[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <09f20441-fec6-7496-2edc-c69db535e441@nvidia.com>
Date: Sat, 25 Apr 2020 18:43:18 -0700
From: Sowjanya Komatineni <skomatineni@...dia.com>
To: Dmitry Osipenko <digetx@...il.com>,
Hans Verkuil <hverkuil@...all.nl>, <thierry.reding@...il.com>,
<jonathanh@...dia.com>, <frankc@...dia.com>, <sakari.ailus@....fi>,
<helen.koike@...labora.com>
CC: <sboyd@...nel.org>, <linux-media@...r.kernel.org>,
<devicetree@...r.kernel.org>, <linux-clk@...r.kernel.org>,
<linux-tegra@...r.kernel.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [RFC PATCH v10 6/9] media: tegra: Add Tegra210 Video input driver
On 4/25/20 6:26 PM, Dmitry Osipenko wrote:
> External email: Use caution opening links or attachments
>
>
> 26.04.2020 04:08, Sowjanya Komatineni пишет:
>> On 4/25/20 5:41 PM, Sowjanya Komatineni wrote:
>>> On 4/25/20 5:36 PM, Dmitry Osipenko wrote:
>>>> External email: Use caution opening links or attachments
>>>>
>>>>
>>>> 25.04.2020 12:36, Hans Verkuil пишет:
>>>> ...
>>>>>> The media/tegra/ sounds a bit too generic, the media/tegra-vi/ path
>>>>>> should better reflect the driver, IMO.
>>>>>>
>>>>>> It also should be better to name the compiled kernel module as
>>>>>> tegra-vi,
>>>>>> IMO.
>>>>>>
>>>>> The driver name and the directory should be in sync, so either
>>>>> tegra-video
>>>>> or tegra-vi for both. I have no preference myself, as long as they
>>>>> are the
>>>>> same.
>>>>>
>>>>> This can be done as a follow-up patch.
>>>> Given that this driver isn't going to be reused by older pre-Tegra210
>>>> SoCs, perhaps it will also be worthwhile to name it as tegra210-vi or
>>>> tegra210-video.
>>> Can you explain what do you meant by can't be reused for pre-tegra210
>>> or for tegra186/194?
> It looks to me that at least all those hardcoded HW format IDs do not
> match the older SoCs.
TPG hard coded formats are supported on prior Tegra.
Other supported formats are SoC dependent and part of soc data in the
driver already.
>>> support for other tegra's can be added to same tegra-video driver.
>>> tegra-video host1x driver can be updated to add other tegra's vi and
>>> csi compatibles to host1x subdevs and vi and csi driver can be updated
>>> to add other tegra soc data and need to add coresponding
>>> tegra186/194/xxx.c file with tegra specific prog sequence
>>>
>> Same tegra-video.ko can be used for all tegra soc as driver supports
>> adding other soc related as well.
> Well, I'm still not sure why you would want to have all the unnecessary
> code of a different SoCs shared within the same kernel module, it will
> be quite be a lot wasted space in comparison to a used part of the driver.
>
> The driver will need to have a bit better separation if it's supposed to
> have a common core for all SoCs. Each incompatible VI/CSI hardware
> version should have its own kernel module.
currently other Tegra host1x driver (drm) also does similar. Single
module for all Tegra SoCs.
With current tegra-video, all the v4l2 related common part of
implementation is same for all tegra's and only
tegra210.c/tegra186.c/tegra194.c will have corresponding tegra soc
specific vi/csi programming sequence.
>> Also was thinking instead of renaming media/tegra as media/tegra-vi,
>> probably we can rename as media/tegra-video so it will be inline with
>> module name we already chosen and also mainly we have vi and csi with in
>> that so instead of tegra-vi probably we can use media/tegra-video?
> The tegra-video should be okay, although the "video" part sounds a bit
> too broad since video could mean a lot of things. I think downstream
> kernel uses (or at least used) the tegra-camera name for the driver,
> perhaps it could be a reasonable variant as well.
prior feedback suggests not to use camera variant instead to use video
Powered by blists - more mailing lists