On Fri, 18 Sep 2015, ville.syrjala@xxxxxxxxxxxxxxx wrote:
> From: Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx>
>
> Rather than computing on demand, store also the aux data reg
> offsets under intel_dp.
>
> Signed-off-by: Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx>
> ---
> drivers/gpu/drm/i915/intel_dp.c | 49 +++++++++++++++++++++++++++-------------
> drivers/gpu/drm/i915/intel_drv.h | 1 +
> 2 files changed, 34 insertions(+), 16 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
> index 27df240..0e5f674 100644
> --- a/drivers/gpu/drm/i915/intel_dp.c
> +++ b/drivers/gpu/drm/i915/intel_dp.c
> @@ -791,7 +791,6 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
> struct drm_device *dev = intel_dig_port->base.base.dev;
> struct drm_i915_private *dev_priv = dev->dev_private;
> uint32_t ch_ctl = intel_dp->aux_ch_ctl_reg;
> - uint32_t ch_data = ch_ctl + 4;
> uint32_t aux_clock_divider;
> int i, ret, recv_bytes;
> uint32_t status;
> @@ -857,7 +856,7 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
> for (try = 0; try < 5; try++) {
> /* Load the send data into the aux channel data registers */
> for (i = 0; i < send_bytes; i += 4)
> - I915_WRITE(ch_data + i,
> + I915_WRITE(intel_dp->aux_ch_data_reg[i >> 2],
> intel_dp_pack_aux(send + i,
> send_bytes - i));
>
> @@ -921,7 +920,7 @@ done:
> recv_bytes = recv_size;
>
> for (i = 0; i < recv_bytes; i += 4)
> - intel_dp_unpack_aux(I915_READ(ch_data + i),
> + intel_dp_unpack_aux(I915_READ(intel_dp->aux_ch_data_reg[i >> 2]),
> recv + i, recv_bytes - i);
>
> ret = recv_bytes;
> @@ -1009,19 +1008,22 @@ intel_dp_aux_transfer(struct drm_dp_aux *aux, struct drm_dp_aux_msg *msg)
> return ret;
> }
>
> -static uint32_t g4x_aux_ctl_reg(struct drm_i915_private *dev_priv,
> - enum port port)
> +static uint32_t g4x_aux_reg(struct drm_i915_private *dev_priv,
> + enum port port, int index)
> {
> - return DP_AUX_CH_CTL(port);
> + return index < 0 ? DP_AUX_CH_CTL(port) :
> + DP_AUX_CH_DATA(port, index);
> }
>
> -static uint32_t ilk_aux_ctl_reg(struct drm_i915_private *dev_priv,
> - enum port port)
> +static uint32_t ilk_aux_reg(struct drm_i915_private *dev_priv,
> + enum port port, int index)
> {
> if (port == PORT_A)
> - return DP_AUX_CH_CTL(port);
> + return index < 0 ? DP_AUX_CH_CTL(port) :
> + DP_AUX_CH_DATA(port, index);
> else
> - return PCH_DP_AUX_CH_CTL(port);
> + return index < 0 ? PCH_DP_AUX_CH_CTL(port) :
> + PCH_DP_AUX_CH_DATA(port, index);
> }
>
> /*
> @@ -1048,13 +1050,28 @@ static enum port skl_porte_aux_port(struct drm_i915_private *dev_priv)
> }
> }
>
> -static uint32_t skl_aux_ctl_reg(struct drm_i915_private *dev_priv,
> - enum port port)
> +static uint32_t skl_aux_reg(struct drm_i915_private *dev_priv,
> + enum port port, int index)
> {
> if (port == PORT_E)
> port = skl_porte_aux_port(dev_priv);
>
> - return DP_AUX_CH_CTL(port);
> + return index < 0 ? DP_AUX_CH_CTL(port) :
> + DP_AUX_CH_DATA(port, index);
> +}
> +
> +static void aux_reg_init(struct intel_dp *intel_dp,
> + uint32_t (*aux_reg)(struct drm_i915_private *dev_priv,
> + enum port port, int index))
> +{
> + struct drm_i915_private *dev_priv = to_i915(intel_dp_to_dev(intel_dp));
> + enum port port = dp_to_dig_port(intel_dp)->port;
> + int i;
> +
> + intel_dp->aux_ch_ctl_reg = aux_reg(dev_priv, port, -1);
> +
> + for (i = 0; i < ARRAY_SIZE(intel_dp->aux_ch_data_reg); i++)
> + intel_dp->aux_ch_data_reg[i] = aux_reg(dev_priv, port, i);
> }
>
> static void
> @@ -1074,11 +1091,11 @@ intel_dp_aux_init(struct intel_dp *intel_dp, struct intel_connector *connector)
> };
>
> if (INTEL_INFO(dev_priv)->gen >= 9)
> - intel_dp->aux_ch_ctl_reg = skl_aux_ctl_reg(dev_priv, port);
> + aux_reg_init(intel_dp, skl_aux_reg);
> else if (HAS_PCH_SPLIT(dev_priv))
> - intel_dp->aux_ch_ctl_reg = ilk_aux_ctl_reg(dev_priv, port);
> + aux_reg_init(intel_dp, ilk_aux_reg);
> else
> - intel_dp->aux_ch_ctl_reg = g4x_aux_ctl_reg(dev_priv, port);
> + aux_reg_init(intel_dp, g4x_aux_reg);
IMO there's too much trickery in this patch to get the registers set
up. -1 for ctl, i >> 2 to get the right index, function pointers for set
up only, etc. I don't think it would be too much duplication to add
separate initializers for each group of platforms, ditching the function
pointers and -1 for ctl.
BR,
Jani.
>
> intel_dp->aux.name = ddc_name[port];
> intel_dp->aux.dev = dev->dev;
> diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
> index 1df6ebf..3fea039 100644
> --- a/drivers/gpu/drm/i915/intel_drv.h
> +++ b/drivers/gpu/drm/i915/intel_drv.h
> @@ -717,6 +717,7 @@ struct sink_crc {
> struct intel_dp {
> uint32_t output_reg;
> uint32_t aux_ch_ctl_reg;
> + uint32_t aux_ch_data_reg[5];
> uint32_t DP;
> int link_rate;
> uint8_t lane_count;
> --
> 2.4.6
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx
--
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
