Skip to contentSkip to conversation

DC Patches Oct 10, 2017

PATCH In review Harry Wentland 2 participants 106 messages Oct 10, 2017 – Oct 11, 2017 3 messages not in the archive

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

cover DC Patches Oct 10, 2017

It's been a while, which means there's about a month of stuff here. Been busy with getting stuff prepped for the DC pull request, XDC, and a bunch of other things.

For major fixes we have:
* Hang on hotplug fix
drm/amd/display: fix multi-display on CZ
* Hotplug not working after S3 drm/amd/display: Fix hotplug after s3 resume. drm/amd/display: Fix irq enable/disable on resume.

Other than that we have
* Bunch of work from Raven enablement, including
* HW object refactor
* Updated DML code (gcc-parseable HW gospel)
* Corrected audio assignment
* An FBC config option (for CZ/ST)

Andrew Jiang (5):
drm/amd/display: Force always scale chroma, set always_scaled to false drm/amd/display: power_down_Hw need signal type to turnoff backlight drm/amd/display: Move power control from link encoder to hwsequencer drm/amd/display: Add missing newlines in pstate wait debug messages drm/amd/display: Fix up plane_states add/remove logic

Bhawanpreet Lakha (5):
drm/amd/display: Refactor dc_update_planes_and_stream. drm/amd/display: fix re-enabling stutter for raven drm/amd/display: increase pstate wait timeout drm/amd/display: Fix underscan not using proper scaling drm/amd/display: change when to set tap_hardcode_coeff

Charlene Liu (9):
drm/amd/display: make sure BL off to mainlink off has enough time drm/amd/display: add hubp/dpp pg debug key drm/amd/display: temp disable DCC on high res. drm/amd/display: soc_bound_box -update DML based on HW. drm/amd/display: audio dynamic resource acquired related drm/amd/display: arbitration find the matching dig-az first. drm/amd/display: Add stereo_support debug flag drm/amd/display: Add debug flag for VSR support drm/amd/display: reject 3D mode based on static debug flag.

Dmytro Laktyushkin (8):
drm/amd/display: Restructuring and cleaning up DML drm/amd/display: add mode support check to dml vba code drm/amd/display: extract global sync params from vba drm/amd/display: change dml vba cursor count define to correct one drm/amd/display: block video planes >4k on dcn10 drm/amd/display: add max_video_width cap to dc drm/amd/display: add performance trace macro to dc drm/amd/display: Set addressable region as active + border

Eric Bernstein (6):
drm/amd/display: update mpc add/remove functions drm/amd/display: update blending mode and set output denorm drm/amd/display: clean up dcn10 dpp after HW review drm/amd/display: group DCN watermark registers drm/amd/display: Add DPP capabilities drm/amd/display: Expose some MPC functions for reuse

Eric Yang (6):
drm/amd/display: fix usb-c dongle regression drm/amd/display: add back removed hack for mpcc add drm/amd/display: Use active + border for bw validation drm/amd/display: Add override for reporting wm ranges drm/amd/display: Handle case when stream not found in set_dpms drm/amd/display: fix null dereference in reset hw ctx

Harry Wentland (2):
drm/amd/display: Don't report fake sink as connected drm/amd/display: Ignore wrong tautological warning

Hersen Wu (4):
drm/amd/display: Driver message to SMU to indicate display off drm/amd/display: screen flickers when connected to ext monitor in
clone
drm/amd/display: add dpms state to DC
drm/amd/display: add dpms state to DC follow up

Jerry Zuo (8):
drm/amd/display: Exclude MST from fake sink drm/amd/display: Fix NULL pointer on MST chained mode drm/amd/display: Fix MST daisy chain SST not light up drm/amd/display: Fix ref_count messed up issue drm/amd/display: Do not access fbdev helper members drm/amd/display: Remove drm_modeset_lock in MST code drm/amd/display: Fix a logic defect in cursor move drm/amd/display: Remove a false-positive error message

Ken Chalmers (1):
drm/amd/display: fix ASSERT() caused by missing registers.

Leo (Sunpeng) Li (4):
drm/amd/display: Match actual state during S3 resume. drm/amd/display: Do not set bandwidth on page flips. drm/amd/display: Report stream as different on scaling change drm/amd/display: ASSERT on dc_create_state failing.

Martin Tsai (1):
drm/amd/display: To prevent detecting new sink from spurious HPD

Roman Li (4):
drm/amd/display: Add FBC on/off to front-end programming drm/amd/display: fix multi-display on CZ drm/amd/display: Fix hotplug after s3 resume. drm/amd/display: Fix irq enable/disable on resume.

ShihChen Chen (1):
drm/amd/display: make tile changing run at ISR

Shirish S (1):
drm/amd/display: make FBC configurable option

SivapiriyanKumarasamy (2):
drm/amd/display: Program stream's csc matrix instead of using default drm/amd/display: Pull transfer func defines into DC

Tony Cheng (10):
drm/amd/display: version 3.1.02
drm/amd/display: enable optional pipe split for single display drm/amd/display: add option to disable DCC for DCC 128b request drm/amd/display: option to maximize lb usage drm/amd/display: fix bug in force_single_disp_pipe_split drm/amd/display: dal 3.1.03
drm/amd/display: align DCLK to voltage level drm/amd/display: wrong VM setting cause invalid DF request from DCN drm/amd/display: dal 3.1.04
drm/amd/display: dal 3.1.05

Vitaly Prosyak (1):
drm/amd/display: [RV] bug in cm programming

Wenjing Liu (3):
drm/amd/display: set cp25201 to use TPS4 drm/amd/display: update link type to mst before topology discovery drm/amd/display: disconnect on HPD low even if edid is the same

Yongqiang Sun (9):
drm/amd/display: Fixed validation return wrong result. drm/amd/display: Fixed incorrect return value for validaton drm/amd/display: Early return when turn off a plane. drm/amd/display: Only reset top pipe back end. drm/amd/display: Enable double buffer as per vertical interrupt
enabled.
drm/amd/display: Program front end first when set mode. drm/amd/display: Fixed extend to second screen mode hang drm/amd/display: wa to program split pipedegamma regs after mode set. drm/amd/display: Fixed switching mode half screen gamma incorrect.

Yue Hin Lau (13):
drm/amd/display: Refactor dcn10_dpp_cm_helper to dcn10_cm_common drm/amd/display: fixing register includes drm/amd/display: removing remaining register definitions work around drm/amd/display: moving backlight registers to hwsequencer drm/amd/display: edp backlight regression fix drm/amd/display: moving cursor functions from ipp to mem_input drm/amd/display: Expose some mem_input functions for reuse drm/amd/display: renaming mem input to hubp drm/amd/display: Making hubp1_program_surface_config public drm/amd/display: rename struct mem_input to hubp for dcn drm/amd/display: renaming filename for hubp drm/amd/display: add null check for cursor functions drm/amd/display: rename transform to dpp for dcn

 drivers/gpu/drm/amd/display/Kconfig                |   10 +
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c  |  110 +-
 .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c  |    6 +-
 .../amd/display/amdgpu_dm/amdgpu_dm_mst_types.c    |  135 +-
 .../drm/amd/display/amdgpu_dm/amdgpu_dm_services.c |    6 +
 drivers/gpu/drm/amd/display/dc/basics/logger.c     |   78 +-
 drivers/gpu/drm/amd/display/dc/basics/logger.h     |   37 -
 drivers/gpu/drm/amd/display/dc/calcs/Makefile      |    2 +-
 .../gpu/drm/amd/display/dc/calcs/dcn_calc_math.c   |   16 +
 drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c   |  224 +-
 drivers/gpu/drm/amd/display/dc/core/dc.c           |  442 +-
 drivers/gpu/drm/amd/display/dc/core/dc_link.c      |   78 +-
 drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c   |   16 +-
 drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c |   17 +-
 drivers/gpu/drm/amd/display/dc/core/dc_resource.c  |  108 +-
 drivers/gpu/drm/amd/display/dc/core/dc_stream.c    |   63 +-
 drivers/gpu/drm/amd/display/dc/core/dc_surface.c   |    1 +
 drivers/gpu/drm/amd/display/dc/dc.h                |  135 +-
 drivers/gpu/drm/amd/display/dc/dc_hw_types.h       |    2 +
 drivers/gpu/drm/amd/display/dc/dc_types.h          |    3 +-
 drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h     |  175 +-
 drivers/gpu/drm/amd/display/dc/dce/dce_ipp.c       |    2 +
 .../gpu/drm/amd/display/dc/dce/dce_link_encoder.c  |  253 +-
 .../gpu/drm/amd/display/dc/dce/dce_link_encoder.h  |   20 +-
 drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c |    3 -
 .../drm/amd/display/dc/dce100/dce100_resource.c    |    2 +-
 .../drm/amd/display/dc/dce100/dce100_resource.h    |    2 +-
 .../drm/amd/display/dc/dce110/dce110_compressor.c  |    2 +-
 .../amd/display/dc/dce110/dce110_hw_sequencer.c    |  316 +-
 .../amd/display/dc/dce110/dce110_hw_sequencer.h    |   11 +-
 .../drm/amd/display/dc/dce110/dce110_resource.c    |   33 +-
 drivers/gpu/drm/amd/display/dc/dcn10/Makefile      |    4 +-
 .../{dcn10_dpp_cm_helper.c => dcn10_cm_common.c}   |    2 +-
 .../{dcn10_dpp_cm_helper.h => dcn10_cm_common.h}   |    4 +-
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c   |  199 +-
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.h   |  127 +-
 .../gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm.c    |  263 +-
 .../gpu/drm/amd/display/dc/dcn10/dcn10_dpp_dscl.c  |  190 +-
 .../dc/dcn10/{dcn10_mem_input.c => dcn10_hubp.c}   |  344 +-
 .../dc/dcn10/{dcn10_mem_input.h => dcn10_hubp.h}   |  139 +-
 .../drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c  |  359 +-
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_ipp.c   |  189 -
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c   |  215 +-
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.h   |   45 +-
 .../gpu/drm/amd/display/dc/dcn10/dcn10_resource.c  |   99 +-
 .../amd/display/dc/dcn10/dcn10_timing_generator.c  |   15 +
 .../amd/display/dc/dcn10/dcn10_timing_generator.h  |   15 +-
 drivers/gpu/drm/amd/display/dc/dm_services.h       |    2 +
 drivers/gpu/drm/amd/display/dc/dml/Makefile        |    8 +-
 drivers/gpu/drm/amd/display/dc/dml/dc_features.h   |    2 +
 .../drm/amd/display/dc/dml/display_mode_enums.h    |   56 +-
 .../gpu/drm/amd/display/dc/dml/display_mode_lib.c  |    7 +-
 .../gpu/drm/amd/display/dc/dml/display_mode_lib.h  |    8 +-
 .../drm/amd/display/dc/dml/display_mode_structs.h  |  903 +--
 .../drm/amd/display/dc/dml/display_mode_support.c  | 2326 --------
 .../drm/amd/display/dc/dml/display_mode_support.h  |  199 -
 .../gpu/drm/amd/display/dc/dml/display_mode_vba.c  | 6104 ++++++++++++++++++++
 .../gpu/drm/amd/display/dc/dml/display_mode_vba.h  |  581 ++
 .../drm/amd/display/dc/dml/display_pipe_clocks.c   |  377 +-
 .../drm/amd/display/dc/dml/display_pipe_clocks.h   |    8 +-
 .../drm/amd/display/dc/dml/display_rq_dlg_calc.c   | 2457 ++++----
 .../drm/amd/display/dc/dml/display_rq_dlg_calc.h   |  151 +-
 .../amd/display/dc/dml/display_rq_dlg_helpers.c    |  482 +-
 .../amd/display/dc/dml/display_rq_dlg_helpers.h    |   41 +-
 .../gpu/drm/amd/display/dc/dml/display_watermark.c | 1281 ----
 .../gpu/drm/amd/display/dc/dml/display_watermark.h |   98 -
 .../amd/display/dc/dml/dml1_display_rq_dlg_calc.c  | 1903 ++++++
 .../amd/display/dc/dml/dml1_display_rq_dlg_calc.h  |   67 +
 .../gpu/drm/amd/display/dc/dml/dml_common_defs.c   |   24 +-
 .../gpu/drm/amd/display/dc/dml/dml_common_defs.h   |   11 +-
 .../gpu/drm/amd/display/dc/dml/soc_bounding_box.c  |   45 +-
 .../gpu/drm/amd/display/dc/dml/soc_bounding_box.h  |    9 +-
 drivers/gpu/drm/amd/display/dc/inc/core_types.h    |   16 +-
 drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h        |  134 +
 drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h       |  105 +
 .../gpu/drm/amd/display/dc/inc/hw/link_encoder.h   |    6 +-
 drivers/gpu/drm/amd/display/dc/inc/hw/mem_input.h  |   11 +-
 drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h        |   15 +-
 drivers/gpu/drm/amd/display/dc/inc/hw/transform.h  |   57 +
 drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h  |   12 +-
 drivers/gpu/drm/amd/display/dc/inc/link_hwss.h     |    4 +
 drivers/gpu/drm/amd/display/dc/inc/resource.h      |    5 +
 .../amd/display/dc/virtual/virtual_link_encoder.c  |   13 +-
 .../gpu/drm/amd/display/include/logger_interface.h |   28 +
 drivers/gpu/drm/amd/display/include/logger_types.h |   36 +-
 85 files changed, 13615 insertions(+), 8494 deletions(-)
 rename drivers/gpu/drm/amd/display/dc/dcn10/{dcn10_dpp_cm_helper.c => dcn10_cm_common.c} (99%)
 rename drivers/gpu/drm/amd/display/dc/dcn10/{dcn10_dpp_cm_helper.h => dcn10_cm_common.h} (97%)
 rename drivers/gpu/drm/amd/display/dc/dcn10/{dcn10_mem_input.c => dcn10_hubp.c} (73%)
 rename drivers/gpu/drm/amd/display/dc/dcn10/{dcn10_mem_input.h => dcn10_hubp.h} (85%)
 delete mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_mode_support.c
 delete mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_mode_support.h
 create mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c
 create mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h
 delete mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_watermark.c
 delete mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_watermark.h
 create mode 100644 drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c
 create mode 100644 drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.h
 create mode 100644 drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h
 create mode 100644 drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): 20171010224112.3023-1-harry.wentland-5C7GfCeVMHo@public.gmane.org

#2 lore raw

PATCH 1/103 drm/amd/display: Driver message to SMU to indicate display off

From: Hersen Wu <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/core/dc.c           |  2 +-
 .../amd/display/dc/dce110/dce110_hw_sequencer.c    |  3 +-
 .../drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c  | 38 ++++++++++++++--------
 drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h  |  2 +-
 4 files changed, 28 insertions(+), 17 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +1 -1
index f41f15faf019..47f16a4b9f27 100644
@@ -763,7 +763,7 @@ static bool dc_commit_state_no_check(struct dc *dc, struct dc_state *context)
 	if (!dcb->funcs->is_accelerated_mode(dcb))
 		dc->hwss.enable_accelerated_mode(dc);
 
-	dc->hwss.ready_shared_resources(dc);
+	dc->hwss.ready_shared_resources(dc, context);
 
 	for (i = 0; i < dc->res_pool->pipe_count; i++) {
 		pipe = &context->res_ctx.pipe_ctx[i];
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +2 -1
index de154329b049..b3448a41718e 100644
@@ -1889,6 +1889,7 @@ enum dc_status dce110_apply_ctx_to_hw(
 			return status;
 	}
 
+	/* pplib is notified if disp_num changed */
 	dc->hwss.set_bandwidth(dc, context, true);
 
 	/* to save power */
@@ -2683,7 +2684,7 @@ static void program_csc_matrix(struct pipe_ctx *pipe_ctx,
 	}
 }
 
-static void ready_shared_resources(struct dc *dc) {}
+static void ready_shared_resources(struct dc *dc, struct dc_state *context) {}
 
 static void optimize_shared_resources(struct dc *dc) {}
 
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +24 -14
index d3fee15aeb79..e60be00abcf2 100644
@@ -802,18 +802,14 @@ static void undo_DEGVIDCN10_253_wa(struct dc *dc)
 			IP_REQUEST_EN, 0);
 }
 
-static void ready_shared_resources(struct dc *dc)
-{
-	if (dc->current_state->stream_count == 0 &&
-			!dc->debug.disable_stutter)
-		undo_DEGVIDCN10_253_wa(dc);
-}
-
 static void apply_DEGVIDCN10_253_wa(struct dc *dc)
 {
 	struct dce_hwseq *hws = dc->hwseq;
 	struct mem_input *mi = dc->res_pool->mis[0];
 
+	if (dc->debug.disable_stutter)
+		return;
+
 	REG_SET(DC_IP_REQUEST_CNTL, 0,
 			IP_REQUEST_EN, 1);
 
@@ -824,13 +820,6 @@ static void apply_DEGVIDCN10_253_wa(struct dc *dc)
 	mi->funcs->set_hubp_blank_en(mi, false);
 }
 
-static void optimize_shared_resources(struct dc *dc)
-{
-	if (dc->current_state->stream_count == 0 &&
-			!dc->debug.disable_stutter)
-		apply_DEGVIDCN10_253_wa(dc);
-}
-
 static void bios_golden_init(struct dc *dc)
 {
 	struct dc_bios *bp = dc->ctx->dc_bios;
@@ -2445,6 +2434,27 @@ static void dcn10_pplib_apply_display_requirements(
 	dc->prev_display_config = *pp_display_cfg;
 }
 
+static void optimize_shared_resources(struct dc *dc)
+{
+	if (dc->current_state->stream_count == 0) {
+		apply_DEGVIDCN10_253_wa(dc);
+		/* S0i2 message */
+		dcn10_pplib_apply_display_requirements(dc, dc->current_state);
+	}
+}
+
+static void ready_shared_resources(struct dc *dc, struct dc_state *context)
+{
+	if (dc->current_state->stream_count == 0 &&
+			!dc->debug.disable_stutter)
+		undo_DEGVIDCN10_253_wa(dc);
+
+	/* S0i2 message */
+	if (dc->current_state->stream_count == 0 &&
+			context->stream_count != 0)
+		dcn10_pplib_apply_display_requirements(dc, context);
+}
+
 static void dcn10_apply_ctx_for_surface(
 		struct dc *dc,
 		const struct dc_stream_state *stream,
drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h +1 -1
index aae7629b1c08..1fa2edc6cfb1 100644
@@ -174,7 +174,7 @@ struct hw_sequencer_funcs {
 			struct resource_pool *res_pool,
 			struct pipe_ctx *pipe_ctx);
 
-	void (*ready_shared_resources)(struct dc *dc);
+	void (*ready_shared_resources)(struct dc *dc, struct dc_state *context);
 	void (*optimize_shared_resources)(struct dc *dc);
 };
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#3 lore raw

PATCH 2/103 drm/amd/display: Force always scale chroma, set always_scaled to false

In reply to a message not in the archive

From: Andrew Jiang <[email protected]>

This allows us to not always have scaling on, which causes issues with validation and causes the text to blur slightly.

---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c      | 9 +++++----
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c | 4 ----
 2 files changed, 5 insertions(+), 8 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c +5 -4
index 9d9604f05095..f9e43e68fc0c 100644
@@ -159,10 +159,11 @@ bool dpp_get_optimal_number_of_taps(
 			scl_data->taps.h_taps = 1;
 		if (IDENTITY_RATIO(scl_data->ratios.vert))
 			scl_data->taps.v_taps = 1;
-		if (IDENTITY_RATIO(scl_data->ratios.horz_c))
-			scl_data->taps.h_taps_c = 1;
-		if (IDENTITY_RATIO(scl_data->ratios.vert_c))
-			scl_data->taps.v_taps_c = 1;
+		/*
+		 * Spreadsheet doesn't handle taps_c is one properly,
+		 * need to force Chroma to always be scaled to pass
+		 * bandwidth validation.
+		 */
 	}
 
 	return true;
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +0 -4
index dc878f77837b..4f124dc303dc 100644
@@ -420,10 +420,6 @@ static const struct dc_debug debug_defaults_drv = {
 		.force_abm_enable = false,
 		.timing_trace = false,
 		.clock_trace = true,
-		/* spread sheet doesn't handle taps_c is one properly,
-		 * need to enable scaler for video surface to pass
-		 * bandwidth validation.*/
-		.always_scale = true,
 		.disable_pplib_clock_request = true,
 		.disable_pplib_wm_range = false,
 #if defined(CONFIG_DRM_AMD_DC_DCN1_0)
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#4 lore raw

PATCH 3/103 drm/amd/display: fix usb-c dongle regression

In reply to a message not in the archive

From: Eric Yang <[email protected]>

A previous change moved link_disconnect_sink after dpcd read, this causes the sink count to be overwritten which in turn causes another detection to be triggered in the short pulse handler, which brings down the display.

