Merge tag 'amd-drm-next-6.11-2024-06-07' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-next-6.11-2024-06-07:

amdgpu:
- DCN 4.0.x support
- DCN 3.5 updates
- GC 12.0 support
- DP MST fixes
- Cursor fixes
- MES11 updates
- MMHUB 4.1 support
- DML2 Updates
- DCN 3.1.5 fixes
- IPS fixes
- Various code cleanups
- GMC 12.0 support
- SDMA 7.0 support
- SMU 13 updates
- SR-IOV fixes
- VCN 5.x fixes
- MES12 support
- SMU 14.x updates
- Devcoredump improvements
- Fixes for HDP flush on platforms with >4k pages
- GC 9.4.3 fixes
- RAS ACA updates
- Silence UBSAN flex array warnings
- MMHUB 3.3 updates

amdkfd:
- Contiguous VRAM allocations
- GC 12.0 support
- SDMA 7.0 support
- SR-IOV fixes

radeon:
- Backlight workaround for iMac
- Silence UBSAN flex array warnings

UAPI:
- GFX12 modifier and DCC support
  Proposed Mesa changes:
  https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/29510
- KFD GFX ALU exceptions
  Proposed ROCdebugger changes:
  https://github.com/ROCm/ROCdbgapi/commit/08c760622b6601abf906f75abbc5e21d9fd425df
  https://github.com/ROCm/ROCgdb/commit/944fe1c1414a68700414e86e32273b6bfa62ba6f
- KFD Contiguous VRAM allocation flag
  Proposed ROCr/HIP changes:
  https://github.com/ROCm/ROCT-Thunk-Interface/commit/f7b4a269914a3ab4f1e2453c2879adb97b5cc9e5
  https://github.com/ROCm/ROCR-Runtime/pull/214/commits/26e8530d05a775872cb06dde6693db72be0c454a
  https://github.com/ROCm/clr/commit/1d48f2a1ab38b632919c4b7274899b3faf4279ff

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240607195900.902537-1-alexander.deucher@amd.com

17 files changed, 351 insertions, 42 deletions

diff --git a/drivers/gpu/drm/amd/display/dc/dml/Makefile b/drivers/gpu/drm/amd/display/dc/dml/Makefile
index a94b6d546cd1..3c0222aa4df1 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/Makefile
+++ b/drivers/gpu/drm/amd/display/dc/dml/Makefile
@@ -108,6 +108,9 @@ CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/calcs/dcn_calcs.o := $(dml_rcflags)
 CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/calcs/dcn_calc_auto.o := $(dml_rcflags)
 CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/calcs/dcn_calc_math.o := $(dml_rcflags)
 
+CFLAGS_$(AMDDALPATH)/dc/dml/dcn401/dcn401_fpu.o := $(dml_ccflags)
+CFLAGS_REMOVE_$(AMDDALPATH)/dc/dml/dcn401/dcn401_fpu.o := $(dml_rcflags)
+
 ifdef CONFIG_DRM_AMD_DC_FP
 DML += display_mode_lib.o display_rq_dlg_helpers.o dml1_display_rq_dlg_calc.o
 DML += dcn10/dcn10_fpu.o
@@ -128,6 +131,7 @@ DML += dcn303/dcn303_fpu.o
 DML += dcn314/dcn314_fpu.o
 DML += dcn35/dcn35_fpu.o
 DML += dcn351/dcn351_fpu.o
+DML += dcn401/dcn401_fpu.o
 DML += dsc/rc_calc_fpu.o
 DML += calcs/dcn_calcs.o calcs/dcn_calc_math.o calcs/dcn_calc_auto.o
 endif
diff --git a/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calcs.c b/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calcs.c
index 0c4a8fe8e5ca..f1cde1e4265f 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calcs.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/calcs/dcn_calcs.c
@@ -1453,10 +1453,9 @@ void dcn_bw_update_from_pplib_fclks(
 	ASSERT(fclks->num_levels);
 
 	vmin0p65_idx = 0;
-	vmid0p72_idx = fclks->num_levels -
-		(fclks->num_levels > 2 ? 3 : (fclks->num_levels > 1 ? 2 : 1));
-	vnom0p8_idx = fclks->num_levels - (fclks->num_levels > 1 ? 2 : 1);
-	vmax0p9_idx = fclks->num_levels - 1;
+	vmid0p72_idx = fclks->num_levels > 2 ? fclks->num_levels - 3 : 0;
+	vnom0p8_idx = fclks->num_levels > 1 ? fclks->num_levels - 2 : 0;
+	vmax0p9_idx = fclks->num_levels > 0 ? fclks->num_levels - 1 : 0;
 
