gatos-devel Mailing List for GATOS

On Fri, 30 Apr 2004, Tommi Sakari Uimonen wrote:

> Hi. I'd like to compile the tv_out branch, but some header files are
> missing from Debian xfree86 packages.
>
> If I compile whole xfree from source, is the 4.3.0 release good for
> getting the tv_out working or should I use cvs.

XFree86 4.3.0 source is what you need.

                    best

                        Vladimir Dergachev

>
> AIW 7500
> kernel 2.4
>
> Tommi
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: Oracle 10g
> Get certified on the hottest thing ever to hit the market... Oracle 10g.
> Take an Oracle 10g class now, and we'll give you the exam FREE.
> http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

Hi. I'd like to compile the tv_out branch, but some header files are
missing from Debian xfree86 packages.

If I compile whole xfree from source, is the 4.3.0 release good for
getting the tv_out working or should I use cvs.

AIW 7500
kernel 2.4

Tommi

Yeah the lastest cvs snapshot for the km kernel modules didn't work for me on a 2.6.5 kernel out
of the box. I was however able to get km to compile and more specifically my AIW to capture with a
few modifications to the km module source code, specifically km_api.c and km.c. 

Hope this helps at all.

-Ian

km.c:
-------------------------------------------------------------------
Line 1025:
 From:
   static int __init init_module(void)
 To:
   static int __init km_init(void)

Line 1053:
 From:
   void cleanup_module(void)
 To:
   static void __exit km_exit(void)

Add to EOF:

	module_init(km_init);
	module_exit(km_exit);
------------------------------------------------------

km_api.c:
----------------------------------------------
Line 581:
 From:
	void cleanup_module(void);
 To:
	static void km_api_exit(void);

Line 583:
 From:
	static int __init init_module(void)
 To:
	static int __init km_api_init(void)

Line 612:
 From:
	void cleanup_module(void)
 To:
	static void __exit km_api_exit(void)
Add to EOF:

	module_init(km_api_init);
	module_exit(km_api_exit);
---------------------------------------------------------

--- Kevin <kf...@si...> wrote:
> I have setup a separate machine with my old AIW Classic (4MB) to test the 
> compilability of km and avview on MDK 10 before I upgrade my main machine 
> that is running MDK 9.2 with avview-0.80.3 and km-0.5.0.
> 
> MDK 10 is using kernel 2.6.3 and when I attempt to compile km I have compile 
> errors.  I have a CVS version from about a month ago but it is still no 
> working.  I am going to try to get a newer CVS version but my CVS only works 
> on a direct dial, not through a proxy.
> 
> Has anyone else had errors compiling km under MDK 10?  Has anyone had success?
> 
> Kevin Foss
> Ottawa, Canada
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: Oracle 10g
> Get certified on the hottest thing ever to hit the market... Oracle 10g. 
> Take an Oracle 10g class now, and we'll give you the exam FREE.
> http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel



	
		
__________________________________
Do you Yahoo!?
Win a $20,000 Career Makeover at Yahoo! HotJobs  
http://hotjobs.sweepstakes.yahoo.com/careermakeover 

On Tue, 27 Apr 2004, Rolf Brusletto wrote:

> Howdy all,
>
> Just curious how active this list is, and if there is support for the
> All in wonder 9600? Has anybody had a good experience getting capture to
> work with this card?

Not yet. Matt Mercer has written some code for it, but we still have not
figured out how to upload the firmware to Rage Theatre 200, without which
none of it works..

                    best

                      Vladimir Dergachev

>
> Thanks,
>
> Rolf Brusletto
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: Oracle 10g
> Get certified on the hottest thing ever to hit the market... Oracle 10g.
> Take an Oracle 10g class now, and we'll give you the exam FREE.
> http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

On Thu, 29 Apr 2004, Spencer wrote:

> Kevin wrote:
>
> >I have setup a separate machine with my old AIW Classic (4MB) to test the
> >compilability of km and avview on MDK 10 before I upgrade my main machine
> >that is running MDK 9.2 with avview-0.80.3 and km-0.5.0.
> >
> >MDK 10 is using kernel 2.6.3 and when I attempt to compile km I have compile
> >errors.  I have a CVS version from about a month ago but it is still no
> >working.  I am going to try to get a newer CVS version but my CVS only works
> >on a direct dial, not through a proxy.
> >
> >Has anyone else had errors compiling km under MDK 10?  Has anyone had success?
> >
> >Kevin Foss
> >Ottawa, Canada
> >
> >
> I had no success compiling km on any of the 2.6 kernels or I would have
> included a RPM in
> Mandrake 10 contribs

Did you try km from CVS ? It is supposed to work with 2.6.x kernels.
(I have not tried it myself for a while, as I was busy and, besides, my
main devel box crashed).


                    best

                      Vladimir Dergachev

>
> Spence
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: Oracle 10g
> Get certified on the hottest thing ever to hit the market... Oracle 10g.
> Take an Oracle 10g class now, and we'll give you the exam FREE.
> http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

Kevin wrote:

>I have setup a separate machine with my old AIW Classic (4MB) to test the 
>compilability of km and avview on MDK 10 before I upgrade my main machine 
>that is running MDK 9.2 with avview-0.80.3 and km-0.5.0.
>
>MDK 10 is using kernel 2.6.3 and when I attempt to compile km I have compile 
>errors.  I have a CVS version from about a month ago but it is still no 
>working.  I am going to try to get a newer CVS version but my CVS only works 
>on a direct dial, not through a proxy.
>
>Has anyone else had errors compiling km under MDK 10?  Has anyone had success?
>
>Kevin Foss
>Ottawa, Canada
>  
>
I had no success compiling km on any of the 2.6 kernels or I would have 
included a RPM in
Mandrake 10 contribs

Spence

I have setup a separate machine with my old AIW Classic (4MB) to test the 
compilability of km and avview on MDK 10 before I upgrade my main machine 
that is running MDK 9.2 with avview-0.80.3 and km-0.5.0.

MDK 10 is using kernel 2.6.3 and when I attempt to compile km I have compile 
errors.  I have a CVS version from about a month ago but it is still no 
working.  I am going to try to get a newer CVS version but my CVS only works 
on a direct dial, not through a proxy.

Has anyone else had errors compiling km under MDK 10?  Has anyone had success?

Kevin Foss
Ottawa, Canada

Hi,
Thursday, 29 April, 2004, 18:55:14, Zachary M. Berke wrote:
> What do I need to do to recompile/reinstall the source after applying  
> retinalburn's patch?
[...]
> [zberke@localhost ati.2]$ make -f Imakefile
> Imakefile:1: *** multiple target patterns.  Stop.

No, Imakefile is for "internal" use only.
From my memory, you should first do "xmkmf
/usr/src/....../XFree86-4.3.0-src/xc" in the
ati.2 directory (put right path to your XFree
source tree here), then just do "make".
After make complete, copy new *_drv.o files
over the older ones. (It might be usefull to
backup or rename old ones)
-- 
Best regards,
 Nikolai Zhubr


> Retinal burn's instructions say that at this point I should "rebuild  
> and install the driver" -- I'm having trouble doing that.  Any help is  
> much appreciated.

