Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1713591 > unrolled thread
| Started by | Wenyou Yang <wenyou.yang@microchip.com> |
|---|---|
| First post | 2017-08-17 09:30 +0200 |
| Last post | 2017-08-24 08:50 +0200 |
| Articles | 9 — 4 participants |
Back to article view | Back to linux.kernel
[PATCH 0/3] media: atmel-isc: Supplement the configuration of formats Wenyou Yang <wenyou.yang@microchip.com> - 2017-08-17 09:30 +0200
[PATCH 1/3] media: atmel-isc: Not support RBG format from sensor. Wenyou Yang <wenyou.yang@microchip.com> - 2017-08-17 09:30 +0200
Re: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor. Hans Verkuil <hverkuil@xs4all.nl> - 2017-08-21 16:20 +0200
Re: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor. "Yang, Wenyou" <Wenyou.Yang@Microchip.com> - 2017-08-22 03:20 +0200
Re: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor. Hans Verkuil <hverkuil@xs4all.nl> - 2017-08-22 09:10 +0200
RE: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor. <Wenyou.Yang@microchip.com> - 2017-08-22 09:40 +0200
Re: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor. Hans Verkuil <hverkuil@xs4all.nl> - 2017-08-23 12:40 +0200
Re: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor. "Yang, Wenyou" <Wenyou.Yang@Microchip.com> - 2017-08-24 08:30 +0200
Re: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor. Hans Verkuil <hverkuil@xs4all.nl> - 2017-08-24 08:50 +0200
| From | Wenyou Yang <wenyou.yang@microchip.com> |
|---|---|
| Date | 2017-08-17 09:30 +0200 |
| Subject | [PATCH 0/3] media: atmel-isc: Supplement the configuration of formats |
| Message-ID | <ufnpD-1zN-3@gated-at.bofh.it> |
The intention of the patch set is to add more configuration of formats: GREY, ARGB444, ARGB555 and ARGB32, and add the checking the format from the extern sensor which doesn't support RGB formats from the external sensor. Wenyou Yang (3): media: atmel-isc: Not support RBG format from sensor. media: atmel-isc: Remove the redundant assignment media: atmel-isc: Add more format configurations drivers/media/platform/atmel/atmel-isc.c | 89 ++++++++++++++++++++------------ 1 file changed, 56 insertions(+), 33 deletions(-) -- 2.13.0
[toc] | [next] | [standalone]
| From | Wenyou Yang <wenyou.yang@microchip.com> |
|---|---|
| Date | 2017-08-17 09:30 +0200 |
| Subject | [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor. |
| Message-ID | <ufnpE-1zN-23@gated-at.bofh.it> |
| In reply to | #1713591 |
The 12-bit parallel interface supports the Raw Bayer, YCbCr,
Monochrome and JPEG Compressed pixel formats from the external
sensor, not support RBG pixel format.
Signed-off-by: Wenyou Yang <wenyou.yang@microchip.com>
---
drivers/media/platform/atmel/atmel-isc.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/drivers/media/platform/atmel/atmel-isc.c b/drivers/media/platform/atmel/atmel-isc.c
index d4df3d4ccd85..535bb03783fe 100644
--- a/drivers/media/platform/atmel/atmel-isc.c
+++ b/drivers/media/platform/atmel/atmel-isc.c
@@ -1478,6 +1478,11 @@ static int isc_formats_init(struct isc_device *isc)
while (!v4l2_subdev_call(subdev, pad, enum_mbus_code,
NULL, &mbus_code)) {
mbus_code.index++;
+
+ /* Not support the RGB pixel formats from sensor */
+ if ((mbus_code.code & 0xf000) == 0x1000)
+ continue;
+
fmt = find_format_by_code(mbus_code.code, &i);
if (!fmt)
continue;
--
2.13.0
[toc] | [prev] | [next] | [standalone]
| From | Hans Verkuil <hverkuil@xs4all.nl> |
|---|---|
| Date | 2017-08-21 16:20 +0200 |
| Subject | Re: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor. |
| Message-ID | <ugVIC-5rU-29@gated-at.bofh.it> |
| In reply to | #1713592 |
On 08/17/2017 09:16 AM, Wenyou Yang wrote:
> The 12-bit parallel interface supports the Raw Bayer, YCbCr,
> Monochrome and JPEG Compressed pixel formats from the external
> sensor, not support RBG pixel format.
