Message ID | 1460124364-31121-1-git-send-email-ander.conselvan.de.oliveira@intel.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
On Fri, Apr 08, 2016 at 05:06:04PM +0300, Ander Conselvan de Oliveira wrote: > The function chv_data_lane_soft_reset() uses the lane count to decide > which lanes to set/reset. However, the HDMI code never sets lane count, > since it always uses the four lanes of the phy. Note that before commit > a8f327fb8464 ("drm/i915: Clean up CHV lane soft reset programming"), all > lanes were reset, regardless of lane count, so this patch restores that > behavior. > > Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> > Cc: Deepak S <deepak.s@linux.intel.com> > Fixes: a8f327fb8464 ("drm/i915: Clean up CHV lane soft reset programming") > Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com> At some point we really should just eliminate the duplicated PHY code by moving it to someting like intel_dpio.c, and then I suppose we should populate lane_count for HDMI as well. But for now this is good enough. Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> > --- > > I noticed this while reading CHV code, so this is only compiled tested. I > don't know if this could cause real issues. > > Ander > > --- > drivers/gpu/drm/i915/intel_hdmi.c | 30 +++++++++++++----------------- > 1 file changed, 13 insertions(+), 17 deletions(-) > > diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c > index b199ede..5410d1a 100644 > --- a/drivers/gpu/drm/i915/intel_hdmi.c > +++ b/drivers/gpu/drm/i915/intel_hdmi.c > @@ -1670,14 +1670,12 @@ static void chv_data_lane_soft_reset(struct intel_encoder *encoder, > val |= DPIO_PCS_TX_LANE2_RESET | DPIO_PCS_TX_LANE1_RESET; > vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW0(ch), val); > > - if (crtc->config->lane_count > 2) { > - val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW0(ch)); > - if (reset) > - val &= ~(DPIO_PCS_TX_LANE2_RESET | DPIO_PCS_TX_LANE1_RESET); > - else > - val |= DPIO_PCS_TX_LANE2_RESET | DPIO_PCS_TX_LANE1_RESET; > - vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW0(ch), val); > - } > + val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW0(ch)); > + if (reset) > + val &= ~(DPIO_PCS_TX_LANE2_RESET | DPIO_PCS_TX_LANE1_RESET); > + else > + val |= DPIO_PCS_TX_LANE2_RESET | DPIO_PCS_TX_LANE1_RESET; > + vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW0(ch), val); > > val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW1(ch)); > val |= CHV_PCS_REQ_SOFTRESET_EN; > @@ -1687,15 +1685,13 @@ static void chv_data_lane_soft_reset(struct intel_encoder *encoder, > val |= DPIO_PCS_CLK_SOFT_RESET; > vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW1(ch), val); > > - if (crtc->config->lane_count > 2) { > - val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW1(ch)); > - val |= CHV_PCS_REQ_SOFTRESET_EN; > - if (reset) > - val &= ~DPIO_PCS_CLK_SOFT_RESET; > - else > - val |= DPIO_PCS_CLK_SOFT_RESET; > - vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW1(ch), val); > - } > + val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW1(ch)); > + val |= CHV_PCS_REQ_SOFTRESET_EN; > + if (reset) > + val &= ~DPIO_PCS_CLK_SOFT_RESET; > + else > + val |= DPIO_PCS_CLK_SOFT_RESET; > + vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW1(ch), val); > } > > static void chv_hdmi_pre_pll_enable(struct intel_encoder *encoder) > -- > 2.4.11
On Fri, 2016-04-08 at 17:11 +0300, Ville Syrjälä wrote: > On Fri, Apr 08, 2016 at 05:06:04PM +0300, Ander Conselvan de Oliveira wrote: > > The function chv_data_lane_soft_reset() uses the lane count to decide > > which lanes to set/reset. However, the HDMI code never sets lane count, > > since it always uses the four lanes of the phy. Note that before commit > > a8f327fb8464 ("drm/i915: Clean up CHV lane soft reset programming"), all > > lanes were reset, regardless of lane count, so this patch restores that > > behavior. > > > > Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> > > Cc: Deepak S <deepak.s@linux.intel.com> > > Fixes: a8f327fb8464 ("drm/i915: Clean up CHV lane soft reset programming") > > Signed-off-by: Ander Conselvan de Oliveira < > > ander.conselvan.de.oliveira@intel.com> > > At some point we really should just eliminate the duplicated PHY code > by moving it to someting like intel_dpio.c, and then I suppose we should > populate lane_count for HDMI as well. But for now this is good enough. I took the bait and did that for CHV, except I couldn't come up with good names and kerneldoc for stuff. I can do the same for VLV, but that will have to wait until Monday. Cheers, Ander > > Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> > > > --- > > > > I noticed this while reading CHV code, so this is only compiled tested. I > > don't know if this could cause real issues. > > > > Ander > > > > --- > > drivers/gpu/drm/i915/intel_hdmi.c | 30 +++++++++++++----------------- > > 1 file changed, 13 