> Thank you,
> Zach

> On Apr 28, 2004, at 3:24 PM, Nikolai Zhubr wrote:

>> Hi,
>> Tuesday, 27 April, 2004, 19:19:05, Zachary M. Berke wrote:
>>> Hi,
>>
>>> I am not sure if my card will work with this project.  Here's the
>> Yes, regular gatos' ati.2 drivers should provide tv-in and
>> accelerated video playback (through xv). These features are
>> reasonably stable and pretty usefull.
>> Gatos' km provides video-capture though it is somewhat buggy
>> currently (especially for this chipset).
>> Patches from retinalburn additionally provide accelerated glx
>> (not very fast) and tv-out (some limited functionality).
>> -- 
>> Best regards,
>>  Nikolai Zhubr
>>
>>> relevant info:
>>> [root@localhost zberke]# /sbin/lspci -vs 00:10.0
>>> 00:10.0 VGA compatible controller: ATI Technologies Inc 3D Rage Pro
>>> 215GP (rev 5c) (prog-if 00 [VGA])
>>>          Subsystem: ATI Technologies Inc Rage Pro Turbo
>>>          Flags: bus master, stepping, medium devsel, latency 64, IRQ  
>>> 11
>>>          Memory at fd000000 (32-bit, prefetchable) [size=16M]
>>>          I/O ports at 7400 [size=256]
>>>          Memory at feafb000 (32-bit, non-prefetchable) [size=4K]
>>>          Expansion ROM at <unassigned> [disabled] [size=128K]
>>
>>> Do I need to patch the ati.2 driver? (
>>> http://www.retinalburn.net/linux/tvout.html)
>>> I believe I do, but I'm having trouble doing so.  I checked out the
>>> ati.2 source from cvs and then applied the patch, but I can't
>>> compile/reinstall the ati.2 source -- there's no Makefile.  Do I need
>>> the entire source tree?
>>
>>> Thanks for any help,
>>> Zach
>>
>>
>>
>>> -------------------------------------------------------
>>> This SF.Net email is sponsored by: Oracle 10g
>>> Get certified on the hottest thing ever to hit the market... Oracle  
>>> 10g.
>>> Take an Oracle 10g class now, and we'll give you the exam FREE.
>>> http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click
>>> _______________________________________________
>>> Gatos-devel mailing list
>>> Gat...@li...
>>> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>>
>>

I am primarily interested in getting TVOut working -- this card doesn't  
have a capture functionality.  So that means I need to install  
retinalburn's patches, yeah?

What do I need to do to recompile/reinstall the source after applying  
retinalburn's patch?

Here's what I've got so far:
[zberke@localhost tmp]$ cvs -z3  
-d:pserver:ano...@cv...:/cvsroot/gatos co ati.2
cvs server: Updating ati.2
[ file list omitted ]
cvs server: Updating ati.2/tvo_set
[zberke@localhost tmp]$ cd ati.2
[zberke@localhost tmp]$ wget  
http://www.retinalburn.net/linux/tvout-patches/mach64-tvout 
-20030328.diff.gz
10:45:55 (41.05 KB/s) - `mach64-tvout-20030328.diff.gz' saved  
[7020/7020]
[zberke@localhost ati.2]$ gunzip -c mach64-tvout-20030328.diff.gz |  
patch -p0
patching file Imakefile
patching file aticonfig.c
patching file aticonsole.c
patching file atilock.c
patching file atimode.c
patching file atioption.c
patching file atioption.h
patching file atipreinit.c
patching file atiscreen.c
patching file atistruct.h
patching file atixv.c
Hunk #1 succeeded at 1833 (offset 6 lines).
[zberke@localhost ati.2]$ make -f Imakefile
Imakefile:1: *** multiple target patterns.  Stop.


Retinal burn's instructions say that at this point I should "rebuild  
and install the driver" -- I'm having trouble doing that.  Any help is  
much appreciated.

Thank you,
Zach

On Apr 28, 2004, at 3:24 PM, Nikolai Zhubr wrote:

> Hi,
> Tuesday, 27 April, 2004, 19:19:05, Zachary M. Berke wrote:
>> Hi,
>
>> I am not sure if my card will work with this project.  Here's the
> Yes, regular gatos' ati.2 drivers should provide tv-in and
> accelerated video playback (through xv). These features are
> reasonably stable and pretty usefull.
> Gatos' km provides video-capture though it is somewhat buggy
> currently (especially for this chipset).
> Patches from retinalburn additionally provide accelerated glx
> (not very fast) and tv-out (some limited functionality).
> -- 
> Best regards,
>  Nikolai Zhubr
>
>> relevant info:
>> [root@localhost zberke]# /sbin/lspci -vs 00:10.0
>> 00:10.0 VGA compatible controller: ATI Technologies Inc 3D Rage Pro
>> 215GP (rev 5c) (prog-if 00 [VGA])
>>          Subsystem: ATI Technologies Inc Rage Pro Turbo
>>          Flags: bus master, stepping, medium devsel, latency 64, IRQ  
>> 11
>>          Memory at fd000000 (32-bit, prefetchable) [size=16M]
>>          I/O ports at 7400 [size=256]
>>          Memory at feafb000 (32-bit, non-prefetchable) [size=4K]
>>          Expansion ROM at <unassigned> [disabled] [size=128K]
>
>> Do I need to patch the ati.2 driver? (
>> http://www.retinalburn.net/linux/tvout.html)
>> I believe I do, but I'm having trouble doing so.  I checked out the
>> ati.2 source from cvs and then applied the patch, but I can't
>> compile/reinstall the ati.2 source -- there's no Makefile.  Do I need
>> the entire source tree?
>
>> Thanks for any help,
>> Zach
>
>
>
>> -------------------------------------------------------
>> This SF.Net email is sponsored by: Oracle 10g
>> Get certified on the hottest thing ever to hit the market... Oracle  
>> 10g.
>> Take an Oracle 10g class now, and we'll give you the exam FREE.
>> http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click
>> _______________________________________________
>> Gatos-devel mailing list
>> Gat...@li...
>> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>
>

Hi,
Tuesday, 27 April, 2004, 19:19:05, Zachary M. Berke wrote:
> Hi,

> I am not sure if my card will work with this project.  Here's the
Yes, regular gatos' ati.2 drivers should provide tv-in and
accelerated video playback (through xv). These features are
reasonably stable and pretty usefull.
Gatos' km provides video-capture though it is somewhat buggy
currently (especially for this chipset).
Patches from retinalburn additionally provide accelerated glx
(not very fast) and tv-out (some limited functionality).
-- 
Best regards,
 Nikolai Zhubr

> relevant info:
> [root@localhost zberke]# /sbin/lspci -vs 00:10.0
> 00:10.0 VGA compatible controller: ATI Technologies Inc 3D Rage Pro 
> 215GP (rev 5c) (prog-if 00 [VGA])
>          Subsystem: ATI Technologies Inc Rage Pro Turbo
>          Flags: bus master, stepping, medium devsel, latency 64, IRQ 11
>          Memory at fd000000 (32-bit, prefetchable) [size=16M]
>          I/O ports at 7400 [size=256]
>          Memory at feafb000 (32-bit, non-prefetchable) [size=4K]
>          Expansion ROM at <unassigned> [disabled] [size=128K]

> Do I need to patch the ati.2 driver? ( 
> http://www.retinalburn.net/linux/tvout.html)
> I believe I do, but I'm having trouble doing so.  I checked out the 
> ati.2 source from cvs and then applied the patch, but I can't 
> compile/reinstall the ati.2 source -- there's no Makefile.  Do I need 
> the entire source tree?

