Re: [PATCH 3/5] drm/i915/dsi: Init panel-enable GPIO to low when the LCD is initially off

[Hans de Goede <hdegoede@redhat.com>]

Hi,

On 16-12-2019 15:14, Ville Syrjälä wrote:
> On Mon, Dec 16, 2019 at 02:51:54PM +0100, Hans de Goede wrote:
>> Hi,
>>
>> Thank you for the reviews.
>>
>> On 16-12-2019 14:45, Ville Syrjälä wrote:
>>> On Sun, Dec 15, 2019 at 05:38:08PM +0100, Hans de Goede wrote:
>>>> When the LCD has not been turned on by the firmware/GOP, because e.g. the
>>>> device was booted with an external monitor connected over HDMI, we should
>>>> not turn on the panel-enable GPIO when we request it.
>>>>
>>>> Turning on the panel-enable GPIO when we request it, means we turn it on
>>>> too early in the init-sequence, which causes some panels to not correctly
>>>> light up.
>>>>
>>>> This commits adds a panel_is_on parameter to intel_dsi_vbt_gpio_init()
>>>> and makes intel_dsi_vbt_gpio_init() set the initial GPIO value accordingly.
>>>>
>>>> This fixes the panel not lighting up on a Thundersoft TST168 tablet when
>>>> booted with an external monitor connected over HDMI.
>>>>
>>>> Signed-off-by: Hans de Goede <hdegoede@redhat.com>
>>>> ---
>>>>    drivers/gpu/drm/i915/display/intel_dsi.h     | 2 +-
>>>>    drivers/gpu/drm/i915/display/intel_dsi_vbt.c | 7 +++----
>>>>    drivers/gpu/drm/i915/display/vlv_dsi.c       | 2 +-
>>>>    3 files changed, 5 insertions(+), 6 deletions(-)
>>>>
>>>> diff --git a/drivers/gpu/drm/i915/display/intel_dsi.h b/drivers/gpu/drm/i915/display/intel_dsi.h
>>>> index de7e51cd3460..675771ea91aa 100644
>>>> --- a/drivers/gpu/drm/i915/display/intel_dsi.h
>>>> +++ b/drivers/gpu/drm/i915/display/intel_dsi.h
>>>> @@ -203,7 +203,7 @@ void bxt_dsi_reset_clocks(struct intel_encoder *encoder, enum port port);
>>>>    
>>>>    /* intel_dsi_vbt.c */
>>>>    bool intel_dsi_vbt_init(struct intel_dsi *intel_dsi, u16 panel_id);
>>>> -void intel_dsi_vbt_gpio_init(struct intel_dsi *intel_dsi);
>>>> +void intel_dsi_vbt_gpio_init(struct intel_dsi *intel_dsi, bool panel_is_on);
>>>>    void intel_dsi_vbt_gpio_cleanup(struct intel_dsi *intel_dsi);
>>>>    void intel_dsi_vbt_exec_sequence(struct intel_dsi *intel_dsi,
>>>>    				 enum mipi_seq seq_id);
>>>> diff --git a/drivers/gpu/drm/i915/display/intel_dsi_vbt.c b/drivers/gpu/drm/i915/display/intel_dsi_vbt.c
>>>> index 5352e8c9eca5..027970348b22 100644
>>>> --- a/drivers/gpu/drm/i915/display/intel_dsi_vbt.c
>>>> +++ b/drivers/gpu/drm/i915/display/intel_dsi_vbt.c
>>>> @@ -688,17 +688,16 @@ bool intel_dsi_vbt_init(struct intel_dsi *intel_dsi, u16 panel_id)
>>>>     * On some BYT/CHT devs some sequences are incomplete and we need to manually
>>>>     * control some GPIOs.
>>>>     */
>>>> -void intel_dsi_vbt_gpio_init(struct intel_dsi *intel_dsi)
>>>> +void intel_dsi_vbt_gpio_init(struct intel_dsi *intel_dsi, bool panel_is_on)
>>>>    {
>>>>    	struct drm_device *dev = intel_dsi->base.base.dev;
>>>>    	struct drm_i915_private *dev_priv = to_i915(dev);
>>>>    	struct mipi_config *mipi_config = dev_priv->vbt.dsi.config;
>>>> +	enum gpiod_flags flags = panel_is_on ? GPIOD_OUT_HIGH : GPIOD_OUT_LOW;
>>>
>>> Can't we just tell it not to change the current setting?
>>
>> We could use GPIOD_ASIS for that, but with the SoC pins (when the PMIC is
>> not used for backlight control) things get a bit muddy, I've seen several
>> instances of this message from drivers/pinctrl/intel/pinctrl-baytrail.c
>> trigger when the GOP did not init the panel:
>>
>> dev_warn(vg->dev, FW_BUG "pin %u forcibly re-configured as GPIO\n", offset);
>>
>> And in that case with GPIOD_ASIS I have no idea which we initially get,
>> so this approach, where we clearly define which initial value we want,
>> seems better.
> 
> I suppose. Probably better to not abuse the current_mode pointer for
> that though and just explicitly call encoder->get_hw_state() instead.

Ok, I'll fix this for v2 of the patch-set.

Regards,

Hans





>> p.s.
>>
>> The intel-gfx CI seems to seriously dislike my patches lately, almost
>> always failing them; and usually on what at least seem to be unrelated
>> test-cases. Any advice on how to deal with this?
> 
> Yeah, CI is snafu. I keep smashing retest until it gets through BAT
> and then just double check the shard results to make sure nothing
> relevant has tripped. If things look OK I recommend replying to the
> result mail and provide a few short log snippets/other details on
> what failed so that it's clear that it's irrelevant.
> 
>>
>>
>>
>>
>>>
>>>>    
>>>>    	if ((IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) &&
>>>>    	    (mipi_config->pwm_blc == PPS_BLC_PMIC)) {
>>>> -		intel_dsi->gpio_panel =
>>>> -			gpiod_get(dev->dev, "panel", GPIOD_OUT_HIGH);
>>>> -
>>>> +		intel_dsi->gpio_panel = gpiod_get(dev->dev, "panel", flags);
>>>>    		if (IS_ERR(intel_dsi->gpio_panel)) {
>>>>    			DRM_ERROR("Failed to own gpio for panel control\n");
>>>>    			intel_dsi->gpio_panel = NULL;
>>>> diff --git a/drivers/gpu/drm/i915/display/vlv_dsi.c b/drivers/gpu/drm/i915/display/vlv_dsi.c
>>>> index 178d0fffba5b..e86e4a11e199 100644
>>>> --- a/drivers/gpu/drm/i915/display/vlv_dsi.c
>>>> +++ b/drivers/gpu/drm/i915/display/vlv_dsi.c
>>>> @@ -1910,7 +1910,7 @@ void vlv_dsi_init(struct drm_i915_private *dev_priv)
>>>>    
>>>>    	vlv_dphy_param_init(intel_dsi);
>>>>    
>>>> -	intel_dsi_vbt_gpio_init(intel_dsi);
>>>> +	intel_dsi_vbt_gpio_init(intel_dsi, current_mode != NULL);
>>>>    
>>>>    	drm_connector_init(dev, connector, &intel_dsi_connector_funcs,
>>>>    			   DRM_MODE_CONNECTOR_DSI);
>>>> -- 
>>>> 2.23.0
>>>
>