From: Ville Syrjälä <ville.syrj...@linux.intel.com>
When we pick a new power sequencer for the port but we're not doing a
full modeset, the power sequencer may have locked on to another port.
So kick it a bit to make sure it controls the port we want.
Signed-off-by: Ville Syrjälä <ville.syrj...@linux.intel.com>
---
drivers/gpu/drm/i915/intel_dp.c | 57 +++++++++++++++++++++++++++++++++++++++++
1 file changed, 57 insertions(+)
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 240bc98..c2b3112 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -288,6 +288,51 @@ intel_dp_init_panel_power_sequencer_registers(struct
drm_device *dev,
struct intel_dp *intel_dp,
struct edp_power_seq *out);
+static void
+vlv_power_sequencer_kick(struct intel_dp *intel_dp,
+ enum pipe pipe)
+{
+ struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp);
+ struct drm_device *dev = intel_dig_port->base.base.dev;
+ struct drm_i915_private *dev_priv = dev->dev_private;
+ uint32_t DP;
+
+ /* Preserve the BIOS-computed detected bit. This is
+ * supposed to be read-only.
+ */
+ DP = I915_READ(intel_dp->output_reg) & DP_DETECTED;
+ DP |= DP_VOLTAGE_0_4 | DP_PRE_EMPHASIS_0;
+ DP |= DP_PORT_WIDTH(intel_dp->lane_count);
+
+ if (!IS_CHERRYVIEW(dev)) {
+ if (pipe == PIPE_B)
+ DP |= DP_PIPEB_SELECT;
+ } else {
+ DP |= DP_PIPE_SELECT_CHV(pipe);
+ }
+
+ /*
+ * Need to enable the port with idle pattern to allow the power
+ * sequencer to lock into the port. Otherwise the power sequence
+ * (including vdd force bit!) doesn't work on this port.
+ *
+ * FIXME do we need a clock from the DPLL?
+ * FIXME and what if the pipe is active, does it matter?
+ */
+ DP |= DP_PORT_EN | DP_LINK_TRAIN_PAT_IDLE;
+
+ I915_WRITE(intel_dp->output_reg, DP);
+ POSTING_READ(intel_dp->output_reg);
+
+ intel_edp_panel_vdd_on(intel_dp);
+ intel_edp_panel_on(intel_dp);
+ intel_edp_panel_off(intel_dp);
+
+ I915_WRITE(intel_dp->output_reg, DP & ~DP_PORT_EN);
+ POSTING_READ(intel_dp->output_reg);
+ msleep(intel_dp->panel_power_down_delay);
+}
+
static enum pipe
vlv_power_sequencer_pipe(struct intel_dp *intel_dp)
{
@@ -333,6 +378,15 @@ vlv_power_sequencer_pipe(struct intel_dp *intel_dp)
intel_dp_init_panel_power_sequencer_registers(dev, intel_dp,
&power_seq);
+ /*
+ * Even vdd force doesn't work until we've made
+ * the power sequencer lock in on the port.
+ */
+ DRM_DEBUG_KMS("kicking pipe %c power sequencer\n",
+ pipe_name(intel_dp->pipe));
+
+ vlv_power_sequencer_kick(intel_dp, intel_dp->pipe);
+
return intel_dp->pipe;
}
@@ -370,6 +424,9 @@ vlv_initial_power_sequencer_setup(struct intel_dp *intel_dp)
intel_dp_init_panel_power_sequencer(dev, intel_dp, &power_seq);
intel_dp_init_panel_power_sequencer_registers(dev, intel_dp,
&power_seq);
+
+ /* kick it just in case someone left it in a stuck state */
+ vlv_power_sequencer_kick(intel_dp, intel_dp->pipe);
}
static u32 _pp_ctrl_reg(struct intel_dp *intel_dp)
--
1.8.5.5
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