>
> Signed-off-by: Wenyou Yang <wenyou.yang@microchip.com>
> ---
>
> drivers/media/platform/atmel/atmel-isc.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/drivers/media/platform/atmel/atmel-isc.c b/drivers/media/platform/atmel/atmel-isc.c
> index d4df3d4ccd85..535bb03783fe 100644
> --- a/drivers/media/platform/atmel/atmel-isc.c
> +++ b/drivers/media/platform/atmel/atmel-isc.c
> @@ -1478,6 +1478,11 @@ static int isc_formats_init(struct isc_device *isc)
> while (!v4l2_subdev_call(subdev, pad, enum_mbus_code,
> NULL, &mbus_code)) {
> mbus_code.index++;
> +
> + /* Not support the RGB pixel formats from sensor */
> + if ((mbus_code.code & 0xf000) == 0x1000)
> + continue;
Am I missing something? Here you skip any RGB mediabus formats, but in patch 3/3
you add RGB mediabus formats. But this patch prevents those new formats from being
selected, right?
Regards,
Hans
> +
> fmt = find_format_by_code(mbus_code.code, &i);
> if (!fmt)
> continue;
>
[toc] | [prev] | [next] | [standalone]
| From | "Yang, Wenyou" <Wenyou.Yang@Microchip.com> |
|---|---|
| Date | 2017-08-22 03:20 +0200 |
| Subject | Re: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor. |
| Message-ID | <uh61k-3xX-5@gated-at.bofh.it> |
| In reply to | #1716540 |
Hi Hans,
On 2017/8/21 22:07, Hans Verkuil wrote:
> On 08/17/2017 09:16 AM, Wenyou Yang wrote:
>> The 12-bit parallel interface supports the Raw Bayer, YCbCr,
>> Monochrome and JPEG Compressed pixel formats from the external
>> sensor, not support RBG pixel format.
>>
>> Signed-off-by: Wenyou Yang <wenyou.yang@microchip.com>
>> ---
>>
>> drivers/media/platform/atmel/atmel-isc.c | 5 +++++
>> 1 file changed, 5 insertions(+)
>>
>> diff --git a/drivers/media/platform/atmel/atmel-isc.c b/drivers/media/platform/atmel/atmel-isc.c
>> index d4df3d4ccd85..535bb03783fe 100644
>> --- a/drivers/media/platform/atmel/atmel-isc.c
>> +++ b/drivers/media/platform/atmel/atmel-isc.c
>> @@ -1478,6 +1478,11 @@ static int isc_formats_init(struct isc_device *isc)
>> while (!v4l2_subdev_call(subdev, pad, enum_mbus_code,
>> NULL, &mbus_code)) {
>> mbus_code.index++;
>> +
>> + /* Not support the RGB pixel formats from sensor */
>> + if ((mbus_code.code & 0xf000) == 0x1000)
>> + continue;
> Am I missing something? Here you skip any RGB mediabus formats, but in patch 3/3
> you add RGB mediabus formats. But this patch prevents those new formats from being
> selected, right?
This patch prevents getting the RGB format from the sensor directly.
The RGB format can be produced by ISC controller by itself.
> Regards,
>
> Hans
>
>> +
>> fmt = find_format_by_code(mbus_code.code, &i);
>> if (!fmt)
>> continue;
>>
Best Regards,
Wenyou Yang
[toc] | [prev] | [next] | [standalone]
| From | Hans Verkuil <hverkuil@xs4all.nl> |
|---|---|
| Date | 2017-08-22 09:10 +0200 |
| Subject | Re: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor. |
| Message-ID | <uhbu2-7on-15@gated-at.bofh.it> |
| In reply to | #1716958 |
On 08/22/2017 03:18 AM, Yang, Wenyou wrote:
> Hi Hans,
>
> On 2017/8/21 22:07, Hans Verkuil wrote:
>> On 08/17/2017 09:16 AM, Wenyou Yang wrote:
>>> The 12-bit parallel interface supports the Raw Bayer, YCbCr,
>>> Monochrome and JPEG Compressed pixel formats from the external
>>> sensor, not support RBG pixel format.