 	dc->dcn_soc->fabric_and_dram_bandwidth_vmin0p65 =
 		32 * (fclks->data[vmin0p65_idx].clocks_in_khz / 1000.0) / 1000.0;
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn10/dcn10_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn10/dcn10_fpu.h
index 63219ecd8478..1bf6b12f5663 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn10/dcn10_fpu.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn10/dcn10_fpu.h
@@ -29,4 +29,4 @@
 
 void dcn10_resource_construct_fp(struct dc *dc);
 
-#endif /* __DCN20_FPU_H__ */
+#endif /* __DCN10_FPU_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn30/dcn30_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn30/dcn30_fpu.c
index 81f7b90849ce..aac0a0ae2966 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn30/dcn30_fpu.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn30/dcn30_fpu.c
@@ -387,13 +387,17 @@ void dcn30_fpu_calculate_wm_and_dlg(
 	double dcfclk = context->bw_ctx.dml.vba.DCFCLKState[vlevel][maxMpcComb];
 	bool pstate_en = context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][maxMpcComb] != dm_dram_clock_change_unsupported;
 	unsigned int dummy_latency_index = 0;
+	struct dc_stream_status *stream_status = NULL;
 
 	dc_assert_fp_enabled();
 
 	context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching = false;
-    for (i = 0; i < context->stream_count; i++) {
+	for (i = 0; i < context->stream_count; i++) {
+		stream_status = NULL;
 		if (context->streams[i])
-			context->streams[i]->fpo_in_use = false;
+			stream_status = dc_state_get_stream_status(context, context->streams[i]);
+		if (stream_status)
+			stream_status->fpo_in_use = false;
 	}
 
 	if (!pstate_en) {
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.c
index 6ce90678b33c..0c0b2d67c9cd 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.c
@@ -320,7 +320,7 @@ static void calculate_wm_set_for_vlevel(int vlevel,
 
 }
 
-void dcn301_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)
+void dcn301_fpu_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params)
 {
 	struct _vcs_dpi_voltage_scaling_st *s = dc->scratch.update_bw_bounding_box.clock_limits;
 	struct dcn301_resource_pool *pool = TO_DCN301_RES_POOL(dc->res_pool);
@@ -409,7 +409,7 @@ void dcn301_fpu_init_soc_bounding_box(struct bp_soc_bb_info bb_info)
 		dcn3_01_soc.sr_exit_time_us = bb_info.dram_sr_exit_latency_100ns * 10;
 }
 
-void dcn301_calculate_wm_and_dlg_fp(struct dc *dc,
+void dcn301_fpu_calculate_wm_and_dlg(struct dc *dc,
 		struct dc_state *context,
 		display_e2e_pipe_params_st *pipes,
 		int pipe_cnt,
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.h
index 774b0fdfc80b..3e103e23dc6f 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn301/dcn301_fpu.h
@@ -26,15 +26,14 @@
 #ifndef __DCN301_FPU_H__
 #define __DCN301_FPU_H__
 
-void dcn301_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params);
+void dcn301_fpu_init_soc_bounding_box(struct bp_soc_bb_info bb_info);
+void dcn301_fpu_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params);
 
 void dcn301_fpu_set_wm_ranges(int i,
 	struct pp_smu_wm_range_sets *ranges,
 	struct _vcs_dpi_soc_bounding_box_st *loaded_bb);
 
-void dcn301_fpu_init_soc_bounding_box(struct bp_soc_bb_info bb_info);
-
-void dcn301_calculate_wm_and_dlg_fp(struct dc *dc,
+void dcn301_fpu_calculate_wm_and_dlg(struct dc *dc,
 		struct dc_state *context,
 		display_e2e_pipe_params_st *pipes,
 		int pipe_cnt,
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn302/dcn302_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn302/dcn302_fpu.c
index e2bcd205aa93..8da97a96b1ce 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn302/dcn302_fpu.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn302/dcn302_fpu.c
@@ -304,6 +304,16 @@ void dcn302_fpu_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_p
 			dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
 		}
 