---
 drivers/gpu/drm/amd/display/dc/core/dc_link.c | 2 --
 1 file changed, 2 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +0 -2
index ee23b609d7dd..d5eaebeaeb89 100644
@@ -652,8 +652,6 @@ bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason)
 		if (link->local_sink)	{
 			sink = link->local_sink;
 		} else {
-			link_disconnect_sink(link);
-
 			sink_init_data.link = link;
 			sink_init_data.sink_signal = sink_caps.signal;
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#5 lore raw

PATCH 4/103 drm/amd/display: Add FBC on/off to front-end programming

In reply to a message not in the archive

From: Roman Li <[email protected]>

- Fixing text mode for cases when VT-switch doesn't result in timing change

---
 drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c | 9 +++++++++
 1 file changed, 9 insertions(+)
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +9 -0
index b3448a41718e..590b2ad37105 100644
@@ -2554,6 +2554,15 @@ static void dce110_program_front_end_for_pipe(
 
 	program_scaler(dc, pipe_ctx);
 
+#ifdef ENABLE_FBC
+	if (dc->fbc_compressor && old_pipe->stream) {
+		if (plane_state->tiling_info.gfx8.array_mode == DC_ARRAY_LINEAR_GENERAL)
+			dc->fbc_compressor->funcs->disable_fbc(dc->fbc_compressor);
+		else
+			enable_fbc(dc, dc->current_state);
+	}
+#endif
+
 	mi->funcs->mem_input_program_surface_config(
 			mi,
 			plane_state->format,
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#6 lore raw

PATCH 5/103 drm/amd/display: To prevent detecting new sink from spurious HPD

In reply to a message not in the archive

From: Martin Tsai <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/core/dc_link.c | 33 ++++++++++++++++-----------
 drivers/gpu/drm/amd/display/dc/dc_types.h     |  1 +
 2 files changed, 21 insertions(+), 13 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +20 -13
index d5eaebeaeb89..c7751a31081a 100644
@@ -649,27 +649,34 @@ bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason)
 		sink_init_data.link = link;
 		sink_init_data.sink_signal = sink_caps.signal;
 
-		if (link->local_sink)	{
-			sink = link->local_sink;
-		} else {
-			sink_init_data.link = link;
-			sink_init_data.sink_signal = sink_caps.signal;
+		sink = dc_sink_create(&sink_init_data);
+		if (!sink) {
+			DC_ERROR("Failed to create sink!\n");
+			return false;
+		}
 
-			sink = dc_sink_create(&sink_init_data);
-			if (!sink) {
-				DC_ERROR("Failed to create sink!\n");
-				return false;
+		if (link->local_sink) {
+			edid_status = dm_helpers_read_local_edid(
+				link->ctx,
+				link,
+				sink);
+
+			if (edid_status == EDID_OK) {
+				// Edid is not the same, to update the local sink with new sink.
+				sink->dongle_max_pix_clk = sink_caps.max_hdmi_pixel_clock;
+				sink->converter_disable_audio = converter_disable_audio;
+				link->local_sink = sink;
 			}
+		} else {
 			sink->dongle_max_pix_clk = sink_caps.max_hdmi_pixel_clock;
 			sink->converter_disable_audio = converter_disable_audio;
-
 			link->local_sink = sink;
-		}
 
-		edid_status = dm_helpers_read_local_edid(
+			edid_status = dm_helpers_read_local_edid(
 				link->ctx,
 				link,
 				sink);
+		}
 
 		switch (edid_status) {
 		case EDID_BAD_CHECKSUM:
@@ -762,7 +769,7 @@ bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason)
 						link->ctx,
 						link,
 						sink);
-			if (edid_status != EDID_OK) {
+			if (edid_status != EDID_OK && edid_status != EDID_THE_SAME) {
 				link_disconnect_sink(link);
 				link->type = dc_connection_none;
 				sink_caps.signal = SIGNAL_TYPE_NONE;
drivers/gpu/drm/amd/display/dc/dc_types.h +1 -0
index 6b891fde400c..4bd74fc6c782 100644
@@ -151,6 +151,7 @@ enum dc_edid_status {
 	EDID_BAD_INPUT,
 	EDID_NO_RESPONSE,
 	EDID_BAD_CHECKSUM,
+	EDID_THE_SAME,
 };
 
 /* audio capability from EDID*/
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#7 lore raw

PATCH 6/103 drm/amd/display: Don't report fake sink as connected

In reply to a message not in the archive

This was missed in the previous fake sink change. The fake sink allows us to enable a pipe with a disconnected display. We shouldn't report it as connected.

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +2 -1
index b9e3360b5655..4e38d69a7489 100644
@@ -2470,7 +2470,8 @@ amdgpu_dm_connector_detect(struct drm_connector *connector, bool force)
 	 * 2. This interface *is called* in context of user-mode ioctl. Which
 	 * makes it a bad place for *any* MST-related activit. */
 
-	if (aconnector->base.force == DRM_FORCE_UNSPECIFIED)
+	if (aconnector->base.force == DRM_FORCE_UNSPECIFIED &&
+	    !aconnector->fake_enable)
 		connected = (aconnector->dc_sink != NULL);
 	else
 		connected = (aconnector->base.force == DRM_FORCE_ON);
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#8 lore raw

PATCH 7/103 drm/amd/display: Refactor dcn10_dpp_cm_helper to dcn10_cm_common

In reply to a message not in the archive

From: Yue Hin Lau <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dcn10/Makefile                         | 2 +-
 .../amd/display/dc/dcn10/{dcn10_dpp_cm_helper.c => dcn10_cm_common.c} | 2 +-
 .../amd/display/dc/dcn10/{dcn10_dpp_cm_helper.h => dcn10_cm_common.h} | 4 ++--
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm.c                   | 2 +-
 4 files changed, 5 insertions(+), 5 deletions(-)
 rename drivers/gpu/drm/amd/display/dc/dcn10/{dcn10_dpp_cm_helper.c => dcn10_cm_common.c} (99%)
 rename drivers/gpu/drm/amd/display/dc/dcn10/{dcn10_dpp_cm_helper.h => dcn10_cm_common.h} (97%)
drivers/gpu/drm/amd/display/dc/dcn10/Makefile +1 -1
index 2d6d3a371858..6b19106f4679 100644
@@ -4,7 +4,7 @@
 DCN10 = dcn10_resource.o dcn10_ipp.o dcn10_hw_sequencer.o \
 		dcn10_dpp.o dcn10_opp.o dcn10_timing_generator.o \
 		dcn10_mem_input.o dcn10_mpc.o \
-		dcn10_dpp_dscl.o dcn10_dpp_cm.o dcn10_dpp_cm_helper.o
+		dcn10_dpp_dscl.o dcn10_dpp_cm.o dcn10_cm_common.o
 
 AMD_DAL_DCN10 = $(addprefix $(AMDDALPATH)/dc/dcn10/,$(DCN10))
 
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.c renamed from drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm_helper.c +1 -1
similarity index 99%
rename from drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm_helper.c
rename to drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.c
index f616e08759de..7f579cb19f4b 100644
@@ -26,7 +26,7 @@
 #include "reg_helper.h"
 #include "dcn10_dpp.h"
 
-#include "dcn10_dpp_cm_helper.h"
+#include "dcn10_cm_common.h"
 
 #define REG(reg) reg
 
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.h renamed from drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm_helper.h +2 -2
similarity index 97%
rename from drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm_helper.h
rename to drivers/gpu/drm/amd/display/dc/dcn10/dcn10_cm_common.h
index 1155ee522898..64836dcf21f2 100644
@@ -23,8 +23,8 @@
  *
  */
 
-#ifndef __DAL_DPP_DCN10_CM_HELPER_H__
-#define __DAL_DPP_DCN10_CM_HELPER_H__
+#ifndef __DAL_DCN10_CM_COMMON_H__
+#define __DAL_DCN10_CM_COMMON_H__
 
 #define TF_HELPER_REG_FIELD_LIST(type) \
 	type exp_region0_lut_offset; \
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm.c +1 -1
index d0e72acfc1d5..61e2a8919e9c 100644
@@ -29,8 +29,8 @@
 
 #include "reg_helper.h"
 #include "dcn10_dpp.h"
-#include "dcn10_dpp_cm_helper.h"
 #include "basics/conversion.h"
+#include "dcn10_cm_common.h"
 
 #define NUM_PHASES    64
 #define HORZ_MAX_TAPS 8
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#9 lore raw

PATCH 8/103 drm/amd/display: fix multi-display on CZ

In reply to a message not in the archive

From: Roman Li <[email protected]>

- fixed wrong index in dce110_validate_surface_sets()
- formatted for better readability

---
 .../drm/amd/display/dc/dce110/dce110_resource.c    | 29 ++++++++++++++++------
 1 file changed, 21 insertions(+), 8 deletions(-)
drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c +21 -8
index 28e768de744c..8e4ec61d59a4 100644
@@ -849,7 +849,7 @@ static bool dce110_validate_bandwidth(
 static bool dce110_validate_surface_sets(
 		struct dc_state *context)
 {
-	int i;
+	int i, j;
 
 	for (i = 0; i < context->stream_count; i++) {
 		if (context->stream_status[i].plane_count == 0)
@@ -858,14 +858,27 @@ static bool dce110_validate_surface_sets(
 		if (context->stream_status[i].plane_count > 2)
 			return false;
 
-		if ((context->stream_status[i].plane_states[i]->format >= SURFACE_PIXEL_FORMAT_VIDEO_BEGIN) &&
-		    (context->stream_status[i].plane_states[i]->src_rect.width > 1920 ||
-		     context->stream_status[i].plane_states[i]->src_rect.height > 1080))
-			return false;
+		for (j = 0; j < context->stream_status[i].plane_count; j++) {
+			struct dc_plane_state *plane =
+				context->stream_status[i].plane_states[j];
 
-		/* irrespective of plane format, stream should be RGB encoded */
-		if (context->streams[i]->timing.pixel_encoding != PIXEL_ENCODING_RGB)
-			return false;
+			/* underlay validation */
+			if (plane->format >= SURFACE_PIXEL_FORMAT_VIDEO_BEGIN) {
+
+				if ((plane->src_rect.width > 1920 ||
+					plane->src_rect.height > 1080))
+					return false;
+
+				/* irrespective of plane format,
+				 * stream should be RGB encoded
+				 */
+				if (context->streams[i]->timing.pixel_encoding
+						!= PIXEL_ENCODING_RGB)
+					return false;
+
+			}
+
+		}
 	}
 
 	return true;
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#10 lore raw

PATCH 9/103 drm/amd/display: update mpc add/remove functions

In reply to a message not in the archive

From: Eric Bernstein <[email protected]>

---
 .../drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c  |  17 +-
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c   | 173 ++++++++++++++-------
 drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h        |  10 +-
 3 files changed, 130 insertions(+), 70 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +11 -6
index e60be00abcf2..82a985a981be 100644
@@ -881,7 +881,8 @@ static void dcn10_init_hw(struct dc *dc)
 
 		xfm->funcs->transform_reset(xfm);
 		dc->res_pool->mpc->funcs->remove(
-				dc->res_pool->mpc, dc->res_pool->opps[i], i);
+				dc->res_pool->mpc, &(dc->res_pool->opps[i]->mpc_tree),
+				dc->res_pool->opps[i]->inst, i);
 
 		/* Blank controller using driver code instead of
 		 * command table.
@@ -1079,7 +1080,8 @@ static void plane_atomic_disconnect(struct dc *dc,
 	if (dc->debug.sanity_checks)
 		verify_allow_pstate_change_high(dc->hwseq);
 
-	mpc->funcs->remove(mpc, dc->res_pool->opps[opp_id], fe_idx);
+	mpc->funcs->remove(mpc, &(dc->res_pool->opps[opp_id]->mpc_tree),
+			dc->res_pool->opps[opp_id]->inst, fe_idx);
 }
 
 /* disable HW used by plane.
@@ -2299,8 +2301,9 @@ static void update_dchubp_dpp(
 			plane_state->format,
 			EXPANSION_MODE_ZERO);
 
-	mpcc_cfg.mi = mi;
-	mpcc_cfg.opp = pipe_ctx->stream_res.opp;
+	mpcc_cfg.dpp_id = mi->inst;
+	mpcc_cfg.opp_id = pipe_ctx->stream_res.opp->inst;
+	mpcc_cfg.tree_cfg = &(pipe_ctx->stream_res.opp->mpc_tree);
 	for (top_pipe = pipe_ctx->top_pipe; top_pipe; top_pipe = top_pipe->top_pipe)
 		mpcc_cfg.z_index++;
 	if (dc->debug.surface_visual_confirm)
@@ -2317,7 +2320,8 @@ static void update_dchubp_dpp(
 	mpcc_cfg.pre_multiplied_alpha = is_rgb_cspace(
 			pipe_ctx->stream->output_color_space)
 					&& per_pixel_alpha;
-	dc->res_pool->mpc->funcs->add(dc->res_pool->mpc, &mpcc_cfg);
+	mi->mpcc_id = dc->res_pool->mpc->funcs->add(dc->res_pool->mpc, &mpcc_cfg);
+	mi->opp_id = mpcc_cfg.opp_id;
 
 	pipe_ctx->plane_res.scl_data.lb_params.alpha_en = per_pixel_alpha;
 	pipe_ctx->plane_res.scl_data.lb_params.depth = LB_PIXEL_DEPTH_30BPP;
@@ -2528,7 +2532,8 @@ static void dcn10_apply_ctx_for_surface(
 			/* reset mpc */
 			dc->res_pool->mpc->funcs->remove(
 					dc->res_pool->mpc,
-					old_pipe_ctx->stream_res.opp,
+					&(old_pipe_ctx->stream_res.opp->mpc_tree),
+					old_pipe_ctx->stream_res.opp->inst,
 					old_pipe_ctx->pipe_idx);
 			old_pipe_ctx->stream_res.opp->mpcc_disconnect_pending[old_pipe_ctx->plane_res.mi->mpcc_id] = true;
 
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c +113 -60

237 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h +6 -4
index 4bbcff48acc8..2d3de5b68e34 100644
@@ -29,8 +29,9 @@
 #include "opp.h"
 
 struct mpcc_cfg {
-	struct mem_input *mi;
-	struct output_pixel_processor *opp;
+	int dpp_id;
+	int opp_id;
+	struct mpc_tree_cfg *tree_cfg;
 	unsigned int z_index;
 
 	struct tg_color black_color;
@@ -44,9 +45,10 @@ struct mpc {
 };
 
 struct mpc_funcs {
-	void (*add)(struct mpc *mpc, struct mpcc_cfg *cfg);
+	int (*add)(struct mpc *mpc, struct mpcc_cfg *cfg);
 	void (*remove)(struct mpc *mpc,
-			struct output_pixel_processor *opp,
+			struct mpc_tree_cfg *tree_cfg,
+			int opp_id,
 			int mpcc_inst);
 	void (*wait_for_idle)(struct mpc *mpc, int id);
 };
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#11 lore raw

PATCH 10/103 drm/amd/display: Match actual state during S3 resume.

In reply to a message not in the archive

From: "Leo (Sunpeng) Li" <[email protected]>

During system suspend, we:

1. Cache a duplicate of the current DRM atomic state, which calls hooks to duplicate our internal states.
2. Call hooks to disable all functionalities.
3. System suspended.

During resume, we attempt to restore the cached state. However, our interal states are now stale, since step 1 was done before step 2. i.e. our cached state does not reflect the disabled nature of things.

This change resolves that by destroying all relevant states to reflect the actual state during resume.

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 36 +++++++++++++++++++++++
 1 file changed, 36 insertions(+)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +36 -0
index 4e38d69a7489..8e1de7af310c 100644
@@ -617,6 +617,11 @@ int amdgpu_dm_display_resume(struct amdgpu_device *adev)
 	struct drm_connector *connector;
 	struct drm_crtc *crtc;
 	struct drm_crtc_state *crtc_state;
+	struct dm_crtc_state *dm_crtc_state;
+	struct drm_plane *plane;
+	struct drm_plane_state *plane_state;
+	struct dm_plane_state *dm_plane_state;
+	struct dm_atomic_state *cached_state;
 	int ret = 0;
 	int i;
 
@@ -655,6 +660,37 @@ int amdgpu_dm_display_resume(struct amdgpu_device *adev)
 	for_each_crtc_in_state(adev->dm.cached_state, crtc, crtc_state, i)
 			crtc_state->active_changed = true;
 
+	cached_state = to_dm_atomic_state(adev->dm.cached_state);
+
+	/*
+	 * During suspend, the cached state is saved before all streams are
+	 * disabled. Refresh cached state to match actual current state before
+	 * restoring it.
+	 */
+	WARN_ON(kref_read(&cached_state->context->refcount) > 1);
+	dc_release_state(cached_state->context);
+
+	cached_state->context = dc_create_state();
+	dc_resource_state_copy_construct_current(adev->dm.dc, cached_state->context);
+
+	for_each_new_crtc_in_state(adev->dm.cached_state, crtc, crtc_state, i) {
+		dm_crtc_state = to_dm_crtc_state(crtc_state);
+		if (dm_crtc_state->stream) {
+			WARN_ON(kref_read(&dm_crtc_state->stream->refcount) > 1);
+			dc_stream_release(dm_crtc_state->stream);
+			dm_crtc_state->stream = NULL;
+		}
+	}
+
+	for_each_new_plane_in_state(adev->dm.cached_state, plane, plane_state, i) {
+		dm_plane_state = to_dm_plane_state(plane_state);
+		if (dm_plane_state->dc_state) {
+			WARN_ON(kref_read(&dm_plane_state->dc_state->refcount) > 1);
+			dc_plane_state_release(dm_plane_state->dc_state);
+			dm_plane_state->dc_state = NULL;
+		}
+	}
+
 	ret = drm_atomic_helper_resume(ddev, adev->dm.cached_state);
 
 	drm_atomic_state_put(adev->dm.cached_state);
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#12 lore raw

PATCH 11/103 drm/amd/display: fixing register includes

In reply to a message not in the archive

From: Yue Hin Lau <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h     | 89 +++++++++++-----------
 .../gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.h |  2 +
 2 files changed, 46 insertions(+), 45 deletions(-)
drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h +44 -45
index 2d3a41f744af..589332019a83 100644
@@ -131,8 +131,6 @@
 	HWSEQ_PHYPLL_REG_LIST(CRTC)
 
 #define HWSEQ_DCN_REG_LIST()\
-	HWSEQ_PIXEL_RATE_REG_LIST(OTG), \
-	HWSEQ_PHYPLL_REG_LIST(OTG), \
 	SRII(OTG_GLOBAL_SYNC_STATUS, OTG, 0), \
 	SRII(OTG_GLOBAL_SYNC_STATUS, OTG, 1), \
 	SRII(OTG_GLOBAL_SYNC_STATUS, OTG, 2), \
@@ -182,22 +180,6 @@
 	SR(DCHUBBUB_TEST_DEBUG_INDEX), \
 	SR(DCHUBBUB_TEST_DEBUG_DATA), \
 	SR(DC_IP_REQUEST_CNTL), \
-	SR(DOMAIN0_PG_CONFIG), \
-	SR(DOMAIN1_PG_CONFIG), \
-	SR(DOMAIN2_PG_CONFIG), \
-	SR(DOMAIN3_PG_CONFIG), \
-	SR(DOMAIN4_PG_CONFIG), \
-	SR(DOMAIN5_PG_CONFIG), \
-	SR(DOMAIN6_PG_CONFIG), \
-	SR(DOMAIN7_PG_CONFIG), \
-	SR(DOMAIN0_PG_STATUS), \
-	SR(DOMAIN1_PG_STATUS), \
-	SR(DOMAIN2_PG_STATUS), \
-	SR(DOMAIN3_PG_STATUS), \
-	SR(DOMAIN4_PG_STATUS), \
-	SR(DOMAIN5_PG_STATUS), \
-	SR(DOMAIN6_PG_STATUS), \
-	SR(DOMAIN7_PG_STATUS), \
 	SR(DIO_MEM_PWR_CTRL), \
 	SR(DCCG_GATE_DISABLE_CNTL), \
 	SR(DCCG_GATE_DISABLE_CNTL2), \
@@ -223,13 +205,30 @@
 
