Hello,
Yet another patch ! This allows xawtv to take snapshots (calling
read()), and xdtv works now
xdtv does VIDIC_STREAMN before making VIDIC_QBUF; this patch removes
a control in VIDIC_STREAMN that was returning EINVAL if no buffer were
queued.
Furthermore, xdtv measures fps in debug mode, and tells 25fps in SECAM mode.
So the driver dynamics is correct, there is a little tuning to do in
saa711x programming to get the dropped frames.
Regards,
Thierry MERLE
Note that this patch requires that my previous patches are applied on
usbvision_v4l2.
The contents of the patch:
- implement read() entry point that works with linux list.h
- rework of
- VIDIC_STREAMN : allows streaming whereas there is no queued buffer
(xdtv does VIDIC_STREAMN before VIDIC_QBUFs)
Signed-off-by: Thierry MERLE <thierry.merle (AT) free (DOT) fr>

Hi Thierry,
Em Dom, 2006-11-12 22:22 +0100, Thierry MERLE escreveu:
Note that this patch requires that my previous patches are applied on
usbvision_v4l2.
, I've applied 3 pending patches at the tree. Please check if I didn't
forgot something ;)
Cheers,
Mauro.

Hi Mauro,
all is perfect, thanks!
With these patches I can run xawtv/xdtv/kdetv successfully

So, here are the remaining tasks to make this driver fully functional,
ordered by priority:
- find the bug about 12 fps: I proved that the NT1004 bridge provides
24fps (progressive mode), something in the chain
tda9887->saa7113->NT1004 is not well configured. Note that the original
usbvision driver was working with the v4l1 saa711x module.
- improve the i2c management: I have little problems with norm settings
(I need to launch kdetv one time before xawtv to get SECAM, tda9887
stays in PAL without this trick)
- implement the radio interface.
- implement the link with the tveeprom module. I decoded the EEPRM by
hand with the help of the tveeprom driver code.
- implement the vbi interface.

Cheers,
Thierry MERLE

Mauro Carvalho Chehab a :
Hi Thierry,

Em Dom, 2006-11-12 22:22 +0100, Thierry MERLE escreveu:

>Note that this patch requires that my previous patches are applied on
>usbvision_v4l2.
>
, I've applied 3 pending patches at the tree. Please check if I didn't
forgot something ;)

Cheers,
Mauro.

Em Seg, 2006-11-13 16:00 +0100, Thierry MERLE escreveu:
Hi Mauro,
all is perfect, thanks!
With these patches I can run xawtv/xdtv/kdetv successfully
Nice.

So, here are the remaining tasks to make this driver fully functional,
ordered by priority:
- find the bug about 12 fps: I proved that the NT1004 bridge provides
24fps (progressive mode), something in the chain
tda9887->saa7113->NT1004 is not well configured. Note that the original
usbvision driver was working with the v4l1 saa711x module.
saa711x and saa7115 send similar programming codes. I think the bug is
not at the decoder. However, we can't discard this possibility. Maybe
for usbvision, some special programming are required there.
- improve the i2c management: I have little problems with norm settings
(I need to launch kdetv one time before xawtv to get SECAM, tda9887
stays in PAL without this trick)
Hmm usbvision honours i2c broadcast?
- implement the radio interface.
- implement the link with the tveeprom module. I decoded the EEPRM by
hand with the help of the tveeprom driver code.
- implement the vbi interface.

Please add into your task list to replace UVICSREG/UVICGREG to the
proper V4L register access for debug
(). Direct register access
should be enabled only when CNFIG_VIDEADV_DEBUG is defined.

Cheers,
Mauro.

Hi,

11/14/06, Mauro Carvalho Chehab <mchehab (AT) infradead (DOT) orgwrote:
Em Seg, 2006-11-13 16:00 +0100, Thierry MERLE escreveu:
Hi Mauro,
all is perfect, thanks!
With these patches I can run xawtv/xdtv/kdetv successfully
Nice.

So, here are the remaining tasks to make this driver fully functional,
ordered by priority:
- find the bug about 12 fps: I proved that the NT1004 bridge provides
24fps (progressive mode), something in the chain
tda9887->saa7113->NT1004 is not well configured. Note that the original
usbvision driver was working with the v4l1 saa711x module.
saa711x and saa7115 send similar programming codes. I think the bug is
not at the decoder. However, we can't discard this possibility. Maybe
for usbvision, some special programming are required there.

hmm can you install the old drivers, and log the saa711x settings that are made?

cheers,
Markus

K, here are the results in the attached /var/log/messages.
The use of saa711x does not resolve my problem, so the saa7115 programs
the saa7113 correctly.

