Re: [PATCH 3/8] drm/loongson: Allow attach drm bridge driver by calling lsdc_output_init()

[Sui Jingfeng]

Hi,


On 2023/10/30 07:10, Dmitry Baryshkov wrote:
> > +
> > +static void lsdc_pipe1_hdmi_encoder_reset(struct drm_encoder *encoder)
> > +{
> > +       struct drm_device *ddev = encoder->dev;
> > +       struct lsdc_device *ldev = to_lsdc(ddev);
> > +       u32 val;
> > +
> > +       val = PHY_CLOCK_POL | PHY_CLOCK_EN | PHY_DATA_EN;
> > +       lsdc_wreg32(ldev, LSDC_CRTC1_DVO_CONF_REG, val);
> > +
> > +       /* Using built-in GPIO emulated I2C instead of the hardware I2C */
> > +       lsdc_ureg32_clr(ldev, LSDC_HDMI1_INTF_CTRL_REG, HW_I2C_EN);
> > +
> > +       /* Help the HDMI phy get out of reset state */
> > +       lsdc_wreg32(ldev, LSDC_HDMI1_PHY_CTRL_REG, HDMI_PHY_RESET_N);
> > +
> > +       drm_dbg(ddev, "%s reset\n", encoder->name);
> > +
> > +       mdelay(20);
> > +}
> > +
> > +const struct drm_encoder_funcs lsdc_pipe1_hdmi_encoder_funcs = {
> > +       .reset = lsdc_pipe1_hdmi_encoder_reset,
> > +       .destroy = drm_encoder_cleanup,
> > +};
> > +
> > +/* Built-in DVO encoder helper funcs */
> > +
> > +static void lsdc_dvo_atomic_disable(struct drm_encoder *encoder,
> > +                                   struct drm_atomic_state *state)
> > +{
> > +}
> > +
> > +static void lsdc_dvo_atomic_enable(struct drm_encoder *encoder,
> > +                                  struct drm_atomic_state *state)
> > +{
> > +}
> > +
> > +static void lsdc_dvo_atomic_modeset(struct drm_encoder *encoder,
> > +                                   struct drm_crtc_state *crtc_state,
> > +                                   struct drm_connector_state *conn_state)
> > +{
> > +}
> You do not need these empty callbacks, do you?


Thanks a lot for the sharpen eyes.

Yes. I do not need these empty callbacks.
I have some program experience with Linux-4.19 kernel.
I clearly remember that these empty callbacks was need when I add lt8619 display bridge driver.
I have just remove it and tested, it seems that those is not need anymore.

Thanks a lot for review, this problem will be fixed at the next version.