1 files changed, 108 insertions, 17 deletions

diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c
index e47828e3b6d5..4548320217fc 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c
@@ -131,7 +131,7 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_2_soc = {
 	.urgent_latency_pixel_data_only_us = 4.0,
 	.urgent_latency_pixel_mixed_with_vm_data_us = 4.0,
 	.urgent_latency_vm_data_only_us = 4.0,
-	.fclk_change_latency_us = 20,
+	.fclk_change_latency_us = 25,
 	.usr_retraining_latency_us = 2,
 	.smn_latency_us = 2,
 	.mall_allocated_for_dcn_mbytes = 64,
@@ -1270,7 +1270,7 @@ static bool is_dtbclk_required(struct dc *dc, struct dc_state *context)
 	for (i = 0; i < dc->res_pool->pipe_count; i++) {
 		if (!context->res_ctx.pipe_ctx[i].stream)
 			continue;
-		if (link_is_dp_128b_132b_signal(&context->res_ctx.pipe_ctx[i]))
+		if (dc->link_srv->dp_is_128b_132b_signal(&context->res_ctx.pipe_ctx[i]))
 			return true;
 	}
 	return false;
@@ -1927,6 +1927,7 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context,
 	unsigned int min_dram_speed_mts_margin;
 	bool need_fclk_lat_as_dummy = false;
 	bool is_subvp_p_drr = false;
+	struct dc_stream_state *fpo_candidate_stream = NULL;
 
 	dc_assert_fp_enabled();
 
@@ -1960,11 +1961,19 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context,
 	}
 
 	context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching = false;
+	for (i = 0; i < context->stream_count; i++) {
+		if (context->streams[i])
+			context->streams[i]->fpo_in_use = false;
+	}
 
-	if (!pstate_en) {
+	if (!pstate_en || (!dc->debug.disable_fpo_optimizations &&
+			pstate_en && vlevel != 0)) {
 		/* only when the mclk switch can not be natural, is the fw based vblank stretch attempted */
-		context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching =
-			dcn30_can_support_mclk_switch_using_fw_based_vblank_stretch(dc, context);
+		fpo_candidate_stream = dcn32_can_support_mclk_switch_using_fw_based_vblank_stretch(dc, context);
+		if (fpo_candidate_stream) {
+			fpo_candidate_stream->fpo_in_use = true;
+			context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching = true;
+		}
 
 		if (context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching) {
 			dummy_latency_index = dcn32_find_dummy_latency_index_for_fw_based_mclk_switch(dc,
@@ -1985,11 +1994,25 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context,
 				context->bw_ctx.dml.soc.fclk_change_latency_us =
 						dc->clk_mgr->bw_params->dummy_pstate_table[dummy_latency_index].dummy_pstate_latency_us;
 			}
-			dcn32_internal_validate_bw(dc, context, pipes, &pipe_cnt, &vlevel, false);
-			maxMpcComb = context->bw_ctx.dml.vba.maxMpcComb;
-			dcfclk_from_fw_based_mclk_switching = context->bw_ctx.dml.vba.DCFCLKState[vlevel][context->bw_ctx.dml.vba.maxMpcComb];
-			pstate_en = context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][maxMpcComb] !=
-					dm_dram_clock_change_unsupported;
+			dcn32_internal_validate_bw(dc, context, pipes, &pipe_cnt, &vlevel_temp, false);
+			if (vlevel_temp < vlevel) {
+				vlevel = vlevel_temp;
+				maxMpcComb = context->bw_ctx.dml.vba.maxMpcComb;
+				dcfclk_from_fw_based_mclk_switching = context->bw_ctx.dml.vba.DCFCLKState[vlevel][context->bw_ctx.dml.vba.maxMpcComb];
+				pstate_en = true;
+			} else {
+				/* Restore FCLK latency and re-run validation to go back to original validation
+				 * output if we find that enabling FPO does not give us any benefit (i.e. lower
+				 * voltage level)
+				 */
+				context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching = false;
+				for (i = 0; i < context->stream_count; i++) {
+					if (context->streams[i])
+						context->streams[i]->fpo_in_use = false;
+				}
+				context->bw_ctx.dml.soc.fclk_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.fclk_change_latency_us;
+				dcn32_internal_validate_bw(dc, context, pipes, &pipe_cnt, &vlevel, false);
+			}
 		}
 	}
 