 #define HWSEQ_DCN1_REG_LIST()\
 	HWSEQ_DCN_REG_LIST(), \
+	HWSEQ_PIXEL_RATE_REG_LIST(OTG), \
+	HWSEQ_PHYPLL_REG_LIST(OTG), \
 	SR(DCHUBBUB_SDPIF_FB_TOP),\
 	SR(DCHUBBUB_SDPIF_FB_BASE),\
 	SR(DCHUBBUB_SDPIF_FB_OFFSET),\
 	SR(DCHUBBUB_SDPIF_AGP_BASE),\
 	SR(DCHUBBUB_SDPIF_AGP_BOT),\
-	SR(DCHUBBUB_SDPIF_AGP_TOP)
-
+	SR(DCHUBBUB_SDPIF_AGP_TOP),\
+	SR(DOMAIN0_PG_CONFIG), \
+	SR(DOMAIN1_PG_CONFIG), \
+	SR(DOMAIN2_PG_CONFIG), \
+	SR(DOMAIN3_PG_CONFIG), \
+	SR(DOMAIN4_PG_CONFIG), \
+	SR(DOMAIN5_PG_CONFIG), \
+	SR(DOMAIN6_PG_CONFIG), \
+	SR(DOMAIN7_PG_CONFIG), \
+	SR(DOMAIN0_PG_STATUS), \
+	SR(DOMAIN1_PG_STATUS), \
+	SR(DOMAIN2_PG_STATUS), \
+	SR(DOMAIN3_PG_STATUS), \
+	SR(DOMAIN4_PG_STATUS), \
+	SR(DOMAIN5_PG_STATUS), \
+	SR(DOMAIN6_PG_STATUS), \
+	SR(DOMAIN7_PG_STATUS)
 
 struct dce_hwseq_registers {
 	uint32_t DCFE_CLOCK_CONTROL[6];
@@ -417,30 +416,6 @@ struct dce_hwseq_registers {
 	HWS_SF(OPP_PIPE0_, OPP_PIPE_CONTROL, OPP_PIPE_CLOCK_EN, mask_sh),\
 	HWS_SF(, DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_ENABLE, mask_sh), \
 	HWS_SF(, DC_IP_REQUEST_CNTL, IP_REQUEST_EN, mask_sh), \
-	HWS_SF(, DOMAIN0_PG_CONFIG, DOMAIN0_POWER_FORCEON, mask_sh), \
-	HWS_SF(, DOMAIN0_PG_CONFIG, DOMAIN0_POWER_GATE, mask_sh), \
-	HWS_SF(, DOMAIN1_PG_CONFIG, DOMAIN1_POWER_FORCEON, mask_sh), \
-	HWS_SF(, DOMAIN1_PG_CONFIG, DOMAIN1_POWER_GATE, mask_sh), \
-	HWS_SF(, DOMAIN2_PG_CONFIG, DOMAIN2_POWER_FORCEON, mask_sh), \
-	HWS_SF(, DOMAIN2_PG_CONFIG, DOMAIN2_POWER_GATE, mask_sh), \
-	HWS_SF(, DOMAIN3_PG_CONFIG, DOMAIN3_POWER_FORCEON, mask_sh), \
-	HWS_SF(, DOMAIN3_PG_CONFIG, DOMAIN3_POWER_GATE, mask_sh), \
-	HWS_SF(, DOMAIN4_PG_CONFIG, DOMAIN4_POWER_FORCEON, mask_sh), \
-	HWS_SF(, DOMAIN4_PG_CONFIG, DOMAIN4_POWER_GATE, mask_sh), \
-	HWS_SF(, DOMAIN5_PG_CONFIG, DOMAIN5_POWER_FORCEON, mask_sh), \
-	HWS_SF(, DOMAIN5_PG_CONFIG, DOMAIN5_POWER_GATE, mask_sh), \
-	HWS_SF(, DOMAIN6_PG_CONFIG, DOMAIN6_POWER_FORCEON, mask_sh), \
-	HWS_SF(, DOMAIN6_PG_CONFIG, DOMAIN6_POWER_GATE, mask_sh), \
-	HWS_SF(, DOMAIN7_PG_CONFIG, DOMAIN7_POWER_FORCEON, mask_sh), \
-	HWS_SF(, DOMAIN7_PG_CONFIG, DOMAIN7_POWER_GATE, mask_sh), \
-	HWS_SF(, DOMAIN0_PG_STATUS, DOMAIN0_PGFSM_PWR_STATUS, mask_sh), \
-	HWS_SF(, DOMAIN1_PG_STATUS, DOMAIN1_PGFSM_PWR_STATUS, mask_sh), \
-	HWS_SF(, DOMAIN2_PG_STATUS, DOMAIN2_PGFSM_PWR_STATUS, mask_sh), \
-	HWS_SF(, DOMAIN3_PG_STATUS, DOMAIN3_PGFSM_PWR_STATUS, mask_sh), \
-	HWS_SF(, DOMAIN4_PG_STATUS, DOMAIN4_PGFSM_PWR_STATUS, mask_sh), \
-	HWS_SF(, DOMAIN5_PG_STATUS, DOMAIN5_PGFSM_PWR_STATUS, mask_sh), \
-	HWS_SF(, DOMAIN6_PG_STATUS, DOMAIN6_PGFSM_PWR_STATUS, mask_sh), \
-	HWS_SF(, DOMAIN7_PG_STATUS, DOMAIN7_PGFSM_PWR_STATUS, mask_sh), \
 	HWS_SF(, DCHUBBUB_ARB_WATERMARK_CHANGE_CNTL, DCHUBBUB_ARB_WATERMARK_CHANGE_REQUEST, mask_sh), \
 	HWS_SF(, DCHUBBUB_ARB_WATERMARK_CHANGE_CNTL, DCHUBBUB_ARB_WATERMARK_CHANGE_DONE_INTERRUPT_DISABLE, mask_sh), \
 	HWS_SF(, DCHUBBUB_ARB_DRAM_STATE_CNTL, DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_VALUE, mask_sh), \
@@ -468,7 +443,31 @@ struct dce_hwseq_registers {
 	HWS_SF(, VM_L2_PROTECTION_FAULT_DEFAULT_ADDR_LO32, PHYSICAL_PAGE_ADDR_LO32, mask_sh),\
 	HWS_SF(, MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB, PHYSICAL_PAGE_NUMBER_MSB, mask_sh),\
 	HWS_SF(, MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB, PHYSICAL_PAGE_NUMBER_LSB, mask_sh),\
-	HWS_SF(, MC_VM_SYSTEM_APERTURE_LOW_ADDR, LOGICAL_ADDR, mask_sh)
+	HWS_SF(, MC_VM_SYSTEM_APERTURE_LOW_ADDR, LOGICAL_ADDR, mask_sh),\
+	HWS_SF(, DOMAIN0_PG_CONFIG, DOMAIN0_POWER_FORCEON, mask_sh), \
+	HWS_SF(, DOMAIN0_PG_CONFIG, DOMAIN0_POWER_GATE, mask_sh), \
+	HWS_SF(, DOMAIN1_PG_CONFIG, DOMAIN1_POWER_FORCEON, mask_sh), \
+	HWS_SF(, DOMAIN1_PG_CONFIG, DOMAIN1_POWER_GATE, mask_sh), \
+	HWS_SF(, DOMAIN2_PG_CONFIG, DOMAIN2_POWER_FORCEON, mask_sh), \
+	HWS_SF(, DOMAIN2_PG_CONFIG, DOMAIN2_POWER_GATE, mask_sh), \
+	HWS_SF(, DOMAIN3_PG_CONFIG, DOMAIN3_POWER_FORCEON, mask_sh), \
+	HWS_SF(, DOMAIN3_PG_CONFIG, DOMAIN3_POWER_GATE, mask_sh), \
+	HWS_SF(, DOMAIN4_PG_CONFIG, DOMAIN4_POWER_FORCEON, mask_sh), \
+	HWS_SF(, DOMAIN4_PG_CONFIG, DOMAIN4_POWER_GATE, mask_sh), \
+	HWS_SF(, DOMAIN5_PG_CONFIG, DOMAIN5_POWER_FORCEON, mask_sh), \
+	HWS_SF(, DOMAIN5_PG_CONFIG, DOMAIN5_POWER_GATE, mask_sh), \
+	HWS_SF(, DOMAIN6_PG_CONFIG, DOMAIN6_POWER_FORCEON, mask_sh), \
+	HWS_SF(, DOMAIN6_PG_CONFIG, DOMAIN6_POWER_GATE, mask_sh), \
+	HWS_SF(, DOMAIN7_PG_CONFIG, DOMAIN7_POWER_FORCEON, mask_sh), \
+	HWS_SF(, DOMAIN7_PG_CONFIG, DOMAIN7_POWER_GATE, mask_sh), \
+	HWS_SF(, DOMAIN0_PG_STATUS, DOMAIN0_PGFSM_PWR_STATUS, mask_sh), \
+	HWS_SF(, DOMAIN1_PG_STATUS, DOMAIN1_PGFSM_PWR_STATUS, mask_sh), \
+	HWS_SF(, DOMAIN2_PG_STATUS, DOMAIN2_PGFSM_PWR_STATUS, mask_sh), \
+	HWS_SF(, DOMAIN3_PG_STATUS, DOMAIN3_PGFSM_PWR_STATUS, mask_sh), \
+	HWS_SF(, DOMAIN4_PG_STATUS, DOMAIN4_PGFSM_PWR_STATUS, mask_sh), \
+	HWS_SF(, DOMAIN5_PG_STATUS, DOMAIN5_PGFSM_PWR_STATUS, mask_sh), \
+	HWS_SF(, DOMAIN6_PG_STATUS, DOMAIN6_PGFSM_PWR_STATUS, mask_sh), \
+	HWS_SF(, DOMAIN7_PG_STATUS, DOMAIN7_PGFSM_PWR_STATUS, mask_sh)
 
 #define HWSEQ_REG_FIELD_LIST(type) \
 	type DCFE_CLOCK_ENABLE; \
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.h +2 -0
index aefd3e7bd7eb..0b7d4faff82f 100644
@@ -120,6 +120,8 @@
 	SR(DCHUBBUB_SDPIF_FB_BASE),\
 	SR(DCHUBBUB_SDPIF_FB_OFFSET)
 
+
+
 struct dcn_mi_registers {
 	uint32_t DCHUBP_CNTL;
 	uint32_t HUBPREQ_DEBUG_DB;
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#13 lore raw

PATCH 12/103 drm/amd/display: Exclude MST from fake sink

In reply to a message not in the archive

From: Jerry Zuo <[email protected]>

Needs effort to take care of the fake sink scenario in downstream daisy chain device. Exclude MST from fake sink feature for now.

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +11 -2
index 8e1de7af310c..fcf77f88a0f3 100644
@@ -2371,8 +2371,16 @@ static struct dc_stream_state *create_stream_for_sink(
 
 	drm_connector = &aconnector->base;
 
-	if (!aconnector->dc_sink)
+	if (!aconnector->dc_sink) {
+		/*
+		 * Exclude MST from creating fake_sink
+		 * TODO: need to enable MST into fake_sink feature
+		 */
+		if (aconnector->mst_port)
+			goto stream_create_fail;
+
 		create_fake_sink(aconnector);
+	}
 
 	stream = dc_create_stream_for_sink(aconnector->dc_sink);
 
@@ -4541,7 +4549,8 @@ static int dm_update_crtcs_state(
 		aconnector = amdgpu_dm_find_first_crct_matching_connector(state, crtc, true);
 
 		/* TODO This hack should go away */
-		if (aconnector) {
+		if (aconnector && enable) {
+			// Make sure fake sink is created in plug-in scenario
 			conn_state = drm_atomic_get_connector_state(state,
 								    &aconnector->base);
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#14 lore raw

PATCH 13/103 drm/amd/display: Fix NULL pointer on MST chained mode

In reply to a message not in the archive

From: Jerry Zuo <[email protected]>

Prevent NULL pointer on new_stream being added to ctx when added MST connectors cannot be found in existing crtc_state in the chained mode

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 7 +++++++
 1 file changed, 7 insertions(+)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +7 -0
index fcf77f88a0f3..fb80e22ae06f 100644
@@ -4628,6 +4628,13 @@ static int dm_update_crtcs_state(
 			*lock_and_validation_needed = true;
 
 		} else {/* Add stream for any updated/enabled CRTC */
+			/*
+			 * Quick fix to prevent NULL pointer on new_stream when
+			 * added MST connectors not found in existing crtc_state in the chained mode
+			 * TODO: need to dig out the root cause of that
+			 */
+			if (!aconnector || (!aconnector->dc_sink && aconnector->mst_port))
+				continue;
 
 			if (modereset_required(crtc_state))
 				goto next_crtc;
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#15 lore raw

PATCH 14/103 drm/amd/display: Refactor dc_update_planes_and_stream.

In reply to a message not in the archive

From: Bhawanpreet Lakha <[email protected]>

Split update_planes_and_stream_state (split Software and Hardware programming) as the state is already build, so we only need to program the hardware

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c |  21 +-
 drivers/gpu/drm/amd/display/dc/core/dc.c          | 256 +++++-----------------
 drivers/gpu/drm/amd/display/dc/dc.h               |  36 +--
 3 files changed, 87 insertions(+), 226 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +16 -5
index fb80e22ae06f..af2c9d1289c0 100644
@@ -3935,7 +3935,8 @@ static void prepare_flip_isr(struct amdgpu_crtc *acrtc)
 static void amdgpu_dm_do_flip(
 				struct drm_crtc *crtc,
 				struct drm_framebuffer *fb,
-				uint32_t target)
+				uint32_t target,
+				struct dc_state *state)
 {
 	unsigned long flags;
 	uint32_t target_vblank;
@@ -4006,7 +4007,13 @@ static void amdgpu_dm_do_flip(
 	surface_updates->flip_addr = &addr;
 
 
-	dc_update_planes_and_stream(adev->dm.dc, surface_updates, 1, acrtc_state->stream, NULL);
+	dc_commit_updates_for_stream(adev->dm.dc,
+					     surface_updates,
+					     1,
+					     acrtc_state->stream,
+					     NULL,
+					     &surface_updates->surface,
+					     state);
 
 	DRM_DEBUG_DRIVER("%s Flipping to hi: 0x%x, low: 0x%x \n",
 			 __func__,
@@ -4030,6 +4037,7 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 	struct dc_plane_state *plane_states_constructed[MAX_SURFACES];
 	struct amdgpu_crtc *acrtc_attach = to_amdgpu_crtc(pcrtc);
 	struct dm_crtc_state *acrtc_state = to_dm_crtc_state(pcrtc->state);
+	struct dm_atomic_state *dm_state = to_dm_atomic_state(state);
 	int planes_count = 0;
 	unsigned long flags;
 
@@ -4086,7 +4094,8 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 			amdgpu_dm_do_flip(
 				crtc,
 				fb,
-				drm_crtc_vblank_count(crtc) + *wait_for_vblank);
+				drm_crtc_vblank_count(crtc) + *wait_for_vblank,
+				dm_state->context);
 		}
 
 	}
@@ -4106,7 +4115,8 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_state *state,
 		if (false == dc_commit_planes_to_stream(dm->dc,
 							plane_states_constructed,
 							planes_count,
-							dc_stream_attach))
+							dc_stream_attach,
+							dm_state->context))
 			dm_error("%s: Failed to attach plane!\n", __func__);
 	} else {
 		/*TODO BUG Here should go disable planes on CRTC. */
@@ -4329,7 +4339,8 @@ void amdgpu_dm_atomic_commit_tail(
 				dm->dc,
 				status->plane_states,
 				status->plane_count,
-				new_acrtc_state->stream))
+				new_acrtc_state->stream,
+				dm_state->context))
 			dm_error("%s: Failed to update stream scaling!\n", __func__);
 	}
 
drivers/gpu/drm/amd/display/dc/core/dc.c +59 -197

296 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dc.h +12 -24
index 4ff543826476..50440731f83c 100644
@@ -449,23 +449,6 @@ struct dc_flip_addrs {
 	/* TODO: add flip duration for FreeSync */
 };
 
-/*
- * Set up surface attributes and associate to a stream
- * The surfaces parameter is an absolute set of all surface active for the stream.
- * If no surfaces are provided, the stream will be blanked; no memory read.
- * Any flip related attribute changes must be done through this interface.
- *
- * After this call:
- *   Surfaces attributes are programmed and configured to be composed into stream.
- *   This does not trigger a flip.  No surface address is programmed.
- */
-
-bool dc_commit_planes_to_stream(
-		struct dc *dc,
-		struct dc_plane_state **plane_states,
-		uint8_t new_plane_count,
-		struct dc_stream_state *stream);
-
 bool dc_post_update_surfaces_to_stream(
 		struct dc *dc);
 
@@ -571,9 +554,6 @@ bool dc_is_stream_unchanged(
 	struct dc_stream_state *old_stream, struct dc_stream_state *stream);
 
 /*
- * Setup stream attributes if no stream updates are provided
- * there will be no impact on the stream parameters
- *
  * Set up surface attributes and associate to a stream
  * The surfaces parameter is an absolute set of all surface active for the stream.
  * If no surfaces are provided, the stream will be blanked; no memory read.
@@ -582,14 +562,22 @@ bool dc_is_stream_unchanged(
  * After this call:
  *   Surfaces attributes are programmed and configured to be composed into stream.
  *   This does not trigger a flip.  No surface address is programmed.
- *
  */
 
-void dc_update_planes_and_stream(struct dc *dc,
-		struct dc_surface_update *surface_updates, int surface_count,
+bool dc_commit_planes_to_stream(
+		struct dc *dc,
+		struct dc_plane_state **plane_states,
+		uint8_t new_plane_count,
 		struct dc_stream_state *dc_stream,
-		struct dc_stream_update *stream_update);
+		struct dc_state *state);
 
+void dc_commit_updates_for_stream(struct dc *dc,
+		struct dc_surface_update *srf_updates,
+		int surface_count,
+		struct dc_stream_state *stream,
+		struct dc_stream_update *stream_update,
+		struct dc_plane_state **plane_states,
+		struct dc_state *state);
 /*
  * Log the current stream state.
  */
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#16 lore raw

PATCH 15/103 drm/amd/display: make tile changing run at ISR

In reply to a message not in the archive

From: ShihChen Chen <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/core/dc.c | 21 ++++++++++++++++++---
 1 file changed, 18 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +18 -3
index 1ec6df9870da..2cea8de9c9be 100644
@@ -1031,7 +1031,6 @@ static enum surface_update_type get_plane_info_update_type(
 	temp_plane_info.plane_size = u->surface->plane_size;
 	temp_plane_info.rotation = u->surface->rotation;
 	temp_plane_info.stereo_format = u->surface->stereo_format;
-	temp_plane_info.tiling_info = u->surface->tiling_info;
 
 	if (surface_index == 0)
 		temp_plane_info.visible = u->plane_info->visible;
@@ -1044,10 +1043,26 @@ static enum surface_update_type get_plane_info_update_type(
 
 	if (pixel_format_to_bpp(u->plane_info->format) !=
 			pixel_format_to_bpp(u->surface->format)) {
+		/* different bytes per element will require full bandwidth
+		 * and DML calculation
+		 */
 		return UPDATE_TYPE_FULL;
-	} else {
-		return UPDATE_TYPE_MED;
 	}
+
+	if (memcmp(&u->plane_info->tiling_info, &u->surface->tiling_info,
+			sizeof(union dc_tiling_info)) != 0) {
+		/* todo: below are HW dependent, we should add a hook to
+		 * DCE/N resource and validated there.
+		 */
+		if (u->plane_info->tiling_info.gfx9.swizzle != DC_SW_LINEAR) {
+			/* swizzled mode requires RQ to be setup properly,
+			 * thus need to run DML to calculate RQ settings
+			 */
+			return UPDATE_TYPE_FULL;
+		}
+	}
+
+	return UPDATE_TYPE_MED;
 }
 
 static enum surface_update_type  get_scaling_info_update_type(
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#17 lore raw

PATCH 16/103 drm/amd/display: power_down_Hw need signal type to turnoff backlight

In reply to a message not in the archive

From: Andrew Jiang <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +4 -1
index 590b2ad37105..d83a330712f7 100644
@@ -1171,6 +1171,7 @@ static void power_down_encoders(struct dc *dc)
 {
 	int i;
 	enum connector_id connector_id;
+	enum signal_type signal = SIGNAL_TYPE_NONE;
 	for (i = 0; i < dc->link_count; i++) {
 		connector_id = dal_graphics_object_id_get_connector_id(dc->links[i]->link_id);
 		if ((connector_id == CONNECTOR_ID_DISPLAY_PORT) ||
@@ -1178,10 +1179,12 @@ static void power_down_encoders(struct dc *dc)
 
 			if (!dc->links[i]->wa_flags.dp_keep_receiver_powered)
 				dp_receiver_power_ctrl(dc->links[i], false);
+			if (connector_id == CONNECTOR_ID_EDP)
+				signal = SIGNAL_TYPE_EDP;
 		}
 
 		dc->links[i]->link_enc->funcs->disable_output(
-				dc->links[i]->link_enc, SIGNAL_TYPE_NONE);
+				dc->links[i]->link_enc, signal);
 	}
 }
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#18 lore raw

PATCH 17/103 drm/amd/display: removing remaining register definitions work around