insertions(+), 17 deletions(-) > > > > diff --git a/drivers/gpu/drm/i915/intel_hdmi.c > > b/drivers/gpu/drm/i915/intel_hdmi.c > > index b199ede..5410d1a 100644 > > --- a/drivers/gpu/drm/i915/intel_hdmi.c > > +++ b/drivers/gpu/drm/i915/intel_hdmi.c > > @@ -1670,14 +1670,12 @@ static void chv_data_lane_soft_reset(struct > > intel_encoder *encoder, > > val |= DPIO_PCS_TX_LANE2_RESET | DPIO_PCS_TX_LANE1_RESET; > > vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW0(ch), val); > > > > - if (crtc->config->lane_count > 2) { > > - val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW0(ch)); > > - if (reset) > > - val &= ~(DPIO_PCS_TX_LANE2_RESET | > > DPIO_PCS_TX_LANE1_RESET); > > - else > > - val |= DPIO_PCS_TX_LANE2_RESET | > > DPIO_PCS_TX_LANE1_RESET; > > - vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW0(ch), val); > > - } > > + val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW0(ch)); > > + if (reset) > > + val &= ~(DPIO_PCS_TX_LANE2_RESET | > > DPIO_PCS_TX_LANE1_RESET); > > + else > > + val |= DPIO_PCS_TX_LANE2_RESET | DPIO_PCS_TX_LANE1_RESET; > > + vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW0(ch), val); > > > > val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW1(ch)); > > val |= CHV_PCS_REQ_SOFTRESET_EN; > > @@ -1687,15 +1685,13 @@ static void chv_data_lane_soft_reset(struct > > intel_encoder *encoder, > > val |= DPIO_PCS_CLK_SOFT_RESET; > > vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW1(ch), val); > > > > - if (crtc->config->lane_count > 2) { > > - val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW1(ch)); > > - val |= CHV_PCS_REQ_SOFTRESET_EN; > > - if (reset) > > - val &= ~DPIO_PCS_CLK_SOFT_RESET; > > - else > > - val |= DPIO_PCS_CLK_SOFT_RESET; > > - vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW1(ch), val); > > - } > > + val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW1(ch)); > > + val |= CHV_PCS_REQ_SOFTRESET_EN; > > + if (reset) > > + val &= ~DPIO_PCS_CLK_SOFT_RESET; > > + else > > + val |= DPIO_PCS_CLK_SOFT_RESET; > > + vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW1(ch), val); > > } > > > > static void chv_hdmi_pre_pll_enable(struct intel_encoder *encoder) > > -- > > 2.4.11 >
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c index b199ede..5410d1a 100644 --- a/drivers/gpu/drm/i915/intel_hdmi.c +++ b/drivers/gpu/drm/i915/intel_hdmi.c @@ -1670,14 +1670,12 @@ static void chv_data_lane_soft_reset(struct intel_encoder *encoder, val |= DPIO_PCS_TX_LANE2_RESET | DPIO_PCS_TX_LANE1_RESET; vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW0(ch), val); - if (crtc->config->lane_count > 2) { - val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW0(ch)); - if (reset) - val &= ~(DPIO_PCS_TX_LANE2_RESET | DPIO_PCS_TX_LANE1_RESET); - else - val |= DPIO_PCS_TX_LANE2_RESET | DPIO_PCS_TX_LANE1_RESET; - vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW0(ch), val); - } + val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW0(ch)); + if (reset) + val &= ~(DPIO_PCS_TX_LANE2_RESET | DPIO_PCS_TX_LANE1_RESET); + else + val |= DPIO_PCS_TX_LANE2_RESET | DPIO_PCS_TX_LANE1_RESET; + vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW0(ch), val); val = vlv_dpio_read(dev_priv, pipe, VLV_PCS01_DW1(ch)); val |= CHV_PCS_REQ_SOFTRESET_EN; @@ -1687,15 +1685,13 @@ static void chv_data_lane_soft_reset(struct intel_encoder *encoder, val |= DPIO_PCS_CLK_SOFT_RESET; vlv_dpio_write(dev_priv, pipe, VLV_PCS01_DW1(ch), val); - if (crtc->config->lane_count > 2) { - val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW1(ch)); - val |= CHV_PCS_REQ_SOFTRESET_EN; - if (reset) - val &= ~DPIO_PCS_CLK_SOFT_RESET; - else - val |= DPIO_PCS_CLK_SOFT_RESET; - vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW1(ch), val); - } + val = vlv_dpio_read(dev_priv, pipe, VLV_PCS23_DW1(ch)); + val |= CHV_PCS_REQ_SOFTRESET_EN; + if (reset) + val &= ~DPIO_PCS_CLK_SOFT_RESET; + else + val |= DPIO_PCS_CLK_SOFT_RESET; + vlv_dpio_write(dev_priv, pipe, VLV_PCS23_DW1(ch), val); } static void chv_hdmi_pre_pll_enable(struct intel_encoder *encoder)
The function chv_data_lane_soft_reset() uses the lane count to decide which lanes to set/reset. However, the HDMI code never sets lane count, since it always uses the four lanes of the phy. Note that before commit a8f327fb8464 ("drm/i915: Clean up CHV lane soft reset programming"), all lanes were reset, regardless of lane count, so this patch restores that behavior. Cc: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: Deepak S <deepak.s@linux.intel.com> Fixes: a8f327fb8464 ("drm/i915: Clean up CHV lane soft reset programming") Signed-off-by: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com> --- I noticed this while reading CHV code, so this is only compiled tested. I don't know if this could cause real issues. Ander --- drivers/gpu/drm/i915/intel_hdmi.c | 30 +++++++++++++----------------- 1 file changed, 13 insertions(+), 17 deletions(-)