Re: [PATCH for 4.5] media.h: use hex values for the range offsets, move connectors base up.

[Hans Verkuil <hverkuil@xs4all.nl>]

On 03/03/16 11:27, Laurent Pinchart wrote:
> Hi Hans,
> 
> On Thursday 03 March 2016 11:18:30 Hans Verkuil wrote:
>> On 03/03/16 11:12, Laurent Pinchart wrote:
>>> On Thursday 03 March 2016 11:03:02 Hans Verkuil wrote:
>>>> On 03/03/16 10:52, Laurent Pinchart wrote:
>>>>> On Monday 29 February 2016 09:02:47 Hans Verkuil wrote:
>>>>>> Make the base offset hexadecimal to simplify debugging since the base
>>>>>> addresses are hex too.
>>>>>>
>>>>>> The offsets for connectors is also changed to start after the
>>>>>> 'reserved' range 0x10000-0x2ffff.
>>>>>>
>>>>>> Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
>>>>>>
>>>>>> diff --git a/include/uapi/linux/media.h b/include/uapi/linux/media.h
>>>>>> index 95e126e..79960ae 100644
>>>>>> --- a/include/uapi/linux/media.h
>>>>>> +++ b/include/uapi/linux/media.h
>>>>>> @@ -66,17 +66,17 @@ struct media_device_info {
>>>>>>  /*
>>>>>>   * DVB entities
>>>>>>   */
>>>>>> -#define MEDIA_ENT_F_DTV_DEMOD		(MEDIA_ENT_F_BASE + 1)
>>>>>> -#define MEDIA_ENT_F_TS_DEMUX		(MEDIA_ENT_F_BASE + 2)
>>>>>> -#define MEDIA_ENT_F_DTV_CA		(MEDIA_ENT_F_BASE + 3)
>>>>>> -#define MEDIA_ENT_F_DTV_NET_DECAP	(MEDIA_ENT_F_BASE + 4)
>>>>>> +#define MEDIA_ENT_F_DTV_DEMOD		(MEDIA_ENT_F_BASE + 0x00001)
>>>>>> +#define MEDIA_ENT_F_TS_DEMUX		(MEDIA_ENT_F_BASE + 0x00002)
>>>>>> +#define MEDIA_ENT_F_DTV_CA		(MEDIA_ENT_F_BASE + 0x00003)
>>>>>> +#define MEDIA_ENT_F_DTV_NET_DECAP	(MEDIA_ENT_F_BASE + 0x00004)
>>>>>>
>>>>>>  /*
>>>>>>   * I/O entities
>>>>>>   */
>>>>>> -#define MEDIA_ENT_F_IO_DTV		(MEDIA_ENT_F_BASE + 1001)
>>>>>> -#define MEDIA_ENT_F_IO_VBI		(MEDIA_ENT_F_BASE + 1002)
>>>>>> -#define MEDIA_ENT_F_IO_SWRADIO		(MEDIA_ENT_F_BASE + 1003)
>>>>>> +#define MEDIA_ENT_F_IO_DTV		(MEDIA_ENT_F_BASE + 0x01001)
>>>>>> +#define MEDIA_ENT_F_IO_VBI		(MEDIA_ENT_F_BASE + 0x01002)
>>>>>> +#define MEDIA_ENT_F_IO_SWRADIO		(MEDIA_ENT_F_BASE + 0x01003)
>>>>>>
>>>>>>  /*
>>>>>>   * Analog TV IF-PLL decoders
>>>>>> @@ -84,23 +84,23 @@ struct media_device_info {
>>>>>>   * It is a responsibility of the master/bridge drivers to create links
>>>>>>   * for MEDIA_ENT_F_IF_VID_DECODER and MEDIA_ENT_F_IF_AUD_DECODER.
>>>>>>   */
>>>>>> -#define MEDIA_ENT_F_IF_VID_DECODER	(MEDIA_ENT_F_BASE + 2001)
>>>>>> -#define MEDIA_ENT_F_IF_AUD_DECODER	(MEDIA_ENT_F_BASE + 2002)
>>>>>> +#define MEDIA_ENT_F_IF_VID_DECODER	(MEDIA_ENT_F_BASE + 0x02001)
>>>>>> +#define MEDIA_ENT_F_IF_AUD_DECODER	(MEDIA_ENT_F_BASE + 0x02002)
>>>>>>
>>>>>>  /*
>>>>>>   * Audio Entity Functions
>>>>>>   */
>>>>>> -#define MEDIA_ENT_F_AUDIO_CAPTURE	(MEDIA_ENT_F_BASE + 3000)