>>>
>>> Signed-off-by: Wenyou Yang <wenyou.yang@microchip.com>
>>> ---
>>>
>>> drivers/media/platform/atmel/atmel-isc.c | 5 +++++
>>> 1 file changed, 5 insertions(+)
>>>
>>> diff --git a/drivers/media/platform/atmel/atmel-isc.c b/drivers/media/platform/atmel/atmel-isc.c
>>> index d4df3d4ccd85..535bb03783fe 100644
>>> --- a/drivers/media/platform/atmel/atmel-isc.c
>>> +++ b/drivers/media/platform/atmel/atmel-isc.c
>>> @@ -1478,6 +1478,11 @@ static int isc_formats_init(struct isc_device *isc)
>>> while (!v4l2_subdev_call(subdev, pad, enum_mbus_code,
>>> NULL, &mbus_code)) {
>>> mbus_code.index++;
>>> +
>>> + /* Not support the RGB pixel formats from sensor */
>>> + if ((mbus_code.code & 0xf000) == 0x1000)
>>> + continue;
>> Am I missing something? Here you skip any RGB mediabus formats, but in patch 3/3
>> you add RGB mediabus formats. But this patch prevents those new formats from being
>> selected, right?
> This patch prevents getting the RGB format from the sensor directly.
> The RGB format can be produced by ISC controller by itself.
OK, I think I see what is going on here. The isc_formats array really is two
arrays in one: up to RAW_FMT_IND_END it describes what it can receive from
the source, and after that it describes what it can convert it to.
But if you can't handle RGB formats from the sensor, then why not make
sure none of the mbus codes in isc_formats uses RGB? That makes much
more sense.
E.g.:
{ V4L2_PIX_FMT_RGB565, MEDIA_BUS_FMT_RGB565_2X8_LE, 16,
ISC_PFE_CFG0_BPS_EIGHT, ISC_BAY_CFG_BGBG, ISC_RLP_CFG_MODE_RGB565,
ISC_DCFG_IMODE_PACKED16, ISC_DCTRL_DVIEW_PACKED, 0x7b,
false, false },
Why use MEDIA_BUS_FMT_RGB565_2X8_LE if this apparently is not supported?
Regards,
Hans
>
>> Regards,
>>
>> Hans
>>
>>> +
>>> fmt = find_format_by_code(mbus_code.code, &i);
>>> if (!fmt)
>>> continue;
>>>
>
> Best Regards,
> Wenyou Yang
>
[toc] | [prev] | [next] | [standalone]
| From | <Wenyou.Yang@microchip.com> |
|---|---|
| Date | 2017-08-22 09:40 +0200 |
| Subject | RE: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor. |
| Message-ID | <uhbX3-7Cz-9@gated-at.bofh.it> |
| In reply to | #1717082 |
Hi Hans,
> -----Original Message-----
> From: Hans Verkuil [mailto:hverkuil@xs4all.nl]
> Sent: 2017年8月22日 15:00
> To: Wenyou Yang - A41535 <Wenyou.Yang@microchip.com>; Mauro Carvalho
> Chehab <mchehab@s-opensource.com>
> Cc: Nicolas Ferre - M43238 <Nicolas.Ferre@microchip.com>; linux-
> kernel@vger.kernel.org; Sakari Ailus <sakari.ailus@iki.fi>; Jonathan Corbet
> <corbet@lwn.net>; linux-arm-kernel@lists.infradead.org; Linux Media Mailing List
> <linux-media@vger.kernel.org>
> Subject: Re: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor.
>
> On 08/22/2017 03:18 AM, Yang, Wenyou wrote:
> > Hi Hans,
> >
> > On 2017/8/21 22:07, Hans Verkuil wrote:
> >> On 08/17/2017 09:16 AM, Wenyou Yang wrote:
> >>> The 12-bit parallel interface supports the Raw Bayer, YCbCr,
> >>> Monochrome and JPEG Compressed pixel formats from the external
> >>> sensor, not support RBG pixel format.