> Thanks for any help,
> Zach



> -------------------------------------------------------
> This SF.Net email is sponsored by: Oracle 10g
> Get certified on the hottest thing ever to hit the market... Oracle 10g. 
> Take an Oracle 10g class now, and we'll give you the exam FREE.
> http://ads.osdn.com/?ad_id=3149&alloc_id=8166&op=click
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel

Howdy all,

Just curious how active this list is, and if there is support for the 
All in wonder 9600? Has anybody had a good experience getting capture to 
work with this card?

Thanks,

Rolf Brusletto

Hi,

I am not sure if my card will work with this project.  Here's the 
relevant info:
[root@localhost zberke]# /sbin/lspci -vs 00:10.0
00:10.0 VGA compatible controller: ATI Technologies Inc 3D Rage Pro 
215GP (rev 5c) (prog-if 00 [VGA])
         Subsystem: ATI Technologies Inc Rage Pro Turbo
         Flags: bus master, stepping, medium devsel, latency 64, IRQ 11
         Memory at fd000000 (32-bit, prefetchable) [size=16M]
         I/O ports at 7400 [size=256]
         Memory at feafb000 (32-bit, non-prefetchable) [size=4K]
         Expansion ROM at <unassigned> [disabled] [size=128K]

Do I need to patch the ati.2 driver? ( 
http://www.retinalburn.net/linux/tvout.html)
I believe I do, but I'm having trouble doing so.  I checked out the 
ati.2 source from cvs and then applied the patch, but I can't 
compile/reinstall the ati.2 source -- there's no Makefile.  Do I need 
the entire source tree?

Thanks for any help,
Zach

Hye ,

has someone successfully installed gatos driver on RHEL 3.0 or
Whiteboxlinux?

I plan to update my RedHat 7.3 to a newer release because 7.3 will be no
longer maintained. 

I got problems using the latest RedHat kernel/XFree4.3 on 7.3. So i'm
not sure if it will work on RHEL 3.0. At hte moment i use
XFree86-4.2.0-8 with kernel 2.4.18-27.7.x successfully.

The hardware is a IBM Thinkpad A30P.

Thanks for suggestions.

regards

Uwe Melcher

Hey everybody I'm new here. My name is Ian Roessle. I'm setting up an open source recording studio and happened to posses an ATI AIW video card.
 
I just spend the last 6 hours or so getting capture support to work under avview and thought you all might benefit somewhat from my experience, specifically with the modifications I had to make to km.c and km_api.c.
 
I used the latest cvs snapshot, copied Makefile-2.6 -> Makefile , and ran the make -C /usr/src/linux SUBDIRES=$PWD modules. When I attempted to load the modules, I would get "-1 Unknown symbol in module".
 
The changes:
 
         1)   Change line 1053 from "void cleanup_module(void)" to "static void __exit cleanup_module(void)" 
         2)   Add "module_init(init_module)" and "module_exit(cleanup_module)" to the EOF.
 
After that the I could load the modules and avview detected the v4l device. 
 
 
Hope this helps some.
 
Ian Roessle
         

		
---------------------------------
Do you Yahoo!?
Yahoo! Photos: High-quality 4x6 digital prints for 25¢

Am Freitag, 16. April 2004 18:03 schrieb Matthias Meyer:
>
> Thanks Ales for pointing me to
> http://sourceforge.net/mailarchive/forum.php?thread_id=4152625&forum_id=5
>014. I will try this patch if I get a little additional help.
> If someone can point me to a .tar which contains all relevant sources or
> if someone can send me a already patched and compiled r128_drv.o it will
> help me so much.
>
> Is this patch the only possible way to get the Rage128pro working with
> TV-out?
> Nobody have running Rage128 at TV-Out?

*g* I am happy! It do work!
First I remove the GATOS-Driver. Finally they do not support TV-Out but only 
TV-In (http://sourceforge.net/mailarchive/message.php?msg_id=4480460)
All what I have to do is to use the vesa instead the r128 driver from 
standard XFree86 4.3.

Here you can see the relevant sections from my /etc/X11/XF86Config-4:
Section "Device"
        Identifier  "ATI Rage128 Pro 32MB"
        Driver      "vesa"
        VendorName  "All"
        BoardName   "All"
EndSection
Section "Screen"
        Identifier "Screen0"
        Device     "ATI Rage128 Pro 32MB"
        Monitor    "Monitor0"
        DefaultColorDepth 24
        SubSection "Display"
                Depth     24
                Modes "800x600"
        EndSubSection
EndSection

At least thanks to Andrej and Ales for there help with drivers for XFree86 
4.2 But for the time being I would like to use XFree86 4.3.
-- 
no panic

The card is a Mach64


Processing "VIDEO_STREAM_ACTIVE"="1" int_value=1
Purging transfer queue
radeon_get_window_parameters: width=720 height=288
Capture buf size=414720
RADEON_BUS_CNTL=0x5133a3b0
Starting GUIDMA queue
km: DMA_GUI_QUEUE stalled after abort
radeon_get_window_parameters: width=720 height=288
radeon_get_window_parameters: width=720 height=288
radeon_get_window_parameters: width=720 height=288
radeon_get_window_parameters: width=720 height=288
radeon_get_window_parameters: width=720 height=288
radeon_get_window_parameters: width=720 height=288
radeon_get_window_parameters: width=720 height=288
radeon_get_window_parameters: width=720 height=288
radeon_get_window_parameters: width=720 height=288
Debug: sleeping function called from invalid context at mm/slab.c:1931
in_atomic():1, irqs_disabled():0
Call Trace:
 [<c011a98b>] __might_sleep+0xab/0xd0
 [<c03998ae>] sock_aio_read+0xbe/0xd0
 [<c013c255>] kmem_cache_alloc+0x65/0x70
 [<e9a515a8>] open_km_device+0x58/0x180 [km_api_drv]
 [<e9a839a8>] km_v4l_open+0x28/0xd0 [km_drv]
 [<c02fb780>] video_open+0x130/0x260
 [<c0159021>] chrdev_open+0xf1/0x220
 [<c0158f30>] chrdev_open+0x0/0x220
 [<c014f162>] dentry_open+0x142/0x210
 [<c014f01b>] filp_open+0x5b/0x60
 [<c014f4a3>] sys_open+0x53/0x90
 [<c0108daf>] syscall_call+0x7/0xb