In reply to a message not in the archive

From: Yue Hin Lau <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h      | 16 ++++++++--------
 .../gpu/drm/amd/display/dc/dce/dce_link_encoder.h   | 21 ++++++++++++++-------
 .../amd/display/dc/dcn10/dcn10_timing_generator.h   | 11 +----------
 3 files changed, 23 insertions(+), 25 deletions(-)
drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h +8 -8
index 589332019a83..5798001da770 100644
@@ -179,16 +179,11 @@
 	SR(DCHUBBUB_GLOBAL_TIMER_CNTL), \
 	SR(DCHUBBUB_TEST_DEBUG_INDEX), \
 	SR(DCHUBBUB_TEST_DEBUG_DATA), \
-	SR(DC_IP_REQUEST_CNTL), \
 	SR(DIO_MEM_PWR_CTRL), \
 	SR(DCCG_GATE_DISABLE_CNTL), \
 	SR(DCCG_GATE_DISABLE_CNTL2), \
 	SR(DCFCLK_CNTL),\
 	SR(DCFCLK_CNTL), \
-	SR(D1VGA_CONTROL), \
-	SR(D2VGA_CONTROL), \
-	SR(D3VGA_CONTROL), \
-	SR(D4VGA_CONTROL), \
 	/* todo:  get these from GVM instead of reading registers ourselves */\
 	MMHUB_SR(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR_HI32),\
 	MMHUB_SR(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR_LO32),\
@@ -228,7 +223,12 @@
 	SR(DOMAIN4_PG_STATUS), \
 	SR(DOMAIN5_PG_STATUS), \
 	SR(DOMAIN6_PG_STATUS), \
-	SR(DOMAIN7_PG_STATUS)
+	SR(DOMAIN7_PG_STATUS), \
+	SR(D1VGA_CONTROL), \
+	SR(D2VGA_CONTROL), \
+	SR(D3VGA_CONTROL), \
+	SR(D4VGA_CONTROL), \
+	SR(DC_IP_REQUEST_CNTL)
 
 struct dce_hwseq_registers {
 	uint32_t DCFE_CLOCK_CONTROL[6];
@@ -415,7 +415,6 @@ struct dce_hwseq_registers {
 	HWS_SF(DPP_TOP0_, DPP_CONTROL, DPP_CLOCK_ENABLE, mask_sh), \
 	HWS_SF(OPP_PIPE0_, OPP_PIPE_CONTROL, OPP_PIPE_CLOCK_EN, mask_sh),\
 	HWS_SF(, DCHUBBUB_GLOBAL_TIMER_CNTL, DCHUBBUB_GLOBAL_TIMER_ENABLE, mask_sh), \
-	HWS_SF(, DC_IP_REQUEST_CNTL, IP_REQUEST_EN, mask_sh), \
 	HWS_SF(, DCHUBBUB_ARB_WATERMARK_CHANGE_CNTL, DCHUBBUB_ARB_WATERMARK_CHANGE_REQUEST, mask_sh), \
 	HWS_SF(, DCHUBBUB_ARB_WATERMARK_CHANGE_CNTL, DCHUBBUB_ARB_WATERMARK_CHANGE_DONE_INTERRUPT_DISABLE, mask_sh), \
 	HWS_SF(, DCHUBBUB_ARB_DRAM_STATE_CNTL, DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_VALUE, mask_sh), \
@@ -467,7 +466,8 @@ struct dce_hwseq_registers {
 	HWS_SF(, DOMAIN4_PG_STATUS, DOMAIN4_PGFSM_PWR_STATUS, mask_sh), \
 	HWS_SF(, DOMAIN5_PG_STATUS, DOMAIN5_PGFSM_PWR_STATUS, mask_sh), \
 	HWS_SF(, DOMAIN6_PG_STATUS, DOMAIN6_PGFSM_PWR_STATUS, mask_sh), \
-	HWS_SF(, DOMAIN7_PG_STATUS, DOMAIN7_PGFSM_PWR_STATUS, mask_sh)
+	HWS_SF(, DOMAIN7_PG_STATUS, DOMAIN7_PGFSM_PWR_STATUS, mask_sh), \
+	HWS_SF(, DC_IP_REQUEST_CNTL, IP_REQUEST_EN, mask_sh)
 
 #define HWSEQ_REG_FIELD_LIST(type) \
 	type DCFE_CLOCK_ENABLE; \
drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.h +14 -7
index be0a45bdc5e2..05463df7e636 100644
@@ -44,8 +44,6 @@
 	SRI(DC_HPD_CONTROL, HPD, id)
 
 #define LE_COMMON_REG_LIST_BASE(id) \
-	SR(LVTMA_PWRSEQ_CNTL), \
-	SR(LVTMA_PWRSEQ_STATE), \
 	SR(DMCU_RAM_ACCESS_CTRL), \
 	SR(DMCU_IRAM_RD_CTRL), \
 	SR(DMCU_IRAM_RD_DATA), \
@@ -71,6 +69,10 @@
 	SRI(DP_DPHY_FAST_TRAINING, DP, id), \
 	SRI(DP_SEC_CNTL1, DP, id)
 
+#define LE_EDP_REG_LIST(id)\
+	SR(LVTMA_PWRSEQ_CNTL), \
+	SR(LVTMA_PWRSEQ_STATE)
+
 #define LE_COMMON_REG_LIST(id)\
 	LE_COMMON_REG_LIST_BASE(id), \
 	SRI(DP_DPHY_BS_SR_SWAP_CNTL, DP, id), \
@@ -79,33 +81,38 @@
 
 #define LE_DCE80_REG_LIST(id)\
 	SRI(DP_DPHY_INTERNAL_CTRL, DP, id), \
-	LE_COMMON_REG_LIST_BASE(id)
+	LE_COMMON_REG_LIST_BASE(id), \
+	LE_EDP_REG_LIST(id)
 
 #define LE_DCE100_REG_LIST(id)\
 	LE_COMMON_REG_LIST_BASE(id), \
 	SRI(DP_DPHY_BS_SR_SWAP_CNTL, DP, id), \
 	SRI(DP_DPHY_INTERNAL_CTRL, DP, id), \
-	SR(DCI_MEM_PWR_STATUS)
+	SR(DCI_MEM_PWR_STATUS), \
+	LE_EDP_REG_LIST(id)
 
 #define LE_DCE110_REG_LIST(id)\
 	LE_COMMON_REG_LIST_BASE(id), \
 	SRI(DP_DPHY_BS_SR_SWAP_CNTL, DP, id), \
 	SRI(DP_DPHY_INTERNAL_CTRL, DP, id), \
 	SRI(DP_DPHY_HBR2_PATTERN_CONTROL, DP, id), \
-	SR(DCI_MEM_PWR_STATUS)
+	SR(DCI_MEM_PWR_STATUS), \
+	LE_EDP_REG_LIST(id)
 
 #define LE_DCE120_REG_LIST(id)\
 	LE_COMMON_REG_LIST_BASE(id), \
 	SRI(DP_DPHY_BS_SR_SWAP_CNTL, DP, id), \
 	SRI(DP_DPHY_HBR2_PATTERN_CONTROL, DP, id), \
-	SR(DCI_MEM_PWR_STATUS)
+	SR(DCI_MEM_PWR_STATUS), \
+	LE_EDP_REG_LIST(id)
 
 #define LE_DCN10_REG_LIST(id)\
 	LE_COMMON_REG_LIST_BASE(id), \
 	SRI(DP_DPHY_BS_SR_SWAP_CNTL, DP, id), \
 	SRI(DP_DPHY_INTERNAL_CTRL, DP, id), \
 	SRI(DP_DPHY_HBR2_PATTERN_CONTROL, DP, id), \
-	SR(DMU_MEM_PWR_CNTL)
+	SR(DMU_MEM_PWR_CNTL), \
+	LE_EDP_REG_LIST(id)
 
 struct dce110_link_enc_aux_registers {
 	uint32_t AUX_CONTROL;
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_timing_generator.h +1 -10
index 69da293e9b4a..0826d73b9809 100644
@@ -71,11 +71,7 @@
 	SRI(OPTC_INPUT_GLOBAL_CONTROL, ODM, inst),\
 	SRI(OPPBUF_CONTROL, OPPBUF, inst),\
 	SRI(OPPBUF_3D_PARAMETERS_0, OPPBUF, inst),\
-	SRI(CONTROL, VTG, inst),\
-	SR(D1VGA_CONTROL),\
-	SR(D2VGA_CONTROL),\
-	SR(D3VGA_CONTROL),\
-	SR(D4VGA_CONTROL)
+	SRI(CONTROL, VTG, inst)
 
 #define TG_COMMON_REG_LIST_DCN1_0(inst) \
 	TG_COMMON_REG_LIST_DCN(inst),\
@@ -128,11 +124,6 @@ struct dcn_tg_registers {
 	uint32_t OPPBUF_CONTROL;
 	uint32_t OPPBUF_3D_PARAMETERS_0;
 	uint32_t CONTROL;
-	/*todo: move VGA to HWSS */
-	uint32_t D1VGA_CONTROL;
-	uint32_t D2VGA_CONTROL;
-	uint32_t D3VGA_CONTROL;
-	uint32_t D4VGA_CONTROL;
 };
 
 #define TG_COMMON_MASK_SH_LIST_DCN(mask_sh)\
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#19 lore raw

PATCH 18/103 drm/amd/display: set cp25201 to use TPS4

In reply to a message not in the archive

From: Wenjing Liu <[email protected]>

[Description]
hbr2 compliance eye output is unstable (toggling on and off) with debugger break. This caueses intermittent PHY automation failure. Need to look into the root cause later

---
 drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c | 16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c +14 -2
index b735782b8fe0..ced42484dcfc 100644
@@ -1700,6 +1700,12 @@ static void dp_test_send_link_training(struct dc_link *link)
 	dp_retrain_link_dp_test(link, &link_settings, false);
 }
 
+/* TODO hbr2 compliance eye output is unstable
+ * (toggling on and off) with debugger break
+ * This caueses intermittent PHY automation failure
+ * Need to look into the root cause */
+static uint8_t force_tps4_for_cp2520 = 1;
+
 static void dp_test_send_phy_test_pattern(struct dc_link *link)
 {
 	union phy_test_pattern dpcd_test_pattern;
@@ -1758,10 +1764,16 @@ static void dp_test_send_phy_test_pattern(struct dc_link *link)
 		test_pattern = DP_TEST_PATTERN_80BIT_CUSTOM;
 		break;
 	case PHY_TEST_PATTERN_CP2520_1:
-		test_pattern = DP_TEST_PATTERN_HBR2_COMPLIANCE_EYE;
+		/* CP2520 pattern is unstable, temporarily use TPS4 instead */
+		test_pattern = (force_tps4_for_cp2520 == 1) ?
+				DP_TEST_PATTERN_TRAINING_PATTERN4 :
+				DP_TEST_PATTERN_HBR2_COMPLIANCE_EYE;
 		break;
 	case PHY_TEST_PATTERN_CP2520_2:
-		test_pattern = DP_TEST_PATTERN_CP2520_2;
+		/* CP2520 pattern is unstable, temporarily use TPS4 instead */
+		test_pattern = (force_tps4_for_cp2520 == 1) ?
+				DP_TEST_PATTERN_TRAINING_PATTERN4 :
+				DP_TEST_PATTERN_HBR2_COMPLIANCE_EYE;
 		break;
 	case PHY_TEST_PATTERN_CP2520_3:
 		test_pattern = DP_TEST_PATTERN_TRAINING_PATTERN4;
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#20 lore raw

PATCH 19/103 drm/amd/display: add back removed hack for mpcc add

In reply to a message not in the archive

From: Eric Yang <[email protected]>

A previous changed removed the hack to match mpcc_idd with mi instance. This causes pstate hang on resume from hibernate for yet unknown reason. Add the hack back for now to work around the issue. More debugging required in init_hw to root cause the hang.

---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c +12 -0
index 8e767c84359c..6e56fa3a135b 100644
@@ -279,6 +279,18 @@ static int mpc10_mpcc_add(struct mpc *mpc, struct mpcc_cfg *cfg)
 	if (z_idx == cfg->tree_cfg->num_pipes) {
 		ASSERT(cfg->z_index <= cfg->tree_cfg->num_pipes);
 		mpcc_id = mpc10_get_idle_mpcc_id(mpc10);
+
+		/*
+		 * TODO: remove hack
+		 * Note: currently there is a bug in init_hw such that
+		 * on resume from hibernate, BIOS sets up MPCC0, and
+		 * we do mpcc_remove but the mpcc cannot go to idle
+		 * after remove. This cause us to pick mpcc1 here,
+		 * which causes a pstate hang for yet unknown reason.
+		 */
+		mpcc_id = cfg->dpp_id;
+		/* end hack*/
+
 		ASSERT(!(mpc10->mpcc_in_use_mask & 1 << mpcc_id));
 
 		if (mpc->ctx->dc->debug.sanity_checks)
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#21 lore raw

PATCH 20/103 drm/amd/display: update blending mode and set output denorm

In reply to a message not in the archive

From: Eric Bernstein <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c | 27 +++++++++++++++++++++++-
 drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h      |  4 ++++
 2 files changed, 30 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c +26 -1
index 6e56fa3a135b..334f072cea1d 100644
@@ -313,10 +313,34 @@ static int mpc10_mpcc_add(struct mpc *mpc, struct mpcc_cfg *cfg)
 	return mpcc_id;
 }
 
+static void mpc10_update_blend_mode(
+		struct mpc *mpc,
+		struct mpcc_cfg *cfg)
+{
+	struct dcn10_mpc *mpc10 = TO_DCN10_MPC(mpc);
+	int mpcc_id, z_idx;
+	int alpha_blnd_mode = cfg->per_pixel_alpha ?
+			BLND_PP_ALPHA : BLND_GLOBAL_ALPHA;
+
+	/* find z_idx for the dpp that requires blending mode update*/
+	for (z_idx = 0; z_idx < cfg->tree_cfg->num_pipes; z_idx++)
+		if (cfg->tree_cfg->dpp[z_idx] == cfg->dpp_id)
+			break;
+
+	ASSERT(z_idx < cfg->tree_cfg->num_pipes);
+	mpcc_id = cfg->tree_cfg->mpcc[z_idx];
+
+	REG_UPDATE_2(MPCC_CONTROL[mpcc_id],
+			MPCC_ALPHA_BLND_MODE, alpha_blnd_mode,
+			MPCC_ALPHA_MULTIPLIED_MODE, cfg->pre_multiplied_alpha);
+}
+
 const struct mpc_funcs dcn10_mpc_funcs = {
 		.add = mpc10_mpcc_add,
 		.remove = mpc10_mpcc_remove,
-		.wait_for_idle = mpc10_assert_idle_mpcc
+		.wait_for_idle = mpc10_assert_idle_mpcc,
+		.set_denorm = NULL,
+		.update_blend_mode = mpc10_update_blend_mode
 };
 
 void dcn10_mpc_construct(struct dcn10_mpc *mpc10,
@@ -337,3 +361,4 @@ void dcn10_mpc_construct(struct dcn10_mpc *mpc10,
 	mpc10->mpcc_in_use_mask = 0;
 	mpc10->num_mpcc = num_mpcc;
 }
+
drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h +4 -0
index 2d3de5b68e34..fb590f5353a0 100644
@@ -51,6 +51,10 @@ struct mpc_funcs {
 			int opp_id,
 			int mpcc_inst);
 	void (*wait_for_idle)(struct mpc *mpc, int id);
+	void (*set_denorm)(struct mpc *mpc,
+			int opp_id,
+			enum dc_color_depth output_depth);
+	void (*update_blend_mode)(struct mpc *mpc, struct mpcc_cfg *cfg);
 };
 
 #endif
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#22 lore raw

PATCH 21/103 drm/amd/display: moving backlight registers to hwsequencer

In reply to a message not in the archive

From: Yue Hin Lau <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/core/dc_link.c      |   2 +-
 drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c |  11 ++-
 drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h     |  47 +++++++---
 .../gpu/drm/amd/display/dc/dce/dce_link_encoder.c  |  79 +---------------
 .../gpu/drm/amd/display/dc/dce/dce_link_encoder.h  |  27 ++----
 .../amd/display/dc/dce110/dce110_hw_sequencer.c    | 103 ++++++++++++++++++++-
 .../gpu/drm/amd/display/dc/inc/hw/link_encoder.h   |   4 +-
 drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h  |   4 +
 .../amd/display/dc/virtual/virtual_link_encoder.c  |   8 +-
 9 files changed, 158 insertions(+), 127 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +1 -1
index c7751a31081a..976f0f136f29 100644
@@ -1807,7 +1807,7 @@ static void disable_link(struct dc_link *link, enum signal_type signal)
 		else
 			dp_disable_link_phy_mst(link, signal);
 	} else
-		link->link_enc->funcs->disable_output(link->link_enc, signal);
+		link->link_enc->funcs->disable_output(link->link_enc, signal, link);
 }
 
 enum dc_status dc_link_validate_mode_timing(
drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c +6 -5
index 616431e25b79..03bf79ce1f85 100644
@@ -94,7 +94,7 @@ void dp_enable_link_phy(
 						link_enc,
 						link_settings,
 						clock_source);
-			link_enc->funcs->backlight_control(link_enc, true);
+			link->dc->hwss.backlight_control(link, true);
 		} else
 			link_enc->funcs->enable_dp_output(
 						link_enc,
@@ -138,12 +138,12 @@ void dp_disable_link_phy(struct dc_link *link, enum signal_type signal)
 		dp_receiver_power_ctrl(link, false);
 
 	if (signal == SIGNAL_TYPE_EDP) {
-		link->link_enc->funcs->backlight_control(link->link_enc, false);
+		link->dc->hwss.backlight_control(link, false);
 		edp_receiver_ready_T9(link);
-		link->link_enc->funcs->disable_output(link->link_enc, signal);
+		link->link_enc->funcs->disable_output(link->link_enc, signal, link);
 		link->link_enc->funcs->power_control(link->link_enc, false);
 	} else
-		link->link_enc->funcs->disable_output(link->link_enc, signal);
+		link->link_enc->funcs->disable_output(link->link_enc, signal, link);
 
 	/* Clear current link setting.*/
 	memset(&link->cur_link_settings, 0,
@@ -286,7 +286,8 @@ void dp_retrain_link_dp_test(struct dc_link *link,
 
 			link->link_enc->funcs->disable_output(
 					link->link_enc,
-					SIGNAL_TYPE_DISPLAY_PORT);
+					SIGNAL_TYPE_DISPLAY_PORT,
+					link);
 
 			/* Clear current link setting. */
 			memset(&link->cur_link_settings, 0,
drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h +35 -12

134 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c +3 -76

109 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.h +7 -20
index 05463df7e636..c65def5f6a44 100644
@@ -69,10 +69,6 @@
 	SRI(DP_DPHY_FAST_TRAINING, DP, id), \
 	SRI(DP_SEC_CNTL1, DP, id)
 
-#define LE_EDP_REG_LIST(id)\
-	SR(LVTMA_PWRSEQ_CNTL), \
-	SR(LVTMA_PWRSEQ_STATE)
-
 #define LE_COMMON_REG_LIST(id)\
 	LE_COMMON_REG_LIST_BASE(id), \
 	SRI(DP_DPHY_BS_SR_SWAP_CNTL, DP, id), \
@@ -81,38 +77,32 @@
 
 #define LE_DCE80_REG_LIST(id)\
 	SRI(DP_DPHY_INTERNAL_CTRL, DP, id), \
-	LE_COMMON_REG_LIST_BASE(id), \
-	LE_EDP_REG_LIST(id)
+	LE_COMMON_REG_LIST_BASE(id)
 
 #define LE_DCE100_REG_LIST(id)\
 	LE_COMMON_REG_LIST_BASE(id), \
 	SRI(DP_DPHY_BS_SR_SWAP_CNTL, DP, id), \
 	SRI(DP_DPHY_INTERNAL_CTRL, DP, id), \
-	SR(DCI_MEM_PWR_STATUS), \
-	LE_EDP_REG_LIST(id)
+	SR(DCI_MEM_PWR_STATUS)
 
 #define LE_DCE110_REG_LIST(id)\
 	LE_COMMON_REG_LIST_BASE(id), \
 	SRI(DP_DPHY_BS_SR_SWAP_CNTL, DP, id), \
 	SRI(DP_DPHY_INTERNAL_CTRL, DP, id), \
 	SRI(DP_DPHY_HBR2_PATTERN_CONTROL, DP, id), \
-	SR(DCI_MEM_PWR_STATUS), \
-	LE_EDP_REG_LIST(id)
+	SR(DCI_MEM_PWR_STATUS)
 
 #define LE_DCE120_REG_LIST(id)\
 	LE_COMMON_REG_LIST_BASE(id), \
 	SRI(DP_DPHY_BS_SR_SWAP_CNTL, DP, id), \
 	SRI(DP_DPHY_HBR2_PATTERN_CONTROL, DP, id), \
-	SR(DCI_MEM_PWR_STATUS), \
-	LE_EDP_REG_LIST(id)
+	SR(DCI_MEM_PWR_STATUS)
 
