Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <09c61b94-1ed1-eb72-9682-1f1f203f6f63@gmail.com>
Date:   Wed, 12 Apr 2023 09:08:11 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.6.1
Content-Language: en-US
To:     AngeloGioacchino Del Regno 
        <angelogioacchino.delregno@collabora.com>, chunkuang.hu@kernel.org
Cc:     p.zabel@pengutronix.de, airlied@gmail.com, daniel@ffwll.ch,
        dri-devel@lists.freedesktop.org,
        linux-mediatek@lists.infradead.org, linux-kernel@vger.kernel.org,
        linux-arm-kernel@lists.infradead.org, kernel@collabora.com,
        wenst@chromium.org
References: <20230404104800.301150-1-angelogioacchino.delregno@collabora.com>
 <20230404104800.301150-2-angelogioacchino.delregno@collabora.com>
From:   Matthias Brugger <matthias.bgg@gmail.com>
Subject: Re: [PATCH v3 1/9] drm/mediatek: dp: Cache EDID for eDP panel
In-Reply-To: <20230404104800.301150-2-angelogioacchino.delregno@collabora.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk


On 04/04/2023 12:47, AngeloGioacchino Del Regno wrote:
> Since eDP panels are not removable it is safe to cache the EDID:
> this will avoid a relatively long read transaction at every PM
> resume that is unnecessary only in the "special" case of eDP,
> hence speeding it up a little, as from now on, as resume operation,
> we will perform only link training.
> 
> Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
> ---
>   drivers/gpu/drm/mediatek/mtk_dp.c | 11 ++++++++++-
>   1 file changed, 10 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/mediatek/mtk_dp.c b/drivers/gpu/drm/mediatek/mtk_dp.c
> index 1f94fcc144d3..84f82cc68672 100644
> --- a/drivers/gpu/drm/mediatek/mtk_dp.c
> +++ b/drivers/gpu/drm/mediatek/mtk_dp.c
> @@ -118,6 +118,7 @@ struct mtk_dp {
>   	const struct mtk_dp_data *data;
>   	struct mtk_dp_info info;
>   	struct mtk_dp_train_info train_info;
> +	struct edid *edid;
>   
>   	struct platform_device *phy_dev;
>   	struct phy *phy;
> @@ -1993,7 +1994,11 @@ static struct edid *mtk_dp_get_edid(struct drm_bridge *bridge,
>   		usleep_range(2000, 5000);
>   	}
>   
> -	new_edid = drm_get_edid(connector, &mtk_dp->aux.ddc);
> +	/* eDP panels aren't removable, so we can return a cached EDID. */
> +	if (mtk_dp->edid && mtk_dp->bridge.type == DRM_MODE_CONNECTOR_eDP)
> +		new_edid = drm_edid_duplicate(mtk_dp->edid);
> +	else
> +		new_edid = drm_get_edid(connector, &mtk_dp->aux.ddc);

Maybe it would make sense to add a macro for the check of mtk_dp->bridge.type == 
DRM_MODE_CONNECTOR_eDP
it would make the code more readable.

>   
>   	/*
>   	 * Parse capability here to let atomic_get_input_bus_fmts and
> @@ -2022,6 +2027,10 @@ static struct edid *mtk_dp_get_edid(struct drm_bridge *bridge,
>   		drm_atomic_bridge_chain_post_disable(bridge, connector->state->state);
>   	}
>   
> +	/* If this is an eDP panel and the read EDID is good, cache it for later */
> +	if (mtk_dp->bridge.type == DRM_MODE_CONNECTOR_eDP && !mtk_dp->edid && new_edid)
> +		mtk_dp->edid = drm_edid_duplicate(new_edid);
> +

How about putting this in an else if branch of mtk_dp_parse_capabilities. At 
least we could get rid of the check regarding if new_edid != NULL.

I was thinking on how to put both if statements in one block, but I think the 
problem is, that we would leak memory if the capability parsing failes due to 
the call to drm_edid_duplicate(). Correct?

Regards,
Matthais

>   	return new_edid;
>   }.	/
>