@@ -2142,7 +2165,13 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context,
 		 * DCFCLK: Min, as reported by PM FW, when available
 		 * UCLK: Min, as reported by PM FW, when available
 		 */
-		dc->res_pool->funcs->update_soc_for_wm_a(dc, context);
+
+		/* For set A set the correct latency values (i.e. non-dummy values) unconditionally
+		 */
+		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us;
+		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.sr_enter_plus_exit_time_us;
+		context->bw_ctx.dml.soc.sr_exit_time_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.sr_exit_time_us;
+
 		context->bw_ctx.bw.dcn.watermarks.a.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
 		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_enter_plus_exit_ns = get_wm_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
 		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
@@ -2188,10 +2217,6 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context,
 		context->bw_ctx.dml.soc.dram_clock_change_latency_us =
 				dc->clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us;
 
-	if (context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching) {
-		dcn30_setup_mclk_switch_using_fw_based_vblank_stretch(dc, context);
-	}
-
 	/* revert fclk lat changes if required */
 	if (need_fclk_lat_as_dummy)
 		context->bw_ctx.dml.soc.fclk_change_latency_us =
@@ -2315,6 +2340,9 @@ static int build_synthetic_soc_states(struct clk_bw_params *bw_params,
 			num_dcfclk_dpms++;
 	}
 
+	if (num_dcfclk_dpms > 0 && bw_params->clk_table.entries[0].fclk_mhz > min_fclk_mhz)
+		min_fclk_mhz = bw_params->clk_table.entries[0].fclk_mhz;
+
 	if (!max_dcfclk_mhz || !max_dispclk_mhz || !max_dtbclk_mhz)
 		return -1;
 
@@ -2423,7 +2451,6 @@ static int build_synthetic_soc_states(struct clk_bw_params *bw_params,
 		for (i = *num_entries - 1; i >= 0 ; i--) {
 			if (table[i].fabricclk_mhz < min_fclk_mhz) {
 				table[i].fabricclk_mhz = min_fclk_mhz;
-				break;
 			}
 		}
 	}
@@ -2432,7 +2459,6 @@ static int build_synthetic_soc_states(struct clk_bw_params *bw_params,
 	for (i = *num_entries - 1; i >= 0 ; i--) {
 		if (table[i].dcfclk_mhz < min_dcfclk_mhz) {
 			table[i].dcfclk_mhz = min_dcfclk_mhz;
-			break;
 		}
 	}
 
@@ -2780,3 +2806,68 @@ double dcn32_determine_max_vratio_prefetch(struct dc *dc, struct dc_state *conte
 	}
 	return max_vratio_pre;
 }
+
+/**
+ * dcn32_assign_fpo_vactive_candidate - Assign the FPO stream candidate for FPO + VActive case
+ *
+ * This function chooses the FPO candidate stream for FPO + VActive cases (2 stream config).
+ * For FPO + VAtive cases, the assumption is that one display has ActiveMargin > 0, and the
+ * other display has ActiveMargin <= 0. This function will choose the pipe/stream that has
+ * ActiveMargin <= 0 to be the FPO stream candidate if found.
+ *
+ *
+ * @param [in]: dc - current dc state
+ * @param [in]: context - new dc state
+ * @param [out]: fpo_candidate_stream - pointer to FPO stream candidate if one is found
+ *
+ * Return: void
+ */
+void dcn32_assign_fpo_vactive_candidate(struct dc *dc, const struct dc_state *context, struct dc_stream_state **fpo_candidate_stream)
+{
+	unsigned int i, pipe_idx;
+	const struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
+
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		const struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (!pipe->stream)
+			continue;
+
+		if (vba->ActiveDRAMClockChangeLatencyMarginPerState[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] <= 0) {
+			*fpo_candidate_stream = pipe->stream;
+			break;
+		}
+		pipe_idx++;
+	}
+}
+
+/**
+ * dcn32_find_vactive_pipe - Determines if the config has a pipe that can switch in VACTIVE
+ *
+ * @param [in]: dc - current dc state
+ * @param [in]: context - new dc state
+ * @param [in]: vactive_margin_req_us - The vactive marign required for a vactive pipe to be
+ *                                      considered "found"
+ *
+ * Return: True if VACTIVE display is found, false otherwise
+ */
+bool dcn32_find_vactive_pipe(struct dc *dc, const struct dc_state *context, uint32_t vactive_margin_req_us)
+{
+	unsigned int i, pipe_idx;
+	const struct vba_vars_st *vba = &context->bw_ctx.dml.vba;
+	bool vactive_found = false;
+
+	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
+		const struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+
+		if (!pipe->stream)
+			continue;
+
+		if (vba->ActiveDRAMClockChangeLatencyMarginPerState[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] >= vactive_margin_req_us) {
+			vactive_found = true;
+			break;
+		}
+		pipe_idx++;
+	}
+	return vactive_found;
+}