[PATCH v8 1/2] drm/i915/panelreplay: Panel replay workaround with VRR
Jani Nikula
jani.nikula at intel.com
Tue Jul 2 08:09:01 UTC 2024
On Tue, 02 Jul 2024, Animesh Manna <animesh.manna at intel.com> wrote:
> Panel Replay VSC SDP not getting sent when VRR is enabled
> and W1 and W2 are 0. So Program Set Context Latency in
> TRANS_SET_CONTEXT_LATENCY register to at least a value of 1.
The patch does a whole lot more than what's explained in the commit
message. Feels like the patch should maybe do less.
>
> HSD: 14015406119
>
> v1: Initial version.
> v2: Update timings stored in adjusted_mode struct. [Ville]
> v3: Add WA in compute_config(). [Ville]
> v4:
> - Add DISPLAY_VER() check and improve code comment. [Rodrigo]
> - Introduce centralized intel_crtc_vblank_delay(). [Ville]
> v5: Move to crtc_compute_config(). [Ville]
> v6: Restrict DISPLAY_VER till 14. [Mitul]
> v7:
> - Corrected code-comment. [Mitul]
> - dev_priv local variable removed. [Jani]
> v8: Introduce late_compute_config() which will take care late
> vblank-delay adjustment. [Ville]
I presume this change is based on [1], but I have a hard time mentally
mapping what's going on in the patch to what Ville asks there.
[1] https://lore.kernel.org/r/ZnlvWk-sOzDfMa0E@intel.com
>
> Signed-off-by: Animesh Manna <animesh.manna at intel.com>
> ---
> drivers/gpu/drm/i915/display/intel_alpm.c | 3 +-
> drivers/gpu/drm/i915/display/intel_alpm.h | 6 ++-
> drivers/gpu/drm/i915/display/intel_display.c | 54 +++++++++++++++++++-
> drivers/gpu/drm/i915/display/intel_display.h | 2 +
> drivers/gpu/drm/i915/display/intel_dp.c | 20 +++++++-
> drivers/gpu/drm/i915/display/intel_dp.h | 2 +
> drivers/gpu/drm/i915/display/intel_psr.c | 12 ++---
> drivers/gpu/drm/i915/display/intel_vrr.c | 13 -----
> 8 files changed, 84 insertions(+), 28 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/display/intel_alpm.c b/drivers/gpu/drm/i915/display/intel_alpm.c
> index 866b3b409c4d..021e970d8209 100644
> --- a/drivers/gpu/drm/i915/display/intel_alpm.c
> +++ b/drivers/gpu/drm/i915/display/intel_alpm.c
> @@ -266,8 +266,7 @@ bool intel_alpm_compute_params(struct intel_dp *intel_dp,
> }
>
> void intel_alpm_lobf_compute_config(struct intel_dp *intel_dp,
> - struct intel_crtc_state *crtc_state,
> - struct drm_connector_state *conn_state)
> + struct intel_crtc_state *crtc_state)
> {
> struct drm_i915_private *i915 = dp_to_i915(intel_dp);
> struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode;
> diff --git a/drivers/gpu/drm/i915/display/intel_alpm.h b/drivers/gpu/drm/i915/display/intel_alpm.h
> index 8c409b10dce6..8abe7bd9bc37 100644
> --- a/drivers/gpu/drm/i915/display/intel_alpm.h
> +++ b/drivers/gpu/drm/i915/display/intel_alpm.h
> @@ -16,9 +16,11 @@ struct intel_connector;
> void intel_alpm_init_dpcd(struct intel_dp *intel_dp);
> bool intel_alpm_compute_params(struct intel_dp *intel_dp,
> const struct intel_crtc_state *crtc_state);
> +bool intel_alpm_config_valid(struct intel_dp *intel_dp,
> + const struct intel_crtc_state *crtc_state,
> + bool aux_less);
> void intel_alpm_lobf_compute_config(struct intel_dp *intel_dp,
> - struct intel_crtc_state *crtc_state,
> - struct drm_connector_state *conn_state);
> + struct intel_crtc_state *crtc_state);
> void intel_alpm_configure(struct intel_dp *intel_dp,
> const struct intel_crtc_state *crtc_state);
> void intel_alpm_lobf_debugfs_add(struct intel_connector *connector);
> diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c
> index c2c388212e2e..cdab71f81eaa 100644
> --- a/drivers/gpu/drm/i915/display/intel_display.c
> +++ b/drivers/gpu/drm/i915/display/intel_display.c
> @@ -2512,9 +2512,21 @@ static int intel_crtc_compute_pipe_mode(struct intel_crtc_state *crtc_state)
> static int intel_crtc_compute_config(struct intel_atomic_state *state,
> struct intel_crtc *crtc)
> {
> + struct drm_connector *connector;
> + struct drm_connector_state *connector_state;
> struct intel_crtc_state *crtc_state =
> intel_atomic_get_new_crtc_state(state, crtc);
> - int ret;
> + int ret, i;
> +
> + for_each_new_connector_in_state(&state->base, connector, connector_state, i) {
> + struct intel_encoder *encoder =
> + to_intel_encoder(connector_state->best_encoder);
> +
> + if (connector_state->crtc != &crtc->base)
> + continue;
> +
> + intel_crtc_adjust_vblank_delay(crtc_state, encoder);
> + }
>
> ret = intel_dpll_crtc_compute_clock(state, crtc);
> if (ret)
> @@ -3925,6 +3937,46 @@ bool intel_crtc_get_pipe_config(struct intel_crtc_state *crtc_state)
> return true;
> }
>
> +void intel_crtc_adjust_vblank_delay(struct intel_crtc_state *crtc_state,
> + struct intel_encoder *encoder)
This could be static.