Processing "VIDEO_STREAM_ACTIVE"="1" int_value=1
radeon_get_window_parameters: width=720 height=288
Debug: sleeping function called from invalid context at mm/slab.c:1931
in_atomic():1, irqs_disabled():0
Call Trace:
 [<c011a98b>] __might_sleep+0xab/0xd0
 [<c03998ae>] sock_aio_read+0xbe/0xd0
 [<c013c255>] kmem_cache_alloc+0x65/0x70
 [<e9a515a8>] open_km_device+0x58/0x180 [km_api_drv]
 [<e9a839a8>] km_v4l_open+0x28/0xd0 [km_drv]
 [<c02fb780>] video_open+0x130/0x260
 [<c0159021>] chrdev_open+0xf1/0x220
 [<c0158f30>] chrdev_open+0x0/0x220
 [<c014f162>] dentry_open+0x142/0x210
 [<c014f01b>] filp_open+0x5b/0x60
 [<c014f4a3>] sys_open+0x53/0x90
 [<c0108daf>] syscall_call+0x7/0xb
Processing "VIDEO_STREAM_ACTIVE"="1" int_value=1
radeon_get_window_parameters: width=720 height=288
radeon_get_window_parameters: width=720 height=288
radeon_get_window_parameters: width=720 height=288
radeon_get_window_parameters: width=720 height=288
radeon_get_window_parameters: width=720 height=288
Unable to handle kernel NULL pointer dereference at virtual address 00000000
 printing eip:
e9a5243f
*pde = 00000000
Oops: 0000 [#1]
PREEMPT
CPU:    0
EIP:    0060:[<e9a5243f>]    Not tainted
EFLAGS: 00210282   (2.6.5)
EIP is at km_data_destroy_kdufpd+0xf/0x60 [km_api_drv]
eax: 00000000   ebx: e9a89d60   ecx: d9516a0c   edx: d9516a0c
esi: 00000000   edi: e7fe0280   ebp: e0547e3c   esp: e0547e2c
ds: 007b   es: 007b   ss: 0068
Process start_avview.no (pid: 4034, threadinfo=e0546000 task=dedfc5e0)
Stack: e6af3740 e0547e4c e9a89d60 e9a83a50 e0547e4c e9a83a75 00000000 db955540
       e0547e70 c0150fc0 d9516a0c db955540 d9516a0c da07fd80 db955540 00000000
       dec06e40 e0547e8c c014f567 db955540 dec06e40 00000009 00000007 dec06e40
Call Trace:
 [<e9a83a50>] km_v4l_close+0x0/0x50 [km_drv]
 [<e9a83a75>] km_v4l_close+0x25/0x50 [km_drv]
 [<c0150fc0>] __fput+0x120/0x140
 [<c014f567>] filp_close+0x57/0x80
 [<c011e54c>] put_files_struct+0x6c/0xe0
 [<c011f1bf>] do_exit+0x13f/0x440
 [<c011f56a>] do_group_exit+0x3a/0xb0
 [<c0127ea1>] get_signal_to_deliver+0x251/0x360
 [<c0108b66>] do_signal+0x66/0xe0
 [<e9a83b0f>] km_v4l_read+0x5f/0xa0 [km_drv]
 [<c014fe5a>] vfs_read+0xca/0x140
 [<c0150129>] sys_read+0x59/0x60
 [<c0108c1b>] do_notify_resume+0x3b/0x40
 [<c0108dfa>] work_notifysig+0x13/0x15

Code: 8b 1e 85 db 74 2d b8 00 e0 ff ff 21 e0 ff 40 14 ff 4b 10 ff
 <1>Unable to handle kernel NULL pointer dereference at virtual address 
00000014
 printing eip:
c011b18d
*pde = 00000000
Oops: 0000 [#2]
PREEMPT
CPU:    0
EIP:    0060:[<c011b18d>]    Not tainted
EFLAGS: 00210202   (2.6.5)
EIP is at mm_release+0x3d/0xc0
eax: 00000000   ebx: 42d09bf8   ecx: dedfc5e0   edx: 00000000
esi: dedfc5e0   edi: dedfc5e0   ebp: e0547cf4   esp: e0547cd8
ds: 007b   es: 007b   ss: 0068
Process start_avview.no (pid: 4034, threadinfo=e0546000 task=dedfc5e0)
Stack: e0547df8 e0546000 c01173a0 e0547d38 c010971c 00000000 00000000 e0547d18
       c011f10a dedfc5e0 00000000 00000000 0000007b e0547df8 e0546000 c01173a0
       e0547d38 c0109e89 0000000b c040c942 00000000 00000001 00000000 00000000
Call Trace:
 [<c01173a0>] do_page_fault+0x0/0x57e
 [<c010971c>] common_interrupt+0x18/0x20
 [<c011f10a>] do_exit+0x8a/0x440
 [<c01173a0>] do_page_fault+0x0/0x57e
 [<c0109e89>] die+0xf9/0x100
 [<c0117596>] do_page_fault+0x1f6/0x57e
 [<c0146f33>] __pte_chain_free+0x63/0x70
 [<c0140e68>] zap_pte_range+0x148/0x190
 [<c01173a0>] do_page_fault+0x0/0x57e
 [<c01097b9>] error_code+0x2d/0x38
 [<e9a5243f>] km_data_destroy_kdufpd+0xf/0x60 [km_api_drv]
 [<e9a83a50>] km_v4l_close+0x0/0x50 [km_drv]
 [<e9a83a75>] km_v4l_close+0x25/0x50 [km_drv]
 [<c0150fc0>] __fput+0x120/0x140
 [<c014f567>] filp_close+0x57/0x80
 [<c011e54c>] put_files_struct+0x6c/0xe0
 [<c011f1bf>] do_exit+0x13f/0x440
 [<c011f56a>] do_group_exit+0x3a/0xb0
 [<c0127ea1>] get_signal_to_deliver+0x251/0x360
 [<c0108b66>] do_signal+0x66/0xe0
 [<e9a83b0f>] km_v4l_read+0x5f/0xa0 [km_drv]
 [<c014fe5a>] vfs_read+0xca/0x140
 [<c0150129>] sys_read+0x59/0x60
 [<c0108c1b>] do_notify_resume+0x3b/0x40
 [<c0108dfa>] work_notifysig+0x13/0x15

-- 
Niels Elgaard Larsen

On Thu, 22 Apr 2004, Harry Helmich wrote:

> Starting with which kernel version?

