[PATCH 00/35] DC Changes Jan 31, 2019

[PATCH 00/35] DC Changes Jan 31, 2019

[Bhawanpreet Lakha]

Summary of Changes
*Fix a DP compliance test
*Refactoring some things (amdgpu_dm, DRR, init_hw, vupdate etc)
*MST fixes

Anthony Koo (8):
  drm/amd/display: fix issue with DC brightness low with VB
  drm/amd/display: link_rate_set should index into table
  drm/amd/display: interface to check if timing can be seamless
  drm/amd/display: refactor out programming of vupdate interrupt
  drm/amd/display: add way to determine if link is active
  drm/amd/display: add seamless boot flag to stream
  drm/amd/display: refactor programming of DRR
  drm/amd/display: refactor init_hw to isolate pipe related init

Chiawen Huang (1):
  drm/amd/display: add gpio lock/unlock

Dmytro Laktyushkin (2):
  drm/amd/display: add a debug flag to force odm combine
  drm/amd/display: add n_vid_mul and half pix_rate for odm

Eryk Brol (1):
  drm/amd/display: DC VM Fixes

Ilya Bakoulin (1):
  drm/amd/display: Check that vrefresh is in freesync range

Jerry (Fangzhi) Zuo (2):
  drm/amd/display: Apply fake sink back to MST sequence
  drm/amd/display: Clear dc_sink after it gets released

John Barberiz (1):
  drm/amd/display: Use udelay when waiting between aux retries

Josip Pavic (1):
  drm/amd/display: Modify ABM 2.2 Max Reduction

Krunoslav Kovac (1):
  drm/amd/display: DGAM enabled for HDR

Murton Liu (1):
  drm/amd/display: PIP overlay corruption

Nicholas Kazlauskas (7):
  drm/amd/display: Apply all surface updates onto surfaces
  drm/amd/display: Use the right surface for flip and FreeSync
  drm/amd/display: Reformat dm_determine_update_type_for_commit
  drm/amd/display: Initialize stream_update to zero
  drm/amd/display: Remove FreeSync timing changed debug output
  drm/amd/display: Disconnect mpcc when changing tg
  drm/amd/display: Don't re-program planes for DPMS changes

Paul Hsieh (1):
  drm/amd/display: dmcu is blocking due to wrong disable ABM command

Steven Chiu (1):
  drm/amd/display: 3.2.16

Su Sung Chung (1):
  drm/amd/display: store timing sync info in dc_stream_status

Wenjing Liu (1):
  drm/amd/display: determine if a pipe is synced by plane state

Wesley Chalmers (1):
  drm/amd/display: Disable Stutter for Stereo 3D

Xiaodong Yan (1):
  drm/amd/display: Add monitor patch for backlight

Yongqiang Sun (2):
  drm/amd/display: make vline_config parameter const.
  drm/amd/display: Calc vline position in dc.

mmcgarri (1):
  drm/amd/display:3.2.17

 .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 182 +++++++---------
 .../display/amdgpu_dm/amdgpu_dm_mst_types.c   |   5 +
 drivers/gpu/drm/amd/display/dc/core/dc.c      | 189 +++++++++++++++--
 drivers/gpu/drm/amd/display/dc/core/dc_link.c |  65 +++---
 .../gpu/drm/amd/display/dc/core/dc_link_dp.c  |  35 +---
 .../drm/amd/display/dc/core/dc_link_hwss.c    |   4 +
 .../gpu/drm/amd/display/dc/core/dc_resource.c |  56 ++++-
 .../drm/amd/display/dc/core/dc_vm_helper.c    |  18 +-
 drivers/gpu/drm/amd/display/dc/dc.h           |   8 +-
 drivers/gpu/drm/amd/display/dc/dc_link.h      |   2 +
 drivers/gpu/drm/amd/display/dc/dc_stream.h    |  25 ++-
 drivers/gpu/drm/amd/display/dc/dc_types.h     |   1 +
 drivers/gpu/drm/amd/display/dc/dce/dce_abm.c  |   4 +-
 drivers/gpu/drm/amd/display/dc/dce/dce_aux.c  |   2 +-
 .../drm/amd/display/dc/dce/dce_clock_source.c |  28 ++-
 .../display/dc/dce110/dce110_hw_sequencer.c   |  50 +++--
 .../drm/amd/display/dc/dcn10/dcn10_hubbub.c   |  19 +-
 .../drm/amd/display/dc/dcn10/dcn10_hubbub.h   |   2 +-
 .../gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c |  23 +-
 .../amd/display/dc/dcn10/dcn10_hw_sequencer.c | 194 ++++++++---------
 .../amd/display/dc/dcn10/dcn10_link_encoder.c |  10 +
 .../amd/display/dc/dcn10/dcn10_link_encoder.h |   2 +
 .../gpu/drm/amd/display/dc/dcn10/dcn10_optc.c | 197 ++++++++++++++++--
 .../gpu/drm/amd/display/dc/dcn10/dcn10_optc.h |   8 +-
 .../gpu/drm/amd/display/dc/gpio/gpio_base.c   |  12 ++
 .../drm/amd/display/dc/gpio/gpio_service.c    |  28 +++
 .../drm/amd/display/dc/gpio/gpio_service.h    |  10 +
 .../gpu/drm/amd/display/dc/inc/clock_source.h |   4 +
 .../drm/amd/display/dc/inc/hw/link_encoder.h  |   1 +
 .../amd/display/dc/inc/hw/timing_generator.h  |  13 +-
 .../gpu/drm/amd/display/dc/inc/hw_sequencer.h |   2 +
 .../gpu/drm/amd/display/dc/inc/vm_helper.h    |   7 +-
 .../drm/amd/display/include/gpio_interface.h  |   8 +
 .../amd/display/modules/color/color_gamma.c   |  89 ++++----
 .../amd/display/modules/power/power_helpers.c |  38 ++--
 35 files changed, 933 insertions(+), 408 deletions(-)

-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 01/35] drm/amd/display: Use udelay when waiting between aux retries

[Bhawanpreet Lakha]

From: John Barberiz <John.Barberiz@amd.com>

[Why]
"IRQ_HPD Pulse Length Test" DP compliance test
fails. Test complains that certain DPCD registers
are not read within 100 ms.

[How]
msleep is inaccurate for small values. Used udelay
instead for accuracy.

Change-Id: I4990fc56c7632fba373f2dbf64ac803d64619529
Signed-off-by: John Barberiz <John.Barberiz@amd.com>
Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/dce/dce_aux.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c
index adbb22224e1a..4febf4ef7240 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_aux.c
@@ -516,7 +516,7 @@ bool dce_aux_transfer_with_retries(struct ddc_service *ddc,
 			}
 		}
 
-		msleep(1);
+		udelay(1000);
 	}
 	return false;
 }
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 02/35] drm/amd/display: Apply all surface updates onto surfaces

[Bhawanpreet Lakha]

From: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>

[Why]
Most surface updates weren't propagated onto the surface during
dc_commit_updates_for_stream. This makes it more difficult for DC to
determine the actual surface update type required.

[How]
Use copy_surface_update_to_plane to propagate the changes. The FreeSync
surface timing information update for BTR has been moved out of
amdgpu_dm.c into this function as well.

Change-Id: I031d3032a56e768560b28d1a76750afd57c0769b
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Sun peng Li <Sunpeng.Li@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c |   9 --
 drivers/gpu/drm/amd/display/dc/core/dc.c      | 104 ++++++++++++++++--
 2 files changed, 96 insertions(+), 17 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index cdda68aba70e..3f50f7f0b8f0 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -4860,15 +4860,6 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 					surface,
 					flip->flip_addrs[flip_count].flip_timestamp_in_us);
 
-			/* Update surface timing information. */
-			surface->time.time_elapsed_in_us[surface->time.index] =
-				flip->flip_addrs[flip_count].flip_timestamp_in_us -
-				surface->time.prev_update_time_in_us;
-			surface->time.prev_update_time_in_us = flip->flip_addrs[flip_count].flip_timestamp_in_us;
-			surface->time.index++;
-			if (surface->time.index >= DC_PLANE_UPDATE_TIMES_MAX)
-				surface->time.index = 0;
-
 			DRM_DEBUG_DRIVER("%s Flipping to hi: 0x%x, low: 0x%x\n",
 					 __func__,
 					 flip->flip_addrs[flip_count].address.grph.addr.high_part,
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index e22be0aefd1b..db8252ec3671 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -1440,6 +1440,101 @@ static struct dc_stream_status *stream_get_status(
 
 static const enum surface_update_type update_surface_trace_level = UPDATE_TYPE_FULL;
 
+static void copy_surface_update_to_plane(
+		struct dc_plane_state *surface,
+		struct dc_surface_update *srf_update)
+{
+	if (srf_update->flip_addr) {
+		surface->address = srf_update->flip_addr->address;
+		surface->flip_immediate =
+			srf_update->flip_addr->flip_immediate;
+		surface->time.time_elapsed_in_us[surface->time.index] =
+			srf_update->flip_addr->flip_timestamp_in_us -
+				surface->time.prev_update_time_in_us;
+		surface->time.prev_update_time_in_us =
+			srf_update->flip_addr->flip_timestamp_in_us;
+		surface->time.index++;
+		if (surface->time.index >= DC_PLANE_UPDATE_TIMES_MAX)
+			surface->time.index = 0;
+	}
+
+	if (srf_update->scaling_info) {
+		surface->scaling_quality =
+				srf_update->scaling_info->scaling_quality;
+		surface->dst_rect =
+				srf_update->scaling_info->dst_rect;
+		surface->src_rect =
+				srf_update->scaling_info->src_rect;
+		surface->clip_rect =
+				srf_update->scaling_info->clip_rect;
+	}
+
+	if (srf_update->plane_info) {
+		surface->color_space =
+				srf_update->plane_info->color_space;
+		surface->format =
+				srf_update->plane_info->format;
+		surface->plane_size =
+				srf_update->plane_info->plane_size;
+		surface->rotation =
+				srf_update->plane_info->rotation;
+		surface->horizontal_mirror =
+				srf_update->plane_info->horizontal_mirror;
+		surface->stereo_format =
+				srf_update->plane_info->stereo_format;
+		surface->tiling_info =
+				srf_update->plane_info->tiling_info;
+		surface->visible =
+				srf_update->plane_info->visible;
+		surface->per_pixel_alpha =
+				srf_update->plane_info->per_pixel_alpha;
+		surface->global_alpha =
+				srf_update->plane_info->global_alpha;
+		surface->global_alpha_value =
+				srf_update->plane_info->global_alpha_value;
+		surface->dcc =
+				srf_update->plane_info->dcc;
+		surface->sdr_white_level =
+				srf_update->plane_info->sdr_white_level;
+	}
+
+	if (srf_update->gamma &&
+			(surface->gamma_correction !=
+					srf_update->gamma)) {
+		memcpy(&surface->gamma_correction->entries,
+			&srf_update->gamma->entries,
+			sizeof(struct dc_gamma_entries));
+		surface->gamma_correction->is_identity =
+			srf_update->gamma->is_identity;
+		surface->gamma_correction->num_entries =
+			srf_update->gamma->num_entries;
+		surface->gamma_correction->type =
+			srf_update->gamma->type;
+	}
+
+	if (srf_update->in_transfer_func &&
+			(surface->in_transfer_func !=
+				srf_update->in_transfer_func)) {
+		surface->in_transfer_func->sdr_ref_white_level =
+			srf_update->in_transfer_func->sdr_ref_white_level;
+		surface->in_transfer_func->tf =
+			srf_update->in_transfer_func->tf;
+		surface->in_transfer_func->type =
+			srf_update->in_transfer_func->type;
+		memcpy(&surface->in_transfer_func->tf_pts,
+			&srf_update->in_transfer_func->tf_pts,
+			sizeof(struct dc_transfer_func_distributed_points));
+	}
+
+	if (srf_update->input_csc_color_matrix)
+		surface->input_csc_color_matrix =
+			*srf_update->input_csc_color_matrix;
+
+	if (srf_update->coeff_reduction_factor)
+		surface->coeff_reduction_factor =
+			*srf_update->coeff_reduction_factor;
+}
+
 static void commit_planes_do_stream_update(struct dc *dc,
 		struct dc_stream_state *stream,
 		struct dc_stream_update *stream_update,
@@ -1645,14 +1740,7 @@ void dc_commit_updates_for_stream(struct dc *dc,
 	for (i = 0; i < surface_count; i++) {
 		struct dc_plane_state *surface = srf_updates[i].surface;
 
-		/* TODO: On flip we don't build the state, so it still has the
-		 * old address. Which is why we are updating the address here
-		 */
-		if (srf_updates[i].flip_addr) {
-			surface->address = srf_updates[i].flip_addr->address;
-			surface->flip_immediate = srf_updates[i].flip_addr->flip_immediate;
-
-		}
+		copy_surface_update_to_plane(surface, &srf_updates[i]);
 
 		if (update_type >= UPDATE_TYPE_MED) {
 			for (j = 0; j < dc->res_pool->pipe_count; j++) {
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 03/35] drm/amd/display: Use the right surface for flip and FreeSync

[Bhawanpreet Lakha]

From: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>

[Why]
We were always passing the first surface on the stream status for
flip updates when we should be using the surface associated with
the plane.

[How]
Use the dc_plane_state from the plane that's being updated.

FreeSync should also only keep track of updates from the primary
plane, so the check needed to be updated.

The acrtc->stream state doesn't need to be checked for NULL before
updating FreeSync either since there needs to be a stream to be
inside this function as a prerequisite.

Change-Id: Iac7f32358d6012d31a25e01053051f1f6623e821
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: David Francis <David.Francis@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c   | 17 ++++-------------
 1 file changed, 4 insertions(+), 13 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 3f50f7f0b8f0..87f43674e6f9 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -4742,7 +4742,6 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 	unsigned long flags;
 	struct amdgpu_bo *abo;
 	uint64_t tiling_flags, dcc_address;
-	struct dc_stream_status *stream_status;
 	uint32_t target, target_vblank;
 
 	struct {
@@ -4773,7 +4772,7 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 		struct drm_framebuffer *fb = new_plane_state->fb;
 		struct amdgpu_framebuffer *afb = to_amdgpu_framebuffer(fb);
 		bool pflip_needed;
-		struct dc_plane_state *surface, *dc_plane;
+		struct dc_plane_state *dc_plane;
 		struct dm_plane_state *dm_new_plane_state = to_dm_plane_state(new_plane_state);
 
 		if (plane->type == DRM_PLANE_TYPE_CURSOR) {
@@ -4836,28 +4835,20 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 			timestamp_ns = ktime_get_ns();
 			flip->flip_addrs[flip_count].flip_timestamp_in_us = div_u64(timestamp_ns, 1000);
 			flip->surface_updates[flip_count].flip_addr = &flip->flip_addrs[flip_count];
+			flip->surface_updates[flip_count].surface = dc_plane;
 
-			stream_status = dc_stream_get_status(acrtc_state->stream);
-			if (!stream_status) {
-				DRM_ERROR("No stream status for CRTC: id=%d\n",
-						acrtc_attach->crtc_id);
-				continue;
-			}
-
-			surface = stream_status->plane_states[0];
-			flip->surface_updates[flip_count].surface = surface;
 			if (!flip->surface_updates[flip_count].surface) {
 				DRM_ERROR("No surface for CRTC: id=%d\n",
 						acrtc_attach->crtc_id);
 				continue;
 			}
 
-			if (acrtc_state->stream)
+			if (plane == pcrtc->primary)
 				update_freesync_state_on_stream(
 					dm,
 					acrtc_state,
 					acrtc_state->stream,
-					surface,
+					dc_plane,
 					flip->flip_addrs[flip_count].flip_timestamp_in_us);
 
 			DRM_DEBUG_DRIVER("%s Flipping to hi: 0x%x, low: 0x%x\n",
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 04/35] drm/amd/display: Reformat dm_determine_update_type_for_commit

[Bhawanpreet Lakha]

From: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>

[Why]
The indenting for this function is a few levels too deep and can be
simplified a fair bit. This patch is in preparation for functional
changes that fix update type determination to occur less frequently
and more accurately.

[How]
Place checks early and exit/continue when possible. This isn't
a functional change.

Change-Id: I1ea41181332e2608f17ac1c608ea657bb3f8cd87
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: David Francis <David.Francis@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 126 +++++++++---------
 1 file changed, 65 insertions(+), 61 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 87f43674e6f9..d7e5391eb432 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -5886,75 +5886,79 @@ dm_determine_update_type_for_commit(struct dc *dc,
 		old_dm_crtc_state = to_dm_crtc_state(old_crtc_state);
 		num_plane = 0;
 
-		if (new_dm_crtc_state->stream) {
-
-			for_each_oldnew_plane_in_state(state, plane, old_plane_state, new_plane_state, j) {
-				new_plane_crtc = new_plane_state->crtc;
-				old_plane_crtc = old_plane_state->crtc;
-				new_dm_plane_state = to_dm_plane_state(new_plane_state);
-				old_dm_plane_state = to_dm_plane_state(old_plane_state);
-
-				if (plane->type == DRM_PLANE_TYPE_CURSOR)
-					continue;
-
-				if (!state->allow_modeset)
-					continue;
-
-				if (crtc == new_plane_crtc) {
-					updates[num_plane].surface = &surface[num_plane];
-
-					if (new_crtc_state->mode_changed) {
-						updates[num_plane].surface->src_rect =
-									new_dm_plane_state->dc_state->src_rect;
-						updates[num_plane].surface->dst_rect =
-									new_dm_plane_state->dc_state->dst_rect;
-						updates[num_plane].surface->rotation =
-									new_dm_plane_state->dc_state->rotation;
-						updates[num_plane].surface->in_transfer_func =
-									new_dm_plane_state->dc_state->in_transfer_func;
-						stream_update.dst = new_dm_crtc_state->stream->dst;
-						stream_update.src = new_dm_crtc_state->stream->src;
-					}
-
-					if (new_crtc_state->color_mgmt_changed) {
-						updates[num_plane].gamma =
-								new_dm_plane_state->dc_state->gamma_correction;
-						updates[num_plane].in_transfer_func =
-								new_dm_plane_state->dc_state->in_transfer_func;
-						stream_update.gamut_remap =
-								&new_dm_crtc_state->stream->gamut_remap_matrix;
-						stream_update.out_transfer_func =
-								new_dm_crtc_state->stream->out_transfer_func;
-					}
-
-					num_plane++;
-				}
+		if (!new_dm_crtc_state->stream) {
+			if (!new_dm_crtc_state->stream && old_dm_crtc_state->stream) {
+				update_type = UPDATE_TYPE_FULL;
+				goto cleanup;
 			}
 
-			if (num_plane > 0) {
-				ret = dm_atomic_get_state(state, &dm_state);
-				if (ret)
-					goto cleanup;
+			continue;
+		}
 
-				old_dm_state = dm_atomic_get_old_state(state);
-				if (!old_dm_state) {
-					ret = -EINVAL;
-					goto cleanup;
-				}
+		for_each_oldnew_plane_in_state(state, plane, old_plane_state, new_plane_state, j) {
+			new_plane_crtc = new_plane_state->crtc;
+			old_plane_crtc = old_plane_state->crtc;
+			new_dm_plane_state = to_dm_plane_state(new_plane_state);
+			old_dm_plane_state = to_dm_plane_state(old_plane_state);
 
-				status = dc_stream_get_status_from_state(old_dm_state->context,
-									 new_dm_crtc_state->stream);
+			if (plane->type == DRM_PLANE_TYPE_CURSOR)
+				continue;
 
-				update_type = dc_check_update_surfaces_for_stream(dc, updates, num_plane,
-										  &stream_update, status);
+			if (!state->allow_modeset)
+				continue;
 
-				if (update_type > UPDATE_TYPE_MED) {
-					update_type = UPDATE_TYPE_FULL;
-					goto cleanup;
-				}
+			if (crtc != new_plane_crtc)
+				continue;
+
+			updates[num_plane].surface = &surface[num_plane];
+
+			if (new_crtc_state->mode_changed) {
+				updates[num_plane].surface->src_rect =
+						new_dm_plane_state->dc_state->src_rect;
+				updates[num_plane].surface->dst_rect =
+						new_dm_plane_state->dc_state->dst_rect;
+				updates[num_plane].surface->rotation =
+						new_dm_plane_state->dc_state->rotation;
+				updates[num_plane].surface->in_transfer_func =
+						new_dm_plane_state->dc_state->in_transfer_func;
+				stream_update.dst = new_dm_crtc_state->stream->dst;
+				stream_update.src = new_dm_crtc_state->stream->src;
+			}
+
+			if (new_crtc_state->color_mgmt_changed) {
+				updates[num_plane].gamma =
+						new_dm_plane_state->dc_state->gamma_correction;
+				updates[num_plane].in_transfer_func =
+						new_dm_plane_state->dc_state->in_transfer_func;
+				stream_update.gamut_remap =
+						&new_dm_crtc_state->stream->gamut_remap_matrix;
+				stream_update.out_transfer_func =
+						new_dm_crtc_state->stream->out_transfer_func;
 			}
 
-		} else if (!new_dm_crtc_state->stream && old_dm_crtc_state->stream) {
+			num_plane++;
+		}
+
+		if (num_plane == 0)
+			continue;
+
+		ret = dm_atomic_get_state(state, &dm_state);
+		if (ret)
+			goto cleanup;
+
+		old_dm_state = dm_atomic_get_old_state(state);
+		if (!old_dm_state) {
+			ret = -EINVAL;
+			goto cleanup;
+		}
+
+		status = dc_stream_get_status_from_state(old_dm_state->context,
+							 new_dm_crtc_state->stream);
+
+		update_type = dc_check_update_surfaces_for_stream(dc, updates, num_plane,
+								  &stream_update, status);
+
+		if (update_type > UPDATE_TYPE_MED) {
 			update_type = UPDATE_TYPE_FULL;
 			goto cleanup;
 		}
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 05/35] drm/amd/display: Initialize stream_update to zero

[Bhawanpreet Lakha]

From: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>

[Why]
The stream_update struct is left unitialized but DC will access
its fields. This usually results in global state validation occur
during any atomic commit with state->allow_modeset = true.