+		/* bw_params->clk_table.entries[MAX_NUM_DPM_LVL].
+		 * MAX_NUM_DPM_LVL is 8.
+		 * dcn3_02_soc.clock_limits[DC__VOLTAGE_STATES].
+		 * DC__VOLTAGE_STATES is 40.
+		 */
+		if (num_states > MAX_NUM_DPM_LVL) {
+			ASSERT(0);
+			return;
+		}
+
 		dcn3_02_soc.num_states = num_states;
 		for (i = 0; i < dcn3_02_soc.num_states; i++) {
 			dcn3_02_soc.clock_limits[i].state = i;
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn303/dcn303_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn303/dcn303_fpu.c
index 3f02bb806d42..e968870a4b81 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn303/dcn303_fpu.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn303/dcn303_fpu.c
@@ -310,6 +310,16 @@ void dcn303_fpu_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_p
 			dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
 		}
 
+		/* bw_params->clk_table.entries[MAX_NUM_DPM_LVL].
+		 * MAX_NUM_DPM_LVL is 8.
+		 * dcn3_02_soc.clock_limits[DC__VOLTAGE_STATES].
+		 * DC__VOLTAGE_STATES is 40.
+		 */
+		if (num_states > MAX_NUM_DPM_LVL) {
+			ASSERT(0);
+			return;
+		}
+
 		dcn3_03_soc.num_states = num_states;
 		for (i = 0; i < dcn3_03_soc.num_states; i++) {
 			dcn3_03_soc.clock_limits[i].state = i;
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.c
index 94317b2e4a85..17a21bcbde17 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn31/dcn31_fpu.c
@@ -647,9 +647,9 @@ void dcn31_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_params
 			dcn3_1_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz;
 		s[i].phyclk_mhz = dcn3_1_soc.clock_limits[closest_clk_lvl].phyclk_mhz;
 	}
-	if (clk_table->num_entries) {
+
+	if (clk_table->num_entries)
 		dcn3_1_soc.num_states = clk_table->num_entries;
-	}
 
 	memcpy(dcn3_1_soc.clock_limits, s, sizeof(dcn3_1_soc.clock_limits));
 
@@ -762,23 +762,11 @@ void dcn316_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_param
 				break;
 			}
 		}
-		// Ported from DCN315
-		if (clk_table->num_entries == 1) {
-			/*smu gives one DPM level, let's take the highest one*/
-			closest_clk_lvl = dcn3_16_soc.num_states - 1;
-		}
 
 		s[i].state = i;
 
 		/* Clocks dependent on voltage level. */
 		s[i].dcfclk_mhz = clk_table->entries[i].dcfclk_mhz;
-		if (clk_table->num_entries == 1 &&
-		    s[i].dcfclk_mhz <
-		    dcn3_16_soc.clock_limits[closest_clk_lvl].dcfclk_mhz) {
-			/*SMU fix not released yet*/
-			s[i].dcfclk_mhz =
-				dcn3_16_soc.clock_limits[closest_clk_lvl].dcfclk_mhz;
-		}
 		s[i].fabricclk_mhz = clk_table->entries[i].fclk_mhz;
 		s[i].socclk_mhz = clk_table->entries[i].socclk_mhz;
 		s[i].dram_speed_mts = clk_table->entries[i].memclk_mhz *
@@ -799,9 +787,9 @@ void dcn316_update_bw_bounding_box(struct dc *dc, struct clk_bw_params *bw_param
 			dcn3_16_soc.clock_limits[closest_clk_lvl].phyclk_d18_mhz;
 		s[i].phyclk_mhz = dcn3_16_soc.clock_limits[closest_clk_lvl].phyclk_mhz;
 	}
-	if (clk_table->num_entries) {
+
+	if (clk_table->num_entries)
 		dcn3_16_soc.num_states = clk_table->num_entries;
-	}
 
 	memcpy(dcn3_16_soc.clock_limits, s, sizeof(dcn3_16_soc.clock_limits));
 
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c
index 3242957d00c5..f52b9e3d2bee 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn314/display_mode_vba_314.c
@@ -24,10 +24,7 @@
  *
  */
 
