drm/amd/display: Enable FPO for configs that could reduce vlevel

[Description]
- On high refresh rate DRR displays that support VBLANK naturally,
  UCLK could be idling at DPM1 instead of DPM0 since it doesn't use
  FPO
- To achieve DPM0, enable FPO on these configs even though it can
  support P-State without FPO
- Default disable for now, have debug option to enable
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alvin Lee <Alvin.Lee2@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/display/dc/dc.h

@@ -873,6 +873,7 @@ struct dc_debug_options {
 	bool dig_fifo_off_in_blank;
 	bool temp_mst_deallocation_sequence;
 	bool override_dispclk_programming;
+	bool disable_fpo_optimizations;
 };
 
 struct gpu_info_soc_bounding_box_v1_0;

drivers/gpu/drm/amd/display/dc/dcn32/dcn32_resource.c

@@ -725,6 +725,7 @@ static const struct dc_debug_options debug_defaults_drv = {
 	.min_prefetch_in_strobe_ns = 60000, // 60us
 	.disable_unbounded_requesting = false,
 	.override_dispclk_programming = true,
+	.disable_fpo_optimizations = true,
 };
 
 static const struct dc_debug_options debug_defaults_diags = {

drivers/gpu/drm/amd/display/dc/dcn321/dcn321_resource.c

@@ -723,6 +723,7 @@ static const struct dc_debug_options debug_defaults_drv = {
 	.min_prefetch_in_strobe_ns = 60000, // 60us
 	.disable_unbounded_requesting = false,
 	.override_dispclk_programming = true,
+	.disable_fpo_optimizations = true,
 };
 
 static const struct dc_debug_options debug_defaults_diags = {

drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c

@@ -1961,7 +1961,8 @@ 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;
 
-	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);
@@ -1985,11 +1986,21 @@ 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;
+				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);
+			}
 		}
 	}