[How]
Initialize the struct to zero for every stream we check.

Change-Id: Ib7f435cc2c045fc43496bea35d08a14ce8193c33
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: David Francis <David.Francis@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index d7e5391eb432..6550d00128dc 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -5868,7 +5868,6 @@ dm_determine_update_type_for_commit(struct dc *dc,
 
 	struct dc_surface_update *updates;
 	struct dc_plane_state *surface;
-	struct dc_stream_update stream_update;
 	enum surface_update_type update_type = UPDATE_TYPE_FAST;
 
 	updates = kcalloc(MAX_SURFACES, sizeof(*updates), GFP_KERNEL);
@@ -5882,6 +5881,8 @@ dm_determine_update_type_for_commit(struct dc *dc,
 	}
 
 	for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state, new_crtc_state, i) {
+		struct dc_stream_update stream_update = { 0 };
+
 		new_dm_crtc_state = to_dm_crtc_state(new_crtc_state);
 		old_dm_crtc_state = to_dm_crtc_state(old_crtc_state);
 		num_plane = 0;
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 06/35] drm/amd/display: Remove FreeSync timing changed debug output

[Bhawanpreet Lakha]

From: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>

[Why]
This provides little debug value and creates a lot of dmesg noise.

[How]
Remove it.

Change-Id: I1cbb8446ab706055cfb771232b6141156bdbe4b6
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: David Francis <David.Francis@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 6 ------
 1 file changed, 6 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 6550d00128dc..634e1ed27383 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -4713,12 +4713,6 @@ static void update_freesync_state_on_stream(
 			      new_crtc_state->base.crtc->base.id,
 			      (int)new_crtc_state->base.vrr_enabled,
 			      (int)vrr_params.state);
-
-	if (new_crtc_state->freesync_timing_changed)
-		DRM_DEBUG_KMS("VRR timing update: crtc=%u min=%u max=%u\n",
-			      new_crtc_state->base.crtc->base.id,
-				  vrr_params.adjust.v_total_min,
-				  vrr_params.adjust.v_total_max);
 }
 
 static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 07/35] drm/amd/display: Add monitor patch for backlight

[Bhawanpreet Lakha]

From: Xiaodong Yan <Xiaodong.Yan@amd.com>

off

[Why]
Different panel need different time from backlight disable to end of valid video data,
if the time is too short, panel will flash when dpms off

[How]
Add monitor patch to control the time from backlight disable to end of valid video data,

Change-Id: Iac651aa22197bce98b1c7c8537333891b87d474e
Signed-off-by: Xiaodong Yan <Xiaodong.Yan@amd.com>
Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c | 4 ++++
 drivers/gpu/drm/amd/display/dc/dc_types.h          | 1 +
 2 files changed, 5 insertions(+)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c
index 16d441d3af8a..f7f7515f65f4 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c
@@ -119,6 +119,10 @@ bool edp_receiver_ready_T9(struct dc_link *link)
 			break;
 		udelay(100); //MAx T9
 	} while (++tries < 50);
+
+	if (link->local_sink->edid_caps.panel_patch.extra_delay_backlight_off > 0)
+		udelay(link->local_sink->edid_caps.panel_patch.extra_delay_backlight_off * 1000);
+
 	return result;
 }
 bool edp_receiver_ready_T7(struct dc_link *link)
diff --git a/drivers/gpu/drm/amd/display/dc/dc_types.h b/drivers/gpu/drm/amd/display/dc/dc_types.h
index 56e7f3dab15a..da2009a108cf 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_types.h
@@ -201,6 +201,7 @@ union display_content_support {
 struct dc_panel_patch {
 	unsigned int dppowerup_delay;
 	unsigned int extra_t12_ms;
+	unsigned int extra_delay_backlight_off;
 };
 
 struct dc_edid_caps {
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 08/35] drm/amd/display: fix issue with DC brightness low with VB

[Bhawanpreet Lakha]

From: Anthony Koo <Anthony.Koo@amd.com>

[Why]
The problem is that we accidentally stopped loading some of the IRAM bytes
used for the backlight ramping mechanism. This happened when we
started reserving some region of IRAM as DMCU FW write only.

[How]
This change will define a start+end region for the IRAM read only region.
So the parameters needed for the backlight operation will be loaded
since it will be defined outside of the read only region.

Change-Id: I6ecd665ee3502f191a0162e30f669c63c1c77630
Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 .../amd/display/modules/power/power_helpers.c | 34 +++++++++++++------
 1 file changed, 24 insertions(+), 10 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c
index baab6c4ae191..27b8cf5460b5 100644
--- a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c
+++ b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c
@@ -67,9 +67,14 @@ static const unsigned char abm_config[abm_defines_max_config][abm_defines_max_le
 #define NUM_AGGR_LEVEL    4
 #define NUM_POWER_FN_SEGS 8
 #define NUM_BL_CURVE_SEGS 16
-#define IRAM_RESERVE_AREA_START 0xF0  // reserve 0xF0~0xFF are write by DMCU only
 #define IRAM_SIZE 256
 
+#define IRAM_RESERVE_AREA_START_V2 0xF0  // reserve 0xF0~0xF6 are write by DMCU only
+#define IRAM_RESERVE_AREA_END_V2 0xF6  // reserve 0xF0~0xF6 are write by DMCU only
+
+#define IRAM_RESERVE_AREA_START_V2_2 0xF0  // reserve 0xF0~0xFF are write by DMCU only
+#define IRAM_RESERVE_AREA_END_V2_2 0xFF  // reserve 0xF0~0xFF are write by DMCU only
+
 #pragma pack(push, 1)
 /* NOTE: iRAM is 256B in size */
 struct iram_table_v_2 {
@@ -148,8 +153,10 @@ struct iram_table_v_2_2 {
 	uint16_t dmcu_version;										/* 0xf4       */
 	uint8_t dmcu_state;											/* 0xf6       */
 
-	uint16_t blRampReduction;									/* 0xf7       */
-	uint16_t blRampStart;										/* 0xf9       */
+	uint8_t dummy1;												/* 0xf7       */
+	uint8_t dummy2;												/* 0xf8       */
+	uint8_t dummy3;												/* 0xf9       */
+	uint8_t dummy4;												/* 0xfa       */
 	uint8_t dummy5;												/* 0xfb       */
 	uint8_t dummy6;												/* 0xfc       */
 	uint8_t dummy7;												/* 0xfd       */
@@ -420,11 +427,6 @@ void fill_iram_v_2_2(struct iram_table_v_2_2 *ram_table, struct dmcu_iram_parame
 	ram_table->deviation_gain[2] = 0xb3;
 	ram_table->deviation_gain[3] = 0xb3;
 
-	ram_table->blRampReduction =
-		cpu_to_be16(params.backlight_ramping_reduction);
-	ram_table->blRampStart =
-		cpu_to_be16(params.backlight_ramping_start);
-
 	ram_table->min_reduction[0][0] = min_reduction_table_v_2_2[abm_config[set][0]];
 	ram_table->min_reduction[1][0] = min_reduction_table_v_2_2[abm_config[set][0]];
 	ram_table->min_reduction[2][0] = min_reduction_table_v_2_2[abm_config[set][0]];
@@ -561,6 +563,7 @@ bool dmcu_load_iram(struct dmcu *dmcu,
 	struct dmcu_iram_parameters params)
 {
 	unsigned char ram_table[IRAM_SIZE];
+	bool result = false;
 
 	if (dmcu == NULL)
 		return false;
@@ -572,10 +575,21 @@ bool dmcu_load_iram(struct dmcu *dmcu,
 
 	if (dmcu->dmcu_version.abm_version == 0x22) {
 		fill_iram_v_2_2((struct iram_table_v_2_2 *)ram_table, params);
+
+		result = dmcu->funcs->load_iram(
+				dmcu, 0, (char *)(&ram_table), IRAM_RESERVE_AREA_START_V2_2);
 	} else {
 		fill_iram_v_2((struct iram_table_v_2 *)ram_table, params);
+
+		result = dmcu->funcs->load_iram(
+				dmcu, 0, (char *)(&ram_table), IRAM_RESERVE_AREA_START_V2);
+
+		if (result)
+			result = dmcu->funcs->load_iram(
+					dmcu, IRAM_RESERVE_AREA_END_V2 + 1,
+					(char *)(&ram_table) + IRAM_RESERVE_AREA_END_V2 + 1,
+					sizeof(ram_table) - IRAM_RESERVE_AREA_END_V2 - 1);
 	}
 
-	return dmcu->funcs->load_iram(
-			dmcu, 0, (char *)(&ram_table), IRAM_RESERVE_AREA_START);
+	return result;
 }
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 09/35] drm/amd/display: dmcu is blocking due to wrong disable ABM command

[Bhawanpreet Lakha]

From: Paul Hsieh <Paul.Hsieh@amd.com>

[Why]
Second screen to clone/extend mode, driver will send ABM pipe
command to DMCU. Change mode from clone/extend to second
screen only, driver send ABM level command to disable ABM but
this command will not clear ABM pipe data. At this time, change
second screen to PC screen only, driver will send first command
"ABM_LEVEL", it will turn on ABM with incorrect ABM pile so that
DMCU is blocking.

[How]
When driver try to disable ABM, change command from "ABM LEVEL"
to "ABM PIPE" so that it will clear ABM pile data.

Change-Id: Ib845fdde7a702b9f97f51367efe71e4b57e965fb
Signed-off-by: Paul Hsieh <Paul.Hsieh@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/dce/dce_abm.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_abm.c b/drivers/gpu/drm/amd/display/dc/dce/dce_abm.c
index 2a342eae80fd..01e56f1a9f34 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_abm.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_abm.c
@@ -314,8 +314,8 @@ static bool dce_abm_immediate_disable(struct abm *abm)
 
 	/* setDMCUParam_ABMLevel */
 	REG_UPDATE_2(MASTER_COMM_CMD_REG,
-			MASTER_COMM_CMD_REG_BYTE0, MCP_ABM_LEVEL_SET,
-			MASTER_COMM_CMD_REG_BYTE2, MCP_DISABLE_ABM_IMMEDIATELY);
+			MASTER_COMM_CMD_REG_BYTE0, MCP_ABM_PIPE_SET,
+			MASTER_COMM_CMD_REG_BYTE1, MCP_DISABLE_ABM_IMMEDIATELY);
 
 	/* notifyDMCUMsg */
 	REG_UPDATE(MASTER_COMM_CNTL_REG, MASTER_COMM_INTERRUPT, 1);
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 10/35] drm/amd/display: Check that vrefresh is in freesync range

[Bhawanpreet Lakha]

From: Ilya Bakoulin <ibakouli@amd.com>

[Why]
Setting monitor refresh rate below freesync range would cause the
monitor to go blank indefinitely with freesync enabled

[How]
Set vrr_supported and ignore_msa_timing_param according to whether
the refresh rate is above or below the minimum freesync frequency.

Change-Id: I7666040400448749d1fe4061da034825b9eb511f
Signed-off-by: Ilya Bakoulin <Ilya.Bakoulin@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 634e1ed27383..5b8a84088f46 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -3113,9 +3113,6 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector,
 
 	update_stream_signal(stream, sink);
 
-	if (dm_state && dm_state->freesync_capable)
-		stream->ignore_msa_timing_param = true;
-
 finish:
 	if (sink && sink->sink_signal == SIGNAL_TYPE_VIRTUAL && aconnector->base.force != DRM_FORCE_ON)
 		dc_sink_release(sink);
@@ -5451,10 +5448,13 @@ static void get_freesync_config_for_crtc(
 	struct mod_freesync_config config = {0};
 	struct amdgpu_dm_connector *aconnector =
 			to_amdgpu_dm_connector(new_con_state->base.connector);
+	struct drm_display_mode *mode = &new_crtc_state->base.mode;
 
-	new_crtc_state->vrr_supported = new_con_state->freesync_capable;
+	new_crtc_state->vrr_supported = new_con_state->freesync_capable &&
+		aconnector->min_vfreq <= drm_mode_vrefresh(mode);
 
-	if (new_con_state->freesync_capable) {
+	if (new_crtc_state->vrr_supported) {
+		new_crtc_state->stream->ignore_msa_timing_param = true;
 		config.state = new_crtc_state->base.vrr_enabled ?
 				VRR_STATE_ACTIVE_VARIABLE :
 				VRR_STATE_INACTIVE;
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 11/35] drm/amd/display: DGAM enabled for HDR

[Bhawanpreet Lakha]

From: Krunoslav Kovac <Krunoslav.Kovac@amd.com>

[Why]
On HW that doesn't have input LUT, we may combine degamma with OS ramp
Problem here is that it assumes DGAM is inverse of PQ or SRGB. It doesn't
handle linear case, it would default to sRGB and always enable DGAM..

[How] Add handling for linear case. Also check for null ramp and instead of
blowing up, assume it's identity.

Change-Id: Ic8061248570d6aaf09ad4c2315a0d82312ebcd71
Signed-off-by: Krunoslav Kovac <Krunoslav.Kovac@amd.com>
Reviewed-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 .../amd/display/modules/color/color_gamma.c   | 89 +++++++++++--------
 1 file changed, 52 insertions(+), 37 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/modules/color/color_gamma.c b/drivers/gpu/drm/amd/display/modules/color/color_gamma.c
index eefb85928298..0fbc8fbc3541 100644
--- a/drivers/gpu/drm/amd/display/modules/color/color_gamma.c
+++ b/drivers/gpu/drm/amd/display/modules/color/color_gamma.c
@@ -1765,68 +1765,85 @@ bool mod_color_calculate_degamma_params(struct dc_transfer_func *input_tf,
 {
 	struct dc_transfer_func_distributed_points *tf_pts = &input_tf->tf_pts;
 	struct dividers dividers;
-
 	struct pwl_float_data *rgb_user = NULL;
 	struct pwl_float_data_ex *curve = NULL;
 	struct gamma_pixel *axis_x = NULL;
 	struct pixel_gamma_point *coeff = NULL;
 	enum dc_transfer_func_predefined tf = TRANSFER_FUNCTION_SRGB;
+	uint32_t i;
 	bool ret = false;
 
 	if (input_tf->type == TF_TYPE_BYPASS)
 		return false;
 
-	/* we can use hardcoded curve for plain SRGB TF */
+	/* we can use hardcoded curve for plain SRGB TF
+	 * If linear, it's bypass if on user ramp
+	 */
 	if (input_tf->type == TF_TYPE_PREDEFINED &&
-			input_tf->tf == TRANSFER_FUNCTION_SRGB &&
-			!mapUserRamp)
+			(input_tf->tf == TRANSFER_FUNCTION_SRGB ||
+					input_tf->tf == TRANSFER_FUNCTION_LINEAR) &&
+					!mapUserRamp)
 		return true;
 
 	input_tf->type = TF_TYPE_DISTRIBUTED_POINTS;
 
-	rgb_user = kvcalloc(ramp->num_entries + _EXTRA_POINTS,
-			    sizeof(*rgb_user),
-			    GFP_KERNEL);
-	if (!rgb_user)
-		goto rgb_user_alloc_fail;
+	if (mapUserRamp && ramp && ramp->type == GAMMA_RGB_256) {
+		rgb_user = kvcalloc(ramp->num_entries + _EXTRA_POINTS,
+				sizeof(*rgb_user),
+				GFP_KERNEL);
+		if (!rgb_user)
+			goto rgb_user_alloc_fail;
+
+		axis_x = kvcalloc(ramp->num_entries + _EXTRA_POINTS, sizeof(*axis_x),
+				GFP_KERNEL);
+		if (!axis_x)
+			goto axis_x_alloc_fail;
+
+		dividers.divider1 = dc_fixpt_from_fraction(3, 2);
+		dividers.divider2 = dc_fixpt_from_int(2);
+		dividers.divider3 = dc_fixpt_from_fraction(5, 2);
+
+		build_evenly_distributed_points(
+				axis_x,
+				ramp->num_entries,
+				dividers);
+
+		scale_gamma(rgb_user, ramp, dividers);
+	}
+
 	curve = kvcalloc(MAX_HW_POINTS + _EXTRA_POINTS, sizeof(*curve),
-			 GFP_KERNEL);
+			GFP_KERNEL);
 	if (!curve)
 		goto curve_alloc_fail;
-	axis_x = kvcalloc(ramp->num_entries + _EXTRA_POINTS, sizeof(*axis_x),
-			  GFP_KERNEL);
-	if (!axis_x)
-		goto axis_x_alloc_fail;
+
 	coeff = kvcalloc(MAX_HW_POINTS + _EXTRA_POINTS, sizeof(*coeff),
-			 GFP_KERNEL);
+			GFP_KERNEL);
 	if (!coeff)
 		goto coeff_alloc_fail;
 
-	dividers.divider1 = dc_fixpt_from_fraction(3, 2);
-	dividers.divider2 = dc_fixpt_from_int(2);
-	dividers.divider3 = dc_fixpt_from_fraction(5, 2);
-
 	tf = input_tf->tf;
 
-	build_evenly_distributed_points(
-			axis_x,
-			ramp->num_entries,
-			dividers);
-
-	if (ramp->type == GAMMA_RGB_256 && mapUserRamp)
-		scale_gamma(rgb_user, ramp, dividers);
-	else if (ramp->type == GAMMA_RGB_FLOAT_1024)
-		scale_gamma_dx(rgb_user, ramp, dividers);
-
 	if (tf == TRANSFER_FUNCTION_PQ)
 		build_de_pq(curve,
 				MAX_HW_POINTS,
 				coordinates_x);
-	else
+	else if (tf == TRANSFER_FUNCTION_SRGB ||
+			tf == TRANSFER_FUNCTION_BT709)
 		build_degamma(curve,
 				MAX_HW_POINTS,
 				coordinates_x,
-				tf == TRANSFER_FUNCTION_SRGB ? true:false);
+				tf == TRANSFER_FUNCTION_SRGB ? true : false);
+	else if (tf == TRANSFER_FUNCTION_LINEAR) {
+		// just copy coordinates_x into curve
+		i = 0;
+		while (i != MAX_HW_POINTS + 1) {
+			curve[i].r = coordinates_x[i].x;
+			curve[i].g = curve[i].r;
+			curve[i].b = curve[i].r;
+			i++;
+		}
+	} else
+		goto invalid_tf_fail;
 
 	tf_pts->end_exponent = 0;
 	tf_pts->x_point_at_y1_red = 1;
@@ -1836,23 +1853,21 @@ bool mod_color_calculate_degamma_params(struct dc_transfer_func *input_tf,
 	map_regamma_hw_to_x_user(ramp, coeff, rgb_user,
 			coordinates_x, axis_x, curve,
 			MAX_HW_POINTS, tf_pts,
-			mapUserRamp && ramp->type != GAMMA_CUSTOM);
-	if (ramp->type == GAMMA_CUSTOM)
-		apply_lut_1d(ramp, MAX_HW_POINTS, tf_pts);
+			mapUserRamp && ramp && ramp->type == GAMMA_RGB_256);
 
 	ret = true;
 
+invalid_tf_fail:
 	kvfree(coeff);
 coeff_alloc_fail:
-	kvfree(axis_x);
-axis_x_alloc_fail:
 	kvfree(curve);
 curve_alloc_fail:
+	kvfree(axis_x);
+axis_x_alloc_fail:
 	kvfree(rgb_user);
 rgb_user_alloc_fail:
 
 	return ret;
-
 }
 
 
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 12/35] drm/amd/display: determine if a pipe is synced by plane state

[Bhawanpreet Lakha]

From: Wenjing Liu <Wenjing.Liu@amd.com>

[why]
is_blanked is not a general indicator of if a pipe is synced
for all asics.
plane state is more accurate and applicable for all asics.

