drm/amd/display: Use per pipe P-State force for FPO

[Description]
*  Pass in pipe index for FPO cmd to DMCUB
- This change will pass in the pipe index for each stream
  that is using FPO
- This change is in preparation to enable FPO + VActive

*  Use per pipe P-State force for FPO
- For FPO, instead of using max watermarks value for P-State disallow,
  use per pipe p-state force instead
- This is in preparation to enable FPO + VActive

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_dmub_srv.c

@@ -302,27 +302,31 @@ static uint8_t dc_dmub_srv_get_pipes_for_stream(struct dc *dc, struct dc_stream_
 	return pipes;
 }
 
-static int dc_dmub_srv_get_timing_generator_offset(struct dc *dc, struct dc_stream_state *stream)
+static void dc_dmub_srv_populate_fams_pipe_info(struct dc *dc, struct dc_state *context,
+		struct pipe_ctx *head_pipe,
+		struct dmub_cmd_fw_assisted_mclk_switch_pipe_data *fams_pipe_data)
 {
-	int  tg_inst = 0;
-	int i = 0;
+	int j;
 
-	for (i = 0; i < MAX_PIPES; i++) {
-		struct pipe_ctx *pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+	// TODO: Uncomment once FW headers are updated in driver
+	//fams_pipe_data->pipe_index[pipe_idx++] = head_pipe->plane_res.hubp->inst;
+	for (j = 0; j < dc->res_pool->pipe_count; j++) {
+		struct pipe_ctx *split_pipe = &context->res_ctx.pipe_ctx[j];
 
-		if (pipe->stream == stream && pipe->stream_res.tg) {
-			tg_inst = pipe->stream_res.tg->inst;
-			break;
+		if (split_pipe->stream == head_pipe->stream && (split_pipe->top_pipe || split_pipe->prev_odm_pipe)) {
+			// TODO: Uncomment once FW headers are updated in driver
+			//fams_pipe_data->pipe_index[pipe_idx++] = split_pipe->plane_res.hubp->inst;
 		}
 	}
-	return tg_inst;
+	// TODO: Uncomment once FW headers are updated in driver
+	//fams_pipe_data->pipe_count = pipe_idx;
 }
 
 bool dc_dmub_srv_p_state_delegate(struct dc *dc, bool should_manage_pstate, struct dc_state *context)
 {
 	union dmub_rb_cmd cmd = { 0 };
 	struct dmub_cmd_fw_assisted_mclk_switch_config *config_data = &cmd.fw_assisted_mclk_switch.config_data;
-	int i = 0;
+	int i = 0, k = 0;
 	int ramp_up_num_steps = 1; // TODO: Ramp is currently disabled. Reenable it.
 	uint8_t visual_confirm_enabled;
 
@@ -337,17 +341,21 @@ bool dc_dmub_srv_p_state_delegate(struct dc *dc, bool should_manage_pstate, stru
 	cmd.fw_assisted_mclk_switch.config_data.fams_enabled = should_manage_pstate;
 	cmd.fw_assisted_mclk_switch.config_data.visual_confirm_enabled = visual_confirm_enabled;
 
-	for (i = 0; context && i < context->stream_count; i++) {
-		struct dc_stream_state *stream = context->streams[i];
-		uint8_t min_refresh_in_hz = (stream->timing.min_refresh_in_uhz + 999999) / 1000000;
-		int  tg_inst = dc_dmub_srv_get_timing_generator_offset(dc, stream);
+	for (i = 0, k = 0; context && i < dc->res_pool->pipe_count; i++) {
+		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
 