 #define LE_DCN10_REG_LIST(id)\
 	LE_COMMON_REG_LIST_BASE(id), \
 	SRI(DP_DPHY_BS_SR_SWAP_CNTL, DP, id), \
 	SRI(DP_DPHY_INTERNAL_CTRL, DP, id), \
-	SRI(DP_DPHY_HBR2_PATTERN_CONTROL, DP, id), \
-	SR(DMU_MEM_PWR_CNTL), \
-	LE_EDP_REG_LIST(id)
+	SRI(DP_DPHY_HBR2_PATTERN_CONTROL, DP, id)
 
 struct dce110_link_enc_aux_registers {
 	uint32_t AUX_CONTROL;
@@ -243,7 +233,8 @@ void dce110_link_encoder_enable_dp_mst_output(
 /* disable PHY output */
 void dce110_link_encoder_disable_output(
 	struct link_encoder *link_enc,
-	enum signal_type signal);
+	enum signal_type signal,
+	struct dc_link *link);
 
 /* set DP lane settings */
 void dce110_link_encoder_dp_set_lane_settings(
@@ -259,10 +250,6 @@ void dce110_link_encoder_update_mst_stream_allocation_table(
 	struct link_encoder *enc,
 	const struct link_mst_stream_allocation_table *table);
 
-void dce110_link_encoder_edp_backlight_control(
-	struct link_encoder *enc,
-	bool enable);
-
 void dce110_link_encoder_edp_power_control(
 	struct link_encoder *enc,
 	bool power_up);
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +99 -4

139 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h +1 -3
index 961bbcc9202c..6cd6bc7d15fb 100644
@@ -111,7 +111,7 @@ struct link_encoder_funcs {
 		const struct dc_link_settings *link_settings,
 		enum clock_source_id clock_source);
 	void (*disable_output)(struct link_encoder *link_enc,
-		enum signal_type signal);
+		enum signal_type signal, struct dc_link *link);
 	void (*dp_set_lane_settings)(struct link_encoder *enc,
 		const struct link_training_settings *link_settings);
 	void (*dp_set_phy_pattern)(struct link_encoder *enc,
@@ -123,8 +123,6 @@ struct link_encoder_funcs {
 			bool exit_link_training_required);
 	void (*psr_program_secondary_packet)(struct link_encoder *enc,
 				unsigned int sdp_transmit_line_num_deadline);
-	void (*backlight_control) (struct link_encoder *enc,
-		bool enable);
 	void (*power_control) (struct link_encoder *enc,
 		bool power_up);
 	void (*connect_dig_be_to_fe)(struct link_encoder *enc,
drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h +4 -0
index 1fa2edc6cfb1..210874f37722 100644
@@ -176,6 +176,10 @@ struct hw_sequencer_funcs {
 
 	void (*ready_shared_resources)(struct dc *dc, struct dc_state *context);
 	void (*optimize_shared_resources)(struct dc *dc);
+
+	void (*backlight_control)(
+			struct dc_link *link,
+			bool enable);
 };
 
 void color_space_to_black_color(
drivers/gpu/drm/amd/display/dc/virtual/virtual_link_encoder.c +2 -6
index db513abd735a..4f405348130f 100644
@@ -58,7 +58,8 @@ static void virtual_link_encoder_enable_dp_mst_output(
 
 static void virtual_link_encoder_disable_output(
 	struct link_encoder *link_enc,
-	enum signal_type signal) {}
+	enum signal_type signal,
+	struct dc_link *link) {}
 
 static void virtual_link_encoder_dp_set_lane_settings(
 	struct link_encoder *enc,
@@ -72,10 +73,6 @@ static void virtual_link_encoder_update_mst_stream_allocation_table(
 	struct link_encoder *enc,
 	const struct link_mst_stream_allocation_table *table) {}
 
-static void virtual_link_encoder_edp_backlight_control(
-	struct link_encoder *enc,
-	bool enable) {}
-
 static void virtual_link_encoder_edp_power_control(
 	struct link_encoder *enc,
 	bool power_up) {}
@@ -105,7 +102,6 @@ static const struct link_encoder_funcs virtual_lnk_enc_funcs = {
 	.dp_set_phy_pattern = virtual_link_encoder_dp_set_phy_pattern,
 	.update_mst_stream_allocation_table =
 		virtual_link_encoder_update_mst_stream_allocation_table,
-	.backlight_control = virtual_link_encoder_edp_backlight_control,
 	.power_control = virtual_link_encoder_edp_power_control,
 	.connect_dig_be_to_fe = virtual_link_encoder_connect_dig_be_to_fe,
 	.destroy = virtual_link_encoder_destroy
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#23 lore raw

PATCH 22/103 drm/amd/display: edp backlight regression fix

In reply to a message not in the archive

From: Yue Hin Lau <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.h | 5 +++++
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c   | 1 +
 2 files changed, 6 insertions(+)
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.h +5 -0
index db6c19cd15eb..3e95f7f92c73 100644
@@ -68,5 +68,10 @@ void dce110_fill_display_configs(
 uint32_t dce110_get_min_vblank_time_us(const struct dc_state *context);
 
 void dp_receiver_power_ctrl(struct dc_link *link, bool on);
+
+void hwss_blacklight_control(
+	struct dc_link *link,
+	bool enable);
+
 #endif /* __DC_HWSS_DCE110_H__ */
 
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +1 -0
index 82a985a981be..bf5c9243ba7a 100644
@@ -2897,6 +2897,7 @@ static const struct hw_sequencer_funcs dcn10_funcs = {
 	.wait_for_mpcc_disconnect = dcn10_wait_for_mpcc_disconnect,
 	.ready_shared_resources = ready_shared_resources,
 	.optimize_shared_resources = optimize_shared_resources,
+	.backlight_control = hwss_blacklight_control
 };
 
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#24 lore raw

PATCH 23/103 drm/amd/display: fix ASSERT() caused by missing registers.

In reply to a message not in the archive

From: Ken Chalmers <[email protected]>

---
 .../drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c  | 42 ++++++++++++----------
 1 file changed, 23 insertions(+), 19 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +23 -19
index bf5c9243ba7a..cdaed0c0a70c 100644
@@ -771,14 +771,16 @@ static void power_on_plane(
 	struct dce_hwseq *hws,
 	int plane_id)
 {
-	REG_SET(DC_IP_REQUEST_CNTL, 0,
-			IP_REQUEST_EN, 1);
-	dpp_pg_control(hws, plane_id, true);
-	hubp_pg_control(hws, plane_id, true);
-	REG_SET(DC_IP_REQUEST_CNTL, 0,
-			IP_REQUEST_EN, 0);
-	dm_logger_write(hws->ctx->logger, LOG_DEBUG,
-			"Un-gated front end for pipe %d\n", plane_id);
+	if (REG(DC_IP_REQUEST_CNTL)) {
+		REG_SET(DC_IP_REQUEST_CNTL, 0,
+				IP_REQUEST_EN, 1);
+		dpp_pg_control(hws, plane_id, true);
+		hubp_pg_control(hws, plane_id, true);
+		REG_SET(DC_IP_REQUEST_CNTL, 0,
+				IP_REQUEST_EN, 0);
+		dm_logger_write(hws->ctx->logger, LOG_DEBUG,
+				"Un-gated front end for pipe %d\n", plane_id);
+	}
 }
 
 static void undo_DEGVIDCN10_253_wa(struct dc *dc)
@@ -1130,18 +1132,20 @@ static void plane_atomic_power_down(struct dc *dc, int fe_idx)
 	struct dce_hwseq *hws = dc->hwseq;
 	struct transform *xfm = dc->res_pool->transforms[fe_idx];
 
-	REG_SET(DC_IP_REQUEST_CNTL, 0,
-			IP_REQUEST_EN, 1);
-	dpp_pg_control(hws, fe_idx, false);
-	hubp_pg_control(hws, fe_idx, false);
-	xfm->funcs->transform_reset(xfm);
-	REG_SET(DC_IP_REQUEST_CNTL, 0,
-			IP_REQUEST_EN, 0);
-	dm_logger_write(dc->ctx->logger, LOG_DEBUG,
-			"Power gated front end %d\n", fe_idx);
+	if (REG(DC_IP_REQUEST_CNTL)) {
+		REG_SET(DC_IP_REQUEST_CNTL, 0,
+				IP_REQUEST_EN, 1);
+		dpp_pg_control(hws, fe_idx, false);
+		hubp_pg_control(hws, fe_idx, false);
+		xfm->funcs->transform_reset(xfm);
+		REG_SET(DC_IP_REQUEST_CNTL, 0,
+				IP_REQUEST_EN, 0);
+		dm_logger_write(dc->ctx->logger, LOG_DEBUG,
+				"Power gated front end %d\n", fe_idx);
 
-	if (dc->debug.sanity_checks)
-		verify_allow_pstate_change_high(dc->hwseq);
+		if (dc->debug.sanity_checks)
+			verify_allow_pstate_change_high(dc->hwseq);
+	}
 }
 
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#25 lore raw

PATCH 24/103 drm/amd/display: make sure BL off to mainlink off has enough time

In reply to a message not in the archive

From: Charlene Liu <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c | 7 -------
 1 file changed, 7 deletions(-)
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +0 -7
index 9fc43bcac5bc..f28fce6e6ba4 100644
@@ -819,13 +819,6 @@ void hwss_blacklight_control(
 		return;
 	}
 
-	if (!enable && !is_panel_backlight_on(hws)) {
-		dm_logger_write(ctx->logger, LOG_HW_RESUME_S3,
-				"%s: panel already powered down. Do nothing.\n",
-				__func__);
-		return;
-	}
-
 	/* Send VBIOS command to control eDP panel backlight */
 
 	dm_logger_write(ctx->logger, LOG_HW_RESUME_S3,
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#26 lore raw

PATCH 25/103 drm/amd/display: add hubp/dpp pg debug key

In reply to a message not in the archive

From: Charlene Liu <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dc.h | 2 ++
 1 file changed, 2 insertions(+)
drivers/gpu/drm/amd/display/dc/dc.h +2 -0
index 50440731f83c..2e77885fc1ea 100644
@@ -191,6 +191,8 @@ struct dc_debug {
 	bool disable_dmcu;
 	bool disable_psr;
 	bool force_abm_enable;
+	bool disable_hbup_pg;
+	bool disable_dpp_pg;
 };
 struct dc_state;
 struct resource_pool;
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#27 lore raw

PATCH 26/103 drm/amd/display: Fixed validation return wrong result.

In reply to a message not in the archive

From: Yongqiang Sun <[email protected]>

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 2 +-
 drivers/gpu/drm/amd/display/dc/core/dc_resource.c | 2 +-
 drivers/gpu/drm/amd/display/dc/dc.h               | 2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +1 -1
index af2c9d1289c0..8f9dcc697aad 100644
@@ -4918,7 +4918,7 @@ int amdgpu_dm_atomic_check(struct drm_device *dev,
 		if (ret)
 			goto fail;
 
-		if (!dc_validate_global_state(dc, dm_state->context)) {
+		if (dc_validate_global_state(dc, dm_state->context) != DC_OK) {
 			ret = -EINVAL;
 			goto fail;
 		}
drivers/gpu/drm/amd/display/dc/core/dc_resource.c +1 -1
index 77b3474a7c9e..d9a2d09a1f4f 100644
@@ -1716,7 +1716,7 @@ void dc_resource_state_construct(
 	dst_ctx->dis_clk = dc->res_pool->display_clock;
 }
 
-bool dc_validate_global_state(
+enum dc_status dc_validate_global_state(
 		struct dc *dc,
 		struct dc_state *new_ctx)
 {
drivers/gpu/drm/amd/display/dc/dc.h +1 -1
index 2e77885fc1ea..1e49b8f9c66e 100644
@@ -654,7 +654,7 @@ bool dc_validate_stream(struct dc *dc, struct dc_stream_state *stream);
 
 bool dc_validate_plane(struct dc *dc, const struct dc_plane_state *plane_state);
 
-bool dc_validate_global_state(
+enum dc_status dc_validate_global_state(
 		struct dc *dc,
 		struct dc_state *new_ctx);
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#28 lore raw

PATCH 27/103 drm/amd/display: version 3.1.02

In reply to a message not in the archive

From: Tony Cheng <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dc.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dc.h +1 -1
index 1e49b8f9c66e..12601ffc0fe4 100644
@@ -38,7 +38,7 @@
 #include "inc/compressor.h"
 #include "dml/display_mode_lib.h"
 
-#define DC_VER "3.1.01"
+#define DC_VER "3.1.02"
 
 #define MAX_SURFACES 3
 #define MAX_STREAMS 6
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#29 lore raw

PATCH 28/103 drm/amd/display: screen flickers when connected to ext monitor in clone

In reply to a message not in the archive

From: Hersen Wu <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c      | 8 ++++++++
 drivers/gpu/drm/amd/display/dc/dc.h                   | 1 +
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c | 3 +++
 3 files changed, 12 insertions(+)
drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c +8 -0
index 8ca6c3e4e65a..cc99073b7a54 100644
@@ -1018,9 +1018,17 @@ bool dcn_validate_bandwidth(
 		context->bw.dcn.calc_clk.min_active_dram_ccm_us = (int)(v->min_active_dram_clock_change_margin);
 		context->bw.dcn.calc_clk.dcfclk_deep_sleep_khz = (int)(v->dcf_clk_deep_sleep * 1000);
 		context->bw.dcn.calc_clk.dcfclk_khz = (int)(v->dcfclk * 1000);
+
 		context->bw.dcn.calc_clk.dispclk_khz = (int)(v->dispclk * 1000);
 		if (dc->debug.max_disp_clk == true)
 			context->bw.dcn.calc_clk.dispclk_khz = (int)(dc->dcn_soc->max_dispclk_vmax0p9 * 1000);
+
+		if (context->bw.dcn.calc_clk.dispclk_khz <
+				dc->debug.min_disp_clk_khz) {
+			context->bw.dcn.calc_clk.dispclk_khz =
+					dc->debug.min_disp_clk_khz;
+		}
+
 		context->bw.dcn.calc_clk.dppclk_div = (int)(v->dispclk_dppclk_ratio) == 2;
 
 		for (i = 0, input_idx = 0; i < pool->pipe_count; i++) {
drivers/gpu/drm/amd/display/dc/dc.h +1 -0
index 12601ffc0fe4..7a3a48179452 100644
@@ -178,6 +178,7 @@ struct dc_debug {
 	bool disable_pplib_wm_range;
 	bool use_dml_wm;
 	bool disable_pipe_split;
+	unsigned int min_disp_clk_khz;
 	int sr_exit_time_dpm0_ns;
 	int sr_enter_plus_exit_time_dpm0_ns;
 	int sr_exit_time_ns;
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +3 -0
index 4f124dc303dc..82c82b6310ca 100644
@@ -420,6 +420,9 @@ static const struct dc_debug debug_defaults_drv = {
 		.force_abm_enable = false,
 		.timing_trace = false,
 		.clock_trace = true,
+
+		.min_disp_clk_khz = 300000,
+
 		.disable_pplib_clock_request = true,
 		.disable_pplib_wm_range = false,
 #if defined(CONFIG_DRM_AMD_DC_DCN1_0)
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#30 lore raw

PATCH 29/103 drm/amd/display: Fixed incorrect return value for validaton

In reply to a message not in the archive

From: Yongqiang Sun <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/core/dc_resource.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_resource.c +5 -4
index d9a2d09a1f4f..bbfec7cb2ad1 100644
@@ -1723,10 +1723,11 @@ enum dc_status dc_validate_global_state(
 	enum dc_status result = DC_ERROR_UNEXPECTED;
 	int i, j;
 
-	if (dc->res_pool->funcs->validate_global &&
-			dc->res_pool->funcs->validate_global(
-			dc, new_ctx) != DC_OK)
-		return false;
+	if (dc->res_pool->funcs->validate_global) {
+			result = dc->res_pool->funcs->validate_global(dc, new_ctx);
+			if (result != DC_OK)
+				return result;
+	}
 
 	for (i = 0; new_ctx && i < new_ctx->stream_count; i++) {
 		struct dc_stream_state *stream = new_ctx->streams[i];
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#31 lore raw

PATCH 30/103 drm/amd/display: Use active + border for bw validation

In reply to a message not in the archive

From: Eric Yang <[email protected]>

When doing SLS, KMD gives us clipped v_addressable with border. This results in bw validation failure.

---
 drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c +5 -3
index cc99073b7a54..319450d9cfc1 100644
@@ -364,7 +364,8 @@ static void pipe_ctx_to_e2e_pipe_params (
 	}
 
 
-	input->dest.vactive        = pipe->stream->timing.v_addressable;
+	input->dest.vactive        = pipe->stream->timing.v_addressable + pipe->stream->timing.v_border_top
+			+ pipe->stream->timing.v_border_bottom;
 
 	input->dest.recout_width   = pipe->plane_res.scl_data.recout.width;
 	input->dest.recout_height  = pipe->plane_res.scl_data.recout.height;
@@ -882,10 +883,11 @@ bool dcn_validate_bandwidth(
 
 		v->htotal[input_idx] = pipe->stream->timing.h_total;
 		v->vtotal[input_idx] = pipe->stream->timing.v_total;
+		v->vactive[input_idx] = pipe->stream->timing.v_addressable +
+				pipe->stream->timing.v_border_top + pipe->stream->timing.v_border_bottom;
 		v->v_sync_plus_back_porch[input_idx] = pipe->stream->timing.v_total
-				- pipe->stream->timing.v_addressable
+				- v->vactive[input_idx]
 				- pipe->stream->timing.v_front_porch;
-		v->vactive[input_idx] = pipe->stream->timing.v_addressable;
 		v->pixel_clock[input_idx] = pipe->stream->timing.pix_clk_khz / 1000.0f;
 
 		if (!pipe->plane_state) {
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#32 lore raw

PATCH 31/103 drm/amd/display: update link type to mst before topology discovery

In reply to a message not in the archive

From: Wenjing Liu <[email protected]>

[Description]
link type is not updated before mst topology discovery.
This causes issue when branch device response to link address after before the start topology discovery event finishes.

[Solution]
update link type to mst before topology discovery

---
 drivers/gpu/drm/amd/display/dc/core/dc_link.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +3 -3
index 976f0f136f29..a58e61b6e9f9 100644
@@ -496,6 +496,7 @@ static void detect_dp(
 		}
 		if (is_mst_supported(link)) {
 			sink_caps->signal = SIGNAL_TYPE_DISPLAY_PORT_MST;
+			link->type = dc_connection_mst_branch;
 
 			/*
 			 * This call will initiate MST topology discovery. Which
@@ -524,12 +525,11 @@ static void detect_dp(
 			if (reason == DETECT_REASON_BOOT)
 				boot = true;
 
-			if (dm_helpers_dp_mst_start_top_mgr(
+			if (!dm_helpers_dp_mst_start_top_mgr(
 				link->ctx,
 				link, boot)) {
-				link->type = dc_connection_mst_branch;
-			} else {
 				/* MST not supported */
+				link->type = dc_connection_single;
 				sink_caps->signal = SIGNAL_TYPE_DISPLAY_PORT;
 			}
 		}
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#33 lore raw

PATCH 32/103 drm/amd/display: Program stream's csc matrix instead of using default

In reply to a message not in the archive

From: SivapiriyanKumarasamy <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +3 -4
index cdaed0c0a70c..4e4f20ba5643 100644
@@ -2259,7 +2259,6 @@ static void update_dchubp_dpp(
 	struct transform *xfm = pipe_ctx->plane_res.xfm;
 	struct dc_plane_state *plane_state = pipe_ctx->plane_state;
 	union plane_size size = plane_state->plane_size;
-	struct default_adjustment ocsc = {0};
 	struct mpcc_cfg mpcc_cfg = {0};
 	struct pipe_ctx *top_pipe;
 	bool per_pixel_alpha = plane_state->per_pixel_alpha && pipe_ctx->bottom_pipe;
@@ -2338,9 +2337,9 @@ static void update_dchubp_dpp(
 	/*gamut remap*/
 	program_gamut_remap(pipe_ctx);
 
-	/*TODO add adjustments parameters*/
-	ocsc.out_color_space = pipe_ctx->stream->output_color_space;
-	pipe_ctx->plane_res.xfm->funcs->opp_set_csc_default(pipe_ctx->plane_res.xfm, &ocsc);
+	program_csc_matrix(pipe_ctx,
+			pipe_ctx->stream->output_color_space,
+			pipe_ctx->stream->csc_color_matrix.matrix);
 
 	mi->funcs->mem_input_program_surface_config(
 		mi,
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#34 lore raw

PATCH 33/103 drm/amd/display: Early return when turn off a plane.

