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;
MIME-Version: 1.0
References: <20220903130833.541463-1-xji@analogixsemi.com>
In-Reply-To: <20220903130833.541463-1-xji@analogixsemi.com>
From:   Robert Foss <robert.foss@linaro.org>
Date:   Mon, 5 Sep 2022 18:48:06 +0200
Message-ID: <CAG3jFyum5M8iGdW++c7rayUn4ogJ6izWaX7VaqMfV7ShAe_kRQ@mail.gmail.com>
Subject: Re: [PATCH] drm/bridge: anx7625: Set HPD irq detect window to 2ms
To:     Xin Ji <xji@analogixsemi.com>
Cc:     Andrzej Hajda <andrzej.hajda@intel.com>,
        Neil Armstrong <narmstrong@baylibre.com>,
        Laurent Pinchart <Laurent.pinchart@ideasonboard.com>,
        Jonas Karlman <jonas@kwiboo.se>,
        Jernej Skrabec <jernej.skrabec@gmail.com>,
        David Airlie <airlied@linux.ie>,
        Daniel Vetter <daniel@ffwll.ch>, bliang@analogixsemi.com,
        qwen@analogixsemi.com, mliu@analogixsemi.com, hsinyi@chromium.org,
        dri-devel@lists.freedesktop.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Hi Xin,

On Sat, 3 Sept 2022 at 15:09, Xin Ji <xji@analogixsemi.com> wrote:
>
> Some panels trigger HPD irq due to noise, the HPD debounce
> may be 1.8ms, exceeding the default irq detect window, ~1.4ms.
> This patch set HPD irq detection window to 2ms to
> tolerate the HPD noise.
>
> Signed-off-by: Xin Ji <xji@analogixsemi.com>
> ---
>  drivers/gpu/drm/bridge/analogix/anx7625.c | 14 ++++++++++++++
>  drivers/gpu/drm/bridge/analogix/anx7625.h |  6 ++++++
>  2 files changed, 20 insertions(+)
>
> diff --git a/drivers/gpu/drm/bridge/analogix/anx7625.c b/drivers/gpu/drm/bridge/analogix/anx7625.c
> index c74b5df4cade..0c323b5a1c99 100644
> --- a/drivers/gpu/drm/bridge/analogix/anx7625.c
> +++ b/drivers/gpu/drm/bridge/analogix/anx7625.c
> @@ -1440,6 +1440,20 @@ static void anx7625_start_dp_work(struct anx7625_data *ctx)
>
>  static int anx7625_read_hpd_status_p0(struct anx7625_data *ctx)
>  {
> +       int ret;
> +
> +       /* Set irq detect window to 2ms */
> +       ret = anx7625_reg_write(ctx, ctx->i2c.tx_p2_client,
> +                               HPD_DET_TIMER_BIT0_7, HPD_TIME & 0xFF);
> +       ret |= anx7625_reg_write(ctx, ctx->i2c.tx_p2_client,
> +                                HPD_DET_TIMER_BIT8_15,
> +                                (HPD_TIME >> 8) & 0xFF);
> +       ret |= anx7625_reg_write(ctx, ctx->i2c.tx_p2_client,
> +                                HPD_DET_TIMER_BIT16_23,
> +                                (HPD_TIME >> 16) & 0xFF);

Does the HPD debounce timer register need to be written for every HPD
status read?

> +       if (ret < 0)
> +               return ret;
> +
>         return anx7625_reg_read(ctx, ctx->i2c.rx_p0_client, SYSTEM_STSTUS);
>  }
>
> diff --git a/drivers/gpu/drm/bridge/analogix/anx7625.h b/drivers/gpu/drm/bridge/analogix/anx7625.h
> index e257a84db962..14f33d6be289 100644
> --- a/drivers/gpu/drm/bridge/analogix/anx7625.h
> +++ b/drivers/gpu/drm/bridge/analogix/anx7625.h
> @@ -132,6 +132,12 @@
>  #define I2S_SLAVE_MODE 0x08
>  #define AUDIO_LAYOUT   0x01
>
> +#define HPD_DET_TIMER_BIT0_7   0xea
> +#define HPD_DET_TIMER_BIT8_15  0xeb
> +#define HPD_DET_TIMER_BIT16_23 0xec
> +/* HPD debounce time 2ms for 27M clock */
> +#define HPD_TIME               54000
> +
>  #define AUDIO_CONTROL_REGISTER 0xe6
>  #define TDM_TIMING_MODE 0x08
>
> --
> 2.25.1
>