[PATCH] drm/vc4: hdmi: Don't poll for the infoframes status on setup

[Dave Stevenson]

Hi Maxime

On Thu, 19 Nov 2020 at 09:20, Maxime Ripard <maxime at cerno.tech> wrote:
>
> The infoframes are sent at a regular interval as a data island packet,
> so we don't need to wait for them to be sent when we're setting them up.
>
> However, we do need to poll when we're disabling since the we can't
> update the packet RAM until it has been sent.
>
> Let's add a boolean flag to tell whether we want to poll or not to
> support both cases.
>
> Suggested-by: Dave Stevenson <dave.stevenson at raspberrypi.com>
> Signed-off-by: Maxime Ripard <maxime at cerno.tech>
> ---
>  drivers/gpu/drm/vc4/vc4_hdmi.c | 10 +++++++---
>  1 file changed, 7 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/vc4/vc4_hdmi.c b/drivers/gpu/drm/vc4/vc4_hdmi.c
> index 285c871b02fb..e1bbdba646db 100644
> --- a/drivers/gpu/drm/vc4/vc4_hdmi.c
> +++ b/drivers/gpu/drm/vc4/vc4_hdmi.c
> @@ -265,7 +265,8 @@ static int vc4_hdmi_connector_init(struct drm_device *dev,
>  }
>
>  static int vc4_hdmi_stop_packet(struct drm_encoder *encoder,
> -                               enum hdmi_infoframe_type type)
> +                               enum hdmi_infoframe_type type,
> +                               bool poll)
>  {
>         struct vc4_hdmi *vc4_hdmi = encoder_to_vc4_hdmi(encoder);
>         u32 packet_id = type - 0x80;
> @@ -273,6 +274,9 @@ static int vc4_hdmi_stop_packet(struct drm_encoder *encoder,
>         HDMI_WRITE(HDMI_RAM_PACKET_CONFIG,
>                    HDMI_READ(HDMI_RAM_PACKET_CONFIG) & ~BIT(packet_id));
>
> +       if (!poll)
> +               return 0;
> +
>         return wait_for(!(HDMI_READ(HDMI_RAM_PACKET_STATUS) &
>                           BIT(packet_id)), 100);
>  }
> @@ -299,7 +303,7 @@ static void vc4_hdmi_write_infoframe(struct drm_encoder *encoder,
>         if (len < 0)
>                 return;
>
> -       ret = vc4_hdmi_stop_packet(encoder, frame->any.type);
> +       ret = vc4_hdmi_stop_packet(encoder, frame->any.type, false);

Have we got the poll flags the right way up here?
When updating we must wait before updating the ram packet registers,
which means we want to wait, so doesn't poll want to be true here?

In vc4_hdmi_audio_reset where we're just stopping the audio we don't
care about waiting, so poll can be false.

  Dave

>         if (ret) {
>                 DRM_ERROR("Failed to wait for infoframe to go idle: %d\n", ret);
>                 return;
> @@ -1055,7 +1059,7 @@ static void vc4_hdmi_audio_reset(struct vc4_hdmi *vc4_hdmi)
>         int ret;
>
>         vc4_hdmi->audio.streaming = false;
> -       ret = vc4_hdmi_stop_packet(encoder, HDMI_INFOFRAME_TYPE_AUDIO);
> +       ret = vc4_hdmi_stop_packet(encoder, HDMI_INFOFRAME_TYPE_AUDIO, true);
>         if (ret)
>                 dev_err(dev, "Failed to stop audio infoframe: %d\n", ret);
>
> --
> 2.28.0
>