[PATCH 1/1] drm/bridge: tc358775: Remove burst mode support

Neil Armstrong neil.armstrong at linaro.org
Mon Nov 18 08:43:56 UTC 2024


On 15/11/2024 17:06, Daniel Semkowicz wrote:
> Current tc358775 bridge driver implementation assumes that DSI clock
> is used as a source for LVDS pixel clock. Moreover, clock divider
> PCLKDIV has a hard-coded value. This means that driver expects DSI
> clock to always match a panel pixel clock. With such assumption, LVDS
> pixel clock is configured to be (DSI pixel clock / 3) for single link.
> This is true for DSI non-burst video modes. In burst mode, DSI host
> is allowed to set transmission rate higher than pixel clock to allow
> shorter data bursts. When the transmission rate is higher than expected
> by the bridge, LVDS output timings are still configured for the lower
> frequency, but the output is clocked with the higher one. In most cases,
> bad LVDS timings cause incorrect panel operation.
> 
> Remove support for burst mode, so the non-burst mode is used by DSI
> host by default. Burst mode is supported by the bridge itself,
> but requires proper implementation in the driver to operate correctly
> in all scenarios.
> 
> Fixes: a4ed72e85c46 ("drm/bridge: tc358775: Add burst and low-power modes")
> Signed-off-by: Daniel Semkowicz <dse at thaumatec.com>
> ---
> 
>   drivers/gpu/drm/bridge/tc358775.c | 3 +--
>   1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/bridge/tc358775.c b/drivers/gpu/drm/bridge/tc358775.c
> index 0b4efaca6d682..7496681c7b883 100644
> --- a/drivers/gpu/drm/bridge/tc358775.c
> +++ b/drivers/gpu/drm/bridge/tc358775.c
> @@ -632,8 +632,7 @@ static int tc_attach_host(struct tc_data *tc)
>   
>   	dsi->lanes = tc->num_dsi_lanes;
>   	dsi->format = MIPI_DSI_FMT_RGB888;
> -	dsi->mode_flags = MIPI_DSI_MODE_VIDEO | MIPI_DSI_MODE_VIDEO_BURST |
> -			  MIPI_DSI_MODE_LPM;
> +	dsi->mode_flags = MIPI_DSI_MODE_VIDEO | MIPI_DSI_MODE_LPM;
>   
>   	/*
>   	 * The hs_rate and lp_rate are data rate values. The HS mode is

Reviewed-by: Neil Armstrong <neil.armstrong at linaro.org>


More information about the dri-devel mailing list