Received-SPF: pass (google.com: domain of linux-kernel+bounces-202295-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
Message-ID: <5dc16b34-d638-4fab-84e1-cb7db08ad80e@linaro.org>
Date: Wed, 5 Jun 2024 13:48:09 +0200
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
From: Neil Armstrong <neil.armstrong@linaro.org>
Reply-To: neil.armstrong@linaro.org
Subject: Re: [PATCH 13/14] drm/bridge: synopsys: Add DW HDMI QP TX controller
 driver
To: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>,
 Sam Ravnborg <sam@ravnborg.org>
Cc: Andrzej Hajda <andrzej.hajda@intel.com>, Robert Foss <rfoss@kernel.org>,
 Laurent Pinchart <Laurent.pinchart@ideasonboard.com>,
 Jonas Karlman <jonas@kwiboo.se>, Jernej Skrabec <jernej.skrabec@gmail.com>,
 Maarten Lankhorst <maarten.lankhorst@linux.intel.com>,
 Maxime Ripard <mripard@kernel.org>, Thomas Zimmermann <tzimmermann@suse.de>,
 David Airlie <airlied@gmail.com>, Daniel Vetter <daniel@ffwll.ch>,
 Sandy Huang <hjc@rock-chips.com>, =?UTF-8?Q?Heiko_St=C3=BCbner?=
 <heiko@sntech.de>, Andy Yan <andy.yan@rock-chips.com>,
 Rob Herring <robh@kernel.org>, Krzysztof Kozlowski <krzk+dt@kernel.org>,
 Conor Dooley <conor+dt@kernel.org>, Mark Yao <markyao0591@gmail.com>,
 dri-devel@lists.freedesktop.org, linux-kernel@vger.kernel.org,
 linux-arm-kernel@lists.infradead.org, linux-rockchip@lists.infradead.org,
 devicetree@vger.kernel.org, kernel@collabora.com,
 Alexandre ARNOUD <aarnoud@me.com>, Luis de Arquer <ldearquer@gmail.com>,
 Algea Cao <algea.cao@rock-chips.com>
References: <20240601-b4-rk3588-bridge-upstream-v1-0-f6203753232b@collabora.com>
 <20240601-b4-rk3588-bridge-upstream-v1-13-f6203753232b@collabora.com>
 <20240601143226.GA2003970@ravnborg.org>
 <59519381-2729-4839-9882-65a981a0c551@collabora.com>
 <20240604204110.GA84949@ravnborg.org>
 <f656c72e-fac8-4345-9b65-1031ebe81c25@collabora.com>
 <304b4602-8722-4ed0-a555-8dada573ee79@collabora.com>