2.6.5.

                best

                    Vladimir Dergachev

>
> On Tuesday 13 April 2004 11:22 am, Vladimir Dergachev wrote:
> > Actually, ati_remote is now part of the Linux kernel,
> > Torrey Hoffman <tho...@ar...> has submitted and maintains it.
> >
> >                      best
> >
> >                        Vladimir Dergachev
> >
> > On Tue, 13 Apr 2004, Vasileios Gkanis wrote:
> > > Hi,
> > > the other day I upgraded the kernel to 2.6.5 and to my surprise (or
> > > disappointment) I found that ati_remote26.c  did not compile. The problem
> > > was in line 382 :
> > > 	interface = &intf->altsetting[intf->act_altsetting];
> > >
> > > However, ati_remote26.c compiles just fine with the kernel 2.6.4
> > >
> > > Vasilis
> > >
> > >
> > > -------------------------------------------------------
> > > This SF.Net email is sponsored by: IBM Linux Tutorials
> > > Free Linux tutorial presented by Daniel Robbins, President and CEO of
> > > GenToo technologies. Learn everything from fundamentals to system
> > > administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> > > _______________________________________________
> > > Gatos-devel mailing list
> > > Gat...@li...
> > > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> >
> > -------------------------------------------------------
> > This SF.Net email is sponsored by: IBM Linux Tutorials
> > Free Linux tutorial presented by Daniel Robbins, President and CEO of
> > GenToo technologies. Learn everything from fundamentals to system
> > administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> > _______________________________________________
> > Gatos-devel mailing list
> > Gat...@li...
> > https://lists.sourceforge.net/lists/listinfo/gatos-devel
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

Starting with which kernel version? 

On Tuesday 13 April 2004 11:22 am, Vladimir Dergachev wrote:
> Actually, ati_remote is now part of the Linux kernel,
> Torrey Hoffman <tho...@ar...> has submitted and maintains it.
>
>                      best
>
>                        Vladimir Dergachev
>
> On Tue, 13 Apr 2004, Vasileios Gkanis wrote:
> > Hi,
> > the other day I upgraded the kernel to 2.6.5 and to my surprise (or
> > disappointment) I found that ati_remote26.c  did not compile. The problem
> > was in line 382 :
> > 	interface = &intf->altsetting[intf->act_altsetting];
> >
> > However, ati_remote26.c compiles just fine with the kernel 2.6.4
> >
> > Vasilis
> >
> >
> > -------------------------------------------------------
> > This SF.Net email is sponsored by: IBM Linux Tutorials
> > Free Linux tutorial presented by Daniel Robbins, President and CEO of
> > GenToo technologies. Learn everything from fundamentals to system
> > administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> > _______________________________________________
> > Gatos-devel mailing list
> > Gat...@li...
> > https://lists.sourceforge.net/lists/listinfo/gatos-devel
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel

Yes, it seems they have support for FI1236/W. 
Note that the AIW9600PRO that I own has a FM1216ME/ H-3 tuner which is a PAL
version tuner(against its NTSC complement FM1236ME/ H-3) and has tv+radio FM
input. I wonder if someone has tried the fi1236 gatos driver with this kind
of tuners.

Regards,
Bogdan

-----Original Message-----
From: lud...@be... [mailto:lud...@be...] 
Sent: Monday, April 19, 2004 9:21 AM
To: vo...@mi...; Bogdan Diaconescu
Cc: gat...@li...
Subject: RE: [GATOS]Reiterate the AIW9600Pro question


Hi,

It seems that the guys of v4l / cx88 drivers are able to make this tuner
working. At least with the ATI TV Wonder Pro card. 

Have a look to this message : http://www.spinics.net/lists/vfl/msg09146.html


-----Original Message-----
From: Vladimir Dergachev [mailto:vo...@mi...]
Sent: 16 April 2004 02:10
To: Diaconescu Bogdan-BDIACON1
Cc: gat...@li...
Subject: Re: [GATOS]Reiterate the AIW9600Pro question




On Thu, 15 Apr 2004, Diaconescu Bogdan-BDIACON1 wrote:

> I'm still trying to find the answer for one question I sent before 
> regarding the AIW9600Pro
>
> I went through the ati.2 mainline sources and I 've seen that RV350 is 
> not supported and also THEATRE 200 is known but not supported yet. Are 
> there any near future plans to support the above chips?

Support is being worked on.. but slowly.

>
> Another question: Is there any documentation available, datasheets  
> for the chips I am interested in? Did ATI provided the specifications 
> to the developers in the case someone want to contribute?

Yes, we do have the documentation - at least part of it. Namely, we have the
docs for Rage Theatre 200, but not for the tuner FI1236/W made by Philips.
It is curious that neither ATI nor Philips are able to come up with
datasheet for it.

One thing you could do to help is to try to dig it up on the web (maybe it
became available since the last attempt) and also try to talk to local
Philips distributors.

Note, that we need the datasheet for FI1236/W  - not the earlier FI1236MK2
which is available on the web.

I am not 100% certain these tuners are different, but I was not successful
trying to talk to W edition, so, perhaps, they are.

                        best

                           Vladimir Dergachev

>
> Regards,
> Bogdan
>
>
>
>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of 
> GenToo technologies. Learn everything from fundamentals to system 
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li... 
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of GenToo
technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
_______________________________________________
Gatos-devel mailing list
Gat...@li...
https://lists.sourceforge.net/lists/listinfo/gatos-devel

Ok, how can I proceed with the NDA? Is it with ATI or with a third party
company? What are the steps a person will do in order to develop for gatos
ATI drivers?

Regards,
Bogdan

-----Original Message-----
From: Vladimir Dergachev [mailto:vo...@mi...] 
Sent: Sunday, April 18, 2004 6:44 PM
To: Diaconescu Bogdan-BDIACON1
Cc: gat...@li...
Subject: RE: [GATOS]Reiterate the AIW9600Pro question




On Fri, 16 Apr 2004, Diaconescu Bogdan-BDIACON1 wrote:

> Thanks for the replay Vladimir,
>
> I'll give a try to find the datasheets for FI1236/W. At the first 
> search on the web, no one seems to have datasheets for it.
>
> Cold you send me the datasheets for R200 and, if you have, for RV350 
> and/or 300?

Sorry, these are only available under NDA.

                      best

                        Vladimir Dergachev

>
> Regards,
> Bogdan
>
>
>
> -----Original Message-----
> From: Vladimir Dergachev [mailto:vo...@mi...]
> Sent: Friday, April 16, 2004 3:10 AM
> To: Diaconescu Bogdan-BDIACON1
> Cc: gat...@li...
> Subject: Re: [GATOS]Reiterate the AIW9600Pro question
>
>
>
>
> On Thu, 15 Apr 2004, Diaconescu Bogdan-BDIACON1 wrote:
>
> > I'm still trying to find the answer for one question I sent before 
> > regarding the AIW9600Pro
> >
> > I went through the ati.2 mainline sources and I 've seen that RV350 
> > is not supported and also THEATRE 200 is known but not supported 
> > yet. Are there any near future plans to support the above chips?
>
> Support is being worked on.. but slowly.
>
> >
> > Another question: Is there any documentation available, datasheets 
> > for the chips I am interested in? Did ATI provided the 
> > specifications to the developers in the case someone want to 
> > contribute?
>
> Yes, we do have the documentation - at least part of it. Namely, we 
> have the docs for Rage Theatre 200, but not for the tuner FI1236/W 
> made by Philips. It is curious that neither ATI nor Philips are able 
> to come up with datasheet for it.
>
> One thing you could do to help is to try to dig it up on the web 
> (maybe it became available since the last attempt) and also try to 
> talk to local Philips distributors.
>
> Note, that we need the datasheet for FI1236/W  - not the earlier 
> FI1236MK2 which is available on the web.
>
> I am not 100% certain these tuners are different, but I was not 
> successful trying to talk to W edition, so, perhaps, they are.
>
>                         best
>
>                            Vladimir Dergachev
>
> >
> > Regards,
> > Bogdan
> >
> >
> >
> >
> >
> >
> >
> > -------------------------------------------------------
> > This SF.Net email is sponsored by: IBM Linux Tutorials
> > Free Linux tutorial presented by Daniel Robbins, President and CEO 
> > of GenToo technologies. Learn everything from fundamentals to system 
> > administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=clic
> > k
> > _______________________________________________
> > Gatos-devel mailing list
> > Gat...@li...
> > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> >
>

