drm/amd/display: add dummy pstate workaround to dcn315

DCN315 has to always allow pstate change or SMU will hang. This
workaround achieves this by applying a low pstate change latency
to be used when pstate is calculated to be unsupported. This lower
latency only accounts for memory retraining; a previous change
handles locking in the highest available pstate allowing us to minimize
required latency hiding to only account for memory retraining.
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/display/dc/dcn30/dcn30_resource.c

@@ -1655,6 +1655,9 @@ noinline bool dcn30_internal_validate_bw(
 	if (!pipes)
 		return false;
 
+	context->bw_ctx.dml.vba.maxMpcComb = 0;
+	context->bw_ctx.dml.vba.VoltageLevel = 0;
+	context->bw_ctx.dml.vba.DRAMClockChangeSupport[0][0] = dm_dram_clock_change_vactive;
 	dc->res_pool->funcs->update_soc_for_wm_a(dc, context);
 	pipe_cnt = dc->res_pool->funcs->populate_dml_pipes(dc, context, pipes, fast_validate);
 
@@ -1873,6 +1876,7 @@ noinline bool dcn30_internal_validate_bw(
 
 	if (repopulate_pipes)
 		pipe_cnt = dc->res_pool->funcs->populate_dml_pipes(dc, context, pipes, fast_validate);
+	context->bw_ctx.dml.vba.VoltageLevel = vlevel;
 	*vlevel_out = vlevel;
 	*pipe_cnt_out = pipe_cnt;
 

drivers/gpu/drm/amd/display/dc/dcn315/dcn315_resource.c

@@ -1731,7 +1731,7 @@ static struct resource_funcs dcn315_res_pool_funcs = {
 	.panel_cntl_create = dcn31_panel_cntl_create,
 	.validate_bandwidth = dcn31_validate_bandwidth,
 	.calculate_wm_and_dlg = dcn31_calculate_wm_and_dlg,
-	.update_soc_for_wm_a = dcn31_update_soc_for_wm_a,
+	.update_soc_for_wm_a = dcn315_update_soc_for_wm_a,
 	.populate_dml_pipes = dcn315_populate_dml_pipes_from_context,
 	.acquire_idle_pipe_for_layer = dcn20_acquire_idle_pipe_for_layer,
 	.add_stream_to_ctx = dcn30_add_stream_to_ctx,

drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.c

@@ -292,6 +292,7 @@ static struct _vcs_dpi_soc_bounding_box_st dcn3_15_soc = {
 	.urgent_latency_adjustment_fabric_clock_component_us = 0,
 	.urgent_latency_adjustment_fabric_clock_reference_mhz = 0,
 	.num_chans = 4,
+	.dummy_pstate_latency_us = 10.0
 };
 
 struct _vcs_dpi_ip_params_st dcn3_16_ip = {
@@ -459,6 +460,23 @@ void dcn31_update_soc_for_wm_a(struct dc *dc, struct dc_state *context)
 	}
 }
 
+void dcn315_update_soc_for_wm_a(struct dc *dc, struct dc_state *context)
+{
+	dc_assert_fp_enabled();
+
+	if (dc->clk_mgr->bw_params->wm_table.entries[WM_A].valid) {
+		/* For 315 pstate change is only supported if possible in vactive */
+		if (context->bw_ctx.dml.vba.DRAMClockChangeSupport[context->bw_ctx.dml.vba.VoltageLevel][context->bw_ctx.dml.vba.maxMpcComb] != dm_dram_clock_change_vactive)
+			context->bw_ctx.dml.soc.dram_clock_change_latency_us = context->bw_ctx.dml.soc.dummy_pstate_latency_us;
+		else
+			context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.entries[WM_A].pstate_latency_us;
+		context->bw_ctx.dml.soc.sr_enter_plus_exit_time_us =
+				dc->clk_mgr->bw_params->wm_table.entries[WM_A].sr_enter_plus_exit_time_us;
+		context->bw_ctx.dml.soc.sr_exit_time_us =
+				dc->clk_mgr->bw_params->wm_table.entries[WM_A].sr_exit_time_us;
+	}
+}
+
 void dcn31_calculate_wm_and_dlg_fp(
 		struct dc *dc, struct dc_state *context,
 		display_e2e_pipe_params_st *pipes,
@@ -486,72 +504,6 @@ void dcn31_calculate_wm_and_dlg_fp(
 	pipes[0].clks_cfg.dcfclk_mhz = dcfclk;
 	pipes[0].clks_cfg.socclk_mhz = context->bw_ctx.dml.soc.clock_limits[vlevel].socclk_mhz;
 