Content-Language: en-US, fr
Autocrypt: addr=neil.armstrong@linaro.org; keydata=
 xsBNBE1ZBs8BCAD78xVLsXPwV/2qQx2FaO/7mhWL0Qodw8UcQJnkrWmgTFRobtTWxuRx8WWP
 GTjuhvbleoQ5Cxjr+v+1ARGCH46MxFP5DwauzPekwJUD5QKZlaw/bURTLmS2id5wWi3lqVH4
 BVF2WzvGyyeV1o4RTCYDnZ9VLLylJ9bneEaIs/7cjCEbipGGFlfIML3sfqnIvMAxIMZrvcl9
 qPV2k+KQ7q+aXavU5W+yLNn7QtXUB530Zlk/d2ETgzQ5FLYYnUDAaRl+8JUTjc0CNOTpCeik
 80TZcE6f8M76Xa6yU8VcNko94Ck7iB4vj70q76P/J7kt98hklrr85/3NU3oti3nrIHmHABEB
 AAHNKk5laWwgQXJtc3Ryb25nIDxuZWlsLmFybXN0cm9uZ0BsaW5hcm8ub3JnPsLAkQQTAQoA
 OwIbIwULCQgHAwUVCgkICwUWAgMBAAIeAQIXgBYhBInsPQWERiF0UPIoSBaat7Gkz/iuBQJk
 Q5wSAhkBAAoJEBaat7Gkz/iuyhMIANiD94qDtUTJRfEW6GwXmtKWwl/mvqQtaTtZID2dos04
 YqBbshiJbejgVJjy+HODcNUIKBB3PSLaln4ltdsV73SBcwUNdzebfKspAQunCM22Mn6FBIxQ
 GizsMLcP/0FX4en9NaKGfK6ZdKK6kN1GR9YffMJd2P08EO8mHowmSRe/ExAODhAs9W7XXExw
 UNCY4pVJyRPpEhv373vvff60bHxc1k/FF9WaPscMt7hlkbFLUs85kHtQAmr8pV5Hy9ezsSRa
 GzJmiVclkPc2BY592IGBXRDQ38urXeM4nfhhvqA50b/nAEXc6FzqgXqDkEIwR66/Gbp0t3+r
 yQzpKRyQif3OwE0ETVkGzwEIALyKDN/OGURaHBVzwjgYq+ZtifvekdrSNl8TIDH8g1xicBYp
 QTbPn6bbSZbdvfeQPNCcD4/EhXZuhQXMcoJsQQQnO4vwVULmPGgtGf8PVc7dxKOeta+qUh6+
 SRh3vIcAUFHDT3f/Zdspz+e2E0hPV2hiSvICLk11qO6cyJE13zeNFoeY3ggrKY+IzbFomIZY
 4yG6xI99NIPEVE9lNBXBKIlewIyVlkOaYvJWSV+p5gdJXOvScNN1epm5YHmf9aE2ZjnqZGoM
 Mtsyw18YoX9BqMFInxqYQQ3j/HpVgTSvmo5ea5qQDDUaCsaTf8UeDcwYOtgI8iL4oHcsGtUX
 oUk33HEAEQEAAcLAXwQYAQIACQUCTVkGzwIbDAAKCRAWmrexpM/4rrXiB/sGbkQ6itMrAIfn
 M7IbRuiSZS1unlySUVYu3SD6YBYnNi3G5EpbwfBNuT3H8//rVvtOFK4OD8cRYkxXRQmTvqa3
 3eDIHu/zr1HMKErm+2SD6PO9umRef8V82o2oaCLvf4WeIssFjwB0b6a12opuRP7yo3E3gTCS
 KmbUuLv1CtxKQF+fUV1cVaTPMyT25Od+RC1K+iOR0F54oUJvJeq7fUzbn/KdlhA8XPGzwGRy
 4zcsPWvwnXgfe5tk680fEKZVwOZKIEuJC3v+/yZpQzDvGYJvbyix0lHnrCzq43WefRHI5XTT
 QbM0WUIBIcGmq38+OgUsMYu4NzLu7uZFAcmp6h8g
Organization: Linaro
In-Reply-To: <304b4602-8722-4ed0-a555-8dada573ee79@collabora.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