Hi,

It seems that the guys of v4l / cx88 drivers are able to make this tuner =
working. At least with the ATI TV Wonder Pro card.=20

Have a look to this message : =
http://www.spinics.net/lists/vfl/msg09146.html


-----Original Message-----
From: Vladimir Dergachev [mailto:vo...@mi...]
Sent: 16 April 2004 02:10
To: Diaconescu Bogdan-BDIACON1
Cc: gat...@li...
Subject: Re: [GATOS]Reiterate the AIW9600Pro question




On Thu, 15 Apr 2004, Diaconescu Bogdan-BDIACON1 wrote:

> I'm still trying to find the answer for one question I sent before =
regarding
> the AIW9600Pro
>
> I went through the ati.2 mainline sources and I 've seen that RV350 is =
not
> supported and also THEATRE 200 is known but not supported yet. Are =
there any
> near future plans to support the above chips?

Support is being worked on.. but slowly.

>
> Another question: Is there any documentation available, datasheets  =
for the
> chips I am interested in? Did ATI provided the specifications to the
> developers in the case someone want to contribute?

Yes, we do have the documentation - at least part of it. Namely, we have
the docs for Rage Theatre 200, but not for the tuner FI1236/W made by
Philips. It is curious that neither ATI nor Philips are able to come up
with datasheet for it.

One thing you could do to help is to try to dig it up on the web (maybe =
it
became available since the last attempt) and also try to talk to local
Philips distributors.

Note, that we need the datasheet for FI1236/W  - not the earlier =
FI1236MK2
which is available on the web.

I am not 100% certain these tuners are different, but I was not =
successful
trying to talk to W edition, so, perhaps, they are.

                        best

                           Vladimir Dergachev

>
> Regards,
> Bogdan
>
>
>
>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> =
administration.http://ads.osdn.com/?ad_id=3D1470&alloc_id=3D3638&op=3Dcli=
ck
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=3D1470&alloc_id=3D3638&op=3Dcli=
ck
_______________________________________________
Gatos-devel mailing list
Gat...@li...
https://lists.sourceforge.net/lists/listinfo/gatos-devel

Hi Francisco,
Sunday, 18 April, 2004, 20:33:15, you wrote:
[...]
> Video: End of filestem load 30.9 %  
> 2:00 (35.5 MB), system load 30.9 %  
> Audio: End of file
> 2:00 (35.5 MB), system load 30.9 %  
> mp1e: Done.
> 1+0 records in
> 1+0 records out

> Results:

> 1. When I start xawtv or avview, the overlay winodow has less pixels that
> the real capture window. I mean that I can't see the hole video
This happens to me sometimes too, I'm too lazy to dig into it as
it can be "cured" by maximizing and then restoring xawtv, on my
box at least. I think it is just overlay scaler gets misconfigured.

> picture/frame, only one part is displayed in the window. If I choose PAL
> encode and S-VIDEO input and close xawtv settings menu, the video overlay
> shows more pixels (The hole PAL frame, altough using the same window size,
> and therefore I suppose that prevousy only part of the frame was shown and
> scaled to fill the window). Well, all this explanation is to say that the
> captured video doesn't cover the full frame size.  I think that is cropped
I'd suppose it is more likely that you need to specify more exact
arguments to mp1e, depending on your video standard. mp1e tries to
automatically correct capture size if the requested/default value
can not be used as is, but typically this automagic will only provide
some very _minimal_ usability.

> in the X axis (horizontally, perhaps 640 pixels instead of 720??). I don't
> know if it is also vertically croped because that effect is less
> noticiable.
Most likely cropping 768 to 384 (horizontally) happened in your case.

> 2. There are no displaced frames. WOW!
> 3. There appear to be some dropped frames as the movement is sometimes
> choppy, but the audio is more or less on sync.
mp1e should have reported framedrops percent then.

> 4. This time, no machine lock-up.
Cool!

> Conclusions:

> 1. Something in your code works avoiding the displacement effect in actual
> km code.
> 2. ffmpeg doesn't work with it.
This is really strange. I haven't used ffmpeg for some time but
still I'd suppose it should work. Perhaps you might need to adjust
arguments you specify for ffmpeg.

> 2. There seem to be dropped frames and some part of the frme is cropped
> out (problem with PAL settings?).
See above.
-- 
Best regards,
 Nikolai Zhubr

Vladimir and Nikolai,

I have tested Nikolai's patched mp1e against last CVS km.
To my surprise, it has -almost- worked.
No displaced frames but cropped and dropped frames, like the results with N=
ikolai's version, perhaps a bit worse (audio sync).
In spite of that, ffmpeg continues to give displaced frames (this time for =
a fraction of seconds).
So perhaps the problem is not related to km, but ffmpeg.

This time I use again my VIA C3 - VIA Apollo 133T setup.

Via C3 Nehemiah 1GHz
Apollo Pro 133T (VIA 694T/686B) Mainboard
512 MB RAM SDRAM CL2
VIA sound codec
Single 120 GB Seagate 7200 rpm HD
AIW PRO 8MB (3D Rage pro chip & BT829) PAL version
Linux Slackware 9.1, kernel 2.4.24 with Con Kolivas patches.=20
km module from CVS=20
Ati.2 experimental 14
avview 0.80.3
ffmpeg 0.48

Best Regards,

Francisco L. Fern=E1ndez

On Tue, 9 Mar 2004 01:55:12 -0500 (EST)
Vladimir Dergachev <vo...@mi...> wrote:

