Re: [PATCH 4/4] drm/sun4i: Enable DRM InfoFrame support on H6

[Andrzej Hajda]

On 24.06.2019 18:07, Chen-Yu Tsai wrote:
> On Tue, Jun 25, 2019 at 12:03 AM Jernej Åkrabec <jernej.skrabec@xxxxxxxx> wrote:
>> Dne ponedeljek, 24. junij 2019 ob 17:56:30 CEST je Chen-Yu Tsai napisal(a):
>>> On Mon, Jun 24, 2019 at 11:49 PM Andrzej Hajda <a.hajda@xxxxxxxxxxx> wrote:
>>>> On 24.06.2019 17:05, Jernej Åkrabec wrote:
>>>>> Dne ponedeljek, 24. junij 2019 ob 17:03:31 CEST je Andrzej Hajda
>> napisal(a):
>>>>>> On 26.05.2019 23:20, Jonas Karlman wrote:
>>>>>>> This patch enables Dynamic Range and Mastering InfoFrame on H6.
>>>>>>>
>>>>>>> Cc: Maxime Ripard <maxime.ripard@xxxxxxxxxxx>
>>>>>>> Cc: Jernej Skrabec <jernej.skrabec@xxxxxxxx>
>>>>>>> Signed-off-by: Jonas Karlman <jonas@xxxxxxxxx>
>>>>>>> ---
>>>>>>>
>>>>>>>  drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c | 2 ++
>>>>>>>  drivers/gpu/drm/sun4i/sun8i_dw_hdmi.h | 1 +
>>>>>>>  2 files changed, 3 insertions(+)
>>>>>>>
>>>>>>> diff --git a/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c
>>>>>>> b/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c index
>>>>>>> 39d8509d96a0..b80164dd8ad8
>>>>>>> 100644
>>>>>>> --- a/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c
>>>>>>> +++ b/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c
>>>>>>> @@ -189,6 +189,7 @@ static int sun8i_dw_hdmi_bind(struct device *dev,
>>>>>>> struct device *master,>
>>>>>>>
>>>>>>>     sun8i_hdmi_phy_init(hdmi->phy);
>>>>>>>
>>>>>>>     plat_data->mode_valid = hdmi->quirks->mode_valid;
>>>>>>>
>>>>>>> +   plat_data->drm_infoframe = hdmi->quirks->drm_infoframe;
>>>>>>>
>>>>>>>     sun8i_hdmi_phy_set_ops(hdmi->phy, plat_data);
>>>>>>>
>>>>>>>     platform_set_drvdata(pdev, hdmi);
>>>>>>>
>>>>>>> @@ -255,6 +256,7 @@ static const struct sun8i_dw_hdmi_quirks
>>>>>>> sun8i_a83t_quirks = {>
>>>>>>>
>>>>>>>  static const struct sun8i_dw_hdmi_quirks sun50i_h6_quirks = {
>>>>>>>
>>>>>>>     .mode_valid = sun8i_dw_hdmi_mode_valid_h6,
>>>>>>>
>>>>>>> +   .drm_infoframe = true,
>>>>>>>
>>>>>>>  };
>>>>>>>
>>>>>>>  static const struct of_device_id sun8i_dw_hdmi_dt_ids[] = {
>>>>>>>
>>>>>>> diff --git a/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.h
>>>>>>> b/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.h index
>>>>>>> 720c5aa8adc1..2a0ec08ee236
>>>>>>> 100644
>>>>>>> --- a/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.h
>>>>>>> +++ b/drivers/gpu/drm/sun4i/sun8i_dw_hdmi.h
>>>>>>> @@ -178,6 +178,7 @@ struct sun8i_dw_hdmi_quirks {
>>>>>>>
>>>>>>>     enum drm_mode_status (*mode_valid)(struct drm_connector
>>>>> *connector,
>>>>>
>>>>>>>                                        const struct
>>>>> drm_display_mode *mode);
>>>>>
>>>>>>>     unsigned int set_rate : 1;
>>>>>>>
>>>>>>> +   unsigned int drm_infoframe : 1;
>>>>>> Again, drm_infoframe suggests it contains inforframe, but in fact it
>>>>>> just informs infoframe can be used, so again my suggestion
>>>>>> use_drm_infoframe.
>>>>>>
>>>>>> Moreover bool type seems more appropriate here.
>>>>> checkpatch will give warning if bool is used.
>>>> Then I would say "fix/ignore checkpatch" :)
>>>>
>>>> But maybe there is a reason.
>>> Here's an old one from Linus: https://lkml.org/lkml/2013/9/1/154
>>>
>>> I'd say that bool in a struct is a waste of space compared to a 1 bit
>>> bitfield, especially when there already are other bitfields in the same
>>> struct.
>>>> Anyway I've tested and I do not see the warning, could you elaborate it.
>>> Maybe checkpatch.pl --strict?
>> It seems they removed that check:
>> https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?
>> id=7967656ffbfa493f5546c0f1
>>
>> After reading that block of text, I'm not sure what would be prefered way for
>> this case.
> This:
>
> +If a structure has many true/false values, consider consolidating them into a
> +bitfield with 1 bit members, or using an appropriate fixed width type, such as
> +u8.
>
> would suggest using a bitfield, or flags within a fixed width type?


OK, I have also guessed what kind of warning Jernej was writing about.
And IMO it rather does not fit here:

- no concurrent writes,

- no need for size/cache optimizations.

But since there are some controversies about bool in struct and file has
already convention of bitfield I do not insist on it, you can keep it as is.


Regards

Andrzej



>
> ChenYu
>
>