-#define UNIT_TEST 0
-#if !UNIT_TEST
 #include "dc.h"
-#endif
 #include "../display_mode_lib.h"
 #include "display_mode_vba_314.h"
 #include "../dml_inline_defs.h"
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 f6fe0a64beac..194422dd979d 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
@@ -2309,6 +2309,7 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context,
 	bool need_fclk_lat_as_dummy = false;
 	bool is_subvp_p_drr = false;
 	struct dc_stream_state *fpo_candidate_stream = NULL;
+	struct dc_stream_status *stream_status = NULL;
 
 	dc_assert_fp_enabled();
 
@@ -2343,8 +2344,11 @@ 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++) {
+		stream_status = NULL;
 		if (context->streams[i])
-			context->streams[i]->fpo_in_use = false;
+			stream_status = dc_state_get_stream_status(context, context->streams[i]);
+		if (stream_status)
+			stream_status->fpo_in_use = false;
 	}
 
 	if (!pstate_en || (!dc->debug.disable_fpo_optimizations &&
@@ -2352,7 +2356,9 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context,
 		/* only when the mclk switch can not be natural, is the fw based vblank stretch attempted */
 		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;
+			stream_status = dc_state_get_stream_status(context, fpo_candidate_stream);
+			if (stream_status)
+				stream_status->fpo_in_use = true;
 			context->bw_ctx.bw.dcn.clk.fw_based_mclk_switching = true;
 		}
 
@@ -2389,8 +2395,11 @@ 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++) {
+					stream_status = NULL;
 					if (context->streams[i])
-						context->streams[i]->fpo_in_use = false;
+						stream_status = dc_state_get_stream_status(context, context->streams[i]);
+					if (stream_status)
+						stream_status->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);
@@ -3232,6 +3241,16 @@ void dcn32_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_pa
 				dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
 			}
 
+			/* bw_params->clk_table.entries[MAX_NUM_DPM_LVL].
+			 * MAX_NUM_DPM_LVL is 8.
+			 * dcn3_02_soc.clock_limits[DC__VOLTAGE_STATES].
+			 * DC__VOLTAGE_STATES is 40.
+			 */
+			if (num_states > MAX_NUM_DPM_LVL) {
+				ASSERT(0);
+				return;
+			}
+
 			dcn3_2_soc.num_states = num_states;
 			for (i = 0; i < dcn3_2_soc.num_states; i++) {
 				dcn3_2_soc.clock_limits[i].state = i;
@@ -3521,15 +3540,16 @@ void dcn32_assign_fpo_vactive_candidate(struct dc *dc, const struct dc_state *co
  *
  * @dc: current dc state
  * @context: new dc state
+ * @fpo_candidate_stream: candidate stream to be chosen for FPO
  * @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)
+bool dcn32_find_vactive_pipe(struct dc *dc, const struct dc_state *context, struct dc_stream_state *fpo_candidate_stream, 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;
+	bool vactive_found = true;
 	unsigned int blank_us = 0;
 
 	for (i = 0, pipe_idx = 0; i < dc->res_pool->pipe_count; i++) {
@@ -3538,11 +3558,20 @@ bool dcn32_find_vactive_pipe(struct dc *dc, const struct dc_state *context, uint
 		if (!pipe->stream)
 			continue;
 
+		/* Don't need to check for vactive margin on the FPO candidate stream */
+		if (fpo_candidate_stream && pipe->stream == fpo_candidate_stream) {
+			pipe_idx++;
+			continue;
+		}
+
+		/* Every plane (apart from the ones driven by the FPO pipes) needs to have active margin
+		 * in order for us to have found a valid "vactive" config for FPO + Vactive
+		 */
 		blank_us = ((pipe->stream->timing.v_total - pipe->stream->timing.v_addressable) * pipe->stream->timing.h_total /
 				(double)(pipe->stream->timing.pix_clk_100hz * 100)) * 1000000;
-		if (vba->ActiveDRAMClockChangeLatencyMarginPerState[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] >= vactive_margin_req_us &&
-				!(pipe->stream->vrr_active_variable || pipe->stream->vrr_active_fixed) && blank_us < dc->debug.fpo_vactive_max_blank_us) {
-			vactive_found = true;
+		if (vba->ActiveDRAMClockChangeLatencyMarginPerState[vba->VoltageLevel][vba->maxMpcComb][vba->pipe_plane[pipe_idx]] < vactive_margin_req_us ||
+				pipe->stream->vrr_active_variable || pipe->stream->vrr_active_fixed || blank_us >= dc->debug.fpo_vactive_max_blank_us) {
+			vactive_found = false;
 			break;
 		}
 		pipe_idx++;
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.h
index d25c3f730a59..276e90e4e0ce 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.h
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.h
@@ -71,7 +71,7 @@ void dcn32_zero_pipe_dcc_fraction(display_e2e_pipe_params_st *pipes,
 
 void dcn32_assign_fpo_vactive_candidate(struct dc *dc, const struct dc_state *context, struct dc_stream_state **fpo_candidate_stream);
 