>=20
>=20
> On Tue, 9 Mar 2004, Nikolai Zhubr wrote:
>=20
> > Hi,
> > Tuesday, 09 March, 2004, 1:24:41, Francisco L Fernandez wrote:
> > [...]
> > > That's why I applied Kon Kolivas patches to my 2.4.24 kernel and I use
> > > schedtool while capturing to set ffmpeg priority to SCHED_RR, that me=
ans
> > > to reserve a fixed amount of CPU to the capture process (i.e. 85%).
> > > But that doesn't seem to help a lot.
> >
> > IMHO the recorder application should not ever receive corrupted
> > frames, no matter what priority it is running at, unless the driver
> > is buggy or the hardware is broken.
> > In the first km versions it could indeed overwrite (refill)
> > buffer(s) with newer frames while user process was reading them,
> > sort of corrupting data. I suppose current version doesn't have
> > this "feature" anymore. Or does it?
>=20
> It still does, however in neither case would you see "displaced frames"
> unless there is an error in km.
>=20
>=20
> > BTW, some kind of displacement sometimes happens here if tv signal
> > is bad quality, but this is then very well visible just without
> > recording and is completely unrelated of km.
>=20
> Yes. Francisco - does the displacement change from frame to frame often or
> not ?
>=20
>                    best
>=20
>                       Vladimir Dergachev
>=20
> > --
> > Best regards,
> >  Nikolai Zhubr
> >
> > > I'm going put into practice all your advices before next Sunday, and =
keep
> > > you informed.
> >
> > > Best regards
> >
> >
> > > -------------------------------------------------------
> > > This SF.Net email is sponsored by: IBM Linux Tutorials
> > > Free Linux tutorial presented by Daniel Robbins, President and CEO of
> > > GenToo technologies. Learn everything from fundamentals to system
> > > administration.http://ads.osdn.com/?ad_id=3D1470&alloc_id=3D3638&op=
=3Dclick
> > > _______________________________________________
> > > Gatos-devel mailing list
> > > Gat...@li...
> > > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> >
> >
> >
> >
> > -------------------------------------------------------
> > This SF.Net email is sponsored by: IBM Linux Tutorials
> > Free Linux tutorial presented by Daniel Robbins, President and CEO of
> > GenToo technologies. Learn everything from fundamentals to system
> > administration.http://ads.osdn.com/?ad_id=3D1470&alloc_id=3D3638&op=3Dc=
lick
> > _______________________________________________
> > Gatos-devel mailing list
> > Gat...@li...
> > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> >
>=20
>=20
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=3D1470&alloc_id=3D3638&op=3Dcli=
ck
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>=20

Nikolai:

On Sun, 7 Mar 2004 03:03:55 +0300
Nikolai Zhubr <s0...@ho...> wrote:

> Hi Francisco,
> Saturday, 06 March, 2004, 21:37:31, you wrote:
> [...]
> > To summarize, the problem of displaced frames is affecting a "low"
> > bandwidth system as well as a "high" bandwidth one, that have very
> > different chipsets and processors, and we have to research more about
> > the cause. An it is affecting equally Mach 64 and Rage 128 (+ Rage
> > Theatre) chipsets
> 
> In case your setup with mach64 card is still available and
> running some 2.4.xx kernel, could you please try this:
> - download http://zhubr.tamb.ru/km-cvs-hackB.tgz
> - unpack it, compile and load km_api_drv.o and km_drv.o
> - open tv window with xawtv (or avview or whatever you like)
> - dd if=/dev/video0 of=/dev/null bs=<your_frame_size>
>   (IIRC, <your_frame_size> = 768*288*2 for PAL)
> - don't close tv-window, leave it alone for several minutes
>   (or until it locks up)
> - if it doesn't yet locks up, let it go on dd-ing and do
>   something else concurrently (like starting other programs,
>   moving other windows around, scrolling something etc..)


I have set up again my AIW PRO PAL workstation with VIA 694T and Nehemiah
1GHz CPU. Then I have been able to build your modified km version and make
some tests.
Interestingly, your version cant be used along with ffmpeg.
Also, as I writed some time ago, once your km is loaded, I can't
right-click the video overlay window to display the menu. But I can do
this in xawtv, that was hence used for the tests as overlay displayer.

Used with test4.sh, it replies:

bash-2.05b# sh test4.sh
ioctl: VIDIOCGCHAN(channel=0;name="";tuners=0;flags=0x0
[];type=unknown;norm=0):
 Invalid argument
ioctl: VIDIOCSCHAN(channel=0;name="";tuners=0;flags=0x0
[];type=unknown;norm=0):
 Invalid argument
ioctl: VIDIOCSCHAN(channel=0;name="";tuners=0;flags=0x0
[];type=unknown;norm=1):
 Invalid argument
ioctl: VIDIOCSCHAN(channel=0;name="";tuners=0;flags=0x0
[];type=unknown;norm=2):
 Invalid argument
ioctl: VIDIOCSCHAN(channel=0;name="";tuners=0;flags=0x0
[];type=unknown;norm=3):
 Invalid argument
ioctl: VIDIOCSCHAN(channel=0;name="";tuners=0;flags=0x0
[];type=unknown;norm=0):
 Invalid argument
ioctl: VIDIOCGFBUF(base=(nil);height=0;width=0;depth=0;bytesperline=0):
Invalid 
argument
ioctl: VIDIOCGCHAN(channel=0;name="";tuners=0;flags=0x0
[];type=unknown;norm=0):
 Invalid argument
ioctl: VIDIOCGCHAN(channel=0;name="";tuners=0;flags=0x0
[];type=unknown;norm=0):
 Invalid argument
ioctl: VIDIOCSCHAN(channel=0;name="";tuners=0;flags=0x0
[];type=unknown;norm=0):
 Invalid argument
ioctl: VIDIOCSCHAN(channel=0;name="";tuners=0;flags=0x0
[];type=unknown;norm=1):
 Invalid argument
ioctl: VIDIOCSCHAN(channel=0;name="";tuners=0;flags=0x0
[];type=unknown;norm=2):
 Invalid argument
ioctl: VIDIOCSCHAN(channel=0;name="";tuners=0;flags=0x0
[];type=unknown;norm=3):
 Invalid argument
ioctl: VIDIOCSCHAN(channel=0;name="";tuners=0;flags=0x0
[];type=unknown;norm=0):
 Invalid argument
ioctl: VIDIOCGFBUF(base=(nil);height=0;width=0;depth=0;bytesperline=0):
Invalid 
argument
ioctl: VIDIOCGCHAN(channel=0;name="";tuners=0;flags=0x0
[];type=unknown;norm=0):
 Invalid argument
invalid value for input: svideo
valid choices for "input": ""
Fatal: grab device does not support suitable format
Could not find video grab device
gv4l.wav: Unknown format
gv4l.wav: Unknown format
1+0 records in
1+0 records out

But I was able to patch (with your patch) and compile mp1e-1.9.4

This time the output was more hopeful:

bash-2.05b# sh test3.sh
Hint: you can write -n h:m[:s] instead of -n 3000
Using SSE optimized routines.
Opened ALSA 0.9 PCM device 'default' (VIA 82C686A/B rev50)
Bounds buffer = 16384 fr; period = 8192 fr; chunk = 8192 B, 0.042667 s
Opened /dev/video0 ('Km')
Apparently the device has no tuner
Failed to query current video input of /dev/video0 (VIDIOCGCHAN),
VIDIOCGTUNER didn't work either to determine the current
video standard. Will try user specified frame rate 25.000000 Hz.
Current video standard #3 unknown, will assume PAL/SECAM.
Source frame rate is 25.000000 Hz.
Assuming kernel jiffies are 100 Hz.
Using mmap interface, 4 capture buffers granted.
Mapping capture buffers
Grab 1st frame and set capture format and dimensions.
Image filter 8, palette 15 not accepted.
Image filter 1, palette 15 not accepted.
WARNING: the driver refused to grab 384x288, trying 768x288 instead.
Image filter 2, palette 8 not accepted.
Filter 'YUYV 4:2:2 fastest'
Allocated 24 bounce buffers.
V4L capture thread launched
Audio compression 48.0 kHz stereo at 160 kbits/s (9.6 : 1)
Psychoacoustic threshold table #2
Macroblocks 24 x 18
Video compression 384 x 288, 25.0 frames/s at 2.30 Mbits/s (14.4 : 1)
Motion compensation disabled
Audio compression thread launched
Video compression thread launched
Video: End of filestem load 30.9 %  
2:00 (35.5 MB), system load 30.9 %  
Audio: End of file
2:00 (35.5 MB), system load 30.9 %  
mp1e: Done.
1+0 records in
1+0 records out