In reply to a message not in the archive

From: Yongqiang Sun <[email protected]>

In case of two monitor connected and turn off one of the monitors, OTG0 is locked after graphic plane off due to redundant programming front end regs.

---
 drivers/gpu/drm/amd/display/dc/core/dc.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +7 -1
index 2cea8de9c9be..ffc36f5ac50c 100644
@@ -1190,8 +1190,14 @@ static void commit_planes_for_stream(struct dc *dc,
 		}
 	}
 
-	if (surface_count == 0)
+	if (surface_count == 0) {
+		/*
+		 * In case of turning off screen, no need to program front end a second time.
+		 * just return after program front end.
+		 */
 		dc->hwss.apply_ctx_for_surface(dc, stream, surface_count, context);
+		return;
+	}
 
 	/* Lock pipes for provided surfaces, or all active if full update*/
 	for (i = 0; i < surface_count; i++) {
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#35 lore raw

PATCH 34/103 drm/amd/display: Only reset top pipe back end.

In reply to a message not in the archive

From: Yongqiang Sun <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +3 -0
index 4e4f20ba5643..014911ebdbd6 100644
@@ -1270,6 +1270,9 @@ static void reset_hw_ctx_wrap(
 		if (!pipe_ctx_old->stream)
 			continue;
 
+		if (pipe_ctx_old->top_pipe)
+			continue;
+
 		if (!pipe_ctx->stream ||
 				pipe_need_reprogram(pipe_ctx_old, pipe_ctx)) {
 			struct clock_source *old_clk = pipe_ctx_old->clock_source;
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#36 lore raw

PATCH 35/103 drm/amd/display: Fix MST daisy chain SST not light up

In reply to a message not in the archive

From: Jerry Zuo <[email protected]>

In SST daisy chain scenario, edid is getting read in mst hotplug routine. It is getting conflict with drm send_enum_path_resources kernel thread in terms of i2c bus which is getting locked up in such case.

Have edid being read in get_mode hook, instead of in hotplug routine.

---
 .../amd/display/amdgpu_dm/amdgpu_dm_mst_types.c    | 110 +++++++++------------
 1 file changed, 49 insertions(+), 61 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c +49 -61
index 343645e4070b..11a10c1c0adb 100644
@@ -172,14 +172,60 @@ static const struct drm_connector_funcs dm_dp_mst_connector_funcs = {
 	.atomic_get_property = amdgpu_dm_connector_atomic_get_property
 };
 
+static int dm_connector_update_modes(struct drm_connector *connector,
+				struct edid *edid)
+{
+	int ret;
+
+	ret = drm_add_edid_modes(connector, edid);
+	drm_edid_to_eld(connector, edid);
+
+	return ret;
+}
+
 static int dm_dp_mst_get_modes(struct drm_connector *connector)
 {
 	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
 	int ret = 0;
 
-	ret = drm_add_edid_modes(&aconnector->base, aconnector->edid);
+	if (!aconnector)
+		return dm_connector_update_modes(connector, NULL);
+
+	if (!aconnector->edid) {
+		struct edid *edid;
+		struct dc_sink *dc_sink;
+		struct dc_sink_init_data init_params = {
+				.link = aconnector->dc_link,
+				.sink_signal = SIGNAL_TYPE_DISPLAY_PORT_MST };
+		edid = drm_dp_mst_get_edid(connector, &aconnector->mst_port->mst_mgr, aconnector->port);
 
-	drm_edid_to_eld(&aconnector->base, aconnector->edid);
+		if (!edid) {
+			drm_mode_connector_update_edid_property(
+				&aconnector->base,
+				NULL);
+			return ret;
+		}
+
+		aconnector->edid = edid;
+
+		dc_sink = dc_link_add_remote_sink(
+			aconnector->dc_link,
+			(uint8_t *)edid,
+			(edid->extensions + 1) * EDID_LENGTH,
+			&init_params);
+
+		dc_sink->priv = aconnector;
+		aconnector->dc_sink = dc_sink;
+
+		if (aconnector->dc_sink)
+			amdgpu_dm_add_sink_to_freesync_module(
+					connector, edid);
+
+		drm_mode_connector_update_edid_property(
+						&aconnector->base, edid);
+	}
+
+	ret = dm_connector_update_modes(connector, aconnector->edid);
 
 	return ret;
 }
@@ -341,66 +387,8 @@ static void dm_dp_mst_hotplug(struct drm_dp_mst_topology_mgr *mgr)
 {
 	struct amdgpu_dm_connector *master = container_of(mgr, struct amdgpu_dm_connector, mst_mgr);
 	struct drm_device *dev = master->base.dev;
-	struct amdgpu_device *adev = dev->dev_private;
-	struct drm_connector *connector;
-	struct amdgpu_dm_connector *aconnector;
-	struct edid *edid;
-	struct dc_sink *dc_sink;
-
-	drm_modeset_lock_all(dev);
-	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
-		aconnector = to_amdgpu_dm_connector(connector);
-		if (aconnector->port &&
-				aconnector->port->pdt != DP_PEER_DEVICE_NONE &&
-				aconnector->port->pdt != DP_PEER_DEVICE_MST_BRANCHING &&
-				!aconnector->dc_sink) {
-			/*
-			 * This is plug in case, where port has been created but
-			 * sink hasn't been created yet
-			 */
-			if (!aconnector->edid) {
-				struct dc_sink_init_data init_params = {
-						.link = aconnector->dc_link,
-						.sink_signal = SIGNAL_TYPE_DISPLAY_PORT_MST};
-				edid = drm_dp_mst_get_edid(connector, &aconnector->mst_port->mst_mgr, aconnector->port);
-
-				if (!edid) {
-					drm_mode_connector_update_edid_property(
-						&aconnector->base,
-						NULL);
-					continue;
-				}
-
-				aconnector->edid = edid;
-
-				dc_sink = dc_link_add_remote_sink(
-					aconnector->dc_link,
-					(uint8_t *)edid,
-					(edid->extensions + 1) * EDID_LENGTH,
-					&init_params);
-
-				dc_sink->priv = aconnector;
-				aconnector->dc_sink = dc_sink;
-
-				if (aconnector->dc_sink)
-					amdgpu_dm_add_sink_to_freesync_module(
-							connector,
-							edid);
-
-				dm_restore_drm_connector_state(connector->dev, connector);
-			} else
-				edid = aconnector->edid;
-
-			DRM_DEBUG_KMS("edid retrieved %p\n", edid);
-
-			drm_mode_connector_update_edid_property(
-				&aconnector->base,
-				aconnector->edid);
-		}
-	}
-	drm_modeset_unlock_all(dev);
 
-	schedule_work(&adev->dm.mst_hotplug_work);
+	drm_kms_helper_hotplug_event(dev);
 }
 
 static void dm_dp_mst_register_connector(struct drm_connector *connector)
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#37 lore raw

PATCH 36/103 drm/amd/display: temp disable DCC on high res.

In reply to a message not in the archive

From: Charlene Liu <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +9 -2
index 82c82b6310ca..dd3ff5114d42 100644
@@ -1133,20 +1133,27 @@ static bool get_dcc_compression_cap(const struct dc *dc,
 		output->grph.rgb.max_uncompressed_blk_size = 256;
 		output->grph.rgb.max_compressed_blk_size = 256;
 		output->grph.rgb.independent_64b_blks = false;
+		output->capable = true;
+		output->const_color_support = false;
 		break;
 	case dcc_control__128_128_xxx:
 		output->grph.rgb.max_uncompressed_blk_size = 128;
 		output->grph.rgb.max_compressed_blk_size = 128;
 		output->grph.rgb.independent_64b_blks = false;
+		/*temp: not allow dcc on high res*/
+		output->capable = false;
+		output->const_color_support = false;
 		break;
 	case dcc_control__256_64_64:
 		output->grph.rgb.max_uncompressed_blk_size = 256;
 		output->grph.rgb.max_compressed_blk_size = 64;
 		output->grph.rgb.independent_64b_blks = true;
+		/*temp: not allow dcc on high res*/
+		output->capable = false;
+		output->const_color_support = false;
 		break;
 	}
-	output->capable = true;
-	output->const_color_support = false;
+
 
 	return true;
 }
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#38 lore raw

PATCH 37/103 drm/amd/display: enable optional pipe split for single display

In reply to a message not in the archive

From: Tony Cheng <[email protected]>

also refactor debug option. now pipe_split_policy are
dynamic = no hack around dcn_calcs. will split based on HW recommendation avoid = avoid split if we can support the config with higher voltage avoid_multi_display = allow split with single display output.

force_single_disp_pipe_split
force single display to pipe split to improve stutter efficiency by using DET buffers using 2 HUBP.

---
 drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c   | 44 ++++++++++++++++++++--
 drivers/gpu/drm/amd/display/dc/dc.h                |  9 ++++-
 .../gpu/drm/amd/display/dc/dcn10/dcn10_resource.c  |  8 +---
 3 files changed, 51 insertions(+), 10 deletions(-)
drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c +41 -3
index 319450d9cfc1..aa56243f2522 100644
@@ -720,6 +720,46 @@ static bool dcn_bw_apply_registry_override(struct dc *dc)
 	return updated;
 }
 
+void hack_disable_optional_pipe_split(struct dcn_bw_internal_vars *v)
+{
+	/*
+	 * disable optional pipe split by lower dispclk bounding box
+	 * at DPM0
+	 */
+	v->max_dispclk[0] = v->max_dppclk_vmin0p65;
+}
+
+void hack_force_pipe_split(struct dcn_bw_internal_vars *v,
+		unsigned int pixel_rate_khz)
+{
+	/*
+	 * force enabling pipe split by lower dpp clock for DPM0 to just
+	 * below the specify pixel_rate, so bw calc would split pipe.
+	 */
+	v->max_dppclk[0] = pixel_rate_khz / 1000;
+}
+
+void hack_bounding_box(struct dcn_bw_internal_vars *v,
+		struct dc_debug *dbg,
+		struct dc_state *context)
+{
+	if (dbg->pipe_split_policy == MPC_SPLIT_AVOID) {
+		hack_disable_optional_pipe_split(v);
+	}
+
+	if (dbg->pipe_split_policy == MPC_SPLIT_AVOID_MULT_DISP &&
+		context->stream_count >= 2) {
+		hack_disable_optional_pipe_split(v);
+	}
+
+	if (context->stream_count == 1 &&
+			dbg->force_single_disp_pipe_split) {
+		struct dc_stream_state *stream0 = context->streams[0];
+
+		hack_force_pipe_split(v, stream0->timing.pix_clk_khz);
+	}
+}
+
 bool dcn_validate_bandwidth(
 		struct dc *dc,
 		struct dc_state *context)
@@ -851,9 +891,7 @@ bool dcn_validate_bandwidth(
 	v->phyclk_per_state[1] = v->phyclkv_mid0p72;
 	v->phyclk_per_state[0] = v->phyclkv_min0p65;
 
-	if (dc->debug.disable_pipe_split) {
-		v->max_dispclk[0] = v->max_dppclk_vmin0p65;
-	}
+	hack_bounding_box(v, &dc->debug, context);
 
 	if (v->voltage_override == dcn_bw_v_max0p9) {
 		v->voltage_override_level = number_of_states - 1;
drivers/gpu/drm/amd/display/dc/dc.h +8 -1
index 7a3a48179452..622488efa9e6 100644
@@ -162,6 +162,12 @@ struct dc_config {
 	bool disable_disp_pll_sharing;
 };
 
+enum pipe_split_policy {
+	MPC_SPLIT_DYNAMIC = 0,
+	MPC_SPLIT_AVOID = 1,
+	MPC_SPLIT_AVOID_MULT_DISP = 2,
+};
+
 struct dc_debug {
 	bool surface_visual_confirm;
 	bool sanity_checks;
@@ -177,7 +183,8 @@ struct dc_debug {
 	bool disable_hubp_power_gate;
 	bool disable_pplib_wm_range;
 	bool use_dml_wm;
-	bool disable_pipe_split;
+	enum pipe_split_policy pipe_split_policy;
+	bool force_single_disp_pipe_split;
 	unsigned int min_disp_clk_khz;
 	int sr_exit_time_dpm0_ns;
 	int sr_enter_plus_exit_time_dpm0_ns;
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +2 -6
index dd3ff5114d42..bd7c86e5800e 100644
@@ -425,10 +425,9 @@ static const struct dc_debug debug_defaults_drv = {
 
 		.disable_pplib_clock_request = true,
 		.disable_pplib_wm_range = false,
-#if defined(CONFIG_DRM_AMD_DC_DCN1_0)
 		.use_dml_wm = false,
-		.disable_pipe_split = true
-#endif
+
+		.pipe_split_policy = MPC_SPLIT_AVOID_MULT_DISP,
 };
 
 static const struct dc_debug debug_defaults_diags = {
@@ -437,12 +436,9 @@ static const struct dc_debug debug_defaults_diags = {
 		.timing_trace = true,
 		.clock_trace = true,
 		.disable_stutter = true,
-#if defined(CONFIG_DRM_AMD_DC_DCN1_0)
 		.disable_pplib_clock_request = true,
 		.disable_pplib_wm_range = true,
 		.use_dml_wm = false,
-		.disable_pipe_split = false
-#endif
 };
 
 static void dcn10_dpp_destroy(struct transform **xfm)
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#39 lore raw

PATCH 38/103 drm/amd/display: add option to disable DCC for DCC 128b request

In reply to a message not in the archive

From: Tony Cheng <[email protected]>

1. reverts commit e67f51012740 ("dc: temp disable DCC on high res.")
- default still DCC enabled

2. add debug options to decide how DCC is disabled
- disable DCC
- disable DCC if DCC requires 128b (aka. half) request
-- observed compressed data corruption result in screen corruption in full (256b) request while half (128b) would cause DCN to hang, result in DF hang

---
 drivers/gpu/drm/amd/display/dc/dc.h                   | 11 ++++++++---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c | 18 ++++++++----------
 2 files changed, 16 insertions(+), 13 deletions(-)
drivers/gpu/drm/amd/display/dc/dc.h +8 -3
index 622488efa9e6..d7afbcfb33ce 100644
@@ -56,11 +56,10 @@ struct dc_caps {
 	uint32_t max_planes;
 	uint32_t max_downscale_ratio;
 	uint32_t i2c_speed_in_khz;
-
 	unsigned int max_cursor_size;
+	bool dcc_const_color;
 };
 
-
 struct dc_dcc_surface_param {
 	struct dc_size surface_size;
 	enum surface_pixel_format format;
@@ -162,6 +161,12 @@ struct dc_config {
 	bool disable_disp_pll_sharing;
 };
 
+enum dcc_option {
+	DCC_ENABLE = 0,
+	DCC_DISABLE = 1,
+	DCC_HALF_REQ_DISALBE = 2,
+};
+
 enum pipe_split_policy {
 	MPC_SPLIT_DYNAMIC = 0,
 	MPC_SPLIT_AVOID = 1,
@@ -177,7 +182,7 @@ struct dc_debug {
 	bool clock_trace;
 	bool validation_trace;
 	bool disable_stutter;
-	bool disable_dcc;
+	enum dcc_option disable_dcc;
 	bool disable_dfs_bypass;
 	bool disable_dpp_power_gate;
 	bool disable_hubp_power_gate;
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +8 -10
index bd7c86e5800e..30c275e05796 100644
@@ -414,7 +414,6 @@ static const struct resource_caps res_cap = {
 };
 
 static const struct dc_debug debug_defaults_drv = {
-		.disable_dcc = false,
 		.sanity_checks = true,
 		.disable_dmcu = true,
 		.force_abm_enable = false,
@@ -428,6 +427,7 @@ static const struct dc_debug debug_defaults_drv = {
 		.use_dml_wm = false,
 
 		.pipe_split_policy = MPC_SPLIT_AVOID_MULT_DISP,
+		.disable_dcc = DCC_ENABLE,
 };
 
 static const struct dc_debug debug_defaults_diags = {
@@ -1080,7 +1080,7 @@ static bool get_dcc_compression_cap(const struct dc *dc,
 
 	memset(output, 0, sizeof(*output));
 
-	if (dc->debug.disable_dcc)
+	if (dc->debug.disable_dcc == DCC_DISABLE)
 		return false;
 
 	if (!dcc_support_pixel_format(input->format,
@@ -1124,32 +1124,30 @@ static bool get_dcc_compression_cap(const struct dc *dc,
 			dcc_control = dcc_control__128_128_xxx;
 	}
 
+	if (dc->debug.disable_dcc == DCC_HALF_REQ_DISALBE &&
+		dcc_control != dcc_control__256_256_xxx)
+		return false;
+
 	switch (dcc_control) {
 	case dcc_control__256_256_xxx:
 		output->grph.rgb.max_uncompressed_blk_size = 256;
 		output->grph.rgb.max_compressed_blk_size = 256;
 		output->grph.rgb.independent_64b_blks = false;
-		output->capable = true;
-		output->const_color_support = false;
 		break;
 	case dcc_control__128_128_xxx:
 		output->grph.rgb.max_uncompressed_blk_size = 128;
 		output->grph.rgb.max_compressed_blk_size = 128;
 		output->grph.rgb.independent_64b_blks = false;
-		/*temp: not allow dcc on high res*/
-		output->capable = false;
-		output->const_color_support = false;
 		break;
 	case dcc_control__256_64_64:
 		output->grph.rgb.max_uncompressed_blk_size = 256;
 		output->grph.rgb.max_compressed_blk_size = 64;
 		output->grph.rgb.independent_64b_blks = true;
-		/*temp: not allow dcc on high res*/
-		output->capable = false;
-		output->const_color_support = false;
 		break;
 	}
 
+	output->capable = true;
+	output->const_color_support = false;
 
 	return true;
 }
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#40 lore raw

PATCH 39/103 drm/amd/display: option to maximize lb usage

In reply to a message not in the archive

From: Tony Cheng <[email protected]>

experimental change for testing if max line buffer result in better stutter efficiency

for 1080p, LB can hold up to 9 line at 10bpcc, potentially add 10 line time of latency hiding.

---
 drivers/gpu/drm/amd/display/dc/dc.h                   | 8 ++++++--
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_dscl.c | 3 +++
 2 files changed, 9 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/dc/dc.h +6 -2
index d7afbcfb33ce..427ccbc28f51 100644
@@ -181,15 +181,19 @@ struct dc_debug {
 	bool timing_trace;
 	bool clock_trace;
 	bool validation_trace;
+
+	/* stutter efficiency related */
 	bool disable_stutter;
+	bool use_max_lb;
 	enum dcc_option disable_dcc;
+	enum pipe_split_policy pipe_split_policy;
+	bool force_single_disp_pipe_split;
+
 	bool disable_dfs_bypass;
 	bool disable_dpp_power_gate;
 	bool disable_hubp_power_gate;
 	bool disable_pplib_wm_range;
 	bool use_dml_wm;
-	enum pipe_split_policy pipe_split_policy;
-	bool force_single_disp_pipe_split;
 	unsigned int min_disp_clk_khz;
 	int sr_exit_time_dpm0_ns;
 	int sr_enter_plus_exit_time_dpm0_ns;
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_dscl.c +3 -0
index 82b8887d4973..fa2ab9952388 100644
@@ -507,6 +507,9 @@ static enum lb_memory_config find_lb_memory_config(struct dcn10_dpp *xfm,
 {
 	enum lb_memory_config mem_cfg = LB_MEMORY_CONFIG_0;
 
+	if (xfm->base.ctx->dc->debug.use_max_lb)
+		return mem_cfg;
+
 	if (xfm->tf_mask->PIXEL_DEPTH) {
 		mem_cfg = dpp10_find_lb_memory_config(scl_data);
 	}
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#41 lore raw

PATCH 40/103 drm/amd/display: fix bug in force_single_disp_pipe_split

In reply to a message not in the archive

From: Tony Cheng <[email protected]>

should only lower dpp clock.

---
 drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c +4 -1
index aa56243f2522..cf474eb1cde8 100644
@@ -732,11 +732,14 @@ void hack_disable_optional_pipe_split(struct dcn_bw_internal_vars *v)
 void hack_force_pipe_split(struct dcn_bw_internal_vars *v,
 		unsigned int pixel_rate_khz)
 {
+	float pixel_rate_mhz = pixel_rate_khz / 1000;
+
 	/*
 	 * force enabling pipe split by lower dpp clock for DPM0 to just
 	 * below the specify pixel_rate, so bw calc would split pipe.
 	 */
-	v->max_dppclk[0] = pixel_rate_khz / 1000;
+	if (pixel_rate_mhz < v->max_dppclk[0])
+		v->max_dppclk[0] = pixel_rate_mhz;
 }
 
 void hack_bounding_box(struct dcn_bw_internal_vars *v,
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#42 lore raw

PATCH 41/103 drm/amd/display: dal 3.1.03

In reply to a message not in the archive

From: Tony Cheng <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dc.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dc.h +1 -1
index 427ccbc28f51..cf9cd108539f 100644
@@ -38,7 +38,7 @@
 #include "inc/compressor.h"
 #include "dml/display_mode_lib.h"
 
-#define DC_VER "3.1.02"
+#define DC_VER "3.1.03"
 
 #define MAX_SURFACES 3
 #define MAX_STREAMS 6
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#43 lore raw

PATCH 42/103 drm/amd/display: Fix ref_count messed up issue

In reply to a message not in the archive

From: Jerry Zuo <[email protected]>

In the full update type, need to add ref_count to the newly created dc_state->stream. It made mistake to add ref_count to dc->current_state->stream which keeps adding up without release.