-		config_data->pipe_data[tg_inst].pix_clk_100hz = stream->timing.pix_clk_100hz;
-		config_data->pipe_data[tg_inst].min_refresh_in_hz = min_refresh_in_hz;
-		config_data->pipe_data[tg_inst].max_ramp_step = ramp_up_num_steps;
-		config_data->pipe_data[tg_inst].pipes = dc_dmub_srv_get_pipes_for_stream(dc, stream);
+		if (!pipe->top_pipe && !pipe->prev_odm_pipe && pipe->stream && pipe->stream->fpo_in_use) {
+			struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+			uint8_t min_refresh_in_hz = (pipe->stream->timing.min_refresh_in_uhz + 999999) / 1000000;
+
+			config_data->pipe_data[k].pix_clk_100hz = pipe->stream->timing.pix_clk_100hz;
+			config_data->pipe_data[k].min_refresh_in_hz = min_refresh_in_hz;
+			config_data->pipe_data[k].max_ramp_step = ramp_up_num_steps;
+			config_data->pipe_data[k].pipes = dc_dmub_srv_get_pipes_for_stream(dc, pipe->stream);
+			dc_dmub_srv_populate_fams_pipe_info(dc, context, pipe, &config_data->pipe_data[k]);
+			k++;
+		}
 	}
-
 	cmd.fw_assisted_mclk_switch.header.payload_bytes =
 		sizeof(cmd.fw_assisted_mclk_switch) - sizeof(cmd.fw_assisted_mclk_switch.header);
 

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

@@ -293,6 +293,7 @@ struct dc_stream_state {
 
 	bool has_non_synchronizable_pclk;
 	bool vblank_synchronized;
+	bool fpo_in_use;
 	struct mall_stream_config mall_stream_config;
 };
 

drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hwseq.c

@@ -1993,6 +1993,16 @@ void dcn20_post_unlock_program_front_end(
 		}
 	}
 
+	/* P-State support transitions:
+	 * Natural -> FPO: 		P-State disabled in prepare, force disallow anytime is safe
+	 * FPO -> Natural: 		Unforce anytime after FW disable is safe (P-State will assert naturally)
+	 * Unsupported -> FPO:	P-State enabled in optimize, force disallow anytime is safe
+	 * FPO -> Unsupported:	P-State disabled in prepare, unforce disallow anytime is safe
+	 * FPO <-> SubVP:		Force disallow is maintained on the FPO / SubVP pipes
+	 */
+	if (hwseq && hwseq->funcs.update_force_pstate)
+		dc->hwseq->funcs.update_force_pstate(dc, context);
+
 	/* Only program the MALL registers after all the main and phantom pipes
 	 * are done programming.
 	 */

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

@@ -2016,6 +2016,8 @@ bool dcn30_can_support_mclk_switch_using_fw_based_vblank_stretch(struct dc *dc,
 	if (context->streams[0]->vrr_active_variable)
 		return false;
 
+	context->streams[0]->fpo_in_use = true;
+
 	return true;
 }
 

drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hubp.c

@@ -47,6 +47,15 @@ void hubp32_update_force_pstate_disallow(struct hubp *hubp, bool pstate_disallow
 			DATA_UCLK_PSTATE_FORCE_VALUE, 0);
 }
 
+void hubp32_update_force_cursor_pstate_disallow(struct hubp *hubp, bool pstate_disallow)
+{
+	struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
+
+	REG_UPDATE_2(UCLK_PSTATE_FORCE,
+			CURSOR_UCLK_PSTATE_FORCE_EN, pstate_disallow,
+			CURSOR_UCLK_PSTATE_FORCE_VALUE, 0);
+}
+
 void hubp32_update_mall_sel(struct hubp *hubp, uint32_t mall_sel, bool c_cursor)
 {
 	struct dcn20_hubp *hubp2 = TO_DCN20_HUBP(hubp);
@@ -188,6 +197,7 @@ static struct hubp_funcs dcn32_hubp_funcs = {
 	.hubp_set_flip_int = hubp1_set_flip_int,
 	.hubp_in_blank = hubp1_in_blank,
 	.hubp_update_force_pstate_disallow = hubp32_update_force_pstate_disallow,
+	.hubp_update_force_cursor_pstate_disallow = hubp32_update_force_cursor_pstate_disallow,
 	.phantom_hubp_post_enable = hubp32_phantom_hubp_post_enable,
 	.hubp_update_mall_sel = hubp32_update_mall_sel,
 	.hubp_prepare_subvp_buffering = hubp32_prepare_subvp_buffering

drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hubp.h

@@ -46,6 +46,8 @@
 
 void hubp32_update_force_pstate_disallow(struct hubp *hubp, bool pstate_disallow);
 
+void hubp32_update_force_cursor_pstate_disallow(struct hubp *hubp, bool pstate_disallow);
+
 void hubp32_update_mall_sel(struct hubp *hubp, uint32_t mall_sel, bool c_cursor);
 
 void hubp32_prepare_subvp_buffering(struct hubp *hubp, bool enable);

drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hwseq.c

@@ -571,40 +571,51 @@ bool dcn32_set_output_transfer_func(struct dc *dc,
 	return ret;
 }
 
-/* Program P-State force value according to if pipe is using SubVP or not:
+/* Program P-State force value according to if pipe is using SubVP / FPO or not:
  * 1. Reset P-State force on all pipes first
  * 2. For each main pipe, force P-State disallow (P-State allow moderated by DMUB)
  */
-void dcn32_subvp_update_force_pstate(struct dc *dc, struct dc_state *context)
+void dcn32_update_force_pstate(struct dc *dc, struct dc_state *context)
 {
 	int i;
-	int num_subvp = 0;
-	/* Unforce p-state for each pipe
+
+	/* Unforce p-state for each pipe if it is not FPO or SubVP.
+	 * For FPO and SubVP, if it's already forced disallow, leave
+	 * it as disallow.
 	 */
 	for (i = 0; i < dc->res_pool->pipe_count; i++) {
 		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
 		struct hubp *hubp = pipe->plane_res.hubp;
 
-		if (hubp && hubp->funcs->hubp_update_force_pstate_disallow)
-			hubp->funcs->hubp_update_force_pstate_disallow(hubp, false);
-		if (pipe->stream && pipe->stream->mall_stream_config.type == SUBVP_MAIN)
-			num_subvp++;
+		if (!pipe->stream || (pipe->stream && !(pipe->stream->mall_stream_config.type == SUBVP_MAIN ||
+						pipe->stream->fpo_in_use))) {
+			if (hubp && hubp->funcs->hubp_update_force_pstate_disallow)
+				hubp->funcs->hubp_update_force_pstate_disallow(hubp, false);
+			if (hubp && hubp->funcs->hubp_update_force_cursor_pstate_disallow)
+				hubp->funcs->hubp_update_force_cursor_pstate_disallow(hubp, false);
+		}
 	}
 
-	if (num_subvp == 0)
-		return;
-
 	/* Loop through each pipe -- for each subvp main pipe force p-state allow equal to false.
 	 */
 	for (i = 0; i < dc->res_pool->pipe_count; i++) {
 		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];
+		struct hubp *hubp = pipe->plane_res.hubp;
 
-		if (pipe->stream && pipe->plane_state && (pipe->stream->mall_stream_config.type == SUBVP_MAIN)) {
-			struct hubp *hubp = pipe->plane_res.hubp;
-
+		if (pipe->stream && pipe->plane_state && pipe->stream->mall_stream_config.type == SUBVP_MAIN) {
 			if (hubp && hubp->funcs->hubp_update_force_pstate_disallow)
 				hubp->funcs->hubp_update_force_pstate_disallow(hubp, true);
 		}
+
+		if (pipe->stream && pipe->stream->fpo_in_use) {
+			if (hubp && hubp->funcs->hubp_update_force_pstate_disallow)
+				hubp->funcs->hubp_update_force_pstate_disallow(hubp, true);
+			/* For now only force cursor p-state disallow for FPO
+			 * Needs to be added for subvp once FW side gets updated
+			 */
+			if (hubp && hubp->funcs->hubp_update_force_cursor_pstate_disallow)
+				hubp->funcs->hubp_update_force_cursor_pstate_disallow(hubp, true);
+		}
 	}
 }
 
@@ -677,10 +688,6 @@ void dcn32_program_mall_pipe_config(struct dc *dc, struct dc_state *context)
 	if (hws && hws->funcs.update_mall_sel)
 		hws->funcs.update_mall_sel(dc, context);
 
-	//update subvp force pstate
-	if (hws && hws->funcs.subvp_update_force_pstate)
-		dc->hwseq->funcs.subvp_update_force_pstate(dc, context);
-
 	// Program FORCE_ONE_ROW_FOR_FRAME and CURSOR_REQ_MODE for main subvp pipes
 	for (i = 0; i < dc->res_pool->pipe_count; i++) {
 		struct pipe_ctx *pipe = &context->res_ctx.pipe_ctx[i];

drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hwseq.h

@@ -67,7 +67,7 @@ void dcn32_program_mall_pipe_config(struct dc *dc, struct dc_state *context);
 
 void dcn32_update_mall_sel(struct dc *dc, struct dc_state *context);
 
-void dcn32_subvp_update_force_pstate(struct dc *dc, struct dc_state *context);
+void dcn32_update_force_pstate(struct dc *dc, struct dc_state *context);
 
 void dcn32_update_odm(struct dc *dc, struct dc_state *context, struct pipe_ctx *pipe_ctx);
 

drivers/gpu/drm/amd/display/dc/dcn32/dcn32_init.c

@@ -149,7 +149,7 @@ static const struct hwseq_private_funcs dcn32_private_funcs = {
 	.dccg_init = dcn20_dccg_init,
 	.set_mcm_luts = dcn32_set_mcm_luts,
 	.program_mall_pipe_config = dcn32_program_mall_pipe_config,
-	.subvp_update_force_pstate = dcn32_subvp_update_force_pstate,
+	.update_force_pstate = dcn32_update_force_pstate,
 	.update_mall_sel = dcn32_update_mall_sel,
 	.calculate_dccg_k1_k2_values = dcn32_calculate_dccg_k1_k2_values,
 	.set_pixels_per_cycle = dcn32_set_pixels_per_cycle,

drivers/gpu/drm/amd/display/dc/dml/dcn30/dcn30_fpu.c

@@ -389,6 +389,10 @@ void dcn30_fpu_calculate_wm_and_dlg(
 	dc_assert_fp_enabled();
 
 	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) {
 		/* only when the mclk switch can not be natural, is the fw based vblank stretch attempted */

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

@@ -1960,6 +1960,10 @@ 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 || (!dc->debug.disable_fpo_optimizations &&
 			pstate_en && vlevel != 0)) {
@@ -2199,10 +2203,6 @@ 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 =

drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h

@@ -203,6 +203,7 @@ struct hubp_funcs {
 	void (*hubp_soft_reset)(struct hubp *hubp, bool reset);
 
 	void (*hubp_update_force_pstate_disallow)(struct hubp *hubp, bool allow);
+	void (*hubp_update_force_cursor_pstate_disallow)(struct hubp *hubp, bool allow);
 	void (*hubp_update_mall_sel)(struct hubp *hubp, uint32_t mall_sel, bool c_cursor);
 	void (*hubp_prepare_subvp_buffering)(struct hubp *hubp, bool enable);
 

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

@@ -154,7 +154,7 @@ struct hwseq_private_funcs {
 	void (*setup_hpo_hw_control)(const struct dce_hwseq *hws, bool enable);
 #ifdef CONFIG_DRM_AMD_DC_FP
 	void (*program_mall_pipe_config)(struct dc *dc, struct dc_state *context);
-	void (*subvp_update_force_pstate)(struct dc *dc, struct dc_state *context);
+	void (*update_force_pstate)(struct dc *dc, struct dc_state *context);
 	void (*update_mall_sel)(struct dc *dc, struct dc_state *context);
 	unsigned int (*calculate_dccg_k1_k2_values)(struct pipe_ctx *pipe_ctx,
 			unsigned int *k1_div,