[how]
Remove is_blanked call and
add checking plane_state against NULL instead.

Change-Id: Ieaef81b60a5fa6ad694f45ff44f9cadc7057f259
Signed-off-by: Wenjing Liu <Wenjing.Liu@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/core/dc.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index db8252ec3671..e0ac009f00ab 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -907,11 +907,11 @@ static void program_timing_sync(
 			}
 		}
 
-		/* set first unblanked pipe as master */
+		/* set first pipe with plane as master */
 		for (j = 0; j < group_size; j++) {
 			struct pipe_ctx *temp;
 
-			if (pipe_set[j]->stream_res.tg->funcs->is_blanked && !pipe_set[j]->stream_res.tg->funcs->is_blanked(pipe_set[j]->stream_res.tg)) {
+			if (pipe_set[j]->plane_state) {
 				if (j == 0)
 					break;
 
@@ -922,9 +922,9 @@ static void program_timing_sync(
 			}
 		}
 
-		/* remove any other unblanked pipes as they have already been synced */
+		/* remove any other pipes with plane as they have already been synced */
 		for (j = j + 1; j < group_size; j++) {
-			if (pipe_set[j]->stream_res.tg->funcs->is_blanked && !pipe_set[j]->stream_res.tg->funcs->is_blanked(pipe_set[j]->stream_res.tg)) {
+			if (pipe_set[j]->plane_state) {
 				group_size--;
 				pipe_set[j] = pipe_set[group_size];
 				j--;
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 13/35] drm/amd/display: 3.2.16

[Bhawanpreet Lakha]

From: Steven Chiu <steven.chiu@amd.com>

Change-Id: Ib0ad307e2fd85725d04c96d9c13756471136143e
Signed-off-by: Steven Chiu <steven.chiu@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/dc.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h
index 8391bc39b7a9..0b2ffc7cfffd 100644
--- a/drivers/gpu/drm/amd/display/dc/dc.h
+++ b/drivers/gpu/drm/amd/display/dc/dc.h
@@ -39,7 +39,7 @@
 #include "inc/hw/dmcu.h"
 #include "dml/display_mode_lib.h"
 
-#define DC_VER "3.2.15"
+#define DC_VER "3.2.16"
 
 #define MAX_SURFACES 3
 #define MAX_STREAMS 6
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 14/35] drm/amd/display: add gpio lock/unlock

[Bhawanpreet Lakha]

From: Chiawen Huang <chiawen.huang@amd.com>

[Why]
When querying HPD via GPIO flow,
it will create a new gpio object then free in the end of query.
There is a irql issue for HPD querying at ISR level.

[How]
Therefore, creating the HPD gpio object in dc_link and set it as unlcok in default.
1. reducing unnecessary malloc/free when HPD querying.
2. reducing init GPIO flow.
3. add lock/unlock to prevent multi gpio service running.

Change-Id: Ibcf95d4d50c37b6831d40530194a7d6f08777c5c
Signed-off-by: Chiawen Huang <chiawen.huang@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/core/dc_link.c | 47 +++++++++----------
 drivers/gpu/drm/amd/display/dc/dc_link.h      |  1 +
 .../gpu/drm/amd/display/dc/gpio/gpio_base.c   | 12 +++++
 .../drm/amd/display/dc/gpio/gpio_service.c    | 28 +++++++++++
 .../drm/amd/display/dc/gpio/gpio_service.h    | 10 ++++
 .../drm/amd/display/include/gpio_interface.h  |  8 ++++
 6 files changed, 81 insertions(+), 25 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link.c b/drivers/gpu/drm/amd/display/dc/core/dc_link.c
index 8ff5d42587c2..137d3c126632 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link.c
@@ -76,6 +76,12 @@ static void destruct(struct dc_link *link)
 {
 	int i;
 
+	if (link->hpd_gpio != NULL) {
+		dal_gpio_close(link->hpd_gpio);
+		dal_gpio_destroy_irq(&link->hpd_gpio);
+		link->hpd_gpio = NULL;
+	}
+
 	if (link->ddc)
 		dal_ddc_service_destroy(&link->ddc);
 
@@ -931,18 +937,11 @@ bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason)
 
 bool dc_link_get_hpd_state(struct dc_link *dc_link)
 {
-	struct gpio *hpd_pin;
 	uint32_t state;
 
-	hpd_pin = get_hpd_gpio(dc_link->ctx->dc_bios,
-					dc_link->link_id, dc_link->ctx->gpio_service);
-	if (hpd_pin == NULL)
-		ASSERT(false);
-
-	dal_gpio_open(hpd_pin, GPIO_MODE_INTERRUPT);
-	dal_gpio_get_value(hpd_pin, &state);
-	dal_gpio_close(hpd_pin);
-	dal_gpio_destroy_irq(&hpd_pin);
+	dal_gpio_lock_pin(dc_link->hpd_gpio);
+	dal_gpio_get_value(dc_link->hpd_gpio, &state);
+	dal_gpio_unlock_pin(dc_link->hpd_gpio);
 
 	return state;
 }
@@ -1098,7 +1097,6 @@ static bool construct(
 	const struct link_init_data *init_params)
 {
 	uint8_t i;
-	struct gpio *hpd_gpio = NULL;
 	struct ddc_service_init_data ddc_service_init_data = { { 0 } };
 	struct dc_context *dc_ctx = init_params->ctx;
 	struct encoder_init_data enc_init_data = { 0 };
@@ -1128,10 +1126,11 @@ static bool construct(
 	if (link->dc->res_pool->funcs->link_init)
 		link->dc->res_pool->funcs->link_init(link);
 
-	hpd_gpio = get_hpd_gpio(link->ctx->dc_bios, link->link_id, link->ctx->gpio_service);
-
-	if (hpd_gpio != NULL)
-		link->irq_source_hpd = dal_irq_get_source(hpd_gpio);
+	link->hpd_gpio = get_hpd_gpio(link->ctx->dc_bios, link->link_id, link->ctx->gpio_service);
+	dal_gpio_open(link->hpd_gpio, GPIO_MODE_INTERRUPT);
+	dal_gpio_unlock_pin(link->hpd_gpio);
+	if (link->hpd_gpio != NULL)
+		link->irq_source_hpd = dal_irq_get_source(link->hpd_gpio);
 
 	switch (link->link_id.id) {
 	case CONNECTOR_ID_HDMI_TYPE_A:
@@ -1149,18 +1148,18 @@ static bool construct(
 	case CONNECTOR_ID_DISPLAY_PORT:
 		link->connector_signal =	SIGNAL_TYPE_DISPLAY_PORT;
 
-		if (hpd_gpio != NULL)
+		if (link->hpd_gpio != NULL)
 			link->irq_source_hpd_rx =
-					dal_irq_get_rx_source(hpd_gpio);
+					dal_irq_get_rx_source(link->hpd_gpio);
 
 		break;
 	case CONNECTOR_ID_EDP:
 		link->connector_signal = SIGNAL_TYPE_EDP;
 
-		if (hpd_gpio != NULL) {
+		if (link->hpd_gpio != NULL) {
 			link->irq_source_hpd = DC_IRQ_SOURCE_INVALID;
 			link->irq_source_hpd_rx =
-					dal_irq_get_rx_source(hpd_gpio);
+					dal_irq_get_rx_source(link->hpd_gpio);
 		}
 		break;
 	case CONNECTOR_ID_LVDS:
@@ -1171,10 +1170,7 @@ static bool construct(
 		goto create_fail;
 	}
 
-	if (hpd_gpio != NULL) {
-		dal_gpio_destroy_irq(&hpd_gpio);
-		hpd_gpio = NULL;
-	}
+
 
 	/* TODO: #DAL3 Implement id to str function.*/
 	LINK_INFO("Connector[%d] description:"
@@ -1277,8 +1273,9 @@ static bool construct(
 ddc_create_fail:
 create_fail:
 
-	if (hpd_gpio != NULL) {
-		dal_gpio_destroy_irq(&hpd_gpio);
+	if (link->hpd_gpio != NULL) {
+		dal_gpio_destroy_irq(&link->hpd_gpio);
+		link->hpd_gpio = NULL;
 	}
 
 	return false;
diff --git a/drivers/gpu/drm/amd/display/dc/dc_link.h b/drivers/gpu/drm/amd/display/dc/dc_link.h
index f249ff9be2a7..d26bbda61ad2 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_link.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_link.h
@@ -125,6 +125,7 @@ struct dc_link {
 	struct dc_link_status link_status;
 
 	struct link_trace link_trace;
+	struct gpio *hpd_gpio;
 };
 
 const struct dc_link_status *dc_link_get_status(const struct dc_link *dc_link);
diff --git a/drivers/gpu/drm/amd/display/dc/gpio/gpio_base.c b/drivers/gpu/drm/amd/display/dc/gpio/gpio_base.c
index 1d1efd72b291..cf76ea2d9f5a 100644
--- a/drivers/gpu/drm/amd/display/dc/gpio/gpio_base.c
+++ b/drivers/gpu/drm/amd/display/dc/gpio/gpio_base.c
@@ -101,6 +101,18 @@ enum gpio_mode dal_gpio_get_mode(
 	return gpio->mode;
 }
 
+enum gpio_result dal_gpio_lock_pin(
+	struct gpio *gpio)
+{
+	return dal_gpio_service_lock(gpio->service, gpio->id, gpio->en);
+}
+
+enum gpio_result dal_gpio_unlock_pin(
+	struct gpio *gpio)
+{
+	return dal_gpio_service_unlock(gpio->service, gpio->id, gpio->en);
+}
+
 enum gpio_result dal_gpio_change_mode(
 	struct gpio *gpio,
 	enum gpio_mode mode)
diff --git a/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c b/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c
index dada04296025..3c63a3c04dbb 100644
--- a/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c
+++ b/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.c
@@ -192,6 +192,34 @@ static void set_pin_free(
 	service->busyness[id][en] = false;
 }
 
+enum gpio_result dal_gpio_service_lock(
+	struct gpio_service *service,
+	enum gpio_id id,
+	uint32_t en)
+{
+	if (!service->busyness[id]) {
+		ASSERT_CRITICAL(false);
+		return GPIO_RESULT_OPEN_FAILED;
+	}
+
+	set_pin_busy(service, id, en);
+	return GPIO_RESULT_OK;
+}
+
+enum gpio_result dal_gpio_service_unlock(
+	struct gpio_service *service,
+	enum gpio_id id,
+	uint32_t en)
+{
+	if (!service->busyness[id]) {
+		ASSERT_CRITICAL(false);
+		return GPIO_RESULT_OPEN_FAILED;
+	}
+
+	set_pin_free(service, id, en);
+	return GPIO_RESULT_OK;
+}
+
 enum gpio_result dal_gpio_service_open(
 	struct gpio_service *service,
 	enum gpio_id id,
diff --git a/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.h b/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.h
index 1d501a43d13b..0c678af75331 100644
--- a/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.h
+++ b/drivers/gpu/drm/amd/display/dc/gpio/gpio_service.h
@@ -52,4 +52,14 @@ void dal_gpio_service_close(
 	struct gpio_service *service,
 	struct hw_gpio_pin **ptr);
 
+enum gpio_result dal_gpio_service_lock(
+	struct gpio_service *service,
+	enum gpio_id id,
+	uint32_t en);
+
+enum gpio_result dal_gpio_service_unlock(
+	struct gpio_service *service,
+	enum gpio_id id,
+	uint32_t en);
+
 #endif
diff --git a/drivers/gpu/drm/amd/display/include/gpio_interface.h b/drivers/gpu/drm/amd/display/include/gpio_interface.h
index e4fd31024b92..7de64195dc33 100644
--- a/drivers/gpu/drm/amd/display/include/gpio_interface.h
+++ b/drivers/gpu/drm/amd/display/include/gpio_interface.h
@@ -59,6 +59,14 @@ enum gpio_result dal_gpio_change_mode(
 	struct gpio *gpio,
 	enum gpio_mode mode);
 
+/* Lock Pin */
+enum gpio_result dal_gpio_lock_pin(
+	struct gpio *gpio);
+
+/* Unlock Pin */
+enum gpio_result dal_gpio_unlock_pin(
+	struct gpio *gpio);
+
 /* Get the GPIO id */
 enum gpio_id dal_gpio_get_id(
 	const struct gpio *gpio);
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 15/35] drm/amd/display: make vline_config parameter const.

[Bhawanpreet Lakha]

From: Yongqiang Sun <yongqiang.sun@amd.com>

Change-Id: Icfe018c7579ad2b3ef65195f578b8e44422d53f3
Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/core/dc.c                 | 4 ++--
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c        | 8 ++++----
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h        | 2 +-
 drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h | 2 +-
 4 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index e0ac009f00ab..d8579b207300 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -1560,11 +1560,11 @@ static void commit_planes_do_stream_update(struct dc *dc,
 
 			if (stream_update->vline0_config && pipe_ctx->stream_res.tg->funcs->program_vline_interrupt)
 				pipe_ctx->stream_res.tg->funcs->program_vline_interrupt(
-					pipe_ctx->stream_res.tg, VLINE0, stream->vline0_config);
+					pipe_ctx->stream_res.tg, VLINE0, &stream->vline0_config);
 
 			if (stream_update->vline1_config && pipe_ctx->stream_res.tg->funcs->program_vline_interrupt)
 				pipe_ctx->stream_res.tg->funcs->program_vline_interrupt(
-					pipe_ctx->stream_res.tg, VLINE1, stream->vline1_config);
+					pipe_ctx->stream_res.tg, VLINE1, &stream->vline1_config);
 
 			if ((stream_update->hdr_static_metadata && !stream->use_dynamic_meta) ||
 					stream_update->vrr_infopacket ||
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
index 1d4f9b48ed7d..cefa322df8a6 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
@@ -95,19 +95,19 @@ static void optc1_disable_stereo(struct timing_generator *optc)
 void optc1_program_vline_interrupt(
 		struct timing_generator *optc,
 		enum vline_select vline,
-		struct vline_config vline_config)
+		const struct vline_config *vline_config)
 {
 	struct optc *optc1 = DCN10TG_FROM_TG(optc);
 
 	switch (vline) {
 	case VLINE0:
 		REG_SET_2(OTG_VERTICAL_INTERRUPT0_POSITION, 0,
-				OTG_VERTICAL_INTERRUPT0_LINE_START, vline_config.start_line,
-				OTG_VERTICAL_INTERRUPT0_LINE_END, vline_config.end_line);
+				OTG_VERTICAL_INTERRUPT0_LINE_START, vline_config->start_line,
+				OTG_VERTICAL_INTERRUPT0_LINE_END, vline_config->end_line);
 		break;
 	case VLINE1:
 		REG_SET(OTG_VERTICAL_INTERRUPT1_POSITION, 0,
-					OTG_VERTICAL_INTERRUPT1_LINE_START, vline_config.start_line);
+					OTG_VERTICAL_INTERRUPT1_LINE_START, vline_config->start_line);
 		break;
 	default:
 		break;
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
index 8eb71c0160a7..b34c8a240598 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
@@ -485,7 +485,7 @@ void optc1_program_timing(
 
 void optc1_program_vline_interrupt(struct timing_generator *optc,
 		enum vline_select vline,
-		struct vline_config vline_config);
+		const struct vline_config *vline_config);
 
 void optc1_program_global_sync(
 		struct timing_generator *optc);
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
index df64cf73ceb9..d22a406c19c0 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
@@ -151,7 +151,7 @@ struct timing_generator_funcs {
 							bool use_vbios);
 	void (*program_vline_interrupt)(struct timing_generator *optc,
 			enum vline_select vline,
-			struct vline_config vline_config);
+			const struct vline_config *vline_config);
 	bool (*enable_crtc)(struct timing_generator *tg);
 	bool (*disable_crtc)(struct timing_generator *tg);
 	bool (*is_counter_moving)(struct timing_generator *tg);
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 16/35] drm/amd/display: Modify ABM 2.2 Max Reduction

[Bhawanpreet Lakha]

From: Josip Pavic <Josip.Pavic@amd.com>

[Why]
Reduced backlight for ABM 2.2 at levels 1 and 2 is desired for
power savings.

[How]
Reduce the max reduction parameters for ABM 2.2

Change-Id: I467b5404ff53878e399bfed1111c24beca7ea5e2
Signed-off-by: Josip Pavic <Josip.Pavic@amd.com>
Reviewed-by: Josip Pavic <Josip.Pavic@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/modules/power/power_helpers.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c
index 27b8cf5460b5..3ba87b076287 100644
--- a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c
+++ b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c
@@ -47,10 +47,10 @@ static const unsigned char min_reduction_table_v_2_2[13] = {
 
 /* Possible ABM 2.2 Max Reduction configs from least aggressive to most aggressive
  *  0    1     2     3     4     5     6     7     8     9     10    11   12
- * 96.1 89.8 85.1  80.3  69.4  64.7  54.9  45.1  30.2  25.1  19.6  12.5  12.5 %
+ * 96.1 89.8 74.9  69.4  64.7  52.2  48.6  39.6  30.2  25.1  19.6  12.5  12.5 %
  */
 static const unsigned char max_reduction_table_v_2_2[13] = {
-0xf5, 0xe5, 0xd9, 0xcd, 0xb1, 0xa5, 0x8c, 0x73, 0x4d, 0x40, 0x32, 0x20, 0x20};
+0xf5, 0xe5, 0xbf, 0xb1, 0xa5, 0x85, 0x7c, 0x65, 0x4d, 0x40, 0x32, 0x20, 0x20};
 
 /* Predefined ABM configuration sets. We may have different configuration sets
  * in order to satisfy different power/quality requirements.
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 17/35] drm/amd/display: store timing sync info in dc_stream_status

[Bhawanpreet Lakha]

From: Su Sung Chung <Su.Chung@amd.com>

in program_timing_sync, after all the pipes are
grouped, store timing sync info in dc_stream_status

Change-Id: I65590b0a4c74f5edb2a97e5706d6588f3dafbe3a
Signed-off-by: Su Sung Chung <Su.Chung@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/core/dc.c   | 16 +++++++++++++++-
 drivers/gpu/drm/amd/display/dc/dc_stream.h |  6 ++++++
 2 files changed, 21 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index d8579b207300..d9f62befd86a 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -869,8 +869,9 @@ static void program_timing_sync(
 		struct dc *dc,
 		struct dc_state *ctx)
 {
-	int i, j;
+	int i, j, k;
 	int group_index = 0;
+	int num_group = 0;
 	int pipe_count = dc->res_pool->pipe_count;
 	struct pipe_ctx *unsynced_pipes[MAX_PIPES] = { NULL };
 
@@ -922,6 +923,18 @@ static void program_timing_sync(
 			}
 		}
 
+
+		for (k = 0; k < group_size; k++) {
+			struct dc_stream_status *status = dc_stream_get_status_from_state(ctx, pipe_set[k]->stream);
+
+			status->timing_sync_info.group_id = num_group;
+			status->timing_sync_info.group_size = group_size;
+			if (k == 0)
+				status->timing_sync_info.master = true;
+			else
+				status->timing_sync_info.master = false;
+
+		}
 		/* remove any other pipes with plane as they have already been synced */
 		for (j = j + 1; j < group_size; j++) {
 			if (pipe_set[j]->plane_state) {
@@ -936,6 +949,7 @@ static void program_timing_sync(
 				dc, group_index, group_size, pipe_set);
 			group_index++;
 		}
+		num_group++;
 	}
 }
 
diff --git a/drivers/gpu/drm/amd/display/dc/dc_stream.h b/drivers/gpu/drm/amd/display/dc/dc_stream.h
index 0de6d7f377a6..2d1f2825de09 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_stream.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_stream.h
@@ -32,11 +32,17 @@
 /*******************************************************************************
  * Stream Interfaces
  ******************************************************************************/
+struct timing_sync_info {
+	int group_id;
+	int group_size;
+	bool master;
+};
 
 struct dc_stream_status {
 	int primary_otg_inst;
 	int stream_enc_inst;
 	int plane_count;
+	struct timing_sync_info timing_sync_info;
 	struct dc_plane_state *plane_states[MAX_SURFACE_NUM];
 };
 
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 18/35] drm/amd/display: add a debug flag to force odm combine

[Bhawanpreet Lakha]

From: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>

Adding a bit vector to allow forcing odm on specific otgs

Change-Id: Ie95a63fe7e1eba541fe5129f5e77b47e96d0530b
Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/dc.h | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h
index 0b2ffc7cfffd..ee2c7e5b6087 100644
--- a/drivers/gpu/drm/amd/display/dc/dc.h
+++ b/drivers/gpu/drm/amd/display/dc/dc.h
@@ -255,6 +255,7 @@ struct dc_debug_options {
 	bool scl_reset_length10;
 	bool hdmi20_disable;
 	bool skip_detection_link_training;
+	unsigned int force_odm_combine; //bit vector based on otg inst
 	unsigned int force_fclk_khz;
 };
 
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 19/35] drm/amd/display: Apply fake sink back to MST sequence

[Bhawanpreet Lakha]

From: "Jerry (Fangzhi) Zuo" <Jerry.Zuo@amd.com>

[Why]
It fixes the failure to create stream for sink in the scenario
when hotplug SST and MST in sequence, and disconnect MST.