> +{
> + struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc);
> + struct drm_display_mode *adjusted_mode = &crtc_state->hw.adjusted_mode;
> +
> + /*
> + * wa_14015401596 for display versions 13, 14.
> + * Program Set Context Latency in TRANS_SET_CONTEXT_LATENCY register
> + * to at least a value of 1 when Panel Replay is enabled with VRR.
> + * Value for TRANS_SET_CONTEXT_LATENCY is calculated by substracting
> + * crtc_vdisplay from crtc_vblank_start, so incrementing crtc_vblank_start
> + * by 1 if both are equal.
> + */
> + if (crtc_state->vrr.enable && crtc_state->has_panel_replay &&
> + adjusted_mode->crtc_vblank_start == adjusted_mode->crtc_vdisplay &&
> + IS_DISPLAY_VER(to_i915(crtc->base.dev), 13, 14))
> + adjusted_mode->crtc_vblank_start += 1;
> +
> + if (crtc_state->vrr.enable) {
> + /*
> + * For XE_LPD+, we use guardband and pipeline override
> + * is deprecated.
> + */
> + if (DISPLAY_VER(to_i915(crtc->base.dev)) >= 13) {
> + crtc_state->vrr.guardband =
> + crtc_state->vrr.vmin + 1 - adjusted_mode->crtc_vblank_start;
> + } else {
> + crtc_state->vrr.pipeline_full =
> + min(255, crtc_state->vrr.vmin - adjusted_mode->crtc_vblank_start -
Misleading indentation (see where you're copy-pasting this from).
> + crtc_state->framestart_delay - 1);
> + }
I think only intel_vrr.c should be modifying vrr.guardband, vrr.pipeline_full.
> + }
> +
> + if (intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DP) ||
> + intel_crtc_has_type(crtc_state, INTEL_OUTPUT_EDP)) {
> + intel_dp_late_compute_config(encoder, crtc_state);
So, ugh. Naming this intel_dp_late_compute_config() is *very* confusing.
We have intel_modeset_pipe_config() calling encoder->compute_config().
Now you're adding intel_modeset_pipe_config() calling
intel_crtc_compute_config() calling intel_crtc_adjust_vblank_delay()
calling intel_dp_late_compute_config().
Yet there's separate intel_modeset_pipe_config_late() and
encoder->compute_config_late().
In general, it's very surprising to have a function called
intel_crtc_adjust_vblank_delay() do *much* more than, well, adjust
vblank delay.
> + }
> +}
> +
> int intel_dotclock_calculate(int link_freq,
> const struct intel_link_m_n *m_n)
> {
> diff --git a/drivers/gpu/drm/i915/display/intel_display.h b/drivers/gpu/drm/i915/display/intel_display.h
> index b0cf6ca70952..21fd330b8834 100644
> --- a/drivers/gpu/drm/i915/display/intel_display.h
> +++ b/drivers/gpu/drm/i915/display/intel_display.h
> @@ -428,6 +428,8 @@ bool intel_crtc_is_joiner_primary(const struct intel_crtc_state *crtc_state);
> u8 intel_crtc_joiner_secondary_pipes(const struct intel_crtc_state *crtc_state);
> struct intel_crtc *intel_primary_crtc(const struct intel_crtc_state *crtc_state);
> bool intel_crtc_get_pipe_config(struct intel_crtc_state *crtc_state);
> +void intel_crtc_adjust_vblank_delay(struct intel_crtc_state *crtc_state,
> + struct intel_encoder *encoder);
> bool intel_pipe_config_compare(const struct intel_crtc_state *current_config,
> const struct intel_crtc_state *pipe_config,
> bool fastset);
> diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c
> index 3903f6ead6e6..e01e72ec48b3 100644
> --- a/drivers/gpu/drm/i915/display/intel_dp.c
> +++ b/drivers/gpu/drm/i915/display/intel_dp.c
> @@ -3009,7 +3009,6 @@ intel_dp_compute_config(struct intel_encoder *encoder,
> intel_vrr_compute_config(pipe_config, conn_state);
> intel_dp_compute_as_sdp(intel_dp, pipe_config);
> intel_psr_compute_config(intel_dp, pipe_config, conn_state);
> - intel_alpm_lobf_compute_config(intel_dp, pipe_config, conn_state);
> intel_dp_drrs_compute_config(connector, pipe_config, link_bpp_x16);
> intel_dp_compute_vsc_sdp(intel_dp, pipe_config, conn_state);
> intel_dp_compute_hdr_metadata_infoframe_sdp(intel_dp, pipe_config, conn_state);
> @@ -3018,6 +3017,25 @@ intel_dp_compute_config(struct intel_encoder *encoder,
> pipe_config);
> }
>
> +void intel_dp_late_compute_config(struct intel_encoder *encoder,
> + struct intel_crtc_state *crtc_state)
> +{
> + struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
> +
> + if (crtc_state->has_sel_update &&
> + !intel_alpm_config_valid(intel_dp, crtc_state, false)) {
> + crtc_state->enable_psr2_sel_fetch = false;
> + crtc_state->has_sel_update = false;
> + }
> +
> + if (crtc_state->has_panel_replay && intel_dp_is_edp(intel_dp) &&
> + !intel_alpm_config_valid(intel_dp, crtc_state, false)) {
> + crtc_state->has_panel_replay = false;
> + }
I think only intel_psr.c should be modifying enable_psr2_sel_fetch,
has_sel_update, has_panel_replay.
