On Tue, Mar 29, 2016 at 04:46:30PM -0400, Lyude wrote:
> On Valleyview, VGA hotplugging is controlled through a seperate register
> than everything else, VLV_ADPA, which must be explicitly set.
>
> While VGA hotplugging worked(ish) before, it looks like that was mainly
> because we'd unintentionally enable it in
> valleyview_crt_detect_hotplug() when we did a force trigger. This
> doesn't work reliably enough because whenever the display powerwell on
> vlv gets disabled, the values set in VLV_ADPA get cleared and
> consequently VGA hotplugging gets disabled. This causes bugs such as one
> we found on an Intel NUC, where doing the following sequence of
> hotplugs:
>
> - Disconnect all monitors
> - Connect VGA
> - Disconnect VGA
> - Connect HDMI
>
> Would result in hotplugging getting disabled, due to the display
> powerwells getting toggled in the process of connecting HDMI.
>
> CC: stable@xxxxxxxxxxxxxxx
> Signed-off-by: Lyude <cpaul@xxxxxxxxxx>
> ---
> drivers/gpu/drm/i915/i915_irq.c | 14 ++++++++++++++
> 1 file changed, 14 insertions(+)
>
> diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
> index 5aa4239..60592a4 100644
> --- a/drivers/gpu/drm/i915/i915_irq.c
> +++ b/drivers/gpu/drm/i915/i915_irq.c
> @@ -3611,6 +3611,7 @@ static void valleyview_display_irqs_install(struct drm_i915_private *dev_priv)
> {
> u32 pipestat_mask;
> u32 iir_mask;
> + u32 adpa_reg;
> enum pipe pipe;
>
> pipestat_mask = PIPESTAT_INT_STATUS_MASK |
> @@ -3627,6 +3628,12 @@ static void valleyview_display_irqs_install(struct drm_i915_private *dev_priv)
> for_each_pipe(dev_priv, pipe)
> i915_enable_pipestat(dev_priv, pipe, pipestat_mask);
>
> + if (IS_VALLEYVIEW(dev_priv)) {
> + adpa_reg = I915_READ(VLV_ADPA);
> + adpa_reg |= ADPA_CRT_HOTPLUG_ENABLE;
> + I915_WRITE(VLV_ADPA, adpa_reg);
> + }
We might not want to enable that when there's no VGA connector.
Seems like we should just be calling intel_crt_reset() here. We
definitely don't want to call the reset for hooks for all the other
connectors so drm_mode_config_reset() is out. Also the connector
locking might be problematic here, so I might suggest adjusting
intel_crt_reset() to take an encoder instead of connector, and then
we should be able to walk the encoder list without any troubles.
> +
> iir_mask = I915_DISPLAY_PORT_INTERRUPT |
> I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
> I915_DISPLAY_PIPE_B_EVENT_INTERRUPT;
> @@ -3645,8 +3652,15 @@ static void valleyview_display_irqs_uninstall(struct drm_i915_private *dev_priv)
> {
> u32 pipestat_mask;
> u32 iir_mask;
> + u32 adpa_reg;
> enum pipe pipe;
>
> + if (IS_VALLEYVIEW(dev_priv)) {
> + adpa_reg = I915_READ(VLV_ADPA);
> + adpa_reg &= ~ADPA_CRT_HOTPLUG_ENABLE;
> + I915_WRITE(VLV_ADPA, adpa_reg);
> + }
> +
> iir_mask = I915_DISPLAY_PORT_INTERRUPT |
> I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
> I915_DISPLAY_PIPE_B_EVENT_INTERRUPT;
> --
> 2.5.5
>
> _______________________________________________
> dri-devel mailing list
> dri-devel@xxxxxxxxxxxxxxxxxxxxx
> https://lists.freedesktop.org/mailman/listinfo/dri-devel
--
Ville Syrjälä
Intel OTC
--
To unsubscribe from this list: send the line "unsubscribe stable" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html