(d, False);
XChangeProperty(d, win2, property, type, 32, PropModeReplace, (unsigned 
char *)data, pib.width*pib.height+2);

i never got too far with xlib but this does not look right

besides from the WM standard in the freedesktop site i found:


      _NET_WM_ICON

_NET_WM_ICON CARDINAL[][2+n]/32

This is an array of possible icons for the client. This specification 
does not stipulate what size these icons should be, but individual 
desktop environments or toolkits may do so. The Window Manager MAY scale 
any of these icons to an appropriate size.

This is an array of 32bit packed CARDINAL ARGB with high byte being A, 
low byte being B. The first two cardinals are width, height. Data is in 
rows, left to right and top to bottom.

so i think he should simply have changed long to int instead...
unless i had a beer too many tonight of course ;-P

regards

Vladimir Dergachev wrote:

>
> Yep, this piece of code is my fault - I was putting long instead of 
> using more generic uint32.
>
> If Annji fixed it the fix might not appear the same day in ViewCVS.
>
>                    best
>
>                       Vladimir Dergachev
>
> On Wed, 5 Oct 2005, nebokazer wrote:
>
>> mmm,  xmisc.c is exactly the same in cvs and avview-0.80.6
>>
>> I got cvs with
>>
>> cvs -d :pserver:ano...@cv...:2401/cvsroot/gatos co 
>> avview
>>
>> and this is the source shown in the gatos home page, right?
>>
>>
>> anyway, the piece of code bombing out is this (line# on the left)
>>
>>  333  for(j=0;j<pib.height;j++){
>>  334          a=&(pib.pixelPtr[j*pib.pitch]);
>>  335          r=&(a[pib.offset[0]]);
>>  336          g=&(a[pib.offset[1]]);
>>  337          b=&(a[pib.offset[2]]);
>>  338          a=&(a[pib.offset[3]]);
>>  339          p=&(data[2+j*pib.width]);
>>  340          for(i=0;i<pib.width;i++){
>>  341                  *p=(*a<<24)|(*r<<16)|(*g<<8)|*b;
>>  342                  p++;
>>  343                  a+=pib.pixelSize;
>>  344                  r+=pib.pixelSize;
>>  345                  g+=pib.pixelSize;
>>  346                  b+=pib.pixelSize;
>>  347                  }
>>  348          }
>>
>> and the killer line is #341,
>> when p is computed on line 339 and incremented in line 342, being 
>> *data* a pointer to long,
>> it goes double speed and eventually fills and overflows *data* buffer 
>> (note that *data* buffer size
>> is computed as: pib.width*pib.height*4+8  and cannot take 
>> width*height longs) that is
>> created using alloca, so it is just under pib on the stack and when j 
>> reaches about a half of
>> the height p is already over &pib
>>
>> in a gdb run i got
>>
>> (gdb) p data
>> $7 = (long unsigned int *) 0x7fffff937ab0
>> (gdb) p &pib
>> $8 = (Tk_PhotoImageBlock *) 0x7fffff937f30
>> (gdb) p pib
>>
>> the difference is 1152
>>
>> $12 = {
>> pixelPtr = 0x626030 "\0052\t204\002?.. meaningless stuff..."...,
>> width = 16, height = 16, pitch = 64, pixelSize = 4, offset = {0, 1, 
>> 2, 3}}
>>
>> data buffer size is 16*16*4+8= 1032 bytes, apparently the size is 
>> round up to 1152
>>
>> but when j reaches 8 p is:    data +  8*(2+8*16) = data + 1040
>> just a few p++ (8 bytes) cycles away from crash
>> being pixelPtr the first field in the of pib is overwritten first
>>
>> 1040+8*15 = 1152+8
>>
>> Program received signal SIGSEGV, Segmentation fault.
>> xmisc_seticon (client_data=0x7fffff937f40, interp=0x7, argc=-1, 
>> argv=0xffffffffffd0f902) at xmisc.c:341
>> 341                     *p=(*a<<24)|(*r<<16)|(*g<<8)|*b;
>>
>> I hope I wasn't too long
>>
>> regards
>>
>>
>> P.S.
>> if there's anybody developing for AIW 9800 please give me some 
>> pointer cause i'm really stuck
>> thanx
>>
>> Antti Ajanki wrote:
>>
>>> On Wed, 5 Oct 2005, nebokazer wrote:
>>>
>>>> I found a bug in avview that was crashing in xmisc_seticon
>>>> in xmisc.c on line 299 : unsigned long *data,*p;
>>>> get pib overwrited on my 64bit box I changed 64bit long to 32bit int
>>>> and avview starts allright  :D
>>>
>>>
>>>
>>> Could you be a little more specific? On which line is the 
>>> overwriting happening?
>>>
>>> Or, better yet, could you try Avview from CVS? I have already fixed 
>>> one 64 bitness bug on xmisc_seticon and this may be the same one.
>>>
>>>>
>>>> so I manages to get avview ``working'', but I get a black picture 
>>>> whatsoever
>>>> the only obvious errors in the Xlog are very oscure to me
>>>>
>>>> *** x log |grep EE
>>>>
>>>> (EE) RADEON(0): Setting pPriv->adjustment to 1
>>>> (EE) RADEON(0): Setting pPriv->adjustment to 2
>>>> (EE) RADEON(0): Setting pPriv->adjustment to 3
>>>> (EE) RADEON(0): Setting pPriv->adjustment to 4
>>>> (EE) RADEON(0): Setting pPriv->adjustment to 5
>>>> (EE) RADEON(0): Setting pPriv->adjustment to 6
>>>> (EE) RADEON(0): Setting pPriv->adjustment to 7
>>>
>>>
>>>
>>> I don't think these are actually errors even though they are marked 
>>> with EE, they are just info messages that you changed debug 
>>> adjustment value.
>>>
>>>> any ideas?
>>>
>>>
>>>
>>> Sorry, can't help you further.
>>>
>>
>>
>>
>> -------------------------------------------------------
>> This SF.Net email is sponsored by:
>> Power Architecture Resource Center: Free content, downloads, 
>> discussions,
>> and more. http://solutions.newsforge.com/ibmarch.tmpl
>> _______________________________________________
>> Gatos-devel mailing list
>> Gat...@li...
>> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>>