Results:

1. When I start xawtv or avview, the overlay winodow has less pixels that
the real capture window. I mean that I can't see the hole video
picture/frame, only one part is displayed in the window. If I choose PAL
encode and S-VIDEO input and close xawtv settings menu, the video overlay
shows more pixels (The hole PAL frame, altough using the same window size,
and therefore I suppose that prevousy only part of the frame was shown and
scaled to fill the window). Well, all this explanation is to say that the
captured video doesn't cover the full frame size.  I think that is cropped
in the X axis (horizontally, perhaps 640 pixels instead of 720??). I don't
know if it is also vertically croped because that effect is less
noticiable.
2. There are no displaced frames. WOW!
3. There appear to be some dropped frames as the movement is sometimes
choppy, but the audio is more or less on sync.
4. This time, no machine lock-up.

Conclusions:

1. Something in your code works avoiding the displacement effect in actual
km code.
2. ffmpeg doesn't work with it.
2. There seem to be dropped frames and some part of the frme is cropped
out (problem with PAL settings?).

> - let me know if it locks up, and also look through your
>   syslog for some relevant messages from km. If you
>   are lucky and the box doesn't lockup then ctrl-C dd will

These were the kernel messages during both tests:

Linux video capture interface: v1.00
Kmultimedia API module version alpha-3.0 loaded
Kmultimedia module version alpha-2.0-unofficial_hack_B loaded
Device ATI Technologies Inc 3D Rage Pro AGP 1X/2X 01:00.0 (0x1002:0x4742)
corresponds to /dev/video0
km: total frames: 0, overrun: 0
km: total frames: 0, overrun: 0
km: total frames: 0, overrun: 1
km: total frames: 0, overrun: 0
km: total frames: 0, overrun: 0
km: total frames: 0, overrun: 1
km: km_vgrab() invalid size 384*288
km: km_vgrab() invalid size 384*288
km: km_vgrab() invalid size 384*288
km: km_vgrab() invalid size 384*288
km: total frames: 0, overrun: 0
km: total frames: 1, overrun: 0
km: total frames: 0, overrun: 0
km: total frames: 0, overrun: 0
km: km_vgrab() invalid size 384*576
km: km_vgrab() invalid size 384*288
km: mach64 FIFO locked up
km: mach64 FIFO locked up
km: total frames: 3054, overrun: 1480
km: total frames: 1, overrun: 0
bash-2.05b# 

>   tell you how many frames were actually transferred and
>   it should be approximately 25*seconds_elapsed.
> 


Here I have to say that with latest km code and ffmpeg, it captures just
the frames I request (300 for 120 second test). Bu that doesn't prevent
from displacement effect.

Finally, the test4.sh an test3.sh scripts:

#!/bin/sh
# test3.sh
mp1e -vv -J 25 -C p3 -f 25 -c /dev/video0 -a 0 -S 48.0 -r 6,30 -p \
alsa/default -s384x288 -n 3000 -o res3.mpg
dd if=/dev/video0 of=frame3.yuv bs=221184 count=1
exit $?

#!/bin/sh
#test4.sh
bitrate=795
abitrate=96
vbitrate=`echo "$bitrate - $abitrate" | bc`
v4lctl capture on
v4lctl setinput svideo
aumix -l R -l 50
ffmpeg -t 00:02:00 -tvstd pal -y -f wav -ad /dev/dsp0 -ac 2 -acodec \
pcm_s16le  -ar 48000 "gv4l.wav" -vd /dev/video0 -s 384x288  -r 25 \
"gv4l.yuv"
for n in 1 2
  do
  ffmpeg -i  "gv4l.wav" -s 384x288 -r 25 -i gv4l.yuv -y -pass $n  -ac 2 \
-acodec mp2 -ab ${abitrate} -b ${vbitrate} -f mpeg -vcodec mpeg1video \
res4.mpg
  done
dd if=/dev/video0 of=frame4.yuv bs=221184 count=1
exit $?

On Fri, 16 Apr 2004, Diaconescu Bogdan-BDIACON1 wrote:

> Thanks for the replay Vladimir,
>
> I'll give a try to find the datasheets for FI1236/W. At the first search on
> the web, no one seems to have datasheets for it.
>
> Cold you send me the datasheets for R200 and, if you have, for RV350 and/or
> 300?

Sorry, these are only available under NDA.

                      best

                        Vladimir Dergachev

>
> Regards,
> Bogdan
>
>
>
> -----Original Message-----
> From: Vladimir Dergachev [mailto:vo...@mi...]
> Sent: Friday, April 16, 2004 3:10 AM
> To: Diaconescu Bogdan-BDIACON1
> Cc: gat...@li...
> Subject: Re: [GATOS]Reiterate the AIW9600Pro question
>
>
>
>
> On Thu, 15 Apr 2004, Diaconescu Bogdan-BDIACON1 wrote:
>
> > I'm still trying to find the answer for one question I sent before
> > regarding the AIW9600Pro
> >
> > I went through the ati.2 mainline sources and I 've seen that RV350 is
> > not supported and also THEATRE 200 is known but not supported yet. Are
> > there any near future plans to support the above chips?
>
> Support is being worked on.. but slowly.
>
> >
> > Another question: Is there any documentation available, datasheets
> > for the chips I am interested in? Did ATI provided the specifications
> > to the developers in the case someone want to contribute?
>
> Yes, we do have the documentation - at least part of it. Namely, we have the
> docs for Rage Theatre 200, but not for the tuner FI1236/W made by Philips.
> It is curious that neither ATI nor Philips are able to come up with
> datasheet for it.
>
> One thing you could do to help is to try to dig it up on the web (maybe it
> became available since the last attempt) and also try to talk to local
> Philips distributors.
>
> Note, that we need the datasheet for FI1236/W  - not the earlier FI1236MK2
> which is available on the web.
>
> I am not 100% certain these tuners are different, but I was not successful
> trying to talk to W edition, so, perhaps, they are.
>
>                         best
>
>                            Vladimir Dergachev
>
> >
> > Regards,
> > Bogdan
> >
> >
> >
> >
> >
> >
> >
> > -------------------------------------------------------
> > This SF.Net email is sponsored by: IBM Linux Tutorials
> > Free Linux tutorial presented by Daniel Robbins, President and CEO of
> > GenToo technologies. Learn everything from fundamentals to system
> > administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> > _______________________________________________
> > Gatos-devel mailing list
> > Gat...@li...
> > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> >
>