-bool dcn32_find_vactive_pipe(struct dc *dc, const struct dc_state *context, uint32_t vactive_margin_req);
+bool dcn32_find_vactive_pipe(struct dc *dc, const struct dc_state *context, struct dc_stream_state *fpo_candidate_stream, uint32_t vactive_margin_req);
 
 void dcn32_override_min_req_memclk(struct dc *dc, struct dc_state *context);
 
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c
index ff4d795c7966..4297402bdab3 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c
@@ -803,6 +803,16 @@ void dcn321_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_p
 			dram_speed_mts[num_states++] = bw_params->clk_table.entries[j++].memclk_mhz * 16;
 		}
 
+		/* bw_params->clk_table.entries[MAX_NUM_DPM_LVL].
+		 * MAX_NUM_DPM_LVL is 8.
+		 * dcn3_02_soc.clock_limits[DC__VOLTAGE_STATES].
+		 * DC__VOLTAGE_STATES is 40.
+		 */
+		if (num_states > MAX_NUM_DPM_LVL) {
+			ASSERT(0);
+			return;
+		}
+
 		dcn3_21_soc.num_states = num_states;
 		for (i = 0; i < dcn3_21_soc.num_states; i++) {
 			dcn3_21_soc.clock_limits[i].state = i;
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn401/dcn401_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn401/dcn401_fpu.c
new file mode 100644
index 000000000000..4fbecb5ff349
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn401/dcn401_fpu.c
@@ -0,0 +1,239 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#include "dcn401_fpu.h"
+#include "dcn401/dcn401_resource.h"
+// We need this includes for WATERMARKS_* defines
+#include "clk_mgr/dcn401/dcn401_smu14_driver_if.h"
+#include "link.h"
+
+#define DC_LOGGER_INIT(logger)
+
+void dcn401_build_wm_range_table_fpu(struct clk_mgr *clk_mgr)
+{
+	/* defaults */
+	double pstate_latency_us = clk_mgr->ctx->dc->dml.soc.dram_clock_change_latency_us;
+	double fclk_change_latency_us = clk_mgr->ctx->dc->dml.soc.fclk_change_latency_us;
+	double sr_exit_time_us = clk_mgr->ctx->dc->dml.soc.sr_exit_time_us;
+	double sr_enter_plus_exit_time_us = clk_mgr->ctx->dc->dml.soc.sr_enter_plus_exit_time_us;
+	/* For min clocks use as reported by PM FW and report those as min */
+	uint16_t min_uclk_mhz			= clk_mgr->bw_params->clk_table.entries[0].memclk_mhz;
+	uint16_t min_dcfclk_mhz			= clk_mgr->bw_params->clk_table.entries[0].dcfclk_mhz;
+	uint16_t setb_min_uclk_mhz		= min_uclk_mhz;
+	uint16_t dcfclk_mhz_for_the_second_state = clk_mgr->ctx->dc->dml.soc.clock_limits[2].dcfclk_mhz;
+
+	dc_assert_fp_enabled();
+
+	/* For Set B ranges use min clocks state 2 when available, and report those to PM FW */
+	if (dcfclk_mhz_for_the_second_state)
+		clk_mgr->bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.min_dcfclk = dcfclk_mhz_for_the_second_state;
+	else
+		clk_mgr->bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.min_dcfclk = clk_mgr->bw_params->clk_table.entries[0].dcfclk_mhz;
+
+	if (clk_mgr->bw_params->clk_table.entries[2].memclk_mhz)
+		setb_min_uclk_mhz = clk_mgr->bw_params->clk_table.entries[2].memclk_mhz;
+
+	/* Set A - Normal - default values */
+	clk_mgr->bw_params->wm_table.nv_entries[WM_A].valid = true;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.pstate_latency_us = pstate_latency_us;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.fclk_change_latency_us = fclk_change_latency_us;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.sr_exit_time_us = sr_exit_time_us;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_A].dml_input.sr_enter_plus_exit_time_us = sr_enter_plus_exit_time_us;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.wm_type = WATERMARKS_CLOCK_RANGE;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.min_dcfclk = min_dcfclk_mhz;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.max_dcfclk = 0xFFFF;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.min_uclk = min_uclk_mhz;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_A].pmfw_breakdown.max_uclk = 0xFFFF;
+
+	/* Set B - Performance - higher clocks, using DPM[2] DCFCLK and UCLK */