> >>>
> >>> Signed-off-by: Wenyou Yang <wenyou.yang@microchip.com>
> >>> ---
> >>>
> >>> drivers/media/platform/atmel/atmel-isc.c | 5 +++++
> >>> 1 file changed, 5 insertions(+)
> >>>
> >>> diff --git a/drivers/media/platform/atmel/atmel-isc.c
> >>> b/drivers/media/platform/atmel/atmel-isc.c
> >>> index d4df3d4ccd85..535bb03783fe 100644
> >>> --- a/drivers/media/platform/atmel/atmel-isc.c
> >>> +++ b/drivers/media/platform/atmel/atmel-isc.c
> >>> @@ -1478,6 +1478,11 @@ static int isc_formats_init(struct isc_device *isc)
> >>> while (!v4l2_subdev_call(subdev, pad, enum_mbus_code,
> >>> NULL, &mbus_code)) {
> >>> mbus_code.index++;
> >>> +
> >>> + /* Not support the RGB pixel formats from sensor */
> >>> + if ((mbus_code.code & 0xf000) == 0x1000)
> >>> + continue;
> >> Am I missing something? Here you skip any RGB mediabus formats, but
> >> in patch 3/3 you add RGB mediabus formats. But this patch prevents
> >> those new formats from being selected, right?
> > This patch prevents getting the RGB format from the sensor directly.
> > The RGB format can be produced by ISC controller by itself.
>
> OK, I think I see what is going on here. The isc_formats array really is two arrays
> in one: up to RAW_FMT_IND_END it describes what it can receive from the
> source, and after that it describes what it can convert it to.
Not exactly.
Yes, up to RAW_FMT_IND_END, these formats must be got from the senor, they are RAW formats.
From ISC_FMT_IND_START to ISC_FMT_IND_END, they can be generated by the ISC controller.
It is possible they can be got from the sensor too, the driver will check it.
If it can be got from both the sensor and the ISC controller, the user can use the "sensor_preferred" parameter to decide from which one to get.
The RBG formats are the exception.
>
> But if you can't handle RGB formats from the sensor, then why not make sure
> none of the mbus codes in isc_formats uses RGB? That makes much more sense.
>
> E.g.:
>
> { V4L2_PIX_FMT_RGB565, MEDIA_BUS_FMT_RGB565_2X8_LE, 16,
> ISC_PFE_CFG0_BPS_EIGHT, ISC_BAY_CFG_BGBG,
> ISC_RLP_CFG_MODE_RGB565,
> ISC_DCFG_IMODE_PACKED16, ISC_DCTRL_DVIEW_PACKED, 0x7b,
> false, false },
>
> Why use MEDIA_BUS_FMT_RGB565_2X8_LE if this apparently is not supported?
This array is also the lists of all formats supported by the ISC(including got from the sensor).
The RGB formats are only generated by the ISC controller, not from the sensor.
>
> Regards,
>
> Hans
>
> >
> >> Regards,
> >>
> >> Hans
> >>
> >>> +
> >>> fmt = find_format_by_code(mbus_code.code, &i);
> >>> if (!fmt)
> >>> continue;
> >>>
> >
> > Best Regards,
> > Wenyou Yang
> >
Best Regards,
Wenyou Yang
[toc] | [prev] | [next] | [standalone]
| From | Hans Verkuil <hverkuil@xs4all.nl> |
|---|---|
| Date | 2017-08-23 12:40 +0200 |
| Subject | Re: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor. |
| Message-ID | <uhBeO-7MM-11@gated-at.bofh.it> |
| In reply to | #1717096 |
On 08/22/17 09:30, Wenyou.Yang@microchip.com wrote:
> Hi Hans,
>
>> -----Original Message-----
>> From: Hans Verkuil [mailto:hverkuil@xs4all.nl]
>> Sent: 2017年8月22日 15:00
>> To: Wenyou Yang - A41535 <Wenyou.Yang@microchip.com>; Mauro Carvalho
>> Chehab <mchehab@s-opensource.com>
>> Cc: Nicolas Ferre - M43238 <Nicolas.Ferre@microchip.com>; linux-
>> kernel@vger.kernel.org; Sakari Ailus <sakari.ailus@iki.fi>; Jonathan Corbet
>> <corbet@lwn.net>; linux-arm-kernel@lists.infradead.org; Linux Media Mailing List
>> <linux-media@vger.kernel.org>
>> Subject: Re: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor.
>>
>> On 08/22/2017 03:18 AM, Yang, Wenyou wrote:
>>> Hi Hans,
>>>
>>> On 2017/8/21 22:07, Hans Verkuil wrote:
>>>> On 08/17/2017 09:16 AM, Wenyou Yang wrote:
>>>>> The 12-bit parallel interface supports the Raw Bayer, YCbCr,
>>>>> Monochrome and JPEG Compressed pixel formats from the external
>>>>> sensor, not support RBG pixel format.
>>>>>
>>>>> Signed-off-by: Wenyou Yang <wenyou.yang@microchip.com>
>>>>> ---
>>>>>
>>>>> drivers/media/platform/atmel/atmel-isc.c | 5 +++++
>>>>> 1 file changed, 5 insertions(+)
>>>>>
>>>>> diff --git a/drivers/media/platform/atmel/atmel-isc.c
>>>>> b/drivers/media/platform/atmel/atmel-isc.c
>>>>> index d4df3d4ccd85..535bb03783fe 100644
>>>>> --- a/drivers/media/platform/atmel/atmel-isc.c
>>>>> +++ b/drivers/media/platform/atmel/atmel-isc.c
>>>>> @@ -1478,6 +1478,11 @@ static int isc_formats_init(struct isc_device *isc)
>>>>> while (!v4l2_subdev_call(subdev, pad, enum_mbus_code,
>>>>> NULL, &mbus_code)) {
>>>>> mbus_code.index++;
>>>>> +
>>>>> + /* Not support the RGB pixel formats from sensor */
>>>>> + if ((mbus_code.code & 0xf000) == 0x1000)
>>>>> + continue;
>>>> Am I missing something? Here you skip any RGB mediabus formats, but
>>>> in patch 3/3 you add RGB mediabus formats. But this patch prevents
>>>> those new formats from being selected, right?
>>> This patch prevents getting the RGB format from the sensor directly.
>>> The RGB format can be produced by ISC controller by itself.
>>
>> OK, I think I see what is going on here. The isc_formats array really is two arrays
>> in one: up to RAW_FMT_IND_END it describes what it can receive from the
>> source, and after that it describes what it can convert it to.
>
> Not exactly.
>
> Yes, up to RAW_FMT_IND_END, these formats must be got from the senor, they are RAW formats.
> From ISC_FMT_IND_START to ISC_FMT_IND_END, they can be generated by the ISC controller.
> It is possible they can be got from the sensor too, the driver will check it.
> If it can be got from both the sensor and the ISC controller, the user can use the "sensor_preferred" parameter to decide from which one to get.
> The RBG formats are the exception.
>
>>
>> But if you can't handle RGB formats from the sensor, then why not make sure
>> none of the mbus codes in isc_formats uses RGB? That makes much more sense.
>>
>> E.g.:
>>
>> { V4L2_PIX_FMT_RGB565, MEDIA_BUS_FMT_RGB565_2X8_LE, 16,
>> ISC_PFE_CFG0_BPS_EIGHT, ISC_BAY_CFG_BGBG,
>> ISC_RLP_CFG_MODE_RGB565,
>> ISC_DCFG_IMODE_PACKED16, ISC_DCTRL_DVIEW_PACKED, 0x7b,
>> false, false },
>>
>> Why use MEDIA_BUS_FMT_RGB565_2X8_LE if this apparently is not supported?
>
> This array is also the lists of all formats supported by the ISC(including got from the sensor).
> The RGB formats are only generated by the ISC controller, not from the sensor.
You're adding code that skips any entries of the table where mbus_code is an
RGB code. But this can also be done by not having RGB mbus codes in the table
in the first place since they make no sense if the HW cannot handle that!
Set the mbus_code to e.g. 0 for such entries, that makes more sense.
I also strongly suggest changing how the table is organized since those
_FMT_IND_ indices are all to easy to get wrong (and frankly hard to understand).
Regards,
Hans
>
>>
>> Regards,
>>
>> Hans
>>
>>>
>>>> Regards,
>>>>
>>>> Hans
>>>>
>>>>> +
>>>>> fmt = find_format_by_code(mbus_code.code, &i);
>>>>> if (!fmt)
>>>>> continue;
>>>>>
>>>
>>> Best Regards,
>>> Wenyou Yang
>>>
>
> Best Regards,
> Wenyou Yang
>
[toc] | [prev] | [next] | [standalone]
| From | "Yang, Wenyou" <Wenyou.Yang@Microchip.com> |
|---|---|
| Date | 2017-08-24 08:30 +0200 |
| Subject | Re: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor. |
| Message-ID | <uhTOp-2Pi-3@gated-at.bofh.it> |
| In reply to | #1718235 |
On 2017/8/23 18:37, Hans Verkuil wrote:
> On 08/22/17 09:30, Wenyou.Yang@microchip.com wrote:
>> Hi Hans,
>>
>>> -----Original Message-----
>>> From: Hans Verkuil [mailto:hverkuil@xs4all.nl]
>>> Sent: 2017年8月22日 15:00
>>> To: Wenyou Yang - A41535 <Wenyou.Yang@microchip.com>; Mauro Carvalho
>>> Chehab <mchehab@s-opensource.com>
>>> Cc: Nicolas Ferre - M43238 <Nicolas.Ferre@microchip.com>; linux-
>>> kernel@vger.kernel.org; Sakari Ailus <sakari.ailus@iki.fi>; Jonathan Corbet
>>> <corbet@lwn.net>; linux-arm-kernel@lists.infradead.org; Linux Media Mailing List
>>> <linux-media@vger.kernel.org>
>>> Subject: Re: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor.
>>>
>>> On 08/22/2017 03:18 AM, Yang, Wenyou wrote:
>>>> Hi Hans,
>>>>
>>>> On 2017/8/21 22:07, Hans Verkuil wrote:
>>>>> On 08/17/2017 09:16 AM, Wenyou Yang wrote:
>>>>>> The 12-bit parallel interface supports the Raw Bayer, YCbCr,
>>>>>> Monochrome and JPEG Compressed pixel formats from the external
>>>>>> sensor, not support RBG pixel format.
>>>>>>
>>>>>> Signed-off-by: Wenyou Yang <wenyou.yang@microchip.com>
>>>>>> ---
>>>>>>
>>>>>> drivers/media/platform/atmel/atmel-isc.c | 5 +++++
>>>>>> 1 file changed, 5 insertions(+)
>>>>>>
>>>>>> diff --git a/drivers/media/platform/atmel/atmel-isc.c
>>>>>> b/drivers/media/platform/atmel/atmel-isc.c
>>>>>> index d4df3d4ccd85..535bb03783fe 100644
>>>>>> --- a/drivers/media/platform/atmel/atmel-isc.c
>>>>>> +++ b/drivers/media/platform/atmel/atmel-isc.c
>>>>>> @@ -1478,6 +1478,11 @@ static int isc_formats_init(struct isc_device *isc)
>>>>>> while (!v4l2_subdev_call(subdev, pad, enum_mbus_code,
>>>>>> NULL, &mbus_code)) {
>>>>>> mbus_code.index++;
>>>>>> +
>>>>>> + /* Not support the RGB pixel formats from sensor */
>>>>>> + if ((mbus_code.code & 0xf000) == 0x1000)
>>>>>> + continue;
>>>>> Am I missing something? Here you skip any RGB mediabus formats, but
>>>>> in patch 3/3 you add RGB mediabus formats. But this patch prevents
>>>>> those new formats from being selected, right?
>>>> This patch prevents getting the RGB format from the sensor directly.
>>>> The RGB format can be produced by ISC controller by itself.
>>> OK, I think I see what is going on here. The isc_formats array really is two arrays
>>> in one: up to RAW_FMT_IND_END it describes what it can receive from the
>>> source, and after that it describes what it can convert it to.
>> Not exactly.
>>
>> Yes, up to RAW_FMT_IND_END, these formats must be got from the senor, they are RAW formats.
>> From ISC_FMT_IND_START to ISC_FMT_IND_END, they can be generated by the ISC controller.
>> It is possible they can be got from the sensor too, the driver will check it.
>> If it can be got from both the sensor and the ISC controller, the user can use the "sensor_preferred" parameter to decide from which one to get.
>> The RBG formats are the exception.
>>
>>> But if you can't handle RGB formats from the sensor, then why not make sure
>>> none of the mbus codes in isc_formats uses RGB? That makes much more sense.
>>>
>>> E.g.:
>>>
>>> { V4L2_PIX_FMT_RGB565, MEDIA_BUS_FMT_RGB565_2X8_LE, 16,
>>> ISC_PFE_CFG0_BPS_EIGHT, ISC_BAY_CFG_BGBG,
>>> ISC_RLP_CFG_MODE_RGB565,
>>> ISC_DCFG_IMODE_PACKED16, ISC_DCTRL_DVIEW_PACKED, 0x7b,
>>> false, false },
>>>
>>> Why use MEDIA_BUS_FMT_RGB565_2X8_LE if this apparently is not supported?
>> This array is also the lists of all formats supported by the ISC(including got from the sensor).
>> The RGB formats are only generated by the ISC controller, not from the sensor.
> You're adding code that skips any entries of the table where mbus_code is an
> RGB code. But this can also be done by not having RGB mbus codes in the table
> in the first place since they make no sense if the HW cannot handle that!
> Set the mbus_code to e.g. 0 for such entries, that makes more sense.
>
> I also strongly suggest changing how the table is organized since those
> _FMT_IND_ indices are all to easy to get wrong (and frankly hard to understand).
Yes, you are right, I will change it. Do you have some advice?
Thank you.
>
> Regards,
>
> Hans
>
>>> Regards,
>>>
>>> Hans
>>>
>>>>> Regards,
>>>>>
>>>>> Hans
>>>>>
>>>>>> +
>>>>>> fmt = find_format_by_code(mbus_code.code, &i);
>>>>>> if (!fmt)
>>>>>> continue;
>>>>>>
>>>> Best Regards,
>>>> Wenyou Yang
>>>>
>> Best Regards,
>> Wenyou Yang
Best Regards,
Wenyou Yang
[toc] | [prev] | [next] | [standalone]
| From | Hans Verkuil <hverkuil@xs4all.nl> |
|---|---|
| Date | 2017-08-24 08:50 +0200 |
| Subject | Re: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor. |
| Message-ID | <uhU7M-2Wi-11@gated-at.bofh.it> |
| In reply to | #1718864 |
On 08/24/2017 08:25 AM, Yang, Wenyou wrote:
>
>
> On 2017/8/23 18:37, Hans Verkuil wrote:
>> On 08/22/17 09:30, Wenyou.Yang@microchip.com wrote:
>>> Hi Hans,
>>>
>>>> -----Original Message-----
>>>> From: Hans Verkuil [mailto:hverkuil@xs4all.nl]
>>>> Sent: 2017年8月22日 15:00
>>>> To: Wenyou Yang - A41535 <Wenyou.Yang@microchip.com>; Mauro Carvalho
>>>> Chehab <mchehab@s-opensource.com>
>>>> Cc: Nicolas Ferre - M43238 <Nicolas.Ferre@microchip.com>; linux-
>>>> kernel@vger.kernel.org; Sakari Ailus <sakari.ailus@iki.fi>; Jonathan Corbet
>>>> <corbet@lwn.net>; linux-arm-kernel@lists.infradead.org; Linux Media Mailing List
>>>> <linux-media@vger.kernel.org>
>>>> Subject: Re: [PATCH 1/3] media: atmel-isc: Not support RBG format from sensor.
>>>>
>>>> On 08/22/2017 03:18 AM, Yang, Wenyou wrote:
>>>>> Hi Hans,
>>>>>
>>>>> On 2017/8/21 22:07, Hans Verkuil wrote:
>>>>>> On 08/17/2017 09:16 AM, Wenyou Yang wrote:
>>>>>>> The 12-bit parallel interface supports the Raw Bayer, YCbCr,
>>>>>>> Monochrome and JPEG Compressed pixel formats from the external
>>>>>>> sensor, not support RBG pixel format.
