Re: [linux-sunxi] Re: [PATCH v3 04/14] drm/sun4i: tcon: Add support for demuxing TCON output on A31

[Julian Calaby]

Hi Chen-Yu,

On Sat, Sep 30, 2017 at 3:58 PM, Chen-Yu Tsai <wens@xxxxxxxx> wrote:
> On Sat, Sep 30, 2017 at 1:35 PM, Julian Calaby <julian.calaby@xxxxxxxxx> wrote:
>> Hi Chen-Yu,
>>
>> On Fri, Sep 29, 2017 at 8:22 PM, Chen-Yu Tsai <wens@xxxxxxxx> wrote:
>>> On Fri, Sep 29, 2017 at 6:20 PM, Maxime Ripard
>>> <maxime.ripard@xxxxxxxxxxxxxxxxxx> wrote:
>>>> On Fri, Sep 29, 2017 at 08:22:56AM +0000, Chen-Yu Tsai wrote:
>>>>> On systems with 2 TCONs such as the A31, it is possible to demux the
>>>>> output of the TCONs to one encoder.
>>>>>
>>>>> Add support for this for the A31.
>>>>>
>>>>> Signed-off-by: Chen-Yu Tsai <wens@xxxxxxxx>
>>>>> ---
>>>>>  drivers/gpu/drm/sun4i/sun4i_tcon.c | 38 ++++++++++++++++++++++++++++++++++++++
>>>>>  1 file changed, 38 insertions(+)
>>>>>
>>>>> diff --git a/drivers/gpu/drm/sun4i/sun4i_tcon.c b/drivers/gpu/drm/sun4i/sun4i_tcon.c
>>>>> index 7bf51abaee97..c949309d4285 100644
>>>>> --- a/drivers/gpu/drm/sun4i/sun4i_tcon.c
>>>>> +++ b/drivers/gpu/drm/sun4i/sun4i_tcon.c
>>>>> @@ -112,6 +112,21 @@ void sun4i_tcon_enable_vblank(struct sun4i_tcon *tcon, bool enable)
>>>>>  }
>>>>>  EXPORT_SYMBOL(sun4i_tcon_enable_vblank);
>>>>>
>>>>> +static struct sun4i_tcon *sun4i_get_first_tcon(struct drm_device *drm)
>>>>
>>>> Would that make sense to make it a bit more generic, and pass the id
>>>> to look for as an argument?
>>>
>>> The reason to look for TCON0 explicitly is to access the muxing registers, which
>>> are only available in TCON0. Other than that, there's nothing else
>>> shared between
>>> the two TCONs. So there's no particular reason to look for TCON1 explicitly.
>>
>> In that case: in the bizarre case where we're trying to use this mux
>> type and there is no TCON0, shouldn't we fail?
>
> It gives out a big warning, indicating something is wrong. If TCON0 is not found
> it is most likely your device tree is broken. There's nothing more the
> driver can do.
> Are you suggesting to return NULL in this case, and also do error
> handling in the
> callers?

You're already returning -EINVAL for other failure cases, so a lack of
TCON0 might as well do the same.

>> (Also, the code doesn't make sense if we have some TCON1 and TCON2 in
>> that order as it'll return TCON2)
>
> I'm guessing you want it to return NULL.

I'm just pointing out the mismatch between getting the "first" TCON
and the actual behaviour.

Thanks,

-- 
Julian Calaby

Email: julian.calaby@xxxxxxxxx
Profile: http://www.google.com/profiles/julian.calaby/