+	clk_mgr->bw_params->wm_table.nv_entries[WM_B].valid = true;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.pstate_latency_us = pstate_latency_us;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.fclk_change_latency_us = fclk_change_latency_us;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.sr_exit_time_us = sr_exit_time_us;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_B].dml_input.sr_enter_plus_exit_time_us = sr_enter_plus_exit_time_us;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.wm_type = WATERMARKS_CLOCK_RANGE;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.max_dcfclk = 0xFFFF;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.min_uclk = setb_min_uclk_mhz;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_B].pmfw_breakdown.max_uclk = 0xFFFF;
+
+	/* Set C - Dummy P-State - P-State latency set to "dummy p-state" value */
+	/* 'DalDummyClockChangeLatencyNs' registry key option set to 0x7FFFFFFF can be used to disable Set C for dummy p-state */
+	if (clk_mgr->ctx->dc->bb_overrides.dummy_clock_change_latency_ns != 0x7FFFFFFF) {
+		clk_mgr->bw_params->wm_table.nv_entries[WM_C].valid = true;
+		clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.pstate_latency_us = 50;
+		clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.fclk_change_latency_us = fclk_change_latency_us;
+		clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.sr_exit_time_us = sr_exit_time_us;
+		clk_mgr->bw_params->wm_table.nv_entries[WM_C].dml_input.sr_enter_plus_exit_time_us = sr_enter_plus_exit_time_us;
+		clk_mgr->bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.wm_type = WATERMARKS_DUMMY_PSTATE;
+		clk_mgr->bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.min_dcfclk = min_dcfclk_mhz;
+		clk_mgr->bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.max_dcfclk = 0xFFFF;
+		clk_mgr->bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.min_uclk = min_uclk_mhz;
+		clk_mgr->bw_params->wm_table.nv_entries[WM_C].pmfw_breakdown.max_uclk = 0xFFFF;
+		clk_mgr->bw_params->dummy_pstate_table[0].dram_speed_mts = clk_mgr->bw_params->clk_table.entries[0].memclk_mhz * 16;
+		clk_mgr->bw_params->dummy_pstate_table[0].dummy_pstate_latency_us = 50;
+		clk_mgr->bw_params->dummy_pstate_table[1].dram_speed_mts = clk_mgr->bw_params->clk_table.entries[1].memclk_mhz * 16;
+		clk_mgr->bw_params->dummy_pstate_table[1].dummy_pstate_latency_us = 9;
+		clk_mgr->bw_params->dummy_pstate_table[2].dram_speed_mts = clk_mgr->bw_params->clk_table.entries[2].memclk_mhz * 16;
+		clk_mgr->bw_params->dummy_pstate_table[2].dummy_pstate_latency_us = 8;
+		clk_mgr->bw_params->dummy_pstate_table[3].dram_speed_mts = clk_mgr->bw_params->clk_table.entries[3].memclk_mhz * 16;
+		clk_mgr->bw_params->dummy_pstate_table[3].dummy_pstate_latency_us = 5;
+	}
+	/* Set D - MALL - SR enter and exit time specific to MALL, TBD after bringup or later phase for now use DRAM values / 2 */
+	/* For MALL DRAM clock change latency is N/A, for watermak calculations use lowest value dummy P state latency */
+	clk_mgr->bw_params->wm_table.nv_entries[WM_D].valid = true;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.pstate_latency_us = clk_mgr->bw_params->dummy_pstate_table[3].dummy_pstate_latency_us;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.fclk_change_latency_us = fclk_change_latency_us;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.sr_exit_time_us = sr_exit_time_us / 2; // TBD
+	clk_mgr->bw_params->wm_table.nv_entries[WM_D].dml_input.sr_enter_plus_exit_time_us = sr_enter_plus_exit_time_us / 2; // TBD
+	clk_mgr->bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.wm_type = WATERMARKS_MALL;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.min_dcfclk = min_dcfclk_mhz;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.max_dcfclk = 0xFFFF;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.min_uclk = min_uclk_mhz;
+	clk_mgr->bw_params->wm_table.nv_entries[WM_D].pmfw_breakdown.max_uclk = 0xFFFF;
+}
+
+/*
+ * dcn401_update_bw_bounding_box
+ *