On 05/06/2024 12:11, Cristian Ciocaltea wrote:
> On 6/5/24 12:34 AM, Cristian Ciocaltea wrote:
>> On 6/4/24 11:41 PM, Sam Ravnborg wrote:
>>> Hi Cristian.
>>>
>>> On Tue, Jun 04, 2024 at 10:32:04PM +0300, Cristian Ciocaltea wrote:
>>>> Hi Sam,
>>>>
>>>> On 6/1/24 5:32 PM, Sam Ravnborg wrote:
>>>>> Hi Cristian,
>>>>>
>>>>> a few drive-by comments below.
>>>>>
>>>>> 	Sam
>>>>>
>>>>>
>>>>>> +
>>>>>> +static const struct drm_connector_funcs dw_hdmi_qp_connector_funcs = {
>>>>>> +	.fill_modes = drm_helper_probe_single_connector_modes,
>>>>>> +	.detect = dw_hdmi_connector_detect,
>>>>>> +	.destroy = drm_connector_cleanup,
>>>>>> +	.force = dw_hdmi_qp_connector_force,
>>>>>> +	.reset = drm_atomic_helper_connector_reset,
>>>>>> +	.atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
>>>>>> +	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
>>>>>> +};
>>>>>> +
>>>>>> +static int dw_hdmi_qp_bridge_attach(struct drm_bridge *bridge,
>>>>>> +				    enum drm_bridge_attach_flags flags)
>>>>>> +{
>>>>>> +	struct dw_hdmi *hdmi = bridge->driver_private;
>>>>>> +
>>>>>> +	if (flags & DRM_BRIDGE_ATTACH_NO_CONNECTOR)
>>>>>> +		return drm_bridge_attach(bridge->encoder, hdmi->next_bridge,
>>>>>> +					 bridge, flags);
>>>>>> +
>>>>>> +	return dw_hdmi_connector_create(hdmi, &dw_hdmi_qp_connector_funcs);
>>>>>> +}
>>>>>
>>>>> Are there any users left that requires the display driver to create the
>>>>> connector?
>>>>> In other words - could this driver fail if DRM_BRIDGE_ATTACH_NO_CONNECTOR
>>>>> is not passed and drop dw_hdmi_connector_create()?
>>>>>
>>>>> I did not try to verify this - just a naive question.
>>>>
>>>> I've just tested this and it doesn't work - dw_hdmi_connector_create()
>>>> is still needed.
>>>
>>> Hmm, seems the display driver or some other bridge driver fails to
>>> support "DRM_BRIDGE_ATTACH_NO_CONNECTOR".
>>> what other drivers are involved?
>>
>> Could it be related to the glue driver (updated in the next patch) which
>> is also responsible for setting up the encoder?
>>
>>> Note that my comments here should be seen as potential future
>>> improvements, and do not block the patch from being used.
>>
>> Thanks for the heads up! Will try to get back to this soon and investigate.
>   
> IIUC, modern bridges should not create the connector but rely on display
> drivers to take care of, which in this case is the VOP2 driver. However,
> it also handles some of the older SoCs relying on the non-QP variant of
> DW HDMI IP. Hence the existing dw-hdmi driver would be also impacted in
> order to come up with a proper solution.
> 
> A quick check shows there are several users of this IP:
> 
> $ git grep -E '= dw_hdmi_(bind|probe)\('
> drivers/gpu/drm/bridge/imx/imx8mp-hdmi-tx.c:    hdmi->dw_hdmi = dw_hdmi_probe(pdev, plat_data);
> drivers/gpu/drm/bridge/synopsys/dw-hdmi.c:      hdmi = dw_hdmi_probe(pdev, plat_data);
> drivers/gpu/drm/imx/ipuv3/dw_hdmi-imx.c:        hdmi->hdmi = dw_hdmi_probe(pdev, match->data);
> drivers/gpu/drm/ingenic/ingenic-dw-hdmi.c:      hdmi = dw_hdmi_probe(pdev, &ingenic_dw_hdmi_plat_data);
> drivers/gpu/drm/meson/meson_dw_hdmi.c:  meson_dw_hdmi->hdmi = dw_hdmi_probe(pdev, &meson_dw_hdmi->dw_plat_data);
> drivers/gpu/drm/renesas/rcar-du/rcar_dw_hdmi.c: hdmi = dw_hdmi_probe(pdev, &rcar_dw_hdmi_plat_data);
> drivers/gpu/drm/rockchip/dw_hdmi-rockchip.c:            hdmi->hdmi = dw_hdmi_bind(pdev, encoder, plat_data);
> drivers/gpu/drm/sun4i/sun8i_dw_hdmi.c:  hdmi->hdmi = dw_hdmi_bind(pdev, encoder, plat_data);
> 
> I didn't check which display drivers would be involved, I'd guess there
> are quite a few of them as well. So it seems this ends up being a pretty
> complex task.

If this would be a brand new driver, then it should only support DRM_BRIDGE_ATTACH_NO_CONNECTOR,
so you should not create a connector from the driver.

The fact dw-hdmi accepts an attach without the flag is for legacy purpose
since some DRM drivers haven't switched to DRM_BRIDGE_ATTACH_NO_CONNECTOR yes,
but it's a requirement for new bridges so at some point you'll need to make
sure the rockchip glue and drm driver supports DRM_BRIDGE_ATTACH_NO_CONNECTOR.

This will greatly simplify the driver!

Neil

> 
> Cristian