[How]
Add the fake sink back after the majority of MST rework is done.

Change-Id: I0e4be01f78e337ab6f9f1813b091c081ebdd54b3
Signed-off-by: Jerry (Fangzhi) Zuo <Jerry.Zuo@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Acked-by: Tony Cheng <Tony.Cheng@amd.com>
---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c         | 8 +++-----
 .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c   | 3 +++
 2 files changed, 6 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 5b8a84088f46..125994b4753e 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -3042,11 +3042,9 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector,
 	drm_connector = &aconnector->base;
 
 	if (!aconnector->dc_sink) {
-		if (!aconnector->mst_port) {
-			sink = create_fake_sink(aconnector);
-			if (!sink)
-				return stream;
-		}
+		sink = create_fake_sink(aconnector);
+		if (!sink)
+			return stream;
 	} else {
 		sink = aconnector->dc_sink;
 	}
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
index 60da1222afaa..73f8e8fa6edb 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
@@ -173,6 +173,9 @@ static int dm_dp_mst_get_modes(struct drm_connector *connector)
 		aconnector->edid = edid;
 	}
 
+	if (aconnector->dc_sink && aconnector->dc_sink->sink_signal == SIGNAL_TYPE_VIRTUAL)
+		dc_sink_release(aconnector->dc_sink);
+
 	if (!aconnector->dc_sink) {
 		struct dc_sink *dc_sink;
 		struct dc_sink_init_data init_params = {
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 20/35] drm/amd/display: Disconnect mpcc when changing tg

[Bhawanpreet Lakha]

From: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>

[Why]
This fixes an mpc programming error for the following sequence of
atomic commits when pipe split is enabled:

Commit 1: CRTC0 (plane 4, plane 3)

Pipe 0: old_plane_state = A0, new_plane_state = A1,   new_tg = T0
Pipe 1: old_plane_state = B0, new_plane_state = B1,   new_tg = T0
Pipe 2: old_plane_state = A0, new_plane_state = A1,   new_tg = T0
Pipe 3: old_plane_state = B0, new_plane_state = B1,   new_tg = T0

Commit 2: CRTC0 (plane 3), CRTC1 (plane 2)

Pipe 0: old_plane_state = A1, new_plane_state = A2,   new_tg = T0
Pipe 1: old_plane_state = B1, new_plane_state = B2,   new_tg = T1
Pipe 2: old_plane_state = A1, new_plane_state = NULL, new_tg = NULL
Pipe 3: old_plane_state = B1, new_plane_state = NULL, new_tg = NULL

In the second commit the assertion for mpcc in use is hit because
mpcc disconnect never occurs for pipe 1. This is because the stream
changes for pipe 1 and the opp_list is empty.

This sequence occurs when running the
"igt@kms_plane_multiple@atomic-pipe-A-tiling-none" test with two
displays connected.

[How]
Expand the reset condition to include:

"old_pipe_ctx->stream_res.tg != new_pipe_ctx->stream_res.tg"

...but only when the plane state is non-NULL for both old and new.

Change-Id: Ida230c0e5248e8d74aaa72561afa42f4206e5c55
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
index 9cde24dbdac8..9f23ea283838 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
@@ -2334,9 +2334,10 @@ static void dcn10_apply_ctx_for_surface(
 			}
 		}
 
-		if (!pipe_ctx->plane_state &&
-			old_pipe_ctx->plane_state &&
-			old_pipe_ctx->stream_res.tg == tg) {
+		if ((!pipe_ctx->plane_state ||
+		     pipe_ctx->stream_res.tg != old_pipe_ctx->stream_res.tg) &&
+		    old_pipe_ctx->plane_state &&
+		    old_pipe_ctx->stream_res.tg == tg) {
 
 			dc->hwss.plane_atomic_disconnect(dc, old_pipe_ctx);
 			removed_pipe[i] = true;
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 21/35] drm/amd/display: Don't re-program planes for DPMS changes

[Bhawanpreet Lakha]

From: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>

[Why]
There are opt1c lock warnings and CRTC read timeouts when running the
"igt@kms_plane@plane-position-hole-dpms-pipe-*" tests. These are
caused by trying to reprogram planes that are not in the current
context.

DPMS off removes the stream from the context. In this case:

new_crtc_state->active_changed = true
new_crtc_state->mode_changed = false

The planes are reprogrammed before the stream is removed from the
context because stream_state->mode_changed = false.

For DPMS adds the stream and planes back to the context:

new_crtc_state->active_changed = true
new_crtc_state->mode_changed = false

The planes are also reprogrammed here before the stream is added to the
context because stream_state->mode_changed = true. They were not
previously in the current context so warnings occur here.

[How]
Set stream_state->mode_changed = true when
new_crtc_state->active_changed = true too.

This prevents reprogramming before the context is applied in DC. The
programming will be done after the context is applied.

Change-Id: I3b405f3081d3ecca68b40e894e681ed0372e92df
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Sun peng Li <Sunpeng.Li@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Acked-by: Tony Cheng <Tony.Cheng@amd.com>
---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 125994b4753e..c6caea1197d4 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -4976,7 +4976,8 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 static void amdgpu_dm_crtc_copy_transient_flags(struct drm_crtc_state *crtc_state,
 						struct dc_stream_state *stream_state)
 {
-	stream_state->mode_changed = crtc_state->mode_changed;
+	stream_state->mode_changed =
+		crtc_state->mode_changed || crtc_state->active_changed;
 }
 
 static int amdgpu_dm_atomic_commit(struct drm_device *dev,
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 22/35] drm/amd/display: Clear dc_sink after it gets released

[Bhawanpreet Lakha]

From: "Jerry (Fangzhi) Zuo" <Jerry.Zuo@amd.com>

[Why]
The dc_sink was released but the pointer on the aconnector was
not cleared.

[How]
Clear it.

Change-Id: Ifc8d72f51b5c6600d63b88f3011aef8ffc957e66
Signed-off-by: Jerry (Fangzhi) Zuo <Jerry.Zuo@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
index 73f8e8fa6edb..44c1a02e6452 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
@@ -173,8 +173,10 @@ static int dm_dp_mst_get_modes(struct drm_connector *connector)
 		aconnector->edid = edid;
 	}
 
-	if (aconnector->dc_sink && aconnector->dc_sink->sink_signal == SIGNAL_TYPE_VIRTUAL)
+	if (aconnector->dc_sink && aconnector->dc_sink->sink_signal == SIGNAL_TYPE_VIRTUAL) {
 		dc_sink_release(aconnector->dc_sink);
+		aconnector->dc_sink = NULL;
+	}
 
 	if (!aconnector->dc_sink) {
 		struct dc_sink *dc_sink;
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 23/35] drm/amd/display: Calc vline position in dc.

[Bhawanpreet Lakha]

From: Yongqiang Sun <yongqiang.sun@amd.com>

We need to calcualte vline position in DC for DCN.

Change-Id: I1632f1d2c597e842f575e0b2979f443ac6d3ffcc
Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/core/dc.c      |  8 +-
 drivers/gpu/drm/amd/display/dc/dc_stream.h    | 15 ++--
 .../gpu/drm/amd/display/dc/dcn10/dcn10_optc.c | 80 ++++++++++++++++++-
 .../gpu/drm/amd/display/dc/dcn10/dcn10_optc.h |  6 +-
 .../amd/display/dc/inc/hw/timing_generator.h  |  8 +-
 5 files changed, 97 insertions(+), 20 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index d9f62befd86a..409097cf58ed 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -1572,13 +1572,13 @@ static void commit_planes_do_stream_update(struct dc *dc,
 					stream_update->adjust->v_total_min,
 					stream_update->adjust->v_total_max);
 
-			if (stream_update->vline0_config && pipe_ctx->stream_res.tg->funcs->program_vline_interrupt)
+			if (stream_update->periodic_vsync_config && pipe_ctx->stream_res.tg->funcs->program_vline_interrupt)
 				pipe_ctx->stream_res.tg->funcs->program_vline_interrupt(
-					pipe_ctx->stream_res.tg, VLINE0, &stream->vline0_config);
+					pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing, VLINE0, &stream->periodic_vsync_config);
 
-			if (stream_update->vline1_config && pipe_ctx->stream_res.tg->funcs->program_vline_interrupt)
+			if (stream_update->enhanced_sync_config && pipe_ctx->stream_res.tg->funcs->program_vline_interrupt)
 				pipe_ctx->stream_res.tg->funcs->program_vline_interrupt(
-					pipe_ctx->stream_res.tg, VLINE1, &stream->vline1_config);
+					pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing, VLINE1, &stream->enhanced_sync_config);
 
 			if ((stream_update->hdr_static_metadata && !stream->use_dynamic_meta) ||
 					stream_update->vrr_infopacket ||
diff --git a/drivers/gpu/drm/amd/display/dc/dc_stream.h b/drivers/gpu/drm/amd/display/dc/dc_stream.h
index 2d1f2825de09..90f019eb54b3 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_stream.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_stream.h
@@ -51,11 +51,12 @@ struct freesync_context {
 	bool dummy;
 };
 
-struct vline_config {
-	unsigned int start_line;
-	unsigned int end_line;
+union vline_config {
+	unsigned int line_number;
+	unsigned long long delta_in_ns;
 };
 
+
 struct dc_stream_state {
 	// sink is deprecated, new code should not reference
 	// this pointer
@@ -105,8 +106,8 @@ struct dc_stream_state {
 	/* DMCU info */
 	unsigned int abm_level;
 
-	struct vline_config vline0_config;
-	struct vline_config vline1_config;
+	union vline_config periodic_vsync_config;
+	union vline_config enhanced_sync_config;
 
 	/* from core_stream struct */
 	struct dc_context *ctx;
@@ -155,8 +156,8 @@ struct dc_stream_update {
 	struct dc_info_packet *hdr_static_metadata;
 	unsigned int *abm_level;
 
-	struct vline_config *vline0_config;
-	struct vline_config *vline1_config;
+	union vline_config *periodic_vsync_config;
+	union vline_config *enhanced_sync_config;
 
 	struct dc_crtc_timing_adjust *adjust;
 	struct dc_info_packet *vrr_infopacket;
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
index cefa322df8a6..0355dcb8554a 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
@@ -92,22 +92,94 @@ static void optc1_disable_stereo(struct timing_generator *optc)
 		OTG_3D_STRUCTURE_STEREO_SEL_OVR, 0);
 }
 
+static uint32_t get_start_vline(struct timing_generator *optc, const struct dc_crtc_timing *dc_crtc_timing)
+{
+	struct dc_crtc_timing patched_crtc_timing;
+	int vesa_sync_start;
+	int asic_blank_end;
+	int interlace_factor;
+	int vertical_line_start;
+
+	patched_crtc_timing = *dc_crtc_timing;
+	optc1_apply_front_porch_workaround(optc, &patched_crtc_timing);
+
+	vesa_sync_start = patched_crtc_timing.h_addressable +
+			patched_crtc_timing.h_border_right +
+			patched_crtc_timing.h_front_porch;
+
+	asic_blank_end = patched_crtc_timing.h_total -
+			vesa_sync_start -
+			patched_crtc_timing.h_border_left;
+
+	interlace_factor = patched_crtc_timing.flags.INTERLACE ? 2 : 1;
+
+	vesa_sync_start = patched_crtc_timing.v_addressable +
+			patched_crtc_timing.v_border_bottom +
+			patched_crtc_timing.v_front_porch;
+
+	asic_blank_end = (patched_crtc_timing.v_total -
+			vesa_sync_start -
+			patched_crtc_timing.v_border_top)
+			* interlace_factor;
+
+	vertical_line_start = asic_blank_end - optc->dlg_otg_param.vstartup_start + 1;
+	if (vertical_line_start < 0) {
+		ASSERT(0);
+		vertical_line_start = 0;
+	}
+
+	return vertical_line_start;
+}
+
+static void calc_vline_position(
+		struct timing_generator *optc,
+		const struct dc_crtc_timing *dc_crtc_timing,
+		unsigned long long vsync_delta,
+		uint32_t *start_line,
+		uint32_t *end_line)
+{
+	unsigned long long req_delta_tens_of_usec = div64_u64((vsync_delta + 9999), 10000);
+	unsigned long long pix_clk_hundreds_khz = div64_u64((dc_crtc_timing->pix_clk_100hz + 999), 1000);
+	uint32_t req_delta_lines = (uint32_t) div64_u64(
+			(req_delta_tens_of_usec * pix_clk_hundreds_khz + dc_crtc_timing->h_total - 1),
+								dc_crtc_timing->h_total);
+
+	uint32_t vsync_line = get_start_vline(optc, dc_crtc_timing);
+
+	if (req_delta_lines != 0)
+			req_delta_lines--;
+
+		if (req_delta_lines > vsync_line)
+			*start_line = dc_crtc_timing->v_total - (req_delta_lines - vsync_line) - 1;
+		else
+			*start_line = vsync_line - req_delta_lines;
+
+		*end_line = *start_line + 2;
+
+		if (*end_line >= dc_crtc_timing->v_total)
+			*end_line = 2;
+}
+
 void optc1_program_vline_interrupt(
 		struct timing_generator *optc,
+		const struct dc_crtc_timing *dc_crtc_timing,
 		enum vline_select vline,
-		const struct vline_config *vline_config)
+		const union vline_config *vline_config)
 {
 	struct optc *optc1 = DCN10TG_FROM_TG(optc);
+	uint32_t start_line = 0;
+	uint32_t end_line = 0;
 
 	switch (vline) {
 	case VLINE0:
+		calc_vline_position(optc, dc_crtc_timing, vline_config->delta_in_ns, &start_line, &end_line);
 		REG_SET_2(OTG_VERTICAL_INTERRUPT0_POSITION, 0,
-				OTG_VERTICAL_INTERRUPT0_LINE_START, vline_config->start_line,
-				OTG_VERTICAL_INTERRUPT0_LINE_END, vline_config->end_line);
+				OTG_VERTICAL_INTERRUPT0_LINE_START, start_line,
+				OTG_VERTICAL_INTERRUPT0_LINE_END, end_line);
 		break;
 	case VLINE1:
 		REG_SET(OTG_VERTICAL_INTERRUPT1_POSITION, 0,
-					OTG_VERTICAL_INTERRUPT1_LINE_START, vline_config->start_line);
+					OTG_VERTICAL_INTERRUPT1_LINE_START, vline_config->line_number);
 		break;
 	default:
 		break;
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
index b34c8a240598..8a4e3e37e894 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
@@ -483,9 +483,11 @@ void optc1_program_timing(
 	const struct dc_crtc_timing *dc_crtc_timing,
 	bool use_vbios);
 
-void optc1_program_vline_interrupt(struct timing_generator *optc,
+void optc1_program_vline_interrupt(
+		struct timing_generator *optc,
+		const struct dc_crtc_timing *dc_crtc_timing,
 		enum vline_select vline,
-		const struct vline_config *vline_config);
+		const union vline_config *vline_config);
 
 void optc1_program_global_sync(
 		struct timing_generator *optc);
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
index d22a406c19c0..39fec0186c10 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
@@ -134,7 +134,7 @@ struct dc_crtc_timing;
 
 struct drr_params;
 
-struct vline_config;
+union vline_config;
 
 
 enum vline_select {
@@ -149,9 +149,11 @@ struct timing_generator_funcs {
 	void (*program_timing)(struct timing_generator *tg,
 							const struct dc_crtc_timing *timing,
 							bool use_vbios);
-	void (*program_vline_interrupt)(struct timing_generator *optc,
+	void (*program_vline_interrupt)(
+			struct timing_generator *optc,
+			const struct dc_crtc_timing *dc_crtc_timing,
 			enum vline_select vline,
-			const struct vline_config *vline_config);
+			const union vline_config *vline_config);
 	bool (*enable_crtc)(struct timing_generator *tg);
 	bool (*disable_crtc)(struct timing_generator *tg);
 	bool (*is_counter_moving)(struct timing_generator *tg);
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 24/35] drm/amd/display: Disable Stutter for Stereo 3D

[Bhawanpreet Lakha]

From: Wesley Chalmers <Wesley.Chalmers@amd.com>

[WHY]
Bandwidth calculation formulas currently do not take Stereo 3D + Stutter
properly into account. Disable stutter feature when we detect a Stereo
3D mode as a temporary workaround.

Change-Id: Icad94f294c662740b988592c6ed0228705b02071
Signed-off-by: Wesley Chalmers <Wesley.Chalmers@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Acked-by: Martin Leung <Martin.Leung@amd.com>
---
 .../drm/amd/display/dc/dcn10/dcn10_hubbub.c   | 19 ++++++++++--------
 .../drm/amd/display/dc/dcn10/dcn10_hubbub.h   |  2 +-
 .../amd/display/dc/dcn10/dcn10_hw_sequencer.c | 20 ++++++++++++++++++-
 3 files changed, 31 insertions(+), 10 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c
index 5a4614c371bc..e161ad836812 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c
@@ -88,11 +88,18 @@ void hubbub1_wm_read_state(struct hubbub *hubbub,
 	s->dram_clk_chanage = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_D);
 }
 
-void hubbub1_disable_allow_self_refresh(struct hubbub *hubbub)
+void hubbub1_allow_self_refresh_control(struct hubbub *hubbub, bool allow)
 {
 	struct dcn10_hubbub *hubbub1 = TO_DCN10_HUBBUB(hubbub);
-	REG_UPDATE(DCHUBBUB_ARB_DRAM_STATE_CNTL,
-			DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE, 0);
+
+	/*
+	 * DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE = 1 means do not allow stutter
+	 * DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE = 0 means allow stutter
+	 */
+
+	REG_UPDATE_2(DCHUBBUB_ARB_DRAM_STATE_CNTL,
+			DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_VALUE, 0,
+			DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE, !allow);
 }
 
 bool hububu1_is_allow_self_refresh_enabled(struct hubbub *hubbub)
@@ -262,8 +269,6 @@ void hubbub1_program_watermarks(
 		bool safe_to_lower)
 {
 	struct dcn10_hubbub *hubbub1 = TO_DCN10_HUBBUB(hubbub);
-
-	uint32_t force_en = hubbub1->base.ctx->dc->debug.disable_stutter ? 1 : 0;
 	/*
 	 * Need to clamp to max of the register values (i.e. no wrap)
 	 * for dcn1, all wm registers are 21-bit wide
@@ -537,9 +542,7 @@ void hubbub1_program_watermarks(
 	REG_UPDATE(DCHUBBUB_ARB_DF_REQ_OUTSTAND,
 			DCHUBBUB_ARB_MIN_REQ_OUTSTAND, 68);
 
-	REG_UPDATE_2(DCHUBBUB_ARB_DRAM_STATE_CNTL,
-			DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_VALUE, 0,
-			DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE, force_en);
+	hubbub1_allow_self_refresh_control(hubbub, !hubbub->ctx->dc->debug.disable_stutter);
 
 #if 0
 	REG_UPDATE_2(DCHUBBUB_ARB_WATERMARK_CHANGE_CNTL,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.h
index c681e1cc9290..9cd4a5194154 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.h
@@ -204,7 +204,7 @@ void hubbub1_program_watermarks(
 		unsigned int refclk_mhz,
 		bool safe_to_lower);
 
-void hubbub1_disable_allow_self_refresh(struct hubbub *hubbub);
+void hubbub1_allow_self_refresh_control(struct hubbub *hubbub, bool allow);
 
 bool hububu1_is_allow_self_refresh_enabled(struct hubbub *hubub);
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
index 9f23ea283838..7cead0398c15 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
@@ -1011,7 +1011,7 @@ static void dcn10_init_hw(struct dc *dc)
 			 */
 			if (allow_self_fresh_force_enable == false &&
 					hububu1_is_allow_self_refresh_enabled(dc->res_pool->hubbub))
-				hubbub1_disable_allow_self_refresh(dc->res_pool->hubbub);
+				hubbub1_allow_self_refresh_control(dc->res_pool->hubbub, true);
 
 			disable_vga(dc->hwseq);
 		}
@@ -2384,6 +2384,22 @@ static void dcn10_apply_ctx_for_surface(
 		hubbub1_wm_change_req_wa(dc->res_pool->hubbub);
 }
 
+static void dcn10_stereo_hw_frame_pack_wa(struct dc *dc, struct dc_state *context)
+{
+	uint8_t i;
+
+	for (i = 0; i < context->stream_count; i++) {
+		if (context->streams[i]->timing.timing_3d_format
+				== TIMING_3D_FORMAT_HW_FRAME_PACKING) {
+			/*
+			 * Disable stutter
+			 */
+			hubbub1_allow_self_refresh_control(dc->res_pool->hubbub, false);
+			break;
+		}
+	}
+}
+
 static void dcn10_prepare_bandwidth(
 		struct dc *dc,
 		struct dc_state *context)
@@ -2405,6 +2421,7 @@ static void dcn10_prepare_bandwidth(
 			&context->bw.dcn.watermarks,
 			dc->res_pool->ref_clock_inKhz / 1000,
 			true);
+	dcn10_stereo_hw_frame_pack_wa(dc, context);
 
 	if (dc->debug.pplib_wm_report_mode == WM_REPORT_OVERRIDE)
 		dcn_bw_notify_pplib_of_wm_ranges(dc);
@@ -2434,6 +2451,7 @@ static void dcn10_optimize_bandwidth(
 			&context->bw.dcn.watermarks,
 			dc->res_pool->ref_clock_inKhz / 1000,
 			true);
+	dcn10_stereo_hw_frame_pack_wa(dc, context);
 
 	if (dc->debug.pplib_wm_report_mode == WM_REPORT_OVERRIDE)
 		dcn_bw_notify_pplib_of_wm_ranges(dc);
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 25/35] drm/amd/display: DC VM Fixes

[Bhawanpreet Lakha]

From: Eryk Brol <eryk.brol@amd.com>

[Why]
VM_helper needs to be intialized with the dc struct in order to fix
an unallocated memory issue. System aperture settings should be
initialized to 0 and guarded with a check to make sure vm_config
is valid.