+ * This would override some dcn4_01 ip_or_soc initial parameters hardcoded from
+ * spreadsheet with actual values as per dGPU SKU:
+ * - with passed few options from dc->config
+ * - with dentist_vco_frequency from Clk Mgr (currently hardcoded, but might
+ *   need to get it from PM FW)
+ * - with passed latency values (passed in ns units) in dc-> bb override for
+ *   debugging purposes
+ * - with passed latencies from VBIOS (in 100_ns units) if available for
+ *   certain dGPU SKU
+ * - with number of DRAM channels from VBIOS (which differ for certain dGPU SKU
+ *   of the same ASIC)
+ * - clocks levels with passed clk_table entries from Clk Mgr as reported by PM
+ *   FW for different clocks (which might differ for certain dGPU SKU of the
+ *   same ASIC)
+ */
+void dcn401_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_params)
+{
+	dc_assert_fp_enabled();
+
+	/* Override from passed dc->bb_overrides if available*/
+	if (dc->bb_overrides.sr_exit_time_ns)
+		dc->dml2_options.bbox_overrides.sr_exit_latency_us =
+				dc->bb_overrides.sr_exit_time_ns / 1000.0;
+
+	if (dc->bb_overrides.sr_enter_plus_exit_time_ns)
+		dc->dml2_options.bbox_overrides.sr_enter_plus_exit_latency_us =
+			dc->bb_overrides.sr_enter_plus_exit_time_ns / 1000.0;
+
+	if (dc->bb_overrides.urgent_latency_ns)
+		dc->dml2_options.bbox_overrides.urgent_latency_us =
+				dc->bb_overrides.urgent_latency_ns / 1000.0;
+
+	if (dc->bb_overrides.dram_clock_change_latency_ns)
+		dc->dml2_options.bbox_overrides.dram_clock_change_latency_us =
+			dc->bb_overrides.dram_clock_change_latency_ns / 1000.0;
+
+	if (dc->bb_overrides.fclk_clock_change_latency_ns)
+		dc->dml2_options.bbox_overrides.fclk_change_latency_us =
+			dc->bb_overrides.fclk_clock_change_latency_ns / 1000;
+
+	/* Override from VBIOS if VBIOS bb_info available */
+	if (dc->ctx->dc_bios->funcs->get_soc_bb_info) {
+		struct bp_soc_bb_info bb_info = {0};
+		if (dc->ctx->dc_bios->funcs->get_soc_bb_info(dc->ctx->dc_bios, &bb_info) == BP_RESULT_OK) {
+			if (bb_info.dram_clock_change_latency_100ns > 0)
+				dc->dml2_options.bbox_overrides.dram_clock_change_latency_us =
+					bb_info.dram_clock_change_latency_100ns * 10;
+
+			if (bb_info.dram_sr_enter_exit_latency_100ns > 0)
+				dc->dml2_options.bbox_overrides.sr_enter_plus_exit_latency_us =
+					bb_info.dram_sr_enter_exit_latency_100ns * 10;
+
+			if (bb_info.dram_sr_exit_latency_100ns > 0)
+				dc->dml2_options.bbox_overrides.sr_exit_latency_us =
+					bb_info.dram_sr_exit_latency_100ns * 10;
+		}
+	}
+
+	/* Override from VBIOS for num_chan */
+	if (dc->ctx->dc_bios->vram_info.num_chans) {
+		dc->dml2_options.bbox_overrides.dram_num_chan =
+				dc->ctx->dc_bios->vram_info.num_chans;
+
+	}
+
+	if (dc->ctx->dc_bios->vram_info.dram_channel_width_bytes)
+		dc->dml2_options.bbox_overrides.dram_chanel_width_bytes =
+				dc->ctx->dc_bios->vram_info.dram_channel_width_bytes;
+
+	dc->dml2_options.bbox_overrides.disp_pll_vco_speed_mhz = dc->clk_mgr->dentist_vco_freq_khz / 1000.0;
+	dc->dml2_options.bbox_overrides.xtalclk_mhz = dc->ctx->dc_bios->fw_info.pll_info.crystal_frequency / 1000.0;
+	dc->dml2_options.bbox_overrides.dchub_refclk_mhz = dc->res_pool->ref_clocks.dchub_ref_clock_inKhz / 1000.0;
+	dc->dml2_options.bbox_overrides.dprefclk_mhz = dc->clk_mgr->dprefclk_khz / 1000.0;
+
+	if (dc->clk_mgr->bw_params->clk_table.num_entries > 1) {
+		unsigned int i = 0;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_states = dc->clk_mgr->bw_params->clk_table.num_entries;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dcfclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dcfclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_fclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_fclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_memclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_memclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_socclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_socclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dtbclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dtbclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dispclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dispclk_levels;
