drm/amd/display: Fix dcn21 num_states

[Why]
DML expects num_states to exclude the duplicate state.

[How]
Set num_states to correct value to prevent array off-by-one error.  Also
refactor max clock level code for diags.

Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Signed-off-by: George Shen <george.shen@amd.com>
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

4 files changed, 18 insertions, 30 deletions

diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c
index 54e096055c6f..5cdbba0cd873 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c
@@ -3343,7 +3343,7 @@ void dcn20_cap_soc_clocks(
 void dcn20_update_bounding_box(struct dc *dc, struct _vcs_dpi_soc_bounding_box_st *bb,
 		struct pp_smu_nv_clock_table *max_clocks, unsigned int *uclk_states, unsigned int num_states)
 {
-	struct _vcs_dpi_voltage_scaling_st calculated_states[MAX_CLOCK_LIMIT_STATES];
+	struct _vcs_dpi_voltage_scaling_st calculated_states[DC__VOLTAGE_STATES];
 	int i;
 	int num_calculated_states = 0;
 	int min_dcfclk = 0;
diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c
index deda1a6b603b..b25484aa8222 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c
@@ -300,7 +300,7 @@ struct _vcs_dpi_soc_bounding_box_st dcn2_1_soc = {
 	.xfc_bus_transport_time_us = 4,
 	.xfc_xbuf_latency_tolerance_us = 4,
 	.use_urgent_burst_bw = 1,
-	.num_states = 9
+	.num_states = 8
 };
 
 #ifndef MAX
@@ -1377,21 +1377,8 @@ static void update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_param
 	unsigned int i, j, k;
 	int closest_clk_lvl;
 
-	// diags does not retrieve proper values from SMU
-	// cap states to 5 and make state 5 the max state
-	if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) || IS_DIAG_DC(dc->ctx->dce_environment)) {
-		dcn2_1_soc.num_states = 5;
-
-		dcn2_1_soc.clock_limits[5].state = 5;
-		dcn2_1_soc.clock_limits[5].dcfclk_mhz = 810.0;
-		dcn2_1_soc.clock_limits[5].fabricclk_mhz = 1600.0;
-		dcn2_1_soc.clock_limits[5].dispclk_mhz = 1395.0;
-		dcn2_1_soc.clock_limits[5].dppclk_mhz = 1285.0;
-		dcn2_1_soc.clock_limits[5].phyclk_mhz = 1325.0;
-		dcn2_1_soc.clock_limits[5].socclk_mhz = 953.0;
-		dcn2_1_soc.clock_limits[5].dscclk_mhz = 489.0;
-		dcn2_1_soc.clock_limits[5].dram_speed_mts = 4266.0;
-	} else {
+	// Default clock levels are used for diags, which may lead to overclocking.
+	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment) && !IS_DIAG_DC(dc->ctx->dce_environment)) {
 		dcn2_1_ip.max_num_otg = pool->base.res_cap->num_timing_generator;
 		dcn2_1_ip.max_num_dpp = pool->base.pipe_count;
 		dcn2_1_soc.num_chans = bw_params->num_channels;
@@ -1404,16 +1391,16 @@ static void update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_param
 		dcn2_1_soc.clock_limits[0].dram_speed_mts = clk_table->entries[0].memclk_mhz * 2;
 
 		/*
-		 * Other levels: find cloest DCN clocks that fit the given clock limit using dcfclk
-		 * as indicater
+		 * Other levels: find closest DCN clocks that fit the given clock limit using dcfclk
+		 * as indicator
 		 */
 
 		closest_clk_lvl = -1;
 		/* index currently being filled */
 		k = 1;
 		for (i = 1; i < clk_table->num_entries; i++) {
-			/* loop backwards, skip duplicate state, +1 because SMU has precision issue */
-			for (j = dcn2_1_soc.num_states - 2; j >= k; j--) {
+			/* loop backwards, skip duplicate state*/
+			for (j = dcn2_1_soc.num_states - 1; j >= k; j--) {
 				if ((unsigned int) dcn2_1_soc.clock_limits[j].dcfclk_mhz <= clk_table->entries[i].dcfclk_mhz) {
 					closest_clk_lvl = j;
 					break;
@@ -1438,13 +1425,13 @@ static void update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_param
 				k++;
 			}
 		}
-
-		/* duplicate last level */
-		dcn2_1_soc.clock_limits[k] = dcn2_1_soc.clock_limits[k - 1];
-		dcn2_1_soc.clock_limits[k].state = k;
-		dcn2_1_soc.num_states = k + 1;
+		dcn2_1_soc.num_states = k;
 	}
 
+	/* duplicate last level */
+	dcn2_1_soc.clock_limits[dcn2_1_soc.num_states] = dcn2_1_soc.clock_limits[dcn2_1_soc.num_states - 1];
+	dcn2_1_soc.clock_limits[dcn2_1_soc.num_states].state = dcn2_1_soc.num_states;
+
 	dml_init_instance(&dc->dml, &dcn2_1_soc, &dcn2_1_ip, DML_PROJECT_DCN21);
 }
 
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dc_features.h b/drivers/gpu/drm/amd/display/dc/dml/dc_features.h
index ea4cde952f4f..2a1983324629 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dc_features.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/dc_features.h
@@ -29,7 +29,7 @@
 #define DC__PRESENT 1
 #define DC__PRESENT__1 1
 #define DC__NUM_DPP 4
-#define DC__VOLTAGE_STATES 7
+#define DC__VOLTAGE_STATES 9
 #define DC__NUM_DPP__4 1
 #define DC__NUM_DPP__0_PRESENT 1
 #define DC__NUM_DPP__1_PRESENT 1
diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h b/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h
index dfd3be452766..687010c17324 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h
@@ -22,11 +22,12 @@
  * Authors: AMD
  *
  */
+
+#include "dc_features.h"
+
 #ifndef __DISPLAY_MODE_STRUCTS_H__
 #define __DISPLAY_MODE_STRUCTS_H__
 
-#define MAX_CLOCK_LIMIT_STATES 9
-
 typedef struct _vcs_dpi_voltage_scaling_st voltage_scaling_st;
 typedef struct _vcs_dpi_soc_bounding_box_st soc_bounding_box_st;
 typedef struct _vcs_dpi_ip_params_st ip_params_st;
@@ -68,7 +69,7 @@ struct _vcs_dpi_voltage_scaling_st {
 };
 
 struct _vcs_dpi_soc_bounding_box_st {
-	struct _vcs_dpi_voltage_scaling_st clock_limits[MAX_CLOCK_LIMIT_STATES];
+	struct _vcs_dpi_voltage_scaling_st clock_limits[DC__VOLTAGE_STATES];
 	unsigned int num_states;
 	double sr_exit_time_us;
 	double sr_enter_plus_exit_time_us;