> +
> + intel_alpm_lobf_compute_config(intel_dp, crtc_state);
> +}
> +
> void intel_dp_set_link_params(struct intel_dp *intel_dp,
> int link_rate, int lane_count)
> {
> diff --git a/drivers/gpu/drm/i915/display/intel_dp.h b/drivers/gpu/drm/i915/display/intel_dp.h
> index a0f990a95ecc..cd473f939941 100644
> --- a/drivers/gpu/drm/i915/display/intel_dp.h
> +++ b/drivers/gpu/drm/i915/display/intel_dp.h
> @@ -74,6 +74,8 @@ void intel_dp_encoder_flush_work(struct drm_encoder *encoder);
> int intel_dp_compute_config(struct intel_encoder *encoder,
> struct intel_crtc_state *pipe_config,
> struct drm_connector_state *conn_state);
> +void intel_dp_late_compute_config(struct intel_encoder *encoder,
> + struct intel_crtc_state *crtc_state);
> int intel_dp_dsc_compute_config(struct intel_dp *intel_dp,
> struct intel_crtc_state *pipe_config,
> struct drm_connector_state *conn_state,
> diff --git a/drivers/gpu/drm/i915/display/intel_psr.c b/drivers/gpu/drm/i915/display/intel_psr.c
> index 9cb1cdaaeefa..4dc917b7f447 100644
> --- a/drivers/gpu/drm/i915/display/intel_psr.c
> +++ b/drivers/gpu/drm/i915/display/intel_psr.c
> @@ -1344,9 +1344,9 @@ static bool wake_lines_fit_into_vblank(struct intel_dp *intel_dp,
> return true;
> }
>
> -static bool alpm_config_valid(struct intel_dp *intel_dp,
> - const struct intel_crtc_state *crtc_state,
> - bool aux_less)
> +bool intel_alpm_config_valid(struct intel_dp *intel_dp,
> + const struct intel_crtc_state *crtc_state,
> + bool aux_less)
> {
> struct drm_i915_private *i915 = dp_to_i915(intel_dp);
>
> @@ -1442,9 +1442,6 @@ static bool intel_psr2_config_valid(struct intel_dp *intel_dp,
> return false;
> }
>
> - if (!alpm_config_valid(intel_dp, crtc_state, false))
> - return false;
> -
The commit message does not explain this.
> if (!crtc_state->enable_psr2_sel_fetch &&
> (crtc_hdisplay > psr_max_h || crtc_vdisplay > psr_max_v)) {
> drm_dbg_kms(&dev_priv->drm,
> @@ -1583,9 +1580,6 @@ _panel_replay_compute_config(struct intel_dp *intel_dp,
> return false;
> }
>
> - if (!alpm_config_valid(intel_dp, crtc_state, true))
> - return false;
> -
Ditto.
> return true;
> }
>
> diff --git a/drivers/gpu/drm/i915/display/intel_vrr.c b/drivers/gpu/drm/i915/display/intel_vrr.c
> index 5a0da64c7db3..46341367d250 100644
> --- a/drivers/gpu/drm/i915/display/intel_vrr.c
> +++ b/drivers/gpu/drm/i915/display/intel_vrr.c
> @@ -242,19 +242,6 @@ intel_vrr_compute_config(struct intel_crtc_state *crtc_state,
> (crtc_state->hw.adjusted_mode.crtc_vtotal -
> crtc_state->hw.adjusted_mode.vsync_end);
> }
> -
> - /*
> - * For XE_LPD+, we use guardband and pipeline override
> - * is deprecated.
> - */
> - if (DISPLAY_VER(i915) >= 13) {
> - crtc_state->vrr.guardband =
> - crtc_state->vrr.vmin + 1 - adjusted_mode->crtc_vblank_start;
> - } else {
> - crtc_state->vrr.pipeline_full =
> - min(255, crtc_state->vrr.vmin - adjusted_mode->crtc_vblank_start -
> - crtc_state->framestart_delay - 1);
> - }
> }
>
> static u32 trans_vrr_ctl(const struct intel_crtc_state *crtc_state)
--
Jani Nikula, Intel
More information about the Intel-gfx
mailing list