+
+		dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dppclk_levels =
+			dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dppclk_levels;
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dcfclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].dcfclk_mhz)
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dcfclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].dcfclk_mhz;
+		}
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_fclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].fclk_mhz)
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].fclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].fclk_mhz;
+		}
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_memclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].memclk_mhz)
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].memclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].memclk_mhz;
+		}
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_socclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].socclk_mhz)
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].socclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].socclk_mhz;
+		}
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dtbclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].dtbclk_mhz)
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dtbclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].dtbclk_mhz;
+		}
+
+		for (i = 0; i < dc->clk_mgr->bw_params->clk_table.num_entries_per_clk.num_dispclk_levels; i++) {
+			if (dc->clk_mgr->bw_params->clk_table.entries[i].dispclk_mhz) {
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dispclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].dispclk_mhz;
+				dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dppclk_mhz =
+					dc->clk_mgr->bw_params->clk_table.entries[i].dispclk_mhz;
+			}
+		}
+	}
+}
+
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn401/dcn401_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn401/dcn401_fpu.h
new file mode 100644
index 000000000000..329f1788843c
--- /dev/null
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn401/dcn401_fpu.h
@@ -0,0 +1,14 @@
+// SPDX-License-Identifier: MIT
+//
+// Copyright 2024 Advanced Micro Devices, Inc.
+
+#ifndef __DCN401_FPU_H__
+#define __DCN401_FPU_H__
+
+#include "clk_mgr.h"
+
+void dcn401_build_wm_range_table_fpu(struct clk_mgr *clk_mgr);
+
+void dcn401_update_bw_bounding_box_fpu(struct dc *dc, struct clk_bw_params *bw_params);
+
+#endif
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 fb17f8868cb4..410e4b671228 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
@@ -632,6 +632,7 @@ struct _vcs_dpi_display_dlg_regs_st {
 	unsigned int ref_freq_to_pix_freq;
 	unsigned int vratio_prefetch;
 	unsigned int vratio_prefetch_c;
+	unsigned int refcyc_per_tdlut_group;
 	unsigned int refcyc_per_pte_group_vblank_l;
 	unsigned int refcyc_per_pte_group_vblank_c;
 	unsigned int refcyc_per_meta_chunk_vblank_l;
diff --git a/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c b/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c
index 9a3ded311195..85453bbb4f9b 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c
@@ -1099,8 +1099,13 @@ void ModeSupportAndSystemConfiguration(struct display_mode_lib *mode_lib)
 
 	// Total Available Pipes Support Check
 	for (k = 0; k < mode_lib->vba.NumberOfActivePlanes; ++k) {
-		total_pipes += mode_lib->vba.DPPPerPlane[k];
 		pipe_idx = get_pipe_idx(mode_lib, k);
+		if (pipe_idx == -1) {
+			ASSERT(0);
+			continue; // skip inactive planes
+		}
+		total_pipes += mode_lib->vba.DPPPerPlane[k];
+
 		if (mode_lib->vba.cache_pipes[pipe_idx].clks_cfg.dppclk_mhz > 0.0)
 			mode_lib->vba.DPPCLK[k] = mode_lib->vba.cache_pipes[pipe_idx].clks_cfg.dppclk_mhz;
 		else