There's still something fishy with the initialization sequence of the video pipeline with respect to the MIPI-DSI bridge.
As a workaround, jump start the pipeline by briefly enabling the video pattern generator and disabling it again. This will lead to a very brief flash of colored bars until that issue is better understood and properly resolved. Signed-off-by: Ahmad Fatoum <[email protected]> --- drivers/video/dw_mipi_dsi.c | 20 +++++++++++++++++++- 1 file changed, 19 insertions(+), 1 deletion(-) diff --git a/drivers/video/dw_mipi_dsi.c b/drivers/video/dw_mipi_dsi.c index bf95b1b9f6ff..681d939e17ea 100644 --- a/drivers/video/dw_mipi_dsi.c +++ b/drivers/video/dw_mipi_dsi.c @@ -906,9 +906,20 @@ static int dw_mipi_dsi_ioctl(struct vpl *vpl, unsigned int port, { struct dw_mipi_dsi *dsi = container_of(vpl, struct dw_mipi_dsi, vpl); struct drm_display_mode mode = {}; + bool vpg_enable_hack; + int ret; + + /* + * HACK: If Video Pattern Generator is not running here, the panel + * initialization on the Sitronix ST7703 (RGB30-Panel) times out... + * FIXME: figure out why Linux doesn't suffer from this + */ + vpg_enable_hack = true; switch (cmd) { case VPL_ENABLE: + if (vpg_enable_hack) + dsi->vpg_mode = VPG_COLORS_V; /* Switch to video mode for panel-bridge enable & panel enable */ dw_mipi_dsi_set_mode(dsi, MIPI_DSI_MODE_VIDEO); break; @@ -934,7 +945,14 @@ static int dw_mipi_dsi_ioctl(struct vpl *vpl, unsigned int port, /* Probes should have occured by now */ WARN_ON(!dsi->panel_bridge); - return vpl_bridge_ioctl(dsi->panel_bridge, cmd, data); + ret = vpl_bridge_ioctl(dsi->panel_bridge, cmd, data); + + if (cmd == VPL_ENABLE && vpg_enable_hack) { + dsi->vpg_mode = VPG_DISABLED; + vpg_param_set(NULL, dsi); + } + + return ret; } static const char *vpg_modes[] = { -- 2.47.3