However, I parsed this file to get the interesting thing about the
saa711x initialization step :
4a >>0 0
4a >>1 8
4a >>2 c0
4a >>3 23
4a >>4 0
4a >>5 0
4a >>6 eb
4a >>7 e0
4a >>8 88
4a >>9 0
4a >>a 80
4a >>b 47
4a >>c 40
4a >>d 0
4a >>e 1
4a >>f aa
4a >>10 0
4a >>11 1c
4a >>12 1
4a >>13 0
4a >>14 0
4a >>15 0
4a >>16 0
4a >>17 0
4a >>0

I did the same thing for the saa7115 module initialization part:
4a >>1 8
4a >>2 c0
4a >>3 23
4a >>4 0
4a >>5 0
4a >>6 eb
4a >>7 e0
4a >>8 88
4a >>9 0
4a >>a 80
4a >>b 47
4a >>c 40
4a >>d 0
4a >>e 1
4a >>f aa
4a >>10 0
4a >>11 1c
4a >>12 1
4a >>13 0
4a >>15 0
4a >>16 0
4a >>17 0
4a >>40 20
4a >>41 ff
4a >>42 ff
4a >>43 ff
4a >>44 ff
4a >>45 ff
4a >>46 ff
4a >>47 ff
4a >>48 ff
4a >>49 ff
4a >>4a ff
4a >>4b ff
4a >>4c ff
4a >>4d ff
4a >>4e ff
4a >>4f ff
4a >>50 ff
4a >>51 ff
4a >>52 ff
4a >>53 ff
4a >>54 ff
4a >>55 ff
4a >>56 ff
4a >>57 ff
4a >>58 40
4a >>59 47
4a >>5b 83
4a >>5e 35
4a >>2 84
4a >>15 3
4a >>16 11
4a >>17 9c
4a >>8 68
4a >>e 7
4a >>5a 6
4a >>e 7
Regards,
Thierry MERLE

Markus Rechberger a :
Hi,

11/14/06, Mauro Carvalho Chehab <mchehab (AT) infradead (DOT) orgwrote:
>Em Seg, 2006-11-13 16:00 +0100, Thierry MERLE escreveu:
>Hi Mauro,
>all is perfect, thanks!
>With these patches I can run xawtv/xdtv/kdetv successfully
>Nice.
>>
>So, here are the remaining tasks to make this driver fully functional,
>ordered by priority:
>- find the bug about 12 fps: I proved that the NT1004 bridge provides
>24fps (progressive mode), something in the chain
>tda9887->saa7113->NT1004 is not well configured. Note that the
>original
>usbvision driver was working with the v4l1 saa711x module.
>saa711x and saa7115 send similar programming codes. I think the bug is
>not at the decoder. However, we can't discard this possibility. Maybe
>for usbvision, some special programming are required there.
>
hmm can you install the old drivers, and log the saa711x settings that
are made?

cheers,
Markus

Doh! Sorry, the output for the saa7115 module initialization step is
_wrong_ !
I did not remember I patched the saa7113_init in saa7115.c with the
saa711x init structure yesterday.

So here is the true I2C debug output for the saa7115:
4a >>1 8
4a >>2 c2
4a >>3 30
4a >>4 0
4a >>5 0
4a >>6 89
4a >>7 d
4a >>8 88
4a >>9 1
4a >>a 80
4a >>b 47
4a >>c 40
4a >>d 0
4a >>e 1
4a >>f 2a
4a >>10 8
4a >>11 c
4a >>12 7
4a >>13 0
4a >>15 0
4a >>16 0
4a >>17 0
4a >>40 20
4a >>41 ff
4a >>42 ff
4a >>43 ff
4a >>44 ff
4a >>45 ff
4a >>46 ff
4a >>47 ff
4a >>48 ff
4a >>49 ff
4a >>4a ff
4a >>4b ff
4a >>4c ff
4a >>4d ff
4a >>4e ff
4a >>4f ff
4a >>50 ff
4a >>51 ff
4a >>52 ff
4a >>53 ff
4a >>54 ff
4a >>55 ff
4a >>56 ff
4a >>57 ff
4a >>58 40
4a >>59 47
4a >>5b 83
4a >>5e 35
4a >>2 84
4a >>15 3
4a >>16 11
4a >>17 9c
4a >>8 68
4a >>e 7
4a >>5a 6
4a >>e

Sorry for this mistake

Thierry MERLE a :
K, here are the results in the attached /var/log/messages.
The use of saa711x does not resolve my problem, so the saa7115
programs the saa7113 correctly.

However, I parsed this file to get the interesting thing about the
saa711x initialization step :
4a >>0 0
4a >>1 8
4a >>2 c0
4a >>3 23
4a >>4 0
4a >>5 0
4a >>6 eb
4a >>7 e0
4a >>8 88
4a >>9 0
4a >>a 80
4a >>b 47
4a >>c 40
4a >>d 0
4a >>e 1
4a >>f aa
4a >>10 0
4a >>11 1c
4a >>12 1
4a >>13 0
4a >>14 0
4a >>15 0
4a >>16 0
4a >>17 0
4a >>0

I did the same thing for the saa7115 module initialization part:

<deleted the wrong output>
Regards,
Thierry MERLE

Markus Rechberger a :
>Hi,
>>
>11/14/06, Mauro Carvalho Chehab <mchehab (AT) infradead (DOT) orgwrote:
Em Seg, 2006-11-13 16:00 +0100, Thierry MERLE escreveu:
Hi Mauro,
all is perfect, thanks!
With these patches I can run xawtv/xdtv/kdetv successfully
Nice.
>
So, here are the remaining tasks to make this driver fully
functional,
ordered by priority:
- find the bug about 12 fps: I proved that the NT1004 bridge provides
24fps (progressive mode), something in the chain
tda9887->saa7113->NT1004 is not well configured. Note that the
original
usbvision driver was working with the v4l1 saa711x module.
saa711x and saa7115 send similar programming codes. I think the bug is
not at the decoder. However, we can't discard this possibility. Maybe
for usbvision, some special programming are required there.
>>
>hmm can you install the old drivers, and log the saa711x settings
>that are made?
>>
>cheers,
>Markus