---
 drivers/gpu/drm/amd/display/dc/core/dc.c | 21 ++++++++++++++++++++-
 1 file changed, 20 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +20 -1
index ffc36f5ac50c..3439bc46a4a2 100644
@@ -1322,6 +1322,7 @@ void dc_commit_updates_for_stream(struct dc *dc,
 	const struct dc_stream_status *stream_status;
 	enum surface_update_type update_type;
 	struct dc_state *context;
+	struct dc_context *dc_ctx = dc->ctx;
 	int i;
 
 	stream_status = dc_stream_get_status(stream);
@@ -1334,8 +1335,17 @@ void dc_commit_updates_for_stream(struct dc *dc,
 		update_surface_trace(dc, srf_updates, surface_count);
 
 
-	if (update_type >= UPDATE_TYPE_FULL)
+	if (update_type >= UPDATE_TYPE_FULL) {
+
+		/* initialize scratch memory for building context */
+		context = dc_create_state();
+		if (context == NULL) {
+			DC_ERROR("Failed to allocate new validate context!\n");
+			return;
+		}
+
 		dc_resource_state_copy_construct(state, context);
+	}
 
 
 	for (i = 0; i < surface_count; i++) {
@@ -1361,6 +1371,15 @@ void dc_commit_updates_for_stream(struct dc *dc,
 
 	dc_post_update_surfaces_to_stream(dc);
 
+	if (dc->current_state != context) {
+
+		struct dc_state *old = dc->current_state;
+
+		dc->current_state = context;
+		dc_release_state(old);
+
+	}
+
 	return;
 
 }
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#44 lore raw

PATCH 43/103 drm/amd/display: moving cursor functions from ipp to mem_input

In reply to a message not in the archive

From: Yue Hin Lau <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/core/dc_stream.c    |  32 +++-
 drivers/gpu/drm/amd/display/dc/dce/dce_ipp.c       |   2 +
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c   |  51 ++++++
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.h   |  15 +-
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_ipp.c   | 189 ---------------------
 .../gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.c | 163 ++++++++++++++++++
 .../gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.h |  57 ++++++-
 drivers/gpu/drm/amd/display/dc/inc/hw/mem_input.h  |  10 ++
 drivers/gpu/drm/amd/display/dc/inc/hw/transform.h  |  11 ++
 9 files changed, 329 insertions(+), 201 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_stream.c +27 -5
index 23df7bc020d2..c19b478bcdc7 100644
@@ -195,13 +195,23 @@ bool dc_stream_set_cursor_attributes(
 	for (i = 0; i < MAX_PIPES; i++) {
 		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
 
-		if (pipe_ctx->stream != stream || !pipe_ctx->plane_res.ipp)
+		if (pipe_ctx->stream != stream || !pipe_ctx->plane_res.mi || !pipe_ctx->plane_res.xfm)
 			continue;
 		if (pipe_ctx->top_pipe && pipe_ctx->plane_state != pipe_ctx->top_pipe->plane_state)
 			continue;
 
-		pipe_ctx->plane_res.ipp->funcs->ipp_cursor_set_attributes(
-				pipe_ctx->plane_res.ipp, attributes);
+
+		if (pipe_ctx->plane_res.ipp->funcs->ipp_cursor_set_attributes != NULL)
+			pipe_ctx->plane_res.ipp->funcs->ipp_cursor_set_attributes(
+						pipe_ctx->plane_res.ipp, attributes);
+
+		if (pipe_ctx->plane_res.mi->funcs->set_cursor_attributes != NULL)
+			pipe_ctx->plane_res.mi->funcs->set_cursor_attributes(
+					pipe_ctx->plane_res.mi, attributes);
+
+		if (pipe_ctx->plane_res.xfm->funcs->set_cursor_attributes != NULL)
+			pipe_ctx->plane_res.xfm->funcs->set_cursor_attributes(
+				pipe_ctx->plane_res.xfm, attributes);
 	}
 
 	return true;
@@ -231,6 +241,8 @@ bool dc_stream_set_cursor_position(
 	for (i = 0; i < MAX_PIPES; i++) {
 		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
 		struct input_pixel_processor *ipp = pipe_ctx->plane_res.ipp;
+		struct mem_input *mi = pipe_ctx->plane_res.mi;
+		struct transform *xfm = pipe_ctx->plane_res.xfm;
 		struct dc_cursor_position pos_cpy = *position;
 		struct dc_cursor_mi_param param = {
 			.pixel_clk_khz = stream->timing.pix_clk_khz,
@@ -241,7 +253,9 @@ bool dc_stream_set_cursor_position(
 		};
 
 		if (pipe_ctx->stream != stream ||
-				!pipe_ctx->plane_res.ipp || !pipe_ctx->plane_state)
+				!pipe_ctx->plane_res.mi ||
+				!pipe_ctx->plane_state ||
+				!pipe_ctx->plane_res.xfm)
 			continue;
 
 		if (pipe_ctx->plane_state->address.type
@@ -251,7 +265,15 @@ bool dc_stream_set_cursor_position(
 		if (pipe_ctx->top_pipe && pipe_ctx->plane_state != pipe_ctx->top_pipe->plane_state)
 			pos_cpy.enable = false;
 
-		ipp->funcs->ipp_cursor_set_position(ipp, &pos_cpy, &param);
+
+		if (ipp->funcs->ipp_cursor_set_position != NULL)
+			ipp->funcs->ipp_cursor_set_position(ipp, &pos_cpy, &param);
+
+		if (mi->funcs->set_cursor_attributes != NULL)
+			mi->funcs->set_cursor_position(mi, &pos_cpy, &param);
+
+		if (xfm->funcs->set_cursor_attributes != NULL)
+			xfm->funcs->set_cursor_position(xfm, &pos_cpy, &param, mi->curs_attr.width);
 	}
 
 	return true;
drivers/gpu/drm/amd/display/dc/dce/dce_ipp.c +2 -0
index fa481d481132..d618fdd0cc82 100644
@@ -37,6 +37,7 @@
 #define CTX \
 	ipp_dce->base.ctx
 
+
 static void dce_ipp_cursor_set_position(
 	struct input_pixel_processor *ipp,
 	const struct dc_cursor_position *position,
@@ -133,6 +134,7 @@ static void dce_ipp_cursor_set_attributes(
 	REG_UPDATE(CUR_UPDATE, CURSOR_UPDATE_LOCK, false);
 }
 
+
 static void dce_ipp_program_prescale(
 	struct input_pixel_processor *ipp,
 	struct ipp_prescale_params *params)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c +51 -0

63 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.h +12 -3
index a1f6b01a2eb4..7fecdb19761d 100644
@@ -112,7 +112,10 @@
 	SRI(CM_DGAM_CONTROL, CM, id), \
 	SRI(FORMAT_CONTROL, CNVC_CFG, id), \
 	SRI(CNVC_SURFACE_PIXEL_FORMAT, CNVC_CFG, id), \
-	SRI(CURSOR0_CONTROL, CNVC_CUR, id)
+	SRI(CURSOR0_CONTROL, CNVC_CUR, id), \
+	SRI(CURSOR0_CONTROL, CNVC_CUR, id), \
+	SRI(CURSOR0_COLOR0, CNVC_CUR, id), \
+	SRI(CURSOR0_COLOR1, CNVC_CUR, id)
 
 
 
@@ -302,7 +305,9 @@
 	TF_SF(CNVC_CFG0_CNVC_SURFACE_PIXEL_FORMAT, CNVC_SURFACE_PIXEL_FORMAT, mask_sh), \
 	TF_SF(CNVC_CUR0_CURSOR0_CONTROL, CUR0_MODE, mask_sh), \
 	TF_SF(CNVC_CUR0_CURSOR0_CONTROL, CUR0_EXPANSION_MODE, mask_sh), \
-	TF_SF(CNVC_CUR0_CURSOR0_CONTROL, CUR0_ENABLE, mask_sh)
+	TF_SF(CNVC_CUR0_CURSOR0_CONTROL, CUR0_ENABLE, mask_sh), \
+	TF_SF(CNVC_CUR0_CURSOR0_COLOR0, CUR0_COLOR0, mask_sh), \
+	TF_SF(CNVC_CUR0_CURSOR0_COLOR1, CUR0_COLOR1, mask_sh)
 
 #define TF_REG_LIST_SH_MASK_DCN10(mask_sh)\
 	TF_REG_LIST_SH_MASK_DCN(mask_sh),\
@@ -989,7 +994,9 @@
 	type CUR0_EXPANSION_MODE; \
 	type CUR0_ENABLE; \
 	type CM_BYPASS; \
-	type FORMAT_CONTROL__ALPHA_EN
+	type FORMAT_CONTROL__ALPHA_EN; \
+	type CUR0_COLOR0; \
+	type CUR0_COLOR1
 
 
 
@@ -1237,6 +1244,8 @@ struct dcn_dpp_registers {
 	uint32_t CNVC_SURFACE_PIXEL_FORMAT;
 	uint32_t CURSOR_CONTROL;
 	uint32_t CURSOR0_CONTROL;
+	uint32_t CURSOR0_COLOR0;
+	uint32_t CURSOR0_COLOR1;
 };
 
 struct dcn10_dpp {
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_ipp.c +0 -189

201 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.c +163 -0

175 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.h +53 -4

87 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/inc/hw/mem_input.h +10 -0
index 6cef9ad0af91..5c8e45ba8f45 100644
@@ -72,6 +72,7 @@ struct mem_input {
 	int opp_id;
 	int mpcc_id;
 	struct stutter_modes stutter_mode;
+	struct dc_cursor_attributes curs_attr;
 };
 
 struct vm_system_aperture_param {
@@ -163,6 +164,15 @@ struct mem_input_funcs {
 	void (*set_blank)(struct mem_input *mi, bool blank);
 	void (*set_hubp_blank_en)(struct mem_input *mi, bool blank);
 
+	void (*set_cursor_attributes)(
+			struct mem_input *mem_input,
+			const struct dc_cursor_attributes *attr);
+
+	void (*set_cursor_position)(
+			struct mem_input *mem_input,
+			const struct dc_cursor_position *pos,
+			const struct dc_cursor_mi_param *param);
+
 };
 
 #endif
drivers/gpu/drm/amd/display/dc/inc/hw/transform.h +11 -0
index 785d39706832..f95621dfd4d4 100644
@@ -238,6 +238,17 @@ struct transform_funcs {
 
 	void (*ipp_full_bypass)(struct transform *xfm_base);
 
+	void (*set_cursor_attributes)(
+			struct transform *xfm_base,
+			const struct dc_cursor_attributes *attr);
+
+	void (*set_cursor_position)(
+			struct transform *xfm_base,
+			const struct dc_cursor_position *pos,
+			const struct dc_cursor_mi_param *param,
+			uint32_t width
+			);
+
 };
 
 const uint16_t *get_filter_2tap_16p(void);
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#45 lore raw

PATCH 44/103 drm/amd/display: Move power control from link encoder to hwsequencer

In reply to a message not in the archive

From: Andrew Jiang <[email protected]>

A recent commit moved the backlight control code along with the register defines, but did not move the power control code. This along with remnant fields in the dce110_link_enc_registers struct made it so that the code still compiled, but any attempts to access the LVTMA_PWRSEQ_STATE register led to reading from an address of 0. This patch corrects that.

Also, rename blacklight_control to edp_backlight_control (Typo fix).

---
 drivers/gpu/drm/amd/display/dc/core/dc_link.c      |  17 +-
 drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c |   8 +-
 drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h     |  18 ++-
 .../gpu/drm/amd/display/dc/dce/dce_link_encoder.c  | 171 +--------------------
 .../gpu/drm/amd/display/dc/dce/dce_link_encoder.h  |   8 -
 .../amd/display/dc/dce110/dce110_hw_sequencer.c    | 155 +++++++++++++++++--
 .../amd/display/dc/dce110/dce110_hw_sequencer.h    |   6 +-
 .../drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c  |   3 +-
 .../gpu/drm/amd/display/dc/inc/hw/link_encoder.h   |   2 -
 drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h  |   7 +-
 drivers/gpu/drm/amd/display/dc/inc/link_hwss.h     |   4 +
 .../amd/display/dc/virtual/virtual_link_encoder.c  |   5 -
 12 files changed, 190 insertions(+), 214 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +9 -8
index a58e61b6e9f9..feb10be0ce4c 100644
@@ -78,14 +78,15 @@ static void destruct(struct dc_link *link)
 		dc_sink_release(link->remote_sinks[i]);
 }
 
-static struct gpio *get_hpd_gpio(const struct dc_link *link)
+struct gpio *get_hpd_gpio(struct dc_bios *dcb,
+		struct graphics_object_id link_id,
+		struct gpio_service *gpio_service)
 {
 	enum bp_result bp_result;
-	struct dc_bios *dcb = link->ctx->dc_bios;
 	struct graphics_object_hpd_info hpd_info;
 	struct gpio_pin_info pin_info;
 
-	if (dcb->funcs->get_hpd_info(dcb, link->link_id, &hpd_info) != BP_RESULT_OK)
+	if (dcb->funcs->get_hpd_info(dcb, link_id, &hpd_info) != BP_RESULT_OK)
 		return NULL;
 
 	bp_result = dcb->funcs->get_gpio_pin_info(dcb,
@@ -97,7 +98,7 @@ static struct gpio *get_hpd_gpio(const struct dc_link *link)
 	}
 
 	return dal_gpio_service_create_irq(
-		link->ctx->gpio_service,
+		gpio_service,
 		pin_info.offset,
 		pin_info.mask);
 }
@@ -153,7 +154,7 @@ static bool program_hpd_filter(
 	}
 
 	/* Obtain HPD handle */
-	hpd = get_hpd_gpio(link);
+	hpd = get_hpd_gpio(link->ctx->dc_bios, link->link_id, link->ctx->gpio_service);
 
 	if (!hpd)
 		return result;
@@ -186,7 +187,7 @@ static bool detect_sink(struct dc_link *link, enum dc_connection_type *type)
 	struct gpio *hpd_pin;
 
 	/* todo: may need to lock gpio access */
-	hpd_pin = get_hpd_gpio(link);
+	hpd_pin = get_hpd_gpio(link->ctx->dc_bios, link->link_id, link->ctx->gpio_service);
 	if (hpd_pin == NULL)
 		goto hpd_gpio_failure;
 
@@ -795,7 +796,7 @@ static enum hpd_source_id get_hpd_line(
 	struct gpio *hpd;
 	enum hpd_source_id hpd_id = HPD_SOURCEID_UNKNOWN;
 
-	hpd = get_hpd_gpio(link);
+	hpd = get_hpd_gpio(link->ctx->dc_bios, link->link_id, link->ctx->gpio_service);
 
 	if (hpd) {
 		switch (dal_irq_get_source(hpd)) {
@@ -965,7 +966,7 @@ static bool construct(
 		goto create_fail;
 	}
 
-	hpd_gpio = get_hpd_gpio(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);
drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c +4 -4
index 03bf79ce1f85..099f3ea7d7e9 100644
@@ -89,12 +89,12 @@ void dp_enable_link_phy(
 
 	if (dc_is_dp_sst_signal(signal)) {
 		if (signal == SIGNAL_TYPE_EDP) {
-			link_enc->funcs->power_control(link_enc, true);
+			link->dc->hwss.edp_power_control(link->link_enc, true);
 			link_enc->funcs->enable_dp_output(
 						link_enc,
 						link_settings,
 						clock_source);
-			link->dc->hwss.backlight_control(link, true);
+			link->dc->hwss.edp_backlight_control(link, true);
 		} else
 			link_enc->funcs->enable_dp_output(
 						link_enc,
@@ -138,10 +138,10 @@ void dp_disable_link_phy(struct dc_link *link, enum signal_type signal)
 		dp_receiver_power_ctrl(link, false);
 
 	if (signal == SIGNAL_TYPE_EDP) {
-		link->dc->hwss.backlight_control(link, false);
+		link->dc->hwss.edp_backlight_control(link, false);
 		edp_receiver_ready_T9(link);
 		link->link_enc->funcs->disable_output(link->link_enc, signal, link);
-		link->link_enc->funcs->power_control(link->link_enc, false);
+		link->dc->hwss.edp_power_control(link->link_enc, false);
 	} else
 		link->link_enc->funcs->disable_output(link->link_enc, signal, link);
 
drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h +13 -5
index 227c9b655b65..0a058e0c3fec 100644
@@ -391,23 +391,27 @@ struct dce_hwseq_registers {
 	HWS_SF(BLND_, V_UPDATE_LOCK, BLND_DCP_GRPH_SURF_V_UPDATE_LOCK, mask_sh),\
 	HWS_SF(BLND_, CONTROL, BLND_MODE, mask_sh),\
 	HWS_SF(, LVTMA_PWRSEQ_CNTL, LVTMA_BLON, mask_sh),\
+	HWS_SF(, LVTMA_PWRSEQ_STATE, LVTMA_PWRSEQ_TARGET_STATE_R, mask_sh),\
 	HWSEQ_PIXEL_RATE_MASK_SH_LIST(mask_sh, CRTC0_)
 
 #define HWSEQ_DCE10_MASK_SH_LIST(mask_sh)\
 	HWSEQ_DCEF_MASK_SH_LIST(mask_sh, DCFE_),\
 	HWSEQ_BLND_MASK_SH_LIST(mask_sh, BLND_),\
 	HWSEQ_PIXEL_RATE_MASK_SH_LIST(mask_sh, CRTC0_), \
-	HWS_SF(, LVTMA_PWRSEQ_CNTL, LVTMA_BLON, mask_sh)
+	HWS_SF(, LVTMA_PWRSEQ_CNTL, LVTMA_BLON, mask_sh), \
+	HWS_SF(, LVTMA_PWRSEQ_STATE, LVTMA_PWRSEQ_TARGET_STATE_R, mask_sh)
 
 #define HWSEQ_DCE11_MASK_SH_LIST(mask_sh)\
 	HWSEQ_DCE10_MASK_SH_LIST(mask_sh),\
 	SF(DCFEV_CLOCK_CONTROL, DCFEV_CLOCK_ENABLE, mask_sh),\
 	HWS_SF(, LVTMA_PWRSEQ_CNTL, LVTMA_BLON, mask_sh),\
+	HWS_SF(, LVTMA_PWRSEQ_STATE, LVTMA_PWRSEQ_TARGET_STATE_R, mask_sh),\
 	HWSEQ_PIXEL_RATE_MASK_SH_LIST(mask_sh, CRTC0_)
 
 #define HWSEQ_DCE112_MASK_SH_LIST(mask_sh)\
 	HWSEQ_DCE10_MASK_SH_LIST(mask_sh),\
 	HWS_SF(, LVTMA_PWRSEQ_CNTL, LVTMA_BLON, mask_sh),\
+	HWS_SF(, LVTMA_PWRSEQ_STATE, LVTMA_PWRSEQ_TARGET_STATE_R, mask_sh),\
 	HWSEQ_PHYPLL_MASK_SH_LIST(mask_sh, CRTC0_)
 
 #define HWSEQ_GFX9_DCHUB_MASK_SH_LIST(mask_sh)\
@@ -416,7 +420,8 @@ struct dce_hwseq_registers {
 	SF(DCHUB_AGP_BASE, AGP_BASE, mask_sh),\
 	SF(DCHUB_AGP_BOT, AGP_BOT, mask_sh),\
 	SF(DCHUB_AGP_TOP, AGP_TOP, mask_sh), \
-	HWS_SF(, LVTMA_PWRSEQ_CNTL, LVTMA_BLON, mask_sh)
+	HWS_SF(, LVTMA_PWRSEQ_CNTL, LVTMA_BLON, mask_sh), \
+	HWS_SF(, LVTMA_PWRSEQ_STATE, LVTMA_PWRSEQ_TARGET_STATE_R, mask_sh)
 