-#if 0 // TODO
-	/* Set B:
-	 * TODO
-	 */
-	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].valid) {
-		if (vlevel == 0) {
-			pipes[0].clks_cfg.voltage = 1;
-			pipes[0].clks_cfg.dcfclk_mhz = context->bw_ctx.dml.soc.clock_limits[0].dcfclk_mhz;
-		}
-		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_B].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_B].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_B].dml_input.sr_exit_time_us;
-	}
-	context->bw_ctx.bw.dcn.watermarks.b.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.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.b.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_enter_plus_exit_z8_ns = get_wm_z8_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.cstate_exit_z8_ns = get_wm_z8_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.b.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-
-	pipes[0].clks_cfg.voltage = vlevel;
-	pipes[0].clks_cfg.dcfclk_mhz = dcfclk;
-
-	/* Set C:
-	 * TODO
-	 */
-	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].valid) {
-		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].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_C].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_C].dml_input.sr_exit_time_us;
-	}
-	context->bw_ctx.bw.dcn.watermarks.c.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.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.c.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_enter_plus_exit_z8_ns = get_wm_z8_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.cstate_exit_z8_ns = get_wm_z8_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.c.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-
-	/* Set D:
-	 * TODO
-	 */
-	if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].valid) {
-		context->bw_ctx.dml.soc.dram_clock_change_latency_us = dc->clk_mgr->bw_params->wm_table.nv_entries[WM_D].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_D].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_D].dml_input.sr_exit_time_us;
-	}
-	context->bw_ctx.bw.dcn.watermarks.d.urgent_ns = get_wm_urgent(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.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.d.cstate_pstate.cstate_exit_ns = get_wm_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.pstate_change_ns = get_wm_dram_clock_change(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_enter_plus_exit_z8_ns = get_wm_z8_stutter_enter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.cstate_pstate.cstate_exit_z8_ns = get_wm_z8_stutter_exit(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.pte_meta_urgent_ns = get_wm_memory_trip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	context->bw_ctx.bw.dcn.watermarks.d.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-#endif
-
 	/* Set A:
 	 * All clocks min required
 	 *
@@ -568,11 +520,9 @@ void dcn31_calculate_wm_and_dlg_fp(
 	context->bw_ctx.bw.dcn.watermarks.a.frac_urg_bw_nom = get_fraction_of_urgent_bandwidth(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
 	context->bw_ctx.bw.dcn.watermarks.a.frac_urg_bw_flip = get_fraction_of_urgent_bandwidth_imm_flip(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
 	context->bw_ctx.bw.dcn.watermarks.a.urgent_latency_ns = get_urgent_latency(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
-	/* TODO: remove: */
 	context->bw_ctx.bw.dcn.watermarks.b = context->bw_ctx.bw.dcn.watermarks.a;
 	context->bw_ctx.bw.dcn.watermarks.c = context->bw_ctx.bw.dcn.watermarks.a;
 	context->bw_ctx.bw.dcn.watermarks.d = context->bw_ctx.bw.dcn.watermarks.a;
-	/* end remove*/
 
 	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
 		if (!context->res_ctx.pipe_ctx[i].stream)
@@ -594,6 +544,9 @@ void dcn31_calculate_wm_and_dlg_fp(
 	}
 
 	dcn20_calculate_dlg_params(dc, context, pipes, pipe_cnt, vlevel);
+	/* For 31x apu pstate change is only supported if possible in vactive */
+	context->bw_ctx.bw.dcn.clk.p_state_change_support =
+			context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb] == dm_dram_clock_change_vactive;
 }
 
 void dcn31_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)

drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.h

@@ -35,6 +35,7 @@ void dcn31_zero_pipe_dcc_fraction(display_e2e_pipe_params_st *pipes,
 				  int pipe_cnt);
 
 void dcn31_update_soc_for_wm_a(struct dc *dc, struct dc_state *context);
+void dcn315_update_soc_for_wm_a(struct dc *dc, struct dc_state *context);
 
 void dcn31_calculate_wm_and_dlg_fp(
 		struct dc *dc, struct dc_state *context,