linux-kernel - Re: [PATCH v3 2/9] media: platform: Add c3 mipi csi2 driver

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <868d13a7-66d2-445c-895f-5cab3e71f1ef@amlogic.com>
Date: Wed, 6 Nov 2024 10:10:58 +0800
From: Keke Li <keke.li@...ogic.com>
To: Tomi Valkeinen <tomi.valkeinen@...asonboard.com>,
 Jacopo Mondi <jacopo.mondi@...asonboard.com>
Cc: Mauro Carvalho Chehab <mchehab@...nel.org>, Rob Herring
 <robh@...nel.org>, Krzysztof Kozlowski <krzk+dt@...nel.org>,
 Conor Dooley <conor+dt@...nel.org>, linux-media@...r.kernel.org,
 devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
 kieran.bingham@...asonboard.com, laurent.pinchart@...asonboard.com,
 dan.scally@...asonboard.com
Subject: Re: [PATCH v3 2/9] media: platform: Add c3 mipi csi2 driver

Hi Tomi

Thanks very much for your reply.

On 2024/11/5 20:08, Tomi Valkeinen wrote:
> [ EXTERNAL EMAIL ]
>
> Hi,
>
> On 05/11/2024 13:39, Jacopo Mondi wrote:
>
>>>> Could you clarify what other streams you plan to support ? As you
>>>> support routing I presume you are preparing to capture
>>>> multiple streams of data like image + embedded data, or to support
>>>> sensors which sends data on different virtual channels ?
>>>>
>>>> How do you see this driver evolve ? Will it be augmented with an
>>>> additional source pad directed to a video device where to capture
>>>> embedded data from ?
>>>>
>>>> I'm wondering because if PAD_SINK become multiplexed, you won't be
>>>> allowed to set a format there. It only works now because you have a
>>>> single stream.
>>>>
>>>> Speaking of which, as you prepare to support multiple streams, I would
>>>> specify the stream number when calling v4l2_subdev_state_get_format().
>>>>
>>>>           fmt = v4l2_subdev_state_get_format(state, format->pad, 0);
>>>>
>>> Thanks for your suggestion.
>>>
>>> But this MIPI CSI2 hardware module doesn't have the ability to 
>>> separate data
>>> , such as image + embedded data.
>>>
>>> So there are no plans to support other streams.
>>
>> I see. Now that I've reviewed the adapter subdevice path I realized
>> that it's the adapter that can split data based on VC/DT to either the
>> ISP direct path or to DDR.
>>
>> The CSI-2 RX then transports streams as received from the sensor
>> directly to the adapter.
>>
>> In order to support capturing embedded data to DDR in the adapter the
>> embedded data stream needs a representation in this subdevice as well,
>> so that the source pad is multiplexed as well and the adapter receives
>> two streams that it can eventually split.
>>
>>
>>        Sensor         CSI-2Rx         Adapter
>>     +-----------+     +------+       +--------+
>>     0-- ED --\  |     |      |       |     /->0---> Embedded Data (DDR)
>>     |         ->0====>0 ==== 0 ====> 0====|   |
>>     0-- I ---/  |     |      |       |     \->0---> Image (ISP)
>>     +-----------+     +------+       +--------+
>>
>> When going to support embedded data capture this driver should create
>> two routes and allow enabling/disabling the embedded data one.
>
> This depends on the hardware. What's the bus between the CSI-2 rx and
> the adapter? If it's a custom bus that basically allows passing forward
> anything received from the CSI-2 bus, then the user should be free to
> configure any routes in the CSI-2 RX subdevice.
>
The bus between CSI-2 rx and the adapter is not a custom bus and can't

be configured by user.

> The adapter subdev, however, as it is limited by the HW, should only
> support two streams if that is what the hardware supports.
>
Yes,  the adapter HW only supports embedded data and image.
>> Tomi in cc for inputs.
>>
>> For now, if you don't support capturing embedded data, I would remove
>> routing support from here and from the adapter.
>
> Right, if the drivers only support a single stream, just drop the
> routing support.
>
>  Tomi
>
Will drop the routing support due to the hardware limit.