 #define HWSEQ_DCE12_MASK_SH_LIST(mask_sh)\
 	HWSEQ_DCEF_MASK_SH_LIST(mask_sh, DCFE0_DCFE_),\
@@ -424,7 +429,8 @@ struct dce_hwseq_registers {
 	HWSEQ_PIXEL_RATE_MASK_SH_LIST(mask_sh, CRTC0_),\
 	HWSEQ_PHYPLL_MASK_SH_LIST(mask_sh, CRTC0_),\
 	HWSEQ_GFX9_DCHUB_MASK_SH_LIST(mask_sh), \
-	HWS_SF(, LVTMA_PWRSEQ_CNTL, LVTMA_BLON, mask_sh)
+	HWS_SF(, LVTMA_PWRSEQ_CNTL, LVTMA_BLON, mask_sh), \
+	HWS_SF(, LVTMA_PWRSEQ_STATE, LVTMA_PWRSEQ_TARGET_STATE_R, mask_sh)
 
 #define HWSEQ_DCN_MASK_SH_LIST(mask_sh)\
 	HWSEQ_PIXEL_RATE_MASK_SH_LIST(mask_sh, OTG0_),\
@@ -489,7 +495,8 @@ struct dce_hwseq_registers {
 	HWS_SF(, DOMAIN6_PG_STATUS, DOMAIN6_PGFSM_PWR_STATUS, mask_sh), \
 	HWS_SF(, DOMAIN7_PG_STATUS, DOMAIN7_PGFSM_PWR_STATUS, mask_sh), \
 	HWS_SF(, DC_IP_REQUEST_CNTL, IP_REQUEST_EN, mask_sh), \
-	HWS_SF(, LVTMA_PWRSEQ_CNTL, LVTMA_BLON, mask_sh)
+	HWS_SF(, LVTMA_PWRSEQ_CNTL, LVTMA_BLON, mask_sh), \
+	HWS_SF(, LVTMA_PWRSEQ_STATE, LVTMA_PWRSEQ_TARGET_STATE_R, mask_sh)
 
 #define HWSEQ_REG_FIELD_LIST(type) \
 	type DCFE_CLOCK_ENABLE; \
@@ -520,7 +527,8 @@ struct dce_hwseq_registers {
 	type LOGICAL_ADDR; \
 	type ENABLE_L1_TLB;\
 	type SYSTEM_ACCESS_MODE;\
-	type LVTMA_BLON;
+	type LVTMA_BLON;\
+	type LVTMA_PWRSEQ_TARGET_STATE_R;
 
 #define HWSEQ_DCN_REG_FIELD_LIST(type) \
 	type VUPDATE_NO_LOCK_EVENT_CLEAR; \
drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c +2 -169

201 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.h +0 -8
index c65def5f6a44..494067dedd03 100644
@@ -114,10 +114,6 @@ struct dce110_link_enc_hpd_registers {
 };
 
 struct dce110_link_enc_registers {
-	/* Backlight registers */
-	uint32_t LVTMA_PWRSEQ_CNTL;
-	uint32_t LVTMA_PWRSEQ_STATE;
-
 	/* DMCU registers */
 	uint32_t MASTER_COMM_DATA_REG1;
 	uint32_t MASTER_COMM_DATA_REG2;
@@ -250,10 +246,6 @@ void dce110_link_encoder_update_mst_stream_allocation_table(
 	struct link_encoder *enc,
 	const struct link_mst_stream_allocation_table *table);
 
-void dce110_link_encoder_edp_power_control(
-	struct link_encoder *enc,
-	bool power_up);
-
 void dce110_link_encoder_connect_dig_be_to_fe(
 	struct link_encoder *enc,
 	enum engine_id engine,
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +146 -9

227 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.h +5 -1
index 3e95f7f92c73..a1e964af60ac 100644
@@ -69,7 +69,11 @@ uint32_t dce110_get_min_vblank_time_us(const struct dc_state *context);
 
 void dp_receiver_power_ctrl(struct dc_link *link, bool on);
 
-void hwss_blacklight_control(
+void hwss_edp_power_control(
+	struct link_encoder *enc,
+	bool power_up);
+
+void hwss_edp_backlight_control(
 	struct dc_link *link,
 	bool enable);
 
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +2 -1
index 014911ebdbd6..efa3f6f97baf 100644
@@ -2903,7 +2903,8 @@ static const struct hw_sequencer_funcs dcn10_funcs = {
 	.wait_for_mpcc_disconnect = dcn10_wait_for_mpcc_disconnect,
 	.ready_shared_resources = ready_shared_resources,
 	.optimize_shared_resources = optimize_shared_resources,
-	.backlight_control = hwss_blacklight_control
+	.edp_backlight_control = hwss_edp_backlight_control,
+	.edp_power_control = hwss_edp_power_control
 };
 
 
drivers/gpu/drm/amd/display/dc/inc/hw/link_encoder.h +0 -2
index 6cd6bc7d15fb..3d33bcda7059 100644
@@ -123,8 +123,6 @@ struct link_encoder_funcs {
 			bool exit_link_training_required);
 	void (*psr_program_secondary_packet)(struct link_encoder *enc,
 				unsigned int sdp_transmit_line_num_deadline);
-	void (*power_control) (struct link_encoder *enc,
-		bool power_up);
 	void (*connect_dig_be_to_fe)(struct link_encoder *enc,
 		enum engine_id engine,
 		bool connect);
drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h +5 -2
index 210874f37722..bf3ab5d7398e 100644
@@ -28,6 +28,7 @@
 #include "dc_types.h"
 #include "clock_source.h"
 #include "inc/hw/timing_generator.h"
+#include "inc/hw/link_encoder.h"
 #include "core_status.h"
 
 enum pipe_gating_control {
@@ -176,8 +177,10 @@ struct hw_sequencer_funcs {
 
 	void (*ready_shared_resources)(struct dc *dc, struct dc_state *context);
 	void (*optimize_shared_resources)(struct dc *dc);
-
-	void (*backlight_control)(
+	void (*edp_power_control)(
+			struct link_encoder *enc,
+			bool enable);
+	void (*edp_backlight_control)(
 			struct dc_link *link,
 			bool enable);
 };
drivers/gpu/drm/amd/display/dc/inc/link_hwss.h +4 -0
index f7994cfc850d..f2b8c9a376d5 100644
@@ -40,6 +40,10 @@ enum dc_status core_link_write_dpcd(
 	const uint8_t *data,
 	uint32_t size);
 
+struct gpio *get_hpd_gpio(struct dc_bios *dcb,
+		struct graphics_object_id link_id,
+		struct gpio_service *gpio_service);
+
 void dp_enable_link_phy(
 	struct dc_link *link,
 	enum signal_type signal,
drivers/gpu/drm/amd/display/dc/virtual/virtual_link_encoder.c +0 -5
index 4f405348130f..88c2bde3f039 100644
@@ -73,10 +73,6 @@ static void virtual_link_encoder_update_mst_stream_allocation_table(
 	struct link_encoder *enc,
 	const struct link_mst_stream_allocation_table *table) {}
 
-static void virtual_link_encoder_edp_power_control(
-	struct link_encoder *enc,
-	bool power_up) {}
-
 static void virtual_link_encoder_connect_dig_be_to_fe(
 	struct link_encoder *enc,
 	enum engine_id engine,
@@ -102,7 +98,6 @@ static const struct link_encoder_funcs virtual_lnk_enc_funcs = {
 	.dp_set_phy_pattern = virtual_link_encoder_dp_set_phy_pattern,
 	.update_mst_stream_allocation_table =
 		virtual_link_encoder_update_mst_stream_allocation_table,
-	.power_control = virtual_link_encoder_edp_power_control,
 	.connect_dig_be_to_fe = virtual_link_encoder_connect_dig_be_to_fe,
 	.destroy = virtual_link_encoder_destroy
 };
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#46 lore raw

PATCH 45/103 drm/amd/display: align DCLK to voltage level

In reply to a message not in the archive

From: Tony Cheng <[email protected]>

in past program SMU will use all voltage headroom. RV does not

if DAL need higher voltage for DCFCLK or DISPCLK, also increase FCLK to improve stutter as voltage is already

---
 drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c      | 4 ++++
 drivers/gpu/drm/amd/display/dc/dc.h                   | 1 +
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c | 1 +
 3 files changed, 6 insertions(+)
drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c +4 -0
index cf474eb1cde8..9337ccadc321 100644
@@ -1049,6 +1049,10 @@ bool dcn_validate_bandwidth(
 		else
 			bw_consumed = v->fabric_and_dram_bandwidth_vmax0p9;
 
+		if (bw_consumed < v->fabric_and_dram_bandwidth)
+			if (dc->debug.voltage_align_fclk)
+				bw_consumed = v->fabric_and_dram_bandwidth;
+
 		display_pipe_configuration(v);
 		calc_wm_sets_and_perf_params(context, v);
 		context->bw.dcn.calc_clk.fclk_khz = (int)(bw_consumed * 1000000 /
drivers/gpu/drm/amd/display/dc/dc.h +1 -0
index cf9cd108539f..dbbe1d621bee 100644
@@ -188,6 +188,7 @@ struct dc_debug {
 	enum dcc_option disable_dcc;
 	enum pipe_split_policy pipe_split_policy;
 	bool force_single_disp_pipe_split;
+	bool voltage_align_fclk;
 
 	bool disable_dfs_bypass;
 	bool disable_dpp_power_gate;
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +1 -0
index 30c275e05796..4018e831eb4b 100644
@@ -428,6 +428,7 @@ static const struct dc_debug debug_defaults_drv = {
 
 		.pipe_split_policy = MPC_SPLIT_AVOID_MULT_DISP,
 		.disable_dcc = DCC_ENABLE,
+		.voltage_align_fclk = true,
 };
 
 static const struct dc_debug debug_defaults_diags = {
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#47 lore raw

PATCH 46/103 drm/amd/display: disconnect on HPD low even if edid is the same

In reply to a message not in the archive

From: Wenjing Liu <[email protected]>

[Description]
There are many occasions we need to retrieve sink capability and notify connectivity change to os even if edid is not changed on a HPD toggle.
(HDMI2.0 display needs re-enable link on every hpd, display changes other capability outside from edid
need to be queried again and possibly reset the mode, etc.)

In these cases we cannot keep the same sink without letting DM know.

---
 drivers/gpu/drm/amd/display/dc/core/dc_link.c | 42 ++++++---------------------
 1 file changed, 9 insertions(+), 33 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +9 -33
index feb10be0ce4c..e414e4770789 100644
@@ -566,6 +566,8 @@ bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason)
 			link->local_sink)
 		return true;
 
+	link_disconnect_sink(link);
+
 	if (new_connection_type != dc_connection_none) {
 		link->type = new_connection_type;
 
@@ -656,28 +658,15 @@ bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason)
 			return false;
 		}
 
-		if (link->local_sink) {
-			edid_status = dm_helpers_read_local_edid(
-				link->ctx,
-				link,
-				sink);
+		sink->dongle_max_pix_clk = sink_caps.max_hdmi_pixel_clock;
+		sink->converter_disable_audio = converter_disable_audio;
 
-			if (edid_status == EDID_OK) {
-				// Edid is not the same, to update the local sink with new sink.
-				sink->dongle_max_pix_clk = sink_caps.max_hdmi_pixel_clock;
-				sink->converter_disable_audio = converter_disable_audio;
-				link->local_sink = sink;
-			}
-		} else {
-			sink->dongle_max_pix_clk = sink_caps.max_hdmi_pixel_clock;
-			sink->converter_disable_audio = converter_disable_audio;
-			link->local_sink = sink;
+		link->local_sink = sink;
 
-			edid_status = dm_helpers_read_local_edid(
+		edid_status = dm_helpers_read_local_edid(
 				link->ctx,
 				link,
 				sink);
-		}
 
 		switch (edid_status) {
 		case EDID_BAD_CHECKSUM:
@@ -758,28 +747,15 @@ bool dc_link_detect(struct dc_link *link, enum dc_detect_reason reason)
 		if (link->type == dc_connection_mst_branch) {
 			LINK_INFO("link=%d, mst branch is now Disconnected\n",
 				link->link_index);
+
 			dm_helpers_dp_mst_stop_top_mgr(link->ctx, link);
 
 			link->mst_stream_alloc_table.stream_count = 0;
 			memset(link->mst_stream_alloc_table.stream_allocations, 0, sizeof(link->mst_stream_alloc_table.stream_allocations));
 		}
 
-		if (link->local_sink) {
-			sink = link->local_sink;
-			edid_status = dm_helpers_read_local_edid(
-						link->ctx,
-						link,
-						sink);
-			if (edid_status != EDID_OK && edid_status != EDID_THE_SAME) {
-				link_disconnect_sink(link);
-				link->type = dc_connection_none;
-				sink_caps.signal = SIGNAL_TYPE_NONE;
-			}
-		} else {
-			link_disconnect_sink(link);
-			link->type = dc_connection_none;
-			sink_caps.signal = SIGNAL_TYPE_NONE;
-		}
+		link->type = dc_connection_none;
+		sink_caps.signal = SIGNAL_TYPE_NONE;
 	}
 
 	LINK_INFO("link=%d, dc_sink_in=%p is now %s\n",
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#48 lore raw

PATCH 47/103 drm/amd/display: Add override for reporting wm ranges

In reply to a message not in the archive

From: Eric Yang <[email protected]>

For verification of watermark select with SMU team, proper implementation will follow

---
 drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c   | 47 ++++++++++++++++++++++
 drivers/gpu/drm/amd/display/dc/dc.h                |  6 +++
 .../drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c  |  3 ++
 .../gpu/drm/amd/display/dc/dcn10/dcn10_resource.c  |  1 +
 4 files changed, 57 insertions(+)
drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c +47 -0
index 9337ccadc321..6318f9f69c92 100644
@@ -1441,6 +1441,53 @@ void dcn_bw_notify_pplib_of_wm_ranges(struct dc *dc)
 	ranges.writer_wm_sets[3].min_drain_clk_khz = max_fclk_khz;
 	ranges.writer_wm_sets[3].max_drain_clk_khz = max_fclk_khz;
 
+	if (dc->debug.pplib_wm_report_mode == WM_REPORT_OVERRIDE) {
+		ranges.reader_wm_sets[0].wm_inst = WM_A;
+		ranges.reader_wm_sets[0].min_drain_clk_khz = 300000;
+		ranges.reader_wm_sets[0].max_drain_clk_khz = 654000;
+		ranges.reader_wm_sets[0].min_fill_clk_khz = 800000;
+		ranges.reader_wm_sets[0].max_fill_clk_khz = 800000;
+		ranges.writer_wm_sets[0].wm_inst = WM_A;
+		ranges.writer_wm_sets[0].min_fill_clk_khz = 200000;
+		ranges.writer_wm_sets[0].max_fill_clk_khz = 757000;
+		ranges.writer_wm_sets[0].min_drain_clk_khz = 800000;
+		ranges.writer_wm_sets[0].max_drain_clk_khz = 800000;
+
+		ranges.reader_wm_sets[1].wm_inst = WM_B;
+		ranges.reader_wm_sets[1].min_drain_clk_khz = 300000;
+		ranges.reader_wm_sets[1].max_drain_clk_khz = 654000;
+		ranges.reader_wm_sets[1].min_fill_clk_khz = 933000;
+		ranges.reader_wm_sets[1].max_fill_clk_khz = 933000;
+		ranges.writer_wm_sets[1].wm_inst = WM_B;
+		ranges.writer_wm_sets[1].min_fill_clk_khz = 200000;
+		ranges.writer_wm_sets[1].max_fill_clk_khz = 757000;
+		ranges.writer_wm_sets[1].min_drain_clk_khz = 933000;
+		ranges.writer_wm_sets[1].max_drain_clk_khz = 933000;
+
+
+		ranges.reader_wm_sets[2].wm_inst = WM_C;
+		ranges.reader_wm_sets[2].min_drain_clk_khz = 300000;
+		ranges.reader_wm_sets[2].max_drain_clk_khz = 654000;
+		ranges.reader_wm_sets[2].min_fill_clk_khz = 1067000;
+		ranges.reader_wm_sets[2].max_fill_clk_khz = 1067000;
+		ranges.writer_wm_sets[2].wm_inst = WM_C;
+		ranges.writer_wm_sets[2].min_fill_clk_khz = 200000;
+		ranges.writer_wm_sets[2].max_fill_clk_khz = 757000;
+		ranges.writer_wm_sets[2].min_drain_clk_khz = 1067000;
+		ranges.writer_wm_sets[2].max_drain_clk_khz = 1067000;
+
+		ranges.reader_wm_sets[3].wm_inst = WM_D;
+		ranges.reader_wm_sets[3].min_drain_clk_khz = 300000;
+		ranges.reader_wm_sets[3].max_drain_clk_khz = 654000;
+		ranges.reader_wm_sets[3].min_fill_clk_khz = 1200000;
+		ranges.reader_wm_sets[3].max_fill_clk_khz = 1200000;
+		ranges.writer_wm_sets[3].wm_inst = WM_D;
+		ranges.writer_wm_sets[3].min_fill_clk_khz = 200000;
+		ranges.writer_wm_sets[3].max_fill_clk_khz = 757000;
+		ranges.writer_wm_sets[3].min_drain_clk_khz = 1200000;
+		ranges.writer_wm_sets[3].max_drain_clk_khz = 1200000;
+	}
+
 	/* Notify PP Lib/SMU which Watermarks to use for which clock ranges */
 	pp->set_wm_ranges(&pp->pp_smu, &ranges);
 }
drivers/gpu/drm/amd/display/dc/dc.h +6 -0
index dbbe1d621bee..d624554efbc7 100644
@@ -173,6 +173,11 @@ enum pipe_split_policy {
 	MPC_SPLIT_AVOID_MULT_DISP = 2,
 };
 
+enum wm_report_mode {
+	WM_REPORT_DEFAULT = 0,
+	WM_REPORT_OVERRIDE = 1,
+};
+
 struct dc_debug {
 	bool surface_visual_confirm;
 	bool sanity_checks;
@@ -194,6 +199,7 @@ struct dc_debug {
 	bool disable_dpp_power_gate;
 	bool disable_hubp_power_gate;
 	bool disable_pplib_wm_range;
+	enum wm_report_mode pplib_wm_report_mode;
 	bool use_dml_wm;
 	unsigned int min_disp_clk_khz;
 	int sr_exit_time_dpm0_ns;
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +3 -0
index efa3f6f97baf..86cfab33feea 100644
@@ -2451,6 +2451,9 @@ static void optimize_shared_resources(struct dc *dc)
 		/* S0i2 message */
 		dcn10_pplib_apply_display_requirements(dc, dc->current_state);
 	}
+
+	if (dc->debug.pplib_wm_report_mode == WM_REPORT_OVERRIDE)
+		dcn_bw_notify_pplib_of_wm_ranges(dc);
 }
 
 static void ready_shared_resources(struct dc *dc, struct dc_state *context)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +1 -0
index 4018e831eb4b..5d618e68a744 100644
@@ -424,6 +424,7 @@ static const struct dc_debug debug_defaults_drv = {
 
 		.disable_pplib_clock_request = true,
 		.disable_pplib_wm_range = false,
+		.pplib_wm_report_mode = WM_REPORT_DEFAULT,
 		.use_dml_wm = false,
 
 		.pipe_split_policy = MPC_SPLIT_AVOID_MULT_DISP,
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#49 lore raw

PATCH 48/103 drm/amd/display: soc_bound_box -update DML based on HW.

In reply to a message not in the archive

From: Charlene Liu <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dml/soc_bounding_box.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/gpu/drm/amd/display/dc/dml/soc_bounding_box.c +1 -2
index cb143d3d40f0..112b0b728b4d 100644
@@ -67,7 +67,6 @@ double dml_socbb_return_bw_mhz(struct _vcs_dpi_soc_bounding_box_st *box, enum vo
 
 	return_bw = dml_min(
 			((double) box->return_bus_width_bytes) * state.dcfclk_mhz,
-			state.dram_bw_per_chan_gbps * 1000.0 * (double) box->num_chans
-					* box->ideal_dram_bw_after_urgent_percent / 100.0);
+			state.dram_bw_per_chan_gbps * 1000.0 * box->ideal_dram_bw_after_urgent_percent / 100.0);
 	return return_bw;
 }
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#50 lore raw

PATCH 49/103 drm/amd/display: Expose some mem_input functions for reuse

In reply to a message not in the archive

From: Yue Hin Lau <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.h   |  1 -
 .../gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.c | 26 +++++-----
 .../gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.h | 60 ++++++++++++++++++++++
 3 files changed, 73 insertions(+), 14 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.h +0 -1
index 7fecdb19761d..889f91cf47cf 100644
@@ -113,7 +113,6 @@
 	SRI(FORMAT_CONTROL, CNVC_CFG, id), \
 	SRI(CNVC_SURFACE_PIXEL_FORMAT, CNVC_CFG, id), \
 	SRI(CURSOR0_CONTROL, CNVC_CUR, id), \
-	SRI(CURSOR0_CONTROL, CNVC_CUR, id), \
 	SRI(CURSOR0_COLOR0, CNVC_CUR, id), \
 	SRI(CURSOR0_COLOR1, CNVC_CUR, id)
 
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.c +13 -13
index efa1aