Re: [PATCH v5 2/7] v4l: Use full 32 bits for buffer flags

[Sakari Ailus <sakari.ailus@iki.fi>]

Hans Verkuil wrote:
> On 02/15/2014 09:53 PM, Sakari Ailus wrote:
>> The buffer flags field is 32 bits but the defined only used 16. This is
>> fine, but as more than 16 bits will be used in the very near future, define
>> them as 32-bit numbers for consistency.
>>
>> Signed-off-by: Sakari Ailus <sakari.ailus@iki.fi>
>> ---
>>  Documentation/DocBook/media/v4l/io.xml |   30 ++++++++++++-------------
>>  include/uapi/linux/videodev2.h         |   38 +++++++++++++++++++-------------
>>  2 files changed, 38 insertions(+), 30 deletions(-)
>>
>> diff --git a/Documentation/DocBook/media/v4l/io.xml b/Documentation/DocBook/media/v4l/io.xml
>> index 8facac4..46d24b3 100644
>> --- a/Documentation/DocBook/media/v4l/io.xml
>> +++ b/Documentation/DocBook/media/v4l/io.xml
> 
> <snip>
> 
>> @@ -1115,7 +1115,7 @@ in which case caches have not been used.</entry>
>>  	  </row>
>>  	  <row>
>>  	    <entry><constant>V4L2_BUF_FLAG_TIMESTAMP_COPY</constant></entry>
>> -	    <entry>0x4000</entry>
>> +	    <entry>0x00004000</entry>
>>  	    <entry>The CAPTURE buffer timestamp has been taken from the
>>  	    corresponding OUTPUT buffer. This flag applies only to mem2mem devices.</entry>
>>  	  </row>
> 
> Should we add here that if TIMESTAMP_COPY is set and the TIMECODE flag is set,
> then drivers should copy the TIMECODE struct as well? This is happening already
> in various drivers and I think that is appropriate. Although to be honest nobody
> is actually using the timecode struct, but we plan to hijack that for hardware
> timestamps in the future anyway.

Is there a single driver which uses the timecode field? The fact is that
many m2m drivers copy it but that's probably mostly copying what one of
them happened to do by accident. :-)

-- 
Sakari Ailus
sakari.ailus@iki.fi