Peter Zijlstra wrote:
> On Mon, 2005-10-10 at 14:42 +0200, Rune Petersen wrote:
> 
> 
>>I have not tested newer versions than 6.8.99.9 if the monolithic source 
>>doesn't work please give the 6.8.99.9 version a try while I see whats wrong.
> 
> 
> I used that patch on yesterdays HEAD (monolithic tree), and all I found
> were a few trivial rejects because the EXA commit inserted some lines.
> 
> Other than that it works just fine here. I could roll a patch using 
> cvs diff -u if anybody cares.

Thank you.
I was hoping to wait until EXA have become stable before switching, but 
since you have done the work, I wouldn't mind you sending it to me.


Rune Petersen

On Mon, 2005-10-10 at 14:42 +0200, Rune Petersen wrote:

> I have not tested newer versions than 6.8.99.9 if the monolithic source 
> doesn't work please give the 6.8.99.9 version a try while I see whats wrong.

I used that patch on yesterdays HEAD (monolithic tree), and all I found
were a few trivial rejects because the EXA commit inserted some lines.

Other than that it works just fine here. I could roll a patch using 
cvs diff -u if anybody cares.

Kind regards,

Peter Zijlstra

-- 
Peter Zijlstra <a.p...@ch...>

Alon Bar-Lev wrote:
> Rune Petersen wrote:
> 
>>
>> TV-in is merged, not TV-out... main reason for it not being merged is 
>> that the code is GPL..
>>
> 
> Hello Rune,
> 
> Thank you for your help!
> 
> After migrating to X Window System Version 6.99.99.900 (7.0.0 RC 0) I 
> tried to apply the patch of TV Output from 
> http://megahurts.dk/rune/tv_output.html, I use modular X so I had to 
> change the Makefile.am (The Imakefile was invalid in the patch).
Please the monolithic source, it should be more stable.
> 
> It seems that there are changes between the 6.8.99.9 and the 
> 6.99.99.900, there was XvExtension preprocessor directive that controled 
> some aspects that is now missing... So I defined this and succeeded in 
> building the theater_out_drv.so.

XvExtension should be defined, something must be wrong with the build.

> 
> But when X is loading it I get:
> ----------------------------------------------
> (II) Loading sub module "theater_out"
> (II) LoadModule: "theater_out"
> (II) Loading /usr/lib/xorg/modules/multimedia/theater_out_drv.so
> (II) Module theater_out: vendor="X.Org Foundation"
>         compiled for 4.3.99.902, module version = 1.0.0
>         ABI class: X.Org Video Driver, version 0.7
> 
>    *** If unresolved symbols were reported above, they might not
>    *** be the reason for the server aborting.
> 
> Backtrace:
> 0: X(xf86SigHandler+0x8a) [0x80b7dea]
> 1: [0xffffe420]
> 2: /usr/lib/xorg/modules/drivers/radeon_drv.so(RADEONScreenInit+0x22b) 
> [0xb783f805]
> 3: X(AddScreen+0x1e3) [0x806e2b3]
> 4: X(InitOutput+0x1df) [0x80a079f]
> 5: X(main+0x2aa) [0x806dc2a]
> 6: /lib/libc.so.6(__libc_start_main+0xa6) [0xb7bde3a6]
> 7: X(FontFileCompleteXLFD+0x81) [0x806d321]

I need you to tell me the lines numbers.

> I will wait until an updated patch version will be available... :(

I have not tested newer versions than 6.8.99.9 if the monolithic source 
doesn't work please give the 6.8.99.9 version a try while I see whats wrong.

You can get it at:
	http://xorg.freedesktop.org/snapshots


Rune Petersen