>>>>>>>
>>>>>>> Signed-off-by: Wenyou Yang <wenyou.yang@microchip.com>
>>>>>>> ---
>>>>>>>
>>>>>>> drivers/media/platform/atmel/atmel-isc.c | 5 +++++
>>>>>>> 1 file changed, 5 insertions(+)
>>>>>>>
>>>>>>> diff --git a/drivers/media/platform/atmel/atmel-isc.c
>>>>>>> b/drivers/media/platform/atmel/atmel-isc.c
>>>>>>> index d4df3d4ccd85..535bb03783fe 100644
>>>>>>> --- a/drivers/media/platform/atmel/atmel-isc.c
>>>>>>> +++ b/drivers/media/platform/atmel/atmel-isc.c
>>>>>>> @@ -1478,6 +1478,11 @@ static int isc_formats_init(struct isc_device *isc)
>>>>>>> while (!v4l2_subdev_call(subdev, pad, enum_mbus_code,
>>>>>>> NULL, &mbus_code)) {
>>>>>>> mbus_code.index++;
>>>>>>> +
>>>>>>> + /* Not support the RGB pixel formats from sensor */
>>>>>>> + if ((mbus_code.code & 0xf000) == 0x1000)
>>>>>>> + continue;
>>>>>> Am I missing something? Here you skip any RGB mediabus formats, but
>>>>>> in patch 3/3 you add RGB mediabus formats. But this patch prevents
>>>>>> those new formats from being selected, right?
>>>>> This patch prevents getting the RGB format from the sensor directly.
>>>>> The RGB format can be produced by ISC controller by itself.
>>>> OK, I think I see what is going on here. The isc_formats array really is two arrays
>>>> in one: up to RAW_FMT_IND_END it describes what it can receive from the
>>>> source, and after that it describes what it can convert it to.
>>> Not exactly.
>>>
>>> Yes, up to RAW_FMT_IND_END, these formats must be got from the senor, they are RAW formats.
>>> From ISC_FMT_IND_START to ISC_FMT_IND_END, they can be generated by the ISC controller.
>>> It is possible they can be got from the sensor too, the driver will check it.
>>> If it can be got from both the sensor and the ISC controller, the user can use the "sensor_preferred" parameter to decide from which one to get.
>>> The RBG formats are the exception.
>>>
>>>> But if you can't handle RGB formats from the sensor, then why not make sure
>>>> none of the mbus codes in isc_formats uses RGB? That makes much more sense.
>>>>
>>>> E.g.:
>>>>
>>>> { V4L2_PIX_FMT_RGB565, MEDIA_BUS_FMT_RGB565_2X8_LE, 16,
>>>> ISC_PFE_CFG0_BPS_EIGHT, ISC_BAY_CFG_BGBG,
>>>> ISC_RLP_CFG_MODE_RGB565,
>>>> ISC_DCFG_IMODE_PACKED16, ISC_DCTRL_DVIEW_PACKED, 0x7b,
>>>> false, false },
>>>>
>>>> Why use MEDIA_BUS_FMT_RGB565_2X8_LE if this apparently is not supported?
>>> This array is also the lists of all formats supported by the ISC(including got from the sensor).
>>> The RGB formats are only generated by the ISC controller, not from the sensor.
>> You're adding code that skips any entries of the table where mbus_code is an
>> RGB code. But this can also be done by not having RGB mbus codes in the table
>> in the first place since they make no sense if the HW cannot handle that!
>> Set the mbus_code to e.g. 0 for such entries, that makes more sense.
>>
>> I also strongly suggest changing how the table is organized since those
>> _FMT_IND_ indices are all to easy to get wrong (and frankly hard to understand).
> Yes, you are right, I will change it. Do you have some advice?
Two options spring to mind: split into two tables or add a bool that tells whether
the format can be created by the isc or not.
Regards,
Hans
[toc] | [prev] | [standalone]
Back to top | Article view | linux.kernel
csiph-web