[How]
Allocate and free memory for vm_helper with other dc members.
Check whether the vm_config valid bit is set before initializing
aperture settings.

Change-Id: I9083eb884492cb16ca767b03eeed41e46277133c
Signed-off-by: Eryk Brol <eryk.brol@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/core/dc.c       |  1 -
 .../gpu/drm/amd/display/dc/core/dc_vm_helper.c | 18 +++++++-----------
 drivers/gpu/drm/amd/display/dc/dc.h            |  1 -
 drivers/gpu/drm/amd/display/dc/inc/vm_helper.h |  7 ++++---
 4 files changed, 11 insertions(+), 16 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index 409097cf58ed..af5f486d24e0 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -621,7 +621,6 @@ static bool construct(struct dc *dc,
 #endif
 
 	enum dce_version dc_version = DCE_VERSION_UNKNOWN;
-
 	dc_dceip = kzalloc(sizeof(*dc_dceip), GFP_KERNEL);
 	if (!dc_dceip) {
 		dm_error("%s: failed to create dceip\n", __func__);
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_vm_helper.c b/drivers/gpu/drm/amd/display/dc/core/dc_vm_helper.c
index e54b8ac339b2..6ce87b682a32 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_vm_helper.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_vm_helper.c
@@ -112,16 +112,12 @@ uint8_t get_vmid_for_ptb(struct vm_helper *vm_helper, int64_t ptb, uint8_t hubp_
 	return vmid;
 }
 
-struct vm_helper init_vm_helper(unsigned int num_vmid, unsigned int num_hubp)
+void init_vm_helper(struct vm_helper *vm_helper, unsigned int num_vmid, unsigned int num_hubp)
 {
-	static uint64_t ptb_assigned_to_vmid[MAX_VMID];
-	static struct vmid_usage hubp_vmid_usage[MAX_HUBP];
-
-	return (struct vm_helper){
-		.num_vmid = num_vmid,
-		.num_hubp = num_hubp,
-		.num_vmids_available = num_vmid - 1,
-		.ptb_assigned_to_vmid = ptb_assigned_to_vmid,
-		.hubp_vmid_usage = hubp_vmid_usage
-	};
+	vm_helper->num_vmid = num_vmid;
+	vm_helper->num_hubp = num_hubp;
+	vm_helper->num_vmids_available = num_vmid - 1;
+
+	memset(vm_helper->hubp_vmid_usage, 0, sizeof(vm_helper->hubp_vmid_usage[0]) * MAX_HUBP);
+	memset(vm_helper->ptb_assigned_to_vmid, 0, sizeof(vm_helper->ptb_assigned_to_vmid[0]) * MAX_VMID);
 }
diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h
index ee2c7e5b6087..2224bba914ec 100644
--- a/drivers/gpu/drm/amd/display/dc/dc.h
+++ b/drivers/gpu/drm/amd/display/dc/dc.h
@@ -265,7 +265,6 @@ struct dc_debug_data {
 	uint32_t auxErrorCount;
 };
 
-
 struct dc_state;
 struct resource_pool;
 struct dce_hwseq;
diff --git a/drivers/gpu/drm/amd/display/dc/inc/vm_helper.h b/drivers/gpu/drm/amd/display/dc/inc/vm_helper.h
index a202206e22a3..193407f76a80 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/vm_helper.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/vm_helper.h
@@ -39,8 +39,8 @@ struct vm_helper {
 	unsigned int num_vmid;
 	unsigned int num_hubp;
 	unsigned int num_vmids_available;
-	uint64_t *ptb_assigned_to_vmid;
-	struct vmid_usage *hubp_vmid_usage;
+	uint64_t ptb_assigned_to_vmid[MAX_VMID];
+	struct vmid_usage hubp_vmid_usage[MAX_HUBP];
 };
 
 uint8_t get_vmid_for_ptb(
@@ -48,7 +48,8 @@ uint8_t get_vmid_for_ptb(
 		int64_t ptb,
 		uint8_t pipe_idx);
 
-struct vm_helper init_vm_helper(
+void init_vm_helper(
+	struct vm_helper *vm_helper,
 	unsigned int num_vmid,
 	unsigned int num_hubp);
 
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 26/35] drm/amd/display: link_rate_set should index into table

[Bhawanpreet Lakha]

From: Anthony Koo <Anthony.Koo@amd.com>

[Why]
Current implementation that maps link_rate_set value to
actual link rate is incorrect.

[How]
Fix this implementation, such that link_rate_set indexes into
the supported_link_rate table.

Change-Id: I5ff7f7e25951c442240698835526de1ff5400b89
Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Reviewed-by: Harry Wentland <Harry.Wentland@amd.com>
Acked-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 .../gpu/drm/amd/display/dc/core/dc_link_dp.c  | 35 +------------------
 1 file changed, 1 insertion(+), 34 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
index 92f565ca1260..8ead8c4d575c 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c
@@ -2538,7 +2538,6 @@ void detect_edp_sink_caps(struct dc_link *link)
 	uint32_t entry;
 	uint32_t link_rate_in_khz;
 	enum dc_link_rate link_rate = LINK_RATE_UNKNOWN;
-	uint8_t link_rate_set = 0;
 
 	retrieve_link_cap(link);
 
@@ -2558,39 +2557,7 @@ void detect_edp_sink_caps(struct dc_link *link)
 				link_rate = linkRateInKHzToLinkRateMultiplier(link_rate_in_khz);
 				if (link->reported_link_cap.link_rate < link_rate) {
 					link->reported_link_cap.link_rate = link_rate;
-
-					switch (link_rate) {
-					case LINK_RATE_LOW:
-						link_rate_set = 1;
-						break;
-					case LINK_RATE_RATE_2:
-						link_rate_set = 2;
-						break;
-					case LINK_RATE_RATE_3:
-						link_rate_set = 3;
-						break;
-					case LINK_RATE_HIGH:
-						link_rate_set = 4;
-						break;
-					case LINK_RATE_RBR2:
-						link_rate_set = 5;
-						break;
-					case LINK_RATE_RATE_6:
-						link_rate_set = 6;
-						break;
-					case LINK_RATE_HIGH2:
-						link_rate_set = 7;
-						break;
-					case LINK_RATE_HIGH3:
-						link_rate_set = 8;
-						break;
-					default:
-						link_rate_set = 0;
-						break;
-					}
-
-					if (link->dpcd_caps.link_rate_set < link_rate_set)
-						link->dpcd_caps.link_rate_set = link_rate_set;
+					link->dpcd_caps.link_rate_set = entry;
 				}
 			}
 		}
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 27/35] drm/amd/display: PIP overlay corruption

[Bhawanpreet Lakha]

From: Murton Liu <murton.liu@amd.com>

[Why]
When moving mouse onto or off of pip plane,
screen would flash briefly due to garbage negative
pos values being programmed for cursor.
Also, text flashes due to PIP flips taking too long.

[How]
When negative pos value seen, default to 0 and adjust by modifying cursor hotspot.
For flip issue, only do post update when optimize required vs all the time.

Change-Id: I38990fe0a916306a781ae06147d5d6bd89f3fe92
Signed-off-by: Murton Liu <murton.liu@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Acked-by: Sivapiriyan Kumarasamy <Sivapiriyan.Kumarasamy@amd.com>
---
 drivers/gpu/drm/amd/display/dc/core/dc.c      |  3 +++
 .../gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c | 23 +++++++++++++++++--
 2 files changed, 24 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index af5f486d24e0..77e1bd18b1dd 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -1125,6 +1125,9 @@ bool dc_post_update_surfaces_to_stream(struct dc *dc)
 	int i;
 	struct dc_state *context = dc->current_state;
 
+	if (dc->optimized_required == false)
+		return true;
+
 	post_surface_trace(dc);
 
 	for (i = 0; i < dc->res_pool->pipe_count; i++)
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c
index 0ba68d41b9c3..683829466a44 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c
@@ -1150,9 +1150,28 @@ void hubp1_cursor_set_position(
 	REG_UPDATE(CURSOR_CONTROL,
 			CURSOR_ENABLE, cur_en);
 
-	REG_SET_2(CURSOR_POSITION, 0,
-			CURSOR_X_POSITION, pos->x,
+	//account for cases where we see negative offset relative to overlay plane
+	if (src_x_offset < 0 && src_y_offset < 0) {
+		REG_SET_2(CURSOR_POSITION, 0,
+			CURSOR_X_POSITION, 0,
+			CURSOR_Y_POSITION, 0);
+		x_hotspot -= src_x_offset;
+		y_hotspot -= src_y_offset;
+	} else if (src_x_offset < 0) {
+		REG_SET_2(CURSOR_POSITION, 0,
+			CURSOR_X_POSITION, 0,
 			CURSOR_Y_POSITION, pos->y);
+		x_hotspot -= src_x_offset;
+	} else if (src_y_offset < 0) {
+		REG_SET_2(CURSOR_POSITION, 0,
+			CURSOR_X_POSITION, pos->x,
+			CURSOR_Y_POSITION, 0);
+		y_hotspot -= src_y_offset;
+	} else {
+		REG_SET_2(CURSOR_POSITION, 0,
+				CURSOR_X_POSITION, pos->x,
+				CURSOR_Y_POSITION, pos->y);
+	}
 
 	REG_SET_2(CURSOR_HOT_SPOT, 0,
 			CURSOR_HOT_SPOT_X, x_hotspot,
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 28/35] drm/amd/display:3.2.17

[Bhawanpreet Lakha]

From: mmcgarri <mark.mcgarrity@amd.com>

Change-Id: I74d794df38e0d7f191220b230416ca6472f74ad5
Signed-off-by: mmcgarri <mark.mcgarrity@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/dc.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h
index 2224bba914ec..bade219d1268 100644
--- a/drivers/gpu/drm/amd/display/dc/dc.h
+++ b/drivers/gpu/drm/amd/display/dc/dc.h
@@ -39,7 +39,7 @@
 #include "inc/hw/dmcu.h"
 #include "dml/display_mode_lib.h"
 
-#define DC_VER "3.2.16"
+#define DC_VER "3.2.17"
 
 #define MAX_SURFACES 3
 #define MAX_STREAMS 6
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 29/35] drm/amd/display: interface to check if timing can be seamless

[Bhawanpreet Lakha]

From: Anthony Koo <Anthony.Koo@amd.com>

[Why]
Need to figure out whether a timing we want to commit matches
something that GOP already programmed, in which case
we can decide to some optimizations

[How]
1. Add way to check for DIG FE
2. Add way to check for matching OTG timing
3. Add way to check for matching pixel clock (if possible)
    - Currently only support DP for pixel clock, since it is easy to calc

Change-Id: Iaf9e0f674f4afd4720e68ccfd8431d6a5d8346ba
Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/core/dc.c      | 46 +++++++++++++++
 drivers/gpu/drm/amd/display/dc/dc.h           |  4 ++
 .../drm/amd/display/dc/dce/dce_clock_source.c | 28 ++++++++-
 .../amd/display/dc/dcn10/dcn10_link_encoder.c | 10 ++++
 .../amd/display/dc/dcn10/dcn10_link_encoder.h |  2 +
 .../gpu/drm/amd/display/dc/dcn10/dcn10_optc.c | 59 +++++++++++++++++++
 .../gpu/drm/amd/display/dc/inc/clock_source.h |  4 ++
 .../drm/amd/display/dc/inc/hw/link_encoder.h  |  1 +
 .../amd/display/dc/inc/hw/timing_generator.h  |  2 +
 9 files changed, 154 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index 77e1bd18b1dd..b620520771a7 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -969,6 +969,52 @@ static bool context_changed(
 	return false;
 }
 
+bool dc_validate_seamless_boot_timing(struct dc *dc,
+				const struct dc_sink *sink,
+				struct dc_crtc_timing *crtc_timing)
+{
+	struct timing_generator *tg;
+	struct dc_link *link = sink->link;
+	unsigned int inst;
+
+	/* Check for enabled DIG to identify enabled display */
+	if (!link->link_enc->funcs->is_dig_enabled(link->link_enc))
+		return false;
+
+	/* Check for which front end is used by this encoder.
+	 * Note the inst is 1 indexed, where 0 is undefined.
+	 * Note that DIG_FE can source from different OTG but our
+	 * current implementation always map 1-to-1, so this code makes
+	 * the same assumption and doesn't check OTG source.
+	 */
+	inst = link->link_enc->funcs->get_dig_frontend(link->link_enc) - 1;
+
+	/* Instance should be within the range of the pool */
+	if (inst >= dc->res_pool->pipe_count)
+		return false;
+
+	tg = dc->res_pool->timing_generators[inst];
+
+	if (!tg->funcs->is_matching_timing)
+		return false;
+
+	if (!tg->funcs->is_matching_timing(tg, crtc_timing))
+		return false;
+
+	if (dc_is_dp_signal(link->connector_signal)) {
+		unsigned int pix_clk_100hz;
+
+		dc->res_pool->dp_clock_source->funcs->get_pixel_clk_frequency_100hz(
+			dc->res_pool->dp_clock_source,
+			inst, &pix_clk_100hz);
+
+		if (crtc_timing->pix_clk_100hz != pix_clk_100hz)
+			return false;
+	}
+
+	return true;
+}
+
 bool dc_enable_stereo(
 	struct dc *dc,
 	struct dc_state *context,
diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h
index bade219d1268..1a7fd6aa77eb 100644
--- a/drivers/gpu/drm/amd/display/dc/dc.h
+++ b/drivers/gpu/drm/amd/display/dc/dc.h
@@ -594,6 +594,10 @@ struct dc_validation_set {
 	uint8_t plane_count;
 };
 
+bool dc_validate_seamless_boot_timing(struct dc *dc,
+				const struct dc_sink *sink,
+				struct dc_crtc_timing *crtc_timing);
+
 enum dc_status dc_validate_plane(struct dc *dc, const struct dc_plane_state *plane_state);
 
 void get_clock_requirements_for_state(struct dc_state *state, struct AsicStateEx *info);
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c
index c67e90e5c339..71d5777de961 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_clock_source.c
@@ -977,6 +977,28 @@ static bool dce110_clock_source_power_down(
 	return bp_result == BP_RESULT_OK;
 }
 
+static bool get_pixel_clk_frequency_100hz(
+		struct clock_source *clock_source,
+		unsigned int inst,
+		unsigned int *pixel_clk_khz)
+{
+	struct dce110_clk_src *clk_src = TO_DCE110_CLK_SRC(clock_source);
+	unsigned int clock_hz = 0;
+
+	if (clock_source->id == CLOCK_SOURCE_ID_DP_DTO) {
+		clock_hz = REG_READ(PHASE[inst]);
+
+		/* NOTE: There is agreement with VBIOS here that MODULO is
+		 * programmed equal to DPREFCLK, in which case PHASE will be
+		 * equivalent to pixel clock.
+		 */
+		*pixel_clk_khz = clock_hz / 100;
+		return true;
+	}
+
+	return false;
+}
+
 /*****************************************/
 /* Constructor                           */
 /*****************************************/
@@ -984,12 +1006,14 @@ static bool dce110_clock_source_power_down(
 static const struct clock_source_funcs dce112_clk_src_funcs = {
 	.cs_power_down = dce110_clock_source_power_down,
 	.program_pix_clk = dce112_program_pix_clk,
-	.get_pix_clk_dividers = dce112_get_pix_clk_dividers
+	.get_pix_clk_dividers = dce112_get_pix_clk_dividers,
+	.get_pixel_clk_frequency_100hz = get_pixel_clk_frequency_100hz
 };
 static const struct clock_source_funcs dce110_clk_src_funcs = {
 	.cs_power_down = dce110_clock_source_power_down,
 	.program_pix_clk = dce110_program_pix_clk,
-	.get_pix_clk_dividers = dce110_get_pix_clk_dividers
+	.get_pix_clk_dividers = dce110_get_pix_clk_dividers,
+	.get_pixel_clk_frequency_100hz = get_pixel_clk_frequency_100hz
 };
 
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c
index 771449f8984f..a9db372688ff 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.c
@@ -85,6 +85,7 @@ static const struct link_encoder_funcs dcn10_lnk_enc_funcs = {
 	.enable_hpd = dcn10_link_encoder_enable_hpd,
 	.disable_hpd = dcn10_link_encoder_disable_hpd,
 	.is_dig_enabled = dcn10_is_dig_enabled,
+	.get_dig_frontend = dcn10_get_dig_frontend,
 	.destroy = dcn10_link_encoder_destroy
 };
 
@@ -495,6 +496,15 @@ bool dcn10_is_dig_enabled(struct link_encoder *enc)
 	return value;
 }
 
+unsigned int dcn10_get_dig_frontend(struct link_encoder *enc)
+{
+	struct dcn10_link_encoder *enc10 = TO_DCN10_LINK_ENC(enc);
+	uint32_t value;
+
+	REG_GET(DIG_BE_CNTL, DIG_FE_SOURCE_SELECT, &value);
+	return value;
+}
+
 static void link_encoder_disable(struct dcn10_link_encoder *enc10)
 {
 	/* reset training pattern */
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.h
index 670b46e887ed..b74b80a247ec 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_link_encoder.h
@@ -336,6 +336,8 @@ void dcn10_psr_program_secondary_packet(struct link_encoder *enc,
 
 bool dcn10_is_dig_enabled(struct link_encoder *enc);
 
+unsigned int dcn10_get_dig_frontend(struct link_encoder *enc);
+
 void dcn10_aux_initialize(struct dcn10_link_encoder *enc10);
 
 #endif /* __DC_LINK_ENCODER__DCN10_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
index 0355dcb8554a..51c98e99237e 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
@@ -1208,6 +1208,64 @@ bool optc1_is_stereo_left_eye(struct timing_generator *optc)
 	return ret;
 }
 
+bool optc1_is_matching_timing(struct timing_generator *tg,
+		const struct dc_crtc_timing *otg_timing)
+{
+	struct dc_crtc_timing hw_crtc_timing = {0};
+	struct dcn_otg_state s = {0};
+
+	if (tg == NULL || otg_timing == NULL)
+		return false;
+
+	optc1_read_otg_state(DCN10TG_FROM_TG(tg), &s);
+
+	hw_crtc_timing.h_total = s.h_total + 1;
+	hw_crtc_timing.h_addressable = s.h_total - ((s.h_total - s.h_blank_start) + s.h_blank_end);
+	hw_crtc_timing.h_front_porch = s.h_total + 1 - s.h_blank_start;
+	hw_crtc_timing.h_sync_width = s.h_sync_a_end - s.h_sync_a_start;
+
+	hw_crtc_timing.v_total = s.v_total + 1;
+	hw_crtc_timing.v_addressable = s.v_total - ((s.v_total - s.v_blank_start) + s.v_blank_end);
+	hw_crtc_timing.v_front_porch = s.v_total + 1 - s.v_blank_start;
+	hw_crtc_timing.v_sync_width = s.v_sync_a_end - s.v_sync_a_start;
+
+	if (otg_timing->h_total != hw_crtc_timing.h_total)
+		return false;
+
+	if (otg_timing->h_border_left != hw_crtc_timing.h_border_left)
+		return false;
+
+	if (otg_timing->h_addressable != hw_crtc_timing.h_addressable)
+		return false;
+
+	if (otg_timing->h_border_right != hw_crtc_timing.h_border_right)
+		return false;
+
+	if (otg_timing->h_front_porch != hw_crtc_timing.h_front_porch)
+		return false;
+
+	if (otg_timing->h_sync_width != hw_crtc_timing.h_sync_width)
+		return false;
+
+	if (otg_timing->v_total != hw_crtc_timing.v_total)
+		return false;
+
+	if (otg_timing->v_border_top != hw_crtc_timing.v_border_top)
+		return false;
+
+	if (otg_timing->v_addressable != hw_crtc_timing.v_addressable)
+		return false;
+
+	if (otg_timing->v_border_bottom != hw_crtc_timing.v_border_bottom)
+		return false;
+
+	if (otg_timing->v_sync_width != hw_crtc_timing.v_sync_width)
+		return false;
+
+	return true;
+}
+
+
 void optc1_read_otg_state(struct optc *optc1,
 		struct dcn_otg_state *s)
 {
@@ -1404,6 +1462,7 @@ static const struct timing_generator_funcs dcn10_tg_funcs = {
 		.get_frame_count = optc1_get_vblank_counter,
 		.get_scanoutpos = optc1_get_crtc_scanoutpos,
 		.get_otg_active_size = optc1_get_otg_active_size,
+		.is_matching_timing = optc1_is_matching_timing,
 		.set_early_control = optc1_set_early_control,
 		/* used by enable_timing_synchronization. Not need for FPGA */
 		.wait_for_state = optc1_wait_for_state,
diff --git a/drivers/gpu/drm/amd/display/dc/inc/clock_source.h b/drivers/gpu/drm/amd/display/dc/inc/clock_source.h
index 43d1fbd8ace5..fe6301cb8681 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/clock_source.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/clock_source.h
@@ -166,6 +166,10 @@ struct clock_source_funcs {
 			struct clock_source *,
 			struct pixel_clk_params *,
 			struct pll_settings *);
+	bool (*get_pixel_clk_frequency_100hz)(
+			struct clock_source *clock_source,
+			unsigned int inst,
+			unsigned int *pixel_clk_khz);
 };
 
 struct clock_source {
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h b/drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h
index c20fdcaac53b..c9d3e37e9531 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h
@@ -153,6 +153,7 @@ struct link_encoder_funcs {
 	void (*enable_hpd)(struct link_encoder *enc);
 	void (*disable_hpd)(struct link_encoder *enc);
 	bool (*is_dig_enabled)(struct link_encoder *enc);
+	unsigned int (*get_dig_frontend)(struct link_encoder *enc);
 	void (*destroy)(struct link_encoder **enc);
 };
 
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
index 39fec0186c10..5d6cca7826f3 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
@@ -170,6 +170,8 @@ struct timing_generator_funcs {
 	bool (*get_otg_active_size)(struct timing_generator *optc,
 			uint32_t *otg_active_width,
 			uint32_t *otg_active_height);
+	bool (*is_matching_timing)(struct timing_generator *tg,
+			const struct dc_crtc_timing *otg_timing);
 	void (*set_early_control)(struct timing_generator *tg,
 							   uint32_t early_cntl);
 	void (*wait_for_state)(struct timing_generator *tg,
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 30/35] drm/amd/display: refactor out programming of vupdate interrupt

[Bhawanpreet Lakha]

From: Anthony Koo <Anthony.Koo@amd.com>

[Why]
More clearly isolate the code that is involved in programming of
vupdate interrupt

[How]
Add function for programming of vupdate interrupt.
Call it after timing is programmed.

Change-Id: Ib05c77800300d5f9bfb43ecfb7b4f2b615778ab5
Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Acked-by: Tony Cheng <Tony.Cheng@amd.com>
---
 .../display/dc/dce110/dce110_hw_sequencer.c   |  5 ++
 .../gpu/drm/amd/display/dc/dcn10/dcn10_optc.c | 51 +++++++++++++++----
 .../amd/display/dc/inc/hw/timing_generator.h  |  3 ++
 3 files changed, 48 insertions(+), 11 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
index db0ef41eb91c..774c51c2d039 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
@@ -1341,6 +1341,11 @@ static enum dc_status apply_single_controller_ctx_to_hw(
 	/*  */
 	dc->hwss.enable_stream_timing(pipe_ctx, context, dc);
 
+	if (pipe_ctx->stream_res.tg->funcs->program_vupdate_interrupt)
+		pipe_ctx->stream_res.tg->funcs->program_vupdate_interrupt(
+						pipe_ctx->stream_res.tg,
+							&stream->timing);
+
 	if (pipe_ctx->stream->signal != SIGNAL_TYPE_VIRTUAL)
 		pipe_ctx->stream_res.stream_enc->funcs->dig_connect_to_otg(
 			pipe_ctx->stream_res.stream_enc,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
index 51c98e99237e..1dbd1d3999e6 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
@@ -186,6 +186,42 @@ void optc1_program_vline_interrupt(
 	}
 }
 
+void optc1_program_vupdate_interrupt(
+		struct timing_generator *optc,
+		const struct dc_crtc_timing *dc_crtc_timing)
+{
+	struct optc *optc1 = DCN10TG_FROM_TG(optc);
+	int32_t vertical_line_start;
+	uint32_t asic_blank_end;
+	uint32_t vesa_sync_start;
+	struct dc_crtc_timing patched_crtc_timing;
+
+	patched_crtc_timing = *dc_crtc_timing;
+	optc1_apply_front_porch_workaround(optc, &patched_crtc_timing);
+
+	/* asic_h_blank_end = HsyncWidth + HbackPorch =
+	 * vesa. usHorizontalTotal - vesa. usHorizontalSyncStart -
+	 * vesa.h_left_border
+	 */
+	vesa_sync_start = patched_crtc_timing.h_addressable +
+			patched_crtc_timing.h_border_right +
+			patched_crtc_timing.h_front_porch;
+
+	asic_blank_end = patched_crtc_timing.h_total -
+			vesa_sync_start -
+			patched_crtc_timing.h_border_left;
+
+	/* Use OTG_VERTICAL_INTERRUPT2 replace VUPDATE interrupt,
+	 * program the reg for interrupt postition.
+	 */
+	vertical_line_start = asic_blank_end - optc->dlg_otg_param.vstartup_start + 1;
+	if (vertical_line_start < 0)
+		vertical_line_start = 0;
+
+	REG_SET(OTG_VERTICAL_INTERRUPT2_POSITION, 0,
+			OTG_VERTICAL_INTERRUPT2_LINE_START, vertical_line_start);
+}
+
 /**
  * program_timing_generator   used by mode timing set
  * Program CRTC Timing Registers - OTG_H_*, OTG_V_*, Pixel repetition.
@@ -288,22 +324,14 @@ void optc1_program_timing(
 			patched_crtc_timing.v_addressable +
 			patched_crtc_timing.v_border_bottom);
 
-	REG_UPDATE_2(OTG_V_BLANK_START_END,
-			OTG_V_BLANK_START, asic_blank_start,
-			OTG_V_BLANK_END, asic_blank_end);
-
-	/* Use OTG_VERTICAL_INTERRUPT2 replace VUPDATE interrupt,
-	 * program the reg for interrupt postition.
-	 */
 	vertical_line_start = asic_blank_end - optc->dlg_otg_param.vstartup_start + 1;
 	v_fp2 = 0;
 	if (vertical_line_start < 0)
 		v_fp2 = -vertical_line_start;
-	if (vertical_line_start < 0)
-		vertical_line_start = 0;
 
-	REG_SET(OTG_VERTICAL_INTERRUPT2_POSITION, 0,
-			OTG_VERTICAL_INTERRUPT2_LINE_START, vertical_line_start);
+	REG_UPDATE_2(OTG_V_BLANK_START_END,
+			OTG_V_BLANK_START, asic_blank_start,
+			OTG_V_BLANK_END, asic_blank_end);
 
 	/* v_sync polarity */
 	v_sync_polarity = patched_crtc_timing.flags.VSYNC_POSITIVE_POLARITY ?
@@ -1453,6 +1481,7 @@ static const struct timing_generator_funcs dcn10_tg_funcs = {
 		.validate_timing = optc1_validate_timing,
 		.program_timing = optc1_program_timing,
 		.program_vline_interrupt = optc1_program_vline_interrupt,
+		.program_vupdate_interrupt = optc1_program_vupdate_interrupt,
 		.program_global_sync = optc1_program_global_sync,
 		.enable_crtc = optc1_enable_crtc,
 		.disable_crtc = optc1_disable_crtc,
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
index 5d6cca7826f3..03ae941895f3 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
@@ -154,6 +154,9 @@ struct timing_generator_funcs {
 			const struct dc_crtc_timing *dc_crtc_timing,
 			enum vline_select vline,
 			const union vline_config *vline_config);
+
+	void (*program_vupdate_interrupt)(struct timing_generator *optc,
+			const struct dc_crtc_timing *dc_crtc_timing);
 	bool (*enable_crtc)(struct timing_generator *tg);
 	bool (*disable_crtc)(struct timing_generator *tg);
 	bool (*is_counter_moving)(struct timing_generator *tg);
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 31/35] drm/amd/display: add way to determine if link is active

[Bhawanpreet Lakha]

From: Anthony Koo <Anthony.Koo@amd.com>

[Why]
Need to understand whether link is active aside from stream state.
This could be used to check what links are enabled by GOP.

[How]
Store link_active state in link status and initialize it by checking
if the DIG is enabled.
Keep it updated on every link enable and disable

Change-Id: I6d985d0ec11404bbaba7e4a23d4ca50d519aa0ac
Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/core/dc_link.c             | 4 ++++
 drivers/gpu/drm/amd/display/dc/dc_link.h                  | 1 +
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c | 5 +++++
 3 files changed, 10 insertions(+)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link.c b/drivers/gpu/drm/amd/display/dc/core/dc_link.c
index 137d3c126632..e2e7bf2ba84b 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link.c
@@ -2612,6 +2612,8 @@ void core_link_enable_stream(
 			}
 		}
 
+		stream->link->link_status.link_active = true;
+
 		core_dc->hwss.enable_audio_stream(pipe_ctx);
 
 		/* turn off otg test pattern if enable */
@@ -2646,6 +2648,8 @@ void core_link_disable_stream(struct pipe_ctx *pipe_ctx, int option)
 	core_dc->hwss.disable_stream(pipe_ctx, option);
 
 	disable_link(pipe_ctx->stream->link, pipe_ctx->stream->signal);
+
+	pipe_ctx->stream->link->link_status.link_active = false;
 }
 
 void core_link_set_avmute(struct pipe_ctx *pipe_ctx, bool enable)
diff --git a/drivers/gpu/drm/amd/display/dc/dc_link.h b/drivers/gpu/drm/amd/display/dc/dc_link.h
index d26bbda61ad2..8fc223defed4 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_link.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_link.h
@@ -30,6 +30,7 @@
 #include "grph_object_defs.h"
 
 struct dc_link_status {
+	bool link_active;
 	struct dpcd_caps *dpcd_caps;
 };
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
index 7cead0398c15..7117144d48bd 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
@@ -1027,6 +1027,11 @@ static void dcn10_init_hw(struct dc *dc)
 				dc->hwss.edp_power_control(link, true);
 
 			link->link_enc->funcs->hw_init(link->link_enc);
+
+			/* Check for enabled DIG to identify enabled display */
+			if (link->link_enc->funcs->is_dig_enabled &&
+				link->link_enc->funcs->is_dig_enabled(link->link_enc))
+				link->link_status.link_active = true;
 		}
 	}
 
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 32/35] drm/amd/display: add seamless boot flag to stream

[Bhawanpreet Lakha]

From: Anthony Koo <Anthony.Koo@amd.com>

[Why]
If we determine the stream we are trying to commit
matches HW, we want to try to optimize.

[How]
Try to acquire the HW resources that are already enabled
and optimize.
Also skip backend reprogramming

Change-Id: Ia3b66c4ccf3e23bf1d91b9d30645abb54306d1ba
Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/core/dc.c      |  3 +
 drivers/gpu/drm/amd/display/dc/core/dc_link.c | 14 ++++-
 .../gpu/drm/amd/display/dc/core/dc_resource.c | 56 ++++++++++++++++++-
 drivers/gpu/drm/amd/display/dc/dc_stream.h    |  4 +-
 .../display/dc/dce110/dce110_hw_sequencer.c   |  4 +-
 5 files changed, 75 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index b620520771a7..52f838442e21 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -1096,6 +1096,9 @@ static enum dc_status dc_commit_state_no_check(struct dc *dc, struct dc_state *c
 		const struct dc_link *link = context->streams[i]->link;
 		struct dc_stream_status *status;
 
+		if (context->streams[i]->apply_seamless_boot_optimization)
+			context->streams[i]->apply_seamless_boot_optimization = false;
+
 		if (!context->streams[i]->mode_changed)
 			continue;
 
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_link.c b/drivers/gpu/drm/amd/display/dc/core/dc_link.c
index e2e7bf2ba84b..6d2cef05b4d7 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_link.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_link.c
@@ -2579,13 +2579,23 @@ void core_link_enable_stream(
 			&stream->timing);
 
 	if (!IS_FPGA_MAXIMUS_DC(core_dc->ctx->dce_environment)) {
+		bool apply_edp_fast_boot_optimization =
+			pipe_ctx->stream->apply_edp_fast_boot_optimization;
+
+		pipe_ctx->stream->apply_edp_fast_boot_optimization = false;
+
 		resource_build_info_frame(pipe_ctx);
 		core_dc->hwss.update_info_frame(pipe_ctx);
 
+		/* Do not touch link on seamless boot optimization. */
+		if (pipe_ctx->stream->apply_seamless_boot_optimization) {
+			pipe_ctx->stream->dpms_off = false;
+			return;
+		}
+
 		/* eDP lit up by bios already, no need to enable again. */
 		if (pipe_ctx->stream->signal == SIGNAL_TYPE_EDP &&
-				pipe_ctx->stream->apply_edp_fast_boot_optimization) {
-			pipe_ctx->stream->apply_edp_fast_boot_optimization = false;
+					apply_edp_fast_boot_optimization) {
 			pipe_ctx->stream->dpms_off = false;
 			return;
 		}
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
index 9888bc7659f3..349ab8017776 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
@@ -1800,6 +1800,51 @@ static void calculate_phy_pix_clks(struct dc_stream_state *stream)
 		stream->phy_pix_clk *= 2;
 }
 
+static int acquire_resource_from_hw_enabled_state(
+		struct resource_context *res_ctx,
+		const struct resource_pool *pool,
+		struct dc_stream_state *stream)
+{
+	struct dc_link *link = stream->link;
+	unsigned int inst;
+
+	/* Check for enabled DIG to identify enabled display */
+	if (!link->link_enc->funcs->is_dig_enabled(link->link_enc))
+		return -1;
+
+	/* Check for which front end is used by this encoder.
+	 * Note the inst is 1 indexed, where 0 is undefined.
+	 * Note that DIG_FE can source from different OTG but our
+	 * current implementation always map 1-to-1, so this code makes
+	 * the same assumption and doesn't check OTG source.
+	 */
+	inst = link->link_enc->funcs->get_dig_frontend(link->link_enc) - 1;
+
+	/* Instance should be within the range of the pool */
+	if (inst >= pool->pipe_count)
+		return -1;
+
+	if (!res_ctx->pipe_ctx[inst].stream) {
+		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[inst];
+
+		pipe_ctx->stream_res.tg = pool->timing_generators[inst];
+		pipe_ctx->plane_res.mi = pool->mis[inst];
+		pipe_ctx->plane_res.hubp = pool->hubps[inst];
+		pipe_ctx->plane_res.ipp = pool->ipps[inst];
+		pipe_ctx->plane_res.xfm = pool->transforms[inst];
+		pipe_ctx->plane_res.dpp = pool->dpps[inst];
+		pipe_ctx->stream_res.opp = pool->opps[inst];
+		if (pool->dpps[inst])
+			pipe_ctx->plane_res.mpcc_inst = pool->dpps[inst]->inst;
+		pipe_ctx->pipe_idx = inst;
+
+		pipe_ctx->stream = stream;
+		return inst;
+	}
+
+	return -1;
+}
+
 enum dc_status resource_map_pool_resources(
 		const struct dc  *dc,
 		struct dc_state *context,
@@ -1824,8 +1869,15 @@ enum dc_status resource_map_pool_resources(
 
 	calculate_phy_pix_clks(stream);
 
-	/* acquire new resources */
-	pipe_idx = acquire_first_free_pipe(&context->res_ctx, pool, stream);
+	if (stream->apply_seamless_boot_optimization)
+		pipe_idx = acquire_resource_from_hw_enabled_state(
+				&context->res_ctx,
+				pool,
+				stream);
+
+	if (pipe_idx < 0)
+		/* acquire new resources */
+		pipe_idx = acquire_first_free_pipe(&context->res_ctx, pool, stream);
 
 #ifdef CONFIG_DRM_AMD_DC_DCN1_0
 	if (pipe_idx < 0)
diff --git a/drivers/gpu/drm/amd/display/dc/dc_stream.h b/drivers/gpu/drm/amd/display/dc/dc_stream.h
index 90f019eb54b3..a798694992b9 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_stream.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_stream.h
@@ -119,7 +119,6 @@ struct dc_stream_state {
 	int phy_pix_clk;
 	enum signal_type signal;
 	bool dpms_off;
-	bool apply_edp_fast_boot_optimization;
 
 	void *dm_stream_context;
 
@@ -146,6 +145,9 @@ struct dc_stream_state {
 		uint8_t otg_offset;
 	} out;
 
+	bool apply_edp_fast_boot_optimization;
+	bool apply_seamless_boot_optimization;
+
 	uint32_t stream_id;
 };
 
diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
index 774c51c2d039..9e67805679a5 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
@@ -1339,7 +1339,9 @@ static enum dc_status apply_single_controller_ctx_to_hw(
 	}
 
 	/*  */
-	dc->hwss.enable_stream_timing(pipe_ctx, context, dc);
+	/* Do not touch stream timing on seamless boot optimization. */
+	if (!pipe_ctx->stream->apply_seamless_boot_optimization)
+		dc->hwss.enable_stream_timing(pipe_ctx, context, dc);
 
 	if (pipe_ctx->stream_res.tg->funcs->program_vupdate_interrupt)
 		pipe_ctx->stream_res.tg->funcs->program_vupdate_interrupt(
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 33/35] drm/amd/display: refactor programming of DRR

[Bhawanpreet Lakha]

From: Anthony Koo <Anthony.Koo@amd.com>

[Why]
Keep enable_stream_timing programming only
timing related stuff.

[How]
Move DRR and static screen mask programming from
enable_stream_timing to outside in
apply_single_controller_ctx_to_hw

Change-Id: I4bf0d94fc3c7dbf52d5a9f91699e6c22b37a0306
Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 .../display/dc/dce110/dce110_hw_sequencer.c   | 31 +++++++++----------
 .../amd/display/dc/dcn10/dcn10_hw_sequencer.c | 15 ---------
 2 files changed, 15 insertions(+), 31 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
index 9e67805679a5..cc81ec6e9134 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
@@ -1250,8 +1250,6 @@ static enum dc_status dce110_enable_stream_timing(
 	struct pipe_ctx *pipe_ctx_old = &dc->current_state->res_ctx.
 			pipe_ctx[pipe_ctx->pipe_idx];
 	struct tg_color black_color = {0};
-	struct drr_params params = {0};
-	unsigned int event_triggers = 0;
 
 	if (!pipe_ctx_old->stream) {
 
@@ -1280,20 +1278,6 @@ static enum dc_status dce110_enable_stream_timing(
 				pipe_ctx->stream_res.tg,
 				&stream->timing,
 				true);
-
-		params.vertical_total_min = stream->adjust.v_total_min;
-		params.vertical_total_max = stream->adjust.v_total_max;
-		if (pipe_ctx->stream_res.tg->funcs->set_drr)
-			pipe_ctx->stream_res.tg->funcs->set_drr(
-				pipe_ctx->stream_res.tg, &params);
-
-		// DRR should set trigger event to monitor surface update event
-		if (stream->adjust.v_total_min != 0 &&
-				stream->adjust.v_total_max != 0)
-			event_triggers = 0x80;
-		if (pipe_ctx->stream_res.tg->funcs->set_static_screen_control)
-			pipe_ctx->stream_res.tg->funcs->set_static_screen_control(
-				pipe_ctx->stream_res.tg, event_triggers);
 	}
 
 	if (!pipe_ctx_old->stream) {
@@ -1313,6 +1297,8 @@ static enum dc_status apply_single_controller_ctx_to_hw(
 		struct dc *dc)
 {
 	struct dc_stream_state *stream = pipe_ctx->stream;
+	struct drr_params params = {0};
+	unsigned int event_triggers = 0;
 
 	if (pipe_ctx->stream_res.audio != NULL) {
 		struct audio_output audio_output;
@@ -1348,6 +1334,19 @@ static enum dc_status apply_single_controller_ctx_to_hw(
 						pipe_ctx->stream_res.tg,
 							&stream->timing);
 
+	params.vertical_total_min = stream->adjust.v_total_min;
+	params.vertical_total_max = stream->adjust.v_total_max;
+	if (pipe_ctx->stream_res.tg->funcs->set_drr)
+		pipe_ctx->stream_res.tg->funcs->set_drr(
+			pipe_ctx->stream_res.tg, &params);
+
+	// DRR should set trigger event to monitor surface update event
+	if (stream->adjust.v_total_min != 0 && stream->adjust.v_total_max != 0)
+		event_triggers = 0x80;
+	if (pipe_ctx->stream_res.tg->funcs->set_static_screen_control)
+		pipe_ctx->stream_res.tg->funcs->set_static_screen_control(
+				pipe_ctx->stream_res.tg, event_triggers);
+
 	if (pipe_ctx->stream->signal != SIGNAL_TYPE_VIRTUAL)
 		pipe_ctx->stream_res.stream_enc->funcs->dig_connect_to_otg(
 			pipe_ctx->stream_res.stream_enc,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
index 7117144d48bd..a03bbbf94d5d 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
@@ -636,8 +636,6 @@ static enum dc_status dcn10_enable_stream_timing(
 	struct dc_stream_state *stream = pipe_ctx->stream;
 	enum dc_color_space color_space;
 	struct tg_color black_color = {0};
-	struct drr_params params = {0};
-	unsigned int event_triggers = 0;
 
 	/* by upper caller loop, pipe0 is parent pipe and be called first.
 	 * back end is set up by for pipe0. Other children pipe share back end
@@ -705,19 +703,6 @@ static enum dc_status dcn10_enable_stream_timing(
 		return DC_ERROR_UNEXPECTED;
 	}
 
-	params.vertical_total_min = stream->adjust.v_total_min;
-	params.vertical_total_max = stream->adjust.v_total_max;
-	if (pipe_ctx->stream_res.tg->funcs->set_drr)
-		pipe_ctx->stream_res.tg->funcs->set_drr(
-			pipe_ctx->stream_res.tg, &params);
-
-	// DRR should set trigger event to monitor surface update event
-	if (stream->adjust.v_total_min != 0 && stream->adjust.v_total_max != 0)
-		event_triggers = 0x80;
-	if (pipe_ctx->stream_res.tg->funcs->set_static_screen_control)
-		pipe_ctx->stream_res.tg->funcs->set_static_screen_control(
-				pipe_ctx->stream_res.tg, event_triggers);
-
 	/* TODO program crtc source select for non-virtual signal*/
 	/* TODO program FMT */
 	/* TODO setup link_enc */
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 34/35] drm/amd/display: refactor init_hw to isolate pipe related init

[Bhawanpreet Lakha]

From: Anthony Koo <Anthony.Koo@amd.com>

[Why]
Pipe related init is possible to optimized if we know what we
intend to program, and if we can determine it matches what is
already programmed for the pipe.

[How]
First step is to isolate the pipe related init code

Change-Id: Idc1c378a0eea78b8d93faced686b19d9b14f77f3
Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 .../display/dc/dce110/dce110_hw_sequencer.c   |   6 +
 .../amd/display/dc/dcn10/dcn10_hw_sequencer.c | 159 +++++++++---------
 .../gpu/drm/amd/display/dc/inc/hw_sequencer.h |   2 +
 3 files changed, 85 insertions(+), 82 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
index cc81ec6e9134..392781bfb7a5 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
@@ -2275,6 +2275,11 @@ static void dce110_enable_per_frame_crtc_position_reset(
 
 }
 
+static void init_pipes(struct dc *dc, struct dc_state *context)
+{
+	// Do nothing
+}
+
 static void init_hw(struct dc *dc)
 {
 	int i;
@@ -2642,6 +2647,7 @@ static const struct hw_sequencer_funcs dce110_funcs = {
 	.program_gamut_remap = program_gamut_remap,
 	.program_output_csc = program_output_csc,
 	.init_hw = init_hw,
+	.init_pipes = init_pipes,
 	.apply_ctx_to_hw = dce110_apply_ctx_to_hw,
 	.apply_ctx_for_surface = dce110_apply_ctx_for_surface,
 	.update_plane_addr = update_plane_addr,
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
index a03bbbf94d5d..117d9d8227f7 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
@@ -956,97 +956,34 @@ static void dcn10_disable_plane(struct dc *dc, struct pipe_ctx *pipe_ctx)
 					pipe_ctx->pipe_idx);
 }
 
-static void dcn10_init_hw(struct dc *dc)
+static void dcn10_init_pipes(struct dc *dc, struct dc_state *context)
 {
 	int i;
-	struct abm *abm = dc->res_pool->abm;
-	struct dmcu *dmcu = dc->res_pool->dmcu;
-	struct dce_hwseq *hws = dc->hwseq;
-	struct dc_bios *dcb = dc->ctx->dc_bios;
-	struct dc_state  *context = dc->current_state;
-
-	if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
-		REG_WRITE(REFCLK_CNTL, 0);
-		REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_ENABLE, 1);
-		REG_WRITE(DIO_MEM_PWR_CTRL, 0);
-
-		if (!dc->debug.disable_clock_gate) {
-			/* enable all DCN clock gating */
-			REG_WRITE(DCCG_GATE_DISABLE_CNTL, 0);
-
-			REG_WRITE(DCCG_GATE_DISABLE_CNTL2, 0);
-
-			REG_UPDATE(DCFCLK_CNTL, DCFCLK_GATE_DIS, 0);
-		}
-
-		enable_power_gating_plane(dc->hwseq, true);
-	} else {
-
-		if (!dcb->funcs->is_accelerated_mode(dcb)) {
-			bool allow_self_fresh_force_enable =
-					hububu1_is_allow_self_refresh_enabled(dc->res_pool->hubbub);
-
-			bios_golden_init(dc);
-
-			/* WA for making DF sleep when idle after resume from S0i3.
-			 * DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE is set to 1 by
-			 * command table, if DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE = 0
-			 * before calling command table and it changed to 1 after,
-			 * it should be set back to 0.
-			 */
-			if (allow_self_fresh_force_enable == false &&
-					hububu1_is_allow_self_refresh_enabled(dc->res_pool->hubbub))
-				hubbub1_allow_self_refresh_control(dc->res_pool->hubbub, true);
-
-			disable_vga(dc->hwseq);
-		}
-
-		for (i = 0; i < dc->link_count; i++) {
-			/* Power up AND update implementation according to the
-			 * required signal (which may be different from the
-			 * default signal on connector).
-			 */
-			struct dc_link *link = dc->links[i];
-
-			if (link->link_enc->connector.id == CONNECTOR_ID_EDP)
-				dc->hwss.edp_power_control(link, true);
-
-			link->link_enc->funcs->hw_init(link->link_enc);
-
-			/* Check for enabled DIG to identify enabled display */
-			if (link->link_enc->funcs->is_dig_enabled &&
-				link->link_enc->funcs->is_dig_enabled(link->link_enc))
-				link->link_status.link_active = true;
-		}
-	}
 
 	for (i = 0; i < dc->res_pool->pipe_count; i++) {
 		struct timing_generator *tg = dc->res_pool->timing_generators[i];
 
 		if (tg->funcs->is_tg_enabled(tg))
 			tg->funcs->lock(tg);
-	}
-
-	/* Blank controller using driver code instead of
-	 * command table.
-	 */
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct timing_generator *tg = dc->res_pool->timing_generators[i];
 
+		/* Blank controller using driver code instead of
+		 * command table.
+		 */
 		if (tg->funcs->is_tg_enabled(tg)) {
 			tg->funcs->set_blank(tg, true);
 			hwss_wait_for_blank_complete(tg);
 		}
 	}
 
-	/* Reset all MPCC muxes */
 	dc->res_pool->mpc->funcs->mpc_init(dc->res_pool->mpc);
 
-	for (i = 0; i < dc->res_pool->timing_generator_count; i++) {
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
 		struct timing_generator *tg = dc->res_pool->timing_generators[i];
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
 		struct hubp *hubp = dc->res_pool->hubps[i];
 		struct dpp *dpp = dc->res_pool->dpps[i];
+		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
+
+		dpp->funcs->dpp_reset(dpp);
 
 		pipe_ctx->stream_res.tg = tg;
 		pipe_ctx->pipe_idx = i;
@@ -1064,18 +1001,9 @@ static void dcn10_init_hw(struct dc *dc)
 		pipe_ctx->stream_res.opp = dc->res_pool->opps[i];
 
 		hwss1_plane_atomic_disconnect(dc, pipe_ctx);
-	}
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct timing_generator *tg = dc->res_pool->timing_generators[i];
 
 		if (tg->funcs->is_tg_enabled(tg))
 			tg->funcs->unlock(tg);
-	}
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		struct timing_generator *tg = dc->res_pool->timing_generators[i];
-		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[i];
 
 		dcn10_disable_plane(dc, pipe_ctx);
 
@@ -1084,10 +1012,73 @@ static void dcn10_init_hw(struct dc *dc)
 
 		tg->funcs->tg_init(tg);
 	}
+}
+
+static void dcn10_init_hw(struct dc *dc)
+{
+	int i;
+	struct abm *abm = dc->res_pool->abm;
+	struct dmcu *dmcu = dc->res_pool->dmcu;
+	struct dce_hwseq *hws = dc->hwseq;
+	struct dc_bios *dcb = dc->ctx->dc_bios;
+
+	if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
+		REG_WRITE(REFCLK_CNTL, 0);
+		REG_UPDATE(DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_ENABLE, 1);
+		REG_WRITE(DIO_MEM_PWR_CTRL, 0);
 
-	/* end of FPGA. Below if real ASIC */
-	if (IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment))
+		if (!dc->debug.disable_clock_gate) {
+			/* enable all DCN clock gating */
+			REG_WRITE(DCCG_GATE_DISABLE_CNTL, 0);
+
+			REG_WRITE(DCCG_GATE_DISABLE_CNTL2, 0);
+
+			REG_UPDATE(DCFCLK_CNTL, DCFCLK_GATE_DIS, 0);
+		}
+
+		enable_power_gating_plane(dc->hwseq, true);
+
+		/* end of FPGA. Below if real ASIC */
 		return;
+	}
+
+	if (!dcb->funcs->is_accelerated_mode(dcb)) {
+		bool allow_self_fresh_force_enable =
+			hububu1_is_allow_self_refresh_enabled(
+						dc->res_pool->hubbub);
+
+		bios_golden_init(dc);
+
+		/* WA for making DF sleep when idle after resume from S0i3.
+		 * DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE is set to 1 by
+		 * command table, if DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE = 0
+		 * before calling command table and it changed to 1 after,
+		 * it should be set back to 0.
+		 */
+		if (allow_self_fresh_force_enable == false &&
+				hububu1_is_allow_self_refresh_enabled(dc->res_pool->hubbub))
+			hubbub1_allow_self_refresh_control(dc->res_pool->hubbub, true);
+
+		disable_vga(dc->hwseq);
+	}
+
+	for (i = 0; i < dc->link_count; i++) {
+		/* Power up AND update implementation according to the
+		 * required signal (which may be different from the
+		 * default signal on connector).
+		 */
+		struct dc_link *link = dc->links[i];
+
+		if (link->link_enc->connector.id == CONNECTOR_ID_EDP)
+			dc->hwss.edp_power_control(link, true);
+
+		link->link_enc->funcs->hw_init(link->link_enc);
+
+		/* Check for enabled DIG to identify enabled display */
+		if (link->link_enc->funcs->is_dig_enabled &&
+			link->link_enc->funcs->is_dig_enabled(link->link_enc))
+			link->link_status.link_active = true;
+	}
 
 	for (i = 0; i < dc->res_pool->audio_count; i++) {
 		struct audio *audio = dc->res_pool->audios[i];
@@ -1118,6 +1109,9 @@ static void dcn10_init_hw(struct dc *dc)
 	enable_power_gating_plane(dc->hwseq, true);
 
 	memset(&dc->res_pool->clk_mgr->clks, 0, sizeof(dc->res_pool->clk_mgr->clks));
+
+	if (dc->hwss.init_pipes)
+		dc->hwss.init_pipes(dc, dc->current_state);
 }
 
 static void reset_hw_ctx_wrap(
@@ -2718,6 +2712,7 @@ static void dcn10_set_cursor_sdr_white_level(struct pipe_ctx *pipe_ctx)
 static const struct hw_sequencer_funcs dcn10_funcs = {
 	.program_gamut_remap = program_gamut_remap,
 	.init_hw = dcn10_init_hw,
+	.init_pipes = dcn10_init_pipes,
 	.apply_ctx_to_hw = dce110_apply_ctx_to_hw,
 	.apply_ctx_for_surface = dcn10_apply_ctx_for_surface,
 	.update_plane_addr = dcn10_update_plane_addr,
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h
index d6a85f48b6d1..341b4810288c 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h
@@ -70,6 +70,8 @@ struct hw_sequencer_funcs {
 
 	void (*init_hw)(struct dc *dc);
 
+	void (*init_pipes)(struct dc *dc, struct dc_state *context);
+
 	enum dc_status (*apply_ctx_to_hw)(
 			struct dc *dc, struct dc_state *context);
 
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 35/35] drm/amd/display: add n_vid_mul and half pix_rate for odm

[Bhawanpreet Lakha]

From: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>

Dp needs half container rate to properly support odm

Change-Id: I5a3053c10f54461c5eefc99a44920abfb3d8900e
Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Nikola Cornij <Nikola.Cornij@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 .../gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c    | 4 ++--
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c          | 7 +++++--
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h          | 2 +-
 3 files changed, 8 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
index 392781bfb7a5..750d6ef8d38a 100644
--- a/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c
@@ -1032,8 +1032,7 @@ void dce110_unblank_stream(struct pipe_ctx *pipe_ctx,
 	struct dc_link *link = stream->link;
 
 	/* only 3 items below are used by unblank */
-	params.pixel_clk_khz =
-		pipe_ctx->stream->timing.pix_clk_100hz / 10;
+	params.pixel_clk_khz = pipe_ctx->stream->timing.pix_clk_100hz / 10;
 	params.link_settings.link_rate = link_settings->link_rate;
 
 	if (dc_is_dp_signal(pipe_ctx->stream->signal))
@@ -1043,6 +1042,7 @@ void dce110_unblank_stream(struct pipe_ctx *pipe_ctx,
 		link->dc->hwss.edp_backlight_control(link, true);
 	}
 }
+
 void dce110_blank_stream(struct pipe_ctx *pipe_ctx)
 {
 	struct dc_stream_state *stream = pipe_ctx->stream;
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
index 1dbd1d3999e6..2f78a84f0dcb 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
@@ -390,7 +390,7 @@ void optc1_program_timing(
 
 	h_div_2 = optc1_is_two_pixels_per_containter(&patched_crtc_timing);
 	REG_UPDATE(OTG_H_TIMING_CNTL,
-			OTG_H_TIMING_DIV_BY2, h_div_2);
+			OTG_H_TIMING_DIV_BY2, h_div_2 || optc1->comb_opp_id != 0xf);
 
 }
 
@@ -1531,10 +1531,13 @@ void dcn10_timing_generator_init(struct optc *optc1)
 	optc1->min_v_blank_interlace = 5;
 	optc1->min_h_sync_width = 8;
 	optc1->min_v_sync_width = 1;
+	optc1->comb_opp_id = 0xf;
 }
 
 bool optc1_is_two_pixels_per_containter(const struct dc_crtc_timing *timing)
 {
-	return timing->pixel_encoding == PIXEL_ENCODING_YCBCR420;
+	bool two_pix = timing->pixel_encoding == PIXEL_ENCODING_YCBCR420;
+
+	return two_pix;
 }
 
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
index 8a4e3e37e894..24452f11c598 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
@@ -435,7 +435,7 @@ struct optc {
 	const struct dcn_optc_shift *tg_shift;
 	const struct dcn_optc_mask *tg_mask;
 
-	enum controller_id controller_id;
+	int comb_opp_id;
 
 	uint32_t max_h_total;
 	uint32_t max_v_total;
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH 01/35] drm/amd/display: Use udelay when waiting between aux retries

[sylvain.bertrand-Re5JQEeQqe8AvxtiuMwx3w]

On Fri, Feb 01, 2019 at 10:28:13AM -0500, Bhawanpreet Lakha wrote:
> From: John Barberiz <John.Barberiz@amd.com>
> [How]
> msleep is inaccurate for small values. Used udelay
> instead for accuracy.

Hi,

Should it be the case for non-DC displayport code too?

regards,

-- 
Sylvain
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH 01/35] drm/amd/display: Use udelay when waiting between aux retries

[Wentland, Harry]

On 2019-02-01 12:31 p.m., sylvain.bertrand@gmail.com wrote:
> On Fri, Feb 01, 2019 at 10:28:13AM -0500, Bhawanpreet Lakha wrote:
>> From: John Barberiz <John.Barberiz@amd.com>
>> [How]
>> msleep is inaccurate for small values. Used udelay
>> instead for accuracy.
> 
> Hi,
> 
> Should it be the case for non-DC displayport code too?
> 

I don't think we're actually using this code on Linux. It's part of the shared codebase with Windows. Apparently there msleep would often sleep longer than 1ms.

drm_dp_dpcd_access already has a tighter bound on the sleep between retries (500-600 us), so I imagine we're fine there.

Harry

> regards,
> 
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH 01/35] drm/amd/display: Use udelay when waiting between aux retries

[sylvain.bertrand-Re5JQEeQqe8AvxtiuMwx3w]

On Fri, Feb 01, 2019 at 07:16:17PM +0000, Wentland, Harry wrote:
> On 2019-02-01 12:31 p.m., sylvain.bertrand@gmail.com wrote:
> > On Fri, Feb 01, 2019 at 10:28:13AM -0500, Bhawanpreet Lakha wrote:
> >> From: John Barberiz <John.Barberiz@amd.com>
> >> [How]
> >> msleep is inaccurate for small values. Used udelay
> >> instead for accuracy.
> > 
> > Hi,
> > 
> > Should it be the case for non-DC displayport code too?
> > 
> 
> I don't think we're actually using this code on Linux. It's part of the shared codebase with Windows. Apparently there msleep would often sleep longer than 1ms.
> 
> drm_dp_dpcd_access already has a tighter bound on the sleep between retries (500-600 us), so I imagine we're fine there.

Do you mean non-DC displayport related code is already using udelay instead of msleep on linux?

-- 
Sylvain
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH 01/35] drm/amd/display: Use udelay when waiting between aux retries

[sylvain.bertrand-Re5JQEeQqe8AvxtiuMwx3w]

On Fri, Feb 01, 2019 at 08:08:30PM +0000, sylvain.bertrand@gmail.com wrote:
> Do you mean non-DC displayport related code is already using udelay instead
> of msleep on linux?

I grep-ed amdgpu displayport atombios code, got udelay(400) only.
Did the same on drm *_dp_* files, got similar udelays with only a very slow
msleep (which seems justified there).

But I did not get any udelays with the range of 1ms for "retries". Maybe it
would be a good idea to a similar delay on non-DC displayport code? That would
give more time to "slow" displayort monitors to timeout on their side.

-- 
Sylvain
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH 01/35] drm/amd/display: Use udelay when waiting between aux retries

[Wentland, Harry]

On 2019-02-01 3:47 p.m., sylvain.bertrand@gmail.com wrote:
> On Fri, Feb 01, 2019 at 08:08:30PM +0000, sylvain.bertrand@gmail.com wrote:
>> Do you mean non-DC displayport related code is already using udelay instead
>> of msleep on linux?
> 
> I grep-ed amdgpu displayport atombios code, got udelay(400) only.
> Did the same on drm *_dp_* files, got similar udelays with only a very slow
> msleep (which seems justified there).
> 
> But I did not get any udelays with the range of 1ms for "retries". Maybe it
> would be a good idea to a similar delay on non-DC displayport code? That would
> give more time to "slow" displayort monitors to timeout on their side.
> 

DRM's AUX code uses usleep_range in drm_dp_dpcd_access.

Harry
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH 01/35] drm/amd/display: Use udelay when waiting between aux retries

[sylvain.bertrand-Re5JQEeQqe8AvxtiuMwx3w]

On Fri, Feb 01, 2019 at 09:20:56PM +0000, Wentland, Harry wrote:
> DRM's AUX code uses usleep_range in drm_dp_dpcd_access.

My bad, forgot about the usleep_range switch. That said AUX_RETRY_INTERVAL is
500 us, with a usleep_range top bound of 600 us.

Then, it would mean DC DP timeout retries would happen after ~1ms, and drm
~600us. Additionaly, it seems that the number of retries are 3 in drm code and
7 in DC code. 

I may be wrong, but it seems DC code is much more "insisting" on auxchannel
(edid retrieval) and much more forgiving on monitor ability to timeout in
time (~1ms).

If I did read the code the right way, it may be more reasonable to have similar
behavior in drm code than in DC code, right?

-- 
Sylvain
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH 05/35] drm/amd/display: Initialize stream_update to zero

[Nils Wallménius]

[-- Attachment #1.1: Type: text/plain, Size: 2361 bytes --]

Hi,

Recently several locations initing structs to zero have been changed to use
memset [1]. Might want to do that here as well.

[1] https://lists.freedesktop.org/archives/amd-gfx/2019-February/031116.html

Regards
Nils

Den fre 1 feb. 2019 16:29 skrev Bhawanpreet Lakha <Bhawanpreet.Lakha-5C7GfCeVMHo@public.gmane.org
>:

> From: Nicholas Kazlauskas <nicholas.kazlauskas-5C7GfCeVMHo@public.gmane.org>
>
> [Why]
> The stream_update struct is left unitialized but DC will access
> its fields. This usually results in global state validation occur
> during any atomic commit with state->allow_modeset = true.
>
> [How]
> Initialize the struct to zero for every stream we check.
>
> Change-Id: Ib7f435cc2c045fc43496bea35d08a14ce8193c33
> Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas-5C7GfCeVMHo@public.gmane.org>
> Reviewed-by: David Francis <David.Francis-5C7GfCeVMHo@public.gmane.org>
> Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha-5C7GfCeVMHo@public.gmane.org>
> ---
>  drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
> b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
> index d7e5391eb432..6550d00128dc 100644
> --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
> +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
> @@ -5868,7 +5868,6 @@ dm_determine_update_type_for_commit(struct dc *dc,
>
>         struct dc_surface_update *updates;
>         struct dc_plane_state *surface;
> -       struct dc_stream_update stream_update;
>         enum surface_update_type update_type = UPDATE_TYPE_FAST;
>
>         updates = kcalloc(MAX_SURFACES, sizeof(*updates), GFP_KERNEL);
> @@ -5882,6 +5881,8 @@ dm_determine_update_type_for_commit(struct dc *dc,
>         }
>
>         for_each_oldnew_crtc_in_state(state, crtc, old_crtc_state,
> new_crtc_state, i) {
> +               struct dc_stream_update stream_update = { 0 };
> +
>                 new_dm_crtc_state = to_dm_crtc_state(new_crtc_state);
>                 old_dm_crtc_state = to_dm_crtc_state(old_crtc_state);
>                 num_plane = 0;
> --
> 2.17.1
>
> _______________________________________________
> amd-gfx mailing list
> amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx
>

[-- Attachment #1.2: Type: text/html, Size: 3693 bytes --]

[-- Attachment #2: Type: text/plain, Size: 154 bytes --]

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH 15/35] drm/amd/display: make vline_config parameter const.

[Nils Wallménius]

[-- Attachment #1.1: Type: text/plain, Size: 5406 bytes --]

Hi,

It seems the commit title is a little confusing. The change is to now pass
by (const) reference rather than by value.

Regards
Nils

Den fre 1 feb. 2019 16:29 skrev Bhawanpreet Lakha <Bhawanpreet.Lakha-5C7GfCeVMHo@public.gmane.org
>:

> From: Yongqiang Sun <yongqiang.sun-5C7GfCeVMHo@public.gmane.org>
>
> Change-Id: Icfe018c7579ad2b3ef65195f578b8e44422d53f3
> Signed-off-by: Yongqiang Sun <yongqiang.sun-5C7GfCeVMHo@public.gmane.org>
> Reviewed-by: Aric Cyr <Aric.Cyr-5C7GfCeVMHo@public.gmane.org>
> Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha-5C7GfCeVMHo@public.gmane.org>
> ---
>  drivers/gpu/drm/amd/display/dc/core/dc.c                 | 4 ++--
>  drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c        | 8 ++++----
>  drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h        | 2 +-
>  drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h | 2 +-
>  4 files changed, 8 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c
> b/drivers/gpu/drm/amd/display/dc/core/dc.c
> index e0ac009f00ab..d8579b207300 100644
> --- a/drivers/gpu/drm/amd/display/dc/core/dc.c
> +++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
> @@ -1560,11 +1560,11 @@ static void commit_planes_do_stream_update(struct
> dc *dc,
>
>                         if (stream_update->vline0_config &&
> pipe_ctx->stream_res.tg->funcs->program_vline_interrupt)
>
> pipe_ctx->stream_res.tg->funcs->program_vline_interrupt(
> -                                       pipe_ctx->stream_res.tg, VLINE0,
> stream->vline0_config);
> +                                       pipe_ctx->stream_res.tg, VLINE0,
> &stream->vline0_config);
>
>                         if (stream_update->vline1_config &&
> pipe_ctx->stream_res.tg->funcs->program_vline_interrupt)
>
> pipe_ctx->stream_res.tg->funcs->program_vline_interrupt(
> -                                       pipe_ctx->stream_res.tg, VLINE1,
> stream->vline1_config);
> +                                       pipe_ctx->stream_res.tg, VLINE1,
> &stream->vline1_config);
>
>                         if ((stream_update->hdr_static_metadata &&
> !stream->use_dynamic_meta) ||
>                                         stream_update->vrr_infopacket ||
> diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
> b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
> index 1d4f9b48ed7d..cefa322df8a6 100644
> --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
> +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
> @@ -95,19 +95,19 @@ static void optc1_disable_stereo(struct
> timing_generator *optc)
>  void optc1_program_vline_interrupt(
>                 struct timing_generator *optc,
>                 enum vline_select vline,
> -               struct vline_config vline_config)
> +               const struct vline_config *vline_config)
>  {
>         struct optc *optc1 = DCN10TG_FROM_TG(optc);
>
>         switch (vline) {
>         case VLINE0:
>                 REG_SET_2(OTG_VERTICAL_INTERRUPT0_POSITION, 0,
> -                               OTG_VERTICAL_INTERRUPT0_LINE_START,
> vline_config.start_line,
> -                               OTG_VERTICAL_INTERRUPT0_LINE_END,
> vline_config.end_line);
> +                               OTG_VERTICAL_INTERRUPT0_LINE_START,
> vline_config->start_line,
> +                               OTG_VERTICAL_INTERRUPT0_LINE_END,
> vline_config->end_line);
>                 break;
>         case VLINE1:
>                 REG_SET(OTG_VERTICAL_INTERRUPT1_POSITION, 0,
> -
>  OTG_VERTICAL_INTERRUPT1_LINE_START, vline_config.start_line);
> +
>  OTG_VERTICAL_INTERRUPT1_LINE_START, vline_config->start_line);
>                 break;
>         default:
>                 break;
> diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
> b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
> index 8eb71c0160a7..b34c8a240598 100644
> --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
> +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
> @@ -485,7 +485,7 @@ void optc1_program_timing(
>
>  void optc1_program_vline_interrupt(struct timing_generator *optc,
>                 enum vline_select vline,
> -               struct vline_config vline_config);
> +               const struct vline_config *vline_config);
>
>  void optc1_program_global_sync(
>                 struct timing_generator *optc);
> diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
> b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
> index df64cf73ceb9..d22a406c19c0 100644
> --- a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
> +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
> @@ -151,7 +151,7 @@ struct timing_generator_funcs {
>                                                         bool use_vbios);
>         void (*program_vline_interrupt)(struct timing_generator *optc,
>                         enum vline_select vline,
> -                       struct vline_config vline_config);
> +                       const struct vline_config *vline_config);
>         bool (*enable_crtc)(struct timing_generator *tg);
>         bool (*disable_crtc)(struct timing_generator *tg);
>         bool (*is_counter_moving)(struct timing_generator *tg);
> --
> 2.17.1
>
> _______________________________________________
> amd-gfx mailing list
> amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org
> https://lists.freedesktop.org/mailman/listinfo/amd-gfx
>

[-- Attachment #1.2: Type: text/html, Size: 7716 bytes --]

[-- Attachment #2: Type: text/plain, Size: 154 bytes --]

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH 01/35] drm/amd/display: Use udelay when waiting between aux retries

[Wentland, Harry]

On 2019-02-01 5:55 p.m., sylvain.bertrand@gmail.com wrote:
> On Fri, Feb 01, 2019 at 09:20:56PM +0000, Wentland, Harry wrote:
>> DRM's AUX code uses usleep_range in drm_dp_dpcd_access.
> 
> My bad, forgot about the usleep_range switch. That said AUX_RETRY_INTERVAL is
> 500 us, with a usleep_range top bound of 600 us.
> 
> Then, it would mean DC DP timeout retries would happen after ~1ms, and drm
> ~600us. Additionaly, it seems that the number of retries are 3 in drm code and
> 7 in DC code. 
> 

DRM actually bumped this to 32 due to an issue with a Dell 4k display.

> I may be wrong, but it seems DC code is much more "insisting" on auxchannel
> (edid retrieval) and much more forgiving on monitor ability to timeout in
> time (~1ms).
> 
> If I did read the code the right way, it may be more reasonable to have similar
> behavior in drm code than in DC code, right?
> 

It depends. I wouldn't call one or the other more correct. I seem to remember that the DP spec is quite vague on these retries but I could be mistaken. Since our driver hasn't show any problems with the DRM code and I believe others (such as i915) also pass DP compliance without issues I wouldn't proactively change this.

Harry
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

[PATCH 15/35] drm/amd/display: pass vline_config parameter by reference.

[Bhawanpreet Lakha]

From: Yongqiang Sun <yongqiang.sun@amd.com>

Change-Id: Icfe018c7579ad2b3ef65195f578b8e44422d53f3
Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
---
 drivers/gpu/drm/amd/display/dc/core/dc.c                 | 4 ++--
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c        | 8 ++++----
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h        | 2 +-
 drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h | 2 +-
 4 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index e0ac009f00ab..d8579b207300 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -1560,11 +1560,11 @@ static void commit_planes_do_stream_update(struct dc *dc,
 
 			if (stream_update->vline0_config && pipe_ctx->stream_res.tg->funcs->program_vline_interrupt)
 				pipe_ctx->stream_res.tg->funcs->program_vline_interrupt(
-					pipe_ctx->stream_res.tg, VLINE0, stream->vline0_config);
+					pipe_ctx->stream_res.tg, VLINE0, &stream->vline0_config);
 
 			if (stream_update->vline1_config && pipe_ctx->stream_res.tg->funcs->program_vline_interrupt)
 				pipe_ctx->stream_res.tg->funcs->program_vline_interrupt(
-					pipe_ctx->stream_res.tg, VLINE1, stream->vline1_config);
+					pipe_ctx->stream_res.tg, VLINE1, &stream->vline1_config);
 
 			if ((stream_update->hdr_static_metadata && !stream->use_dynamic_meta) ||
 					stream_update->vrr_infopacket ||
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
index 1d4f9b48ed7d..cefa322df8a6 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
@@ -95,19 +95,19 @@ static void optc1_disable_stereo(struct timing_generator *optc)
 void optc1_program_vline_interrupt(
 		struct timing_generator *optc,
 		enum vline_select vline,
-		struct vline_config vline_config)
+		const struct vline_config *vline_config)
 {
 	struct optc *optc1 = DCN10TG_FROM_TG(optc);
 
 	switch (vline) {
 	case VLINE0:
 		REG_SET_2(OTG_VERTICAL_INTERRUPT0_POSITION, 0,
-				OTG_VERTICAL_INTERRUPT0_LINE_START, vline_config.start_line,
-				OTG_VERTICAL_INTERRUPT0_LINE_END, vline_config.end_line);
+				OTG_VERTICAL_INTERRUPT0_LINE_START, vline_config->start_line,
+				OTG_VERTICAL_INTERRUPT0_LINE_END, vline_config->end_line);
 		break;
 	case VLINE1:
 		REG_SET(OTG_VERTICAL_INTERRUPT1_POSITION, 0,
-					OTG_VERTICAL_INTERRUPT1_LINE_START, vline_config.start_line);
+					OTG_VERTICAL_INTERRUPT1_LINE_START, vline_config->start_line);
 		break;
 	default:
 		break;
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
index 8eb71c0160a7..b34c8a240598 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
@@ -485,7 +485,7 @@ void optc1_program_timing(
 
 void optc1_program_vline_interrupt(struct timing_generator *optc,
 		enum vline_select vline,
-		struct vline_config vline_config);
+		const struct vline_config *vline_config);
 
 void optc1_program_global_sync(
 		struct timing_generator *optc);
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
index df64cf73ceb9..d22a406c19c0 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
@@ -151,7 +151,7 @@ struct timing_generator_funcs {
 							bool use_vbios);
 	void (*program_vline_interrupt)(struct timing_generator *optc,
 			enum vline_select vline,
-			struct vline_config vline_config);
+			const struct vline_config *vline_config);
 	bool (*enable_crtc)(struct timing_generator *tg);
 	bool (*disable_crtc)(struct timing_generator *tg);
 	bool (*is_counter_moving)(struct timing_generator *tg);
-- 
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Fw: [PATCH 15/35] drm/amd/display: pass vline_config parameter by reference.

[Lakha, Bhawanpreet]

[-- Attachment #1.1: Type: text/plain, Size: 6032 bytes --]




________________________________
From: amd-gfx <amd-gfx-bounces-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org> on behalf of Bhawanpreet Lakha <Bhawanpreet.Lakha-5C7GfCeVMHo@public.gmane.org>
Sent: February 4, 2019 10:45 AM
To: amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org
Cc: Sun, Yongqiang
Subject: [PATCH 15/35] drm/amd/display: pass vline_config parameter by reference.

From: Yongqiang Sun <yongqiang.sun-5C7GfCeVMHo@public.gmane.org>

Change-Id: Icfe018c7579ad2b3ef65195f578b8e44422d53f3
Signed-off-by: Yongqiang Sun <yongqiang.sun-5C7GfCeVMHo@public.gmane.org>
Reviewed-by: Aric Cyr <Aric.Cyr-5C7GfCeVMHo@public.gmane.org>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha-5C7GfCeVMHo@public.gmane.org>
---
 drivers/gpu/drm/amd/display/dc/core/dc.c                 | 4 ++--
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c        | 8 ++++----
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h        | 2 +-
 drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h | 2 +-
 4 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index e0ac009f00ab..d8579b207300 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -1560,11 +1560,11 @@ static void commit_planes_do_stream_update(struct dc *dc,

                         if (stream_update->vline0_config && pipe_ctx->stream_res.tg->funcs->program_vline_interrupt)
                                 pipe_ctx->stream_res.tg->funcs->program_vline_interrupt(
-                                       pipe_ctx->stream_res.tg, VLINE0, stream->vline0_config);
+                                       pipe_ctx->stream_res.tg, VLINE0, &stream->vline0_config);

                         if (stream_update->vline1_config && pipe_ctx->stream_res.tg->funcs->program_vline_interrupt)
                                 pipe_ctx->stream_res.tg->funcs->program_vline_interrupt(
-                                       pipe_ctx->stream_res.tg, VLINE1, stream->vline1_config);
+                                       pipe_ctx->stream_res.tg, VLINE1, &stream->vline1_config);

                         if ((stream_update->hdr_static_metadata && !stream->use_dynamic_meta) ||
                                         stream_update->vrr_infopacket ||
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
index 1d4f9b48ed7d..cefa322df8a6 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.c
@@ -95,19 +95,19 @@ static void optc1_disable_stereo(struct timing_generator *optc)
 void optc1_program_vline_interrupt(
                 struct timing_generator *optc,
                 enum vline_select vline,
-               struct vline_config vline_config)
+               const struct vline_config *vline_config)
 {
         struct optc *optc1 = DCN10TG_FROM_TG(optc);

         switch (vline) {
         case VLINE0:
                 REG_SET_2(OTG_VERTICAL_INTERRUPT0_POSITION, 0,
-                               OTG_VERTICAL_INTERRUPT0_LINE_START, vline_config.start_line,
-                               OTG_VERTICAL_INTERRUPT0_LINE_END, vline_config.end_line);
+                               OTG_VERTICAL_INTERRUPT0_LINE_START, vline_config->start_line,
+                               OTG_VERTICAL_INTERRUPT0_LINE_END, vline_config->end_line);
                 break;
         case VLINE1:
                 REG_SET(OTG_VERTICAL_INTERRUPT1_POSITION, 0,
-                                       OTG_VERTICAL_INTERRUPT1_LINE_START, vline_config.start_line);
+                                       OTG_VERTICAL_INTERRUPT1_LINE_START, vline_config->start_line);
                 break;
         default:
                 break;
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
index 8eb71c0160a7..b34c8a240598 100644
--- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
+++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h
@@ -485,7 +485,7 @@ void optc1_program_timing(

 void optc1_program_vline_interrupt(struct timing_generator *optc,
                 enum vline_select vline,
-               struct vline_config vline_config);
+               const struct vline_config *vline_config);

 void optc1_program_global_sync(
                 struct timing_generator *optc);
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
index df64cf73ceb9..d22a406c19c0 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h
@@ -151,7 +151,7 @@ struct timing_generator_funcs {
                                                         bool use_vbios);
         void (*program_vline_interrupt)(struct timing_generator *optc,
                         enum vline_select vline,
-                       struct vline_config vline_config);
+                       const struct vline_config *vline_config);
         bool (*enable_crtc)(struct timing_generator *tg);
         bool (*disable_crtc)(struct timing_generator *tg);
         bool (*is_counter_moving)(struct timing_generator *tg);
--
2.17.1

_______________________________________________
amd-gfx mailing list
amd-gfx-PD4FTy7X32lNgt0PjOBp9y5qC8QIuHrW@public.gmane.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
amd-gfx Info Page - freedesktop.org<https://lists.freedesktop.org/mailman/listinfo/amd-gfx>
lists.freedesktop.org
To see the collection of prior postings to the list, visit the amd-gfx Archives.. Using amd-gfx: To post a message to all the list members, send email to amd-gfx-PD4FTy7X32lNgt0PjOBp9xlNPtJONSTn@public.gmane.org You can subscribe to the list, or change your existing subscription, in the sections below.



[-- Attachment #1.2: Type: text/html, Size: 14288 bytes --]

[-- Attachment #2: Type: text/plain, Size: 154 bytes --]

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH 01/35] drm/amd/display: Use udelay when waiting between aux retries

[sylvain.bertrand-Re5JQEeQqe8AvxtiuMwx3w]

On Mon, Feb 04, 2019 at 03:43:36PM +0000, Wentland, Harry wrote:
> DRM actually bumped this to 32 due to an issue with a Dell 4k display.

As I feared there is a retry counter higher in the code. My bad.

> It depends. I wouldn't call one or the other more correct. I seem to remember
> that the DP spec is quite vague on these retries but I could be mistaken.
> Since our driver hasn't show any problems with the DRM code and I believe
> others (such as i915) also pass DP compliance without issues I wouldn't
> proactively change this.

I reacted to this matter since my DP monitor had troubles getting out of DP off
state (iiyama) with logs which did seem to point towards a failure of getting
the EDID. It happens rarely, and I could not figure out a context to reproduce
for sure. Could be gone for good with the latest fixes.

I guess fixes for popular and defective hardware (for instance that dell 4k)
from a compliance stand point have to go in the code unfortunately.

regards,

-- 
Sylvain
_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx