RE: [RFC 4/7] ARM: DaVinci: DM646x Video: Defintions for standards supported by display

2009-03-19 Thread Subrahmanya, Chaithrika
Hi Hans,

As of now we will provide support for SD, soon we will add support for HD too. 
I am working on the review comments and will submit the revised patches soon.
This time the patch set will be divided into two
Set 1 - support for ADV7343 and THS7303 drivers.
Set 2 - VPIF display driver related patches

Thanks,
Chaithrika
 

From: Hans Verkuil [hverk...@xs4all.nl]
Sent: Thursday, March 19, 2009 4:05 AM
To: davinci-linux-open-source@linux.davincidsp.com; Subrahmanya, Chaithrika
Cc: linux-me...@vger.kernel.org
Subject: Re: [RFC 4/7] ARM: DaVinci: DM646x Video: Defintions for standards 
supported by display

Hi Chaithrika,

Just a quick note: getting these digital TV standards sorted out is the main
blocking issue for the DM646x. Without this nothing can get merged. In
addition, extending the V4L2 API will take time and usually takes several
review cycles. So I recommend that this gets a high priority.

I'm not sure if the importance of this came across in my initial review. My
apologies if that wasn't clear.

Regards,

Hans

On Saturday 14 March 2009 14:27:09 Hans Verkuil wrote:
 On Friday 13 March 2009 10:01:37 chaithr...@ti.com wrote:
  From: Chaithrika U S chaithr...@ti.com
 
  Add defintions for Digital TV Standards supported by display driver
 
  Signed-off-by: Chaithrika U S chaithr...@ti.com
  ---
  Applies to v4l-dvb repository located at
  http://linuxtv.org/hg/v4l-dvb/rev/1fd54a62abde
 
   include/linux/videodev2.h |   12 
   1 files changed, 12 insertions(+), 0 deletions(-)
 
  diff --git a/include/linux/videodev2.h b/include/linux/videodev2.h
  index 7a8eafd..df4a622 100644
  --- a/include/linux/videodev2.h
  +++ b/include/linux/videodev2.h
  @@ -704,6 +704,18 @@ typedef __u64 v4l2_std_id;
   #define V4L2_STD_ALL(V4L2_STD_525_60   |\
   V4L2_STD_625_50)
 
  +#define V4L2_STD_720P_60((v4l2_std_id)(0x0001ULL))
  +#define V4L2_STD_1080I_30   ((v4l2_std_id)(0x0002ULL))
  +#define V4L2_STD_1080I_25   ((v4l2_std_id)(0x0004ULL))
  +#define V4L2_STD_480P_60((v4l2_std_id)(0x0008ULL))
  +#define V4L2_STD_576P_50((v4l2_std_id)(0x0010ULL))
  +#define V4L2_STD_720P_25((v4l2_std_id)(0x0020ULL))
  +#define V4L2_STD_720P_30((v4l2_std_id)(0x0040ULL))
  +#define V4L2_STD_720P_50((v4l2_std_id)(0x0080ULL))
  +#define V4L2_STD_1080P_25   ((v4l2_std_id)(0x0100ULL))
  +#define V4L2_STD_1080P_30   ((v4l2_std_id)(0x0200ULL))
  +#define V4L2_STD_1080P_24   ((v4l2_std_id)(0x0400ULL))
  +
   struct v4l2_standard {
  __u32index;
  v4l2_std_id  id;

 This requires an RFC. I'm not convinced that using v4l2_std_id is the
 best approach. If you look at the CEA-861-D you see a lot more standards
 (and E adds even more). Not to mention that when the DM646x is used in
 combination with e.g. an FPGA then it should be possible to supply the
 driver with custom timings as well. The v4l2_std_id type was never
 designed for that.

 My gut feeling is that v4l2_std_id should be effectively frozen and used
 for the old TV broadcast standards only, and that a new API should be
 created to setup these digital formats.

 I've discussed this with Manju in the past, and I suggest that TI should
 make a proposal in the form of an RFC that we can then discuss on the
 mailinglists. One of the disadvantages of being the first who needs these
 HDTV formats. The advantage of being the first is that you can design it
 yourself, of course!

 Regards,

   Hans



--
Hans Verkuil - video4linux developer - sponsored by 
TANDBERG___
Davinci-linux-open-source mailing list
Davinci-linux-open-source@linux.davincidsp.com
http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source


Re: [RFC 4/7] ARM: DaVinci: DM646x Video: Defintions for standards supported by display

2009-03-18 Thread Hans Verkuil
Hi Chaithrika,

Just a quick note: getting these digital TV standards sorted out is the main 
blocking issue for the DM646x. Without this nothing can get merged. In 
addition, extending the V4L2 API will take time and usually takes several 
review cycles. So I recommend that this gets a high priority.

I'm not sure if the importance of this came across in my initial review. My 
apologies if that wasn't clear.

Regards,

Hans

On Saturday 14 March 2009 14:27:09 Hans Verkuil wrote:
 On Friday 13 March 2009 10:01:37 chaithr...@ti.com wrote:
  From: Chaithrika U S chaithr...@ti.com
 
  Add defintions for Digital TV Standards supported by display driver
 
  Signed-off-by: Chaithrika U S chaithr...@ti.com
  ---
  Applies to v4l-dvb repository located at
  http://linuxtv.org/hg/v4l-dvb/rev/1fd54a62abde
 
   include/linux/videodev2.h |   12 
   1 files changed, 12 insertions(+), 0 deletions(-)
 
  diff --git a/include/linux/videodev2.h b/include/linux/videodev2.h
  index 7a8eafd..df4a622 100644
  --- a/include/linux/videodev2.h
  +++ b/include/linux/videodev2.h
  @@ -704,6 +704,18 @@ typedef __u64 v4l2_std_id;
   #define V4L2_STD_ALL(V4L2_STD_525_60   |\
   V4L2_STD_625_50)
 
  +#define V4L2_STD_720P_60((v4l2_std_id)(0x0001ULL))
  +#define V4L2_STD_1080I_30   ((v4l2_std_id)(0x0002ULL))
  +#define V4L2_STD_1080I_25   ((v4l2_std_id)(0x0004ULL))
  +#define V4L2_STD_480P_60((v4l2_std_id)(0x0008ULL))
  +#define V4L2_STD_576P_50((v4l2_std_id)(0x0010ULL))
  +#define V4L2_STD_720P_25((v4l2_std_id)(0x0020ULL))
  +#define V4L2_STD_720P_30((v4l2_std_id)(0x0040ULL))
  +#define V4L2_STD_720P_50((v4l2_std_id)(0x0080ULL))
  +#define V4L2_STD_1080P_25   ((v4l2_std_id)(0x0100ULL))
  +#define V4L2_STD_1080P_30   ((v4l2_std_id)(0x0200ULL))
  +#define V4L2_STD_1080P_24   ((v4l2_std_id)(0x0400ULL))
  +
   struct v4l2_standard {
  __u32index;
  v4l2_std_id  id;

 This requires an RFC. I'm not convinced that using v4l2_std_id is the
 best approach. If you look at the CEA-861-D you see a lot more standards
 (and E adds even more). Not to mention that when the DM646x is used in
 combination with e.g. an FPGA then it should be possible to supply the
 driver with custom timings as well. The v4l2_std_id type was never
 designed for that.

 My gut feeling is that v4l2_std_id should be effectively frozen and used
 for the old TV broadcast standards only, and that a new API should be
 created to setup these digital formats.

 I've discussed this with Manju in the past, and I suggest that TI should
 make a proposal in the form of an RFC that we can then discuss on the
 mailinglists. One of the disadvantages of being the first who needs these
 HDTV formats. The advantage of being the first is that you can design it
 yourself, of course!

 Regards,

   Hans



-- 
Hans Verkuil - video4linux developer - sponsored by TANDBERG

___
Davinci-linux-open-source mailing list
Davinci-linux-open-source@linux.davincidsp.com
http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source


Re: [RFC 4/7] ARM: DaVinci: DM646x Video: Defintions for standards supported by display

2009-03-14 Thread Hans Verkuil
On Friday 13 March 2009 10:01:37 chaithr...@ti.com wrote:
 From: Chaithrika U S chaithr...@ti.com
 
 Add defintions for Digital TV Standards supported by display driver
 
 Signed-off-by: Chaithrika U S chaithr...@ti.com
 ---
 Applies to v4l-dvb repository located at
 http://linuxtv.org/hg/v4l-dvb/rev/1fd54a62abde
 
  include/linux/videodev2.h |   12 
  1 files changed, 12 insertions(+), 0 deletions(-)
 
 diff --git a/include/linux/videodev2.h b/include/linux/videodev2.h
 index 7a8eafd..df4a622 100644
 --- a/include/linux/videodev2.h
 +++ b/include/linux/videodev2.h
 @@ -704,6 +704,18 @@ typedef __u64 v4l2_std_id;
  #define V4L2_STD_ALL(V4L2_STD_525_60 |\
V4L2_STD_625_50)
  
 +#define V4L2_STD_720P_60((v4l2_std_id)(0x0001ULL))
 +#define V4L2_STD_1080I_30   ((v4l2_std_id)(0x0002ULL))
 +#define V4L2_STD_1080I_25   ((v4l2_std_id)(0x0004ULL))
 +#define V4L2_STD_480P_60((v4l2_std_id)(0x0008ULL))
 +#define V4L2_STD_576P_50((v4l2_std_id)(0x0010ULL))
 +#define V4L2_STD_720P_25((v4l2_std_id)(0x0020ULL))
 +#define V4L2_STD_720P_30((v4l2_std_id)(0x0040ULL))
 +#define V4L2_STD_720P_50((v4l2_std_id)(0x0080ULL))
 +#define V4L2_STD_1080P_25   ((v4l2_std_id)(0x0100ULL))
 +#define V4L2_STD_1080P_30   ((v4l2_std_id)(0x0200ULL))
 +#define V4L2_STD_1080P_24   ((v4l2_std_id)(0x0400ULL))
 +
  struct v4l2_standard {
   __u32index;
   v4l2_std_id  id;

This requires an RFC. I'm not convinced that using v4l2_std_id is the best
approach. If you look at the CEA-861-D you see a lot more standards (and E
adds even more). Not to mention that when the DM646x is used in combination
with e.g. an FPGA then it should be possible to supply the driver with
custom timings as well. The v4l2_std_id type was never designed for that.

My gut feeling is that v4l2_std_id should be effectively frozen and used for
the old TV broadcast standards only, and that a new API should be created
to setup these digital formats.

I've discussed this with Manju in the past, and I suggest that TI should make
a proposal in the form of an RFC that we can then discuss on the mailinglists.
One of the disadvantages of being the first who needs these HDTV formats. The
advantage of being the first is that you can design it yourself, of course!

Regards,

Hans

-- 
Hans Verkuil - video4linux developer - sponsored by TANDBERG

___
Davinci-linux-open-source mailing list
Davinci-linux-open-source@linux.davincidsp.com
http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source