>>>>>> -#define MEDIA_ENT_F_AUDIO_PLAYBACK	(MEDIA_ENT_F_BASE + 3001)
>>>>>> -#define MEDIA_ENT_F_AUDIO_MIXER		(MEDIA_ENT_F_BASE + 3002)
>>>>>> +#define MEDIA_ENT_F_AUDIO_CAPTURE	(MEDIA_ENT_F_BASE + 0x03000)
>>>>>
>>>>> Why does this one start at 0x*000 while the others start at 0x*0001 ? I
>>>>> know that the problem predates your patch.
>>>>
>>>> I hadn't noticed. It is my personal preference not to start with 0.
>>>> But it is not needed in this case, at least not today.
>>>>
>>>> I think starting with 1 will help if we ever want to do AND operations
>>>> on the ID. Then it is nice that the lower 16 bits can't be 0 for valid
>>>> IDs.
>>>
>>> I'm fine starting at 1, would you like to resubmit this patch to change
>>> that ?
>>>
>>>>>> +#define MEDIA_ENT_F_AUDIO_PLAYBACK	(MEDIA_ENT_F_BASE + 0x03001)
>>>>>> +#define MEDIA_ENT_F_AUDIO_MIXER		(MEDIA_ENT_F_BASE + 0x03002)
>>>>>>
>>>>>>  /*
>>>>>>  
>>>>>>   * Connectors
>>>>>>   */
>>>>>>  
>>>>>>  /* It is a responsibility of the entity drivers to add connectors and
>>>>>>  links
>>>>>>
>>>>>> */ -#define MEDIA_ENT_F_CONN_RF		(MEDIA_ENT_F_BASE + 10001)
>>>>>> -#define MEDIA_ENT_F_CONN_SVIDEO		(MEDIA_ENT_F_BASE + 10002)
>>>>>> -#define MEDIA_ENT_F_CONN_COMPOSITE	(MEDIA_ENT_F_BASE + 10003)
>>>>>> +#define MEDIA_ENT_F_CONN_RF		(MEDIA_ENT_F_BASE + 0x30001)
>>>>>
>>>>> Anything wrong with 0x4xxx ?
>>>>
>>>> Possibly overkill, but Sakari preferred to make more generous use of the
>>>> 32 bit space. And since there may potentially be a lot of connector types
>>>> I thought I gave it plenty of space. Of course, if we ever need that
>>>> many, then something is seriously wrong...
>>>
>>> And you'd need space *after* the existing connector IDs, not before. Using
>>> 0x30000 instead of 0x4000 has the effect of reserving plenty of space for
>>> audio functions, not for connectors. I think 0x4000 should be fine.
>>
>> Huh? Connectors start at 0x30000, audio entities are 0x03000. I think you
>> got confused by that. Connectors can now go from 0x30000-0xffffffff :-)
>> Definitely more space than 0x4000-0xffff.
> 
> But less than 0x4000-0xffffffff :-) There's no information about how the range 
> is supposed to be split. And 4096 connector functions should be more than 
> enough.

The range 0x10000-0x2ffff is blocked due to the old MCv1 defines.
Hmm, that could be stated more clearly in the code.

Regards,

	Hans

> 
> On a side note, writing "connector functions" really doesn't feel right. We're 
> not there yet, this API isn't correct.
> 
>>>>>> +#define MEDIA_ENT_F_CONN_SVIDEO		(MEDIA_ENT_F_BASE + 0x30002)
>>>>>> +#define MEDIA_ENT_F_CONN_COMPOSITE	(MEDIA_ENT_F_BASE + 0x30003)
>>>>>>
>>>>>>  /*
>>>>>>   * Don't touch on those. The ranges MEDIA_ENT_F_OLD_BASE and
>