drm/amd/display: Separate pipe disconnect from rest of progrmaming

[Why]
When changing pixel formats for HDR (e.g. ARGB -> FP16)
there are configurations that change from 2 pipes to 1 pipe.
In these cases, it seems that disconnecting MPCC and doing
a surface update at the same time(after unlocking) causes
some registers to be updated slightly faster than others
after unlocking (e.g. if the pixel format is updated to FP16
before the new surface address is programmed, we get
corruption on the screen because the pixel formats aren't
matching). We separate disconnecting MPCC from the rest
of  the  pipe programming sequence to prevent this.

[How]
Move MPCC disconnect into separate operation than the
rest of the pipe programming.

Signed-off-by: Alvin Lee <alvin.lee2@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/display/dc/core/dc.c

@@ -2300,6 +2300,7 @@ static void commit_planes_for_stream(struct dc *dc,
 		enum surface_update_type update_type,
 		struct dc_state *context)
 {
+	bool mpcc_disconnected = false;
 	int i, j;
 	struct pipe_ctx *top_pipe_to_program = NULL;
 
@@ -2330,6 +2331,15 @@ static void commit_planes_for_stream(struct dc *dc,
 		context_clock_trace(dc, context);
 	}
 
+	if (update_type != UPDATE_TYPE_FAST && dc->hwss.interdependent_update_lock &&
+		dc->hwss.disconnect_pipes && dc->hwss.wait_for_pending_cleared){
+		dc->hwss.interdependent_update_lock(dc, context, true);
+		mpcc_disconnected = dc->hwss.disconnect_pipes(dc, context);
+		dc->hwss.interdependent_update_lock(dc, context, false);
+		if (mpcc_disconnected)
+			dc->hwss.wait_for_pending_cleared(dc, context);
+	}
+
 	for (j = 0; j < dc->res_pool->pipe_count; j++) {
 		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j];
 

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

@@ -1624,6 +1624,120 @@ static void dcn20_program_pipe(
 	}
 }
 
+bool dcn20_disconnect_pipes(
+		struct dc *dc,
+		struct dc_state *context)
+{
+		int i;
+		struct dce_hwseq *hws = dc->hwseq;
+		bool mpcc_disconnected = false;
+		DC_LOGGER_INIT(dc->ctx->logger);
+
+		/* Set pipe update flags and lock pipes */
+		for (i = 0; i < dc->res_pool->pipe_count; i++)
+			dcn20_detect_pipe_changes(&dc->current_state->res_ctx.pipe_ctx[i],
+					&context->res_ctx.pipe_ctx[i]);
+
+		if (!IS_DIAG_DC(dc->ctx->dce_environment)) {
+			/* OTG blank before disabling all front ends */
+			for (i = 0; i < dc->res_pool->pipe_count; i++) {
+				if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable
+					&& !context->res_ctx.pipe_ctx[i].top_pipe
+					&& !context->res_ctx.pipe_ctx[i].prev_odm_pipe
+					&& context->res_ctx.pipe_ctx[i].stream) {
+					hws->funcs.blank_pixel_data(dc, &context->res_ctx.pipe_ctx[i], true);
+				}
+			}
+
+			/* Disconnect mpcc */
+			for (i = 0; i < dc->res_pool->pipe_count; i++) {
+				if (context->res_ctx.pipe_ctx[i].update_flags.bits.disable) {
+					hws->funcs.plane_atomic_disconnect(dc, &dc->current_state->res_ctx.pipe_ctx[i]);
+					DC_LOG_DC("Reset mpcc for pipe %d\n", dc->current_state->res_ctx.pipe_ctx[i].pipe_idx);
+					mpcc_disconnected = true;
+				}
+			}
+		}
+
+		if (mpcc_disconnected) {
+			for (i = 0; i < dc->res_pool->pipe_count; i++) {
+				struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
+				struct pipe_ctx *old_pipe = &dc->current_state->res_ctx.pipe_ctx[i];
+				struct dc_plane_state *plane_state = pipe_ctx->plane_state;
+				struct hubp *hubp = pipe_ctx->plane_res.hubp;
+
+				if (!pipe_ctx || !plane_state || !pipe_ctx->stream)
+					continue;
+
+			// Only update scaler and viewport here if we lose a pipe split.
+			// This is to prevent half the screen from being black when we
+			// unlock after disconnecting MPCC.
+			if (!(old_pipe && !pipe_ctx->top_pipe &&
+				!pipe_ctx->bottom_pipe && old_pipe->bottom_pipe))
+				continue;
+
+			if (pipe_ctx->update_flags.raw || pipe_ctx->plane_state->update_flags.raw || pipe_ctx->stream->update_flags.raw) {
+				if (pipe_ctx->update_flags.bits.scaler ||
+					plane_state->update_flags.bits.scaling_change ||
+					plane_state->update_flags.bits.position_change ||
+					plane_state->update_flags.bits.per_pixel_alpha_change ||
+					pipe_ctx->stream->update_flags.bits.scaling) {
+
+					pipe_ctx->plane_res.scl_data.lb_params.alpha_en = pipe_ctx->plane_state->per_pixel_alpha;
+					ASSERT(pipe_ctx->plane_res.scl_data.lb_params.depth == LB_PIXEL_DEPTH_30BPP);
+					/* scaler configuration */
+					pipe_ctx->plane_res.dpp->funcs->dpp_set_scaler(
+					pipe_ctx->plane_res.dpp, &pipe_ctx->plane_res.scl_data);
+				}
+
+				if (pipe_ctx->update_flags.bits.viewport ||
+					(context == dc->current_state && plane_state->update_flags.bits.position_change) ||
+					(context == dc->current_state && plane_state->update_flags.bits.scaling_change) ||
+					(context == dc->current_state && pipe_ctx->stream->update_flags.bits.scaling)) {
+
+					hubp->funcs->mem_program_viewport(
+						hubp,
+						&pipe_ctx->plane_res.scl_data.viewport,
+						&pipe_ctx->plane_res.scl_data.viewport_c);
+				}
+			}
+		}
+	}
+	return mpcc_disconnected;
+}
+
+void dcn20_wait_for_pending_cleared(struct dc *dc,
+		struct dc_state *context)
+{
+		struct pipe_ctx *pipe_ctx;
+		struct timing_generator *tg;
+		int i;
+
+		for (i = 0; i < dc->res_pool->pipe_count; i++) {
+			pipe_ctx = &context->res_ctx.pipe_ctx[i];
+			tg = pipe_ctx->stream_res.tg;
+
+			/*
+			 * Only wait for top pipe's tg penindg bit
+			 * Also skip if pipe is disabled.
+			 */
+			if (pipe_ctx->top_pipe ||
+			    !pipe_ctx->stream || !pipe_ctx->plane_state ||
+			    !tg->funcs->is_tg_enabled(tg))
+				continue;
+
+			/*
+			 * Wait for VBLANK then VACTIVE to ensure we get VUPDATE.
+			 * For some reason waiting for OTG_UPDATE_PENDING cleared
+			 * seems to not trigger the update right away, and if we
+			 * lock again before VUPDATE then we don't get a separated
+			 * operation.
+			 */
+			pipe_ctx->stream_res.tg->funcs->wait_for_state(pipe_ctx->stream_res.tg, CRTC_STATE_VBLANK);
+			pipe_ctx->stream_res.tg->funcs->wait_for_state(pipe_ctx->stream_res.tg, CRTC_STATE_VACTIVE);
+		}
+}
+
 void dcn20_program_front_end_for_ctx(
 		struct dc *dc,
 		struct dc_state *context)

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

@@ -131,6 +131,13 @@ void dcn20_dccg_init(struct dce_hwseq *hws);
 int dcn20_init_sys_ctx(struct dce_hwseq *hws,
 		struct dc *dc,
 		struct dc_phy_addr_space_config *pa_config);
+bool dcn20_disconnect_pipes(
+		struct dc *dc,
+		struct dc_state *context);
+
+void dcn20_wait_for_pending_cleared(struct dc *dc,
+		struct dc_state *context);
+                
 
 #ifndef TRIM_FSFT
 bool dcn20_optimize_timing_for_fsft(struct dc *dc,

drivers/gpu/drm/amd/display/dc/dcn20/dcn20_init.c

@@ -34,6 +34,8 @@ static const struct hw_sequencer_funcs dcn20_funcs = {
 	.apply_ctx_to_hw = dce110_apply_ctx_to_hw,
 	.apply_ctx_for_surface = NULL,
 	.program_front_end_for_ctx = dcn20_program_front_end_for_ctx,
+	.disconnect_pipes = dcn20_disconnect_pipes,
+	.wait_for_pending_cleared = dcn20_wait_for_pending_cleared,
 	.post_unlock_program_front_end = dcn20_post_unlock_program_front_end,
 	.update_plane_addr = dcn20_update_plane_addr,
 	.update_dchub = dcn10_update_dchub,

drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c

@@ -35,6 +35,8 @@ static const struct hw_sequencer_funcs dcn21_funcs = {
 	.apply_ctx_to_hw = dce110_apply_ctx_to_hw,
 	.apply_ctx_for_surface = NULL,
 	.program_front_end_for_ctx = dcn20_program_front_end_for_ctx,
+	.disconnect_pipes = dcn20_disconnect_pipes,
+	.wait_for_pending_cleared = dcn20_wait_for_pending_cleared,
 	.post_unlock_program_front_end = dcn20_post_unlock_program_front_end,
 	.update_plane_addr = dcn20_update_plane_addr,
 	.update_dchub = dcn10_update_dchub,

drivers/gpu/drm/amd/display/dc/dcn30/dcn30_init.c

@@ -35,6 +35,8 @@ static const struct hw_sequencer_funcs dcn30_funcs = {
 	.apply_ctx_to_hw = dce110_apply_ctx_to_hw,
 	.apply_ctx_for_surface = NULL,
 	.program_front_end_for_ctx = dcn20_program_front_end_for_ctx,
+	.disconnect_pipes = dcn20_disconnect_pipes,
+	.wait_for_pending_cleared = dcn20_wait_for_pending_cleared,
 	.post_unlock_program_front_end = dcn20_post_unlock_program_front_end,
 	.update_plane_addr = dcn20_update_plane_addr,
 	.update_dchub = dcn10_update_dchub,

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

@@ -67,6 +67,10 @@ struct hw_sequencer_funcs {
 			int num_planes, struct dc_state *context);
 	void (*program_front_end_for_ctx)(struct dc *dc,
 			struct dc_state *context);
+	bool (*disconnect_pipes)(struct dc *dc,
+			struct dc_state *context);
+	void (*wait_for_pending_cleared)(struct dc *dc,
+			struct dc_state *context);
 	void (*post_unlock_program_front_end)(struct dc *dc,
 			struct dc_state *context);
 	void (*update_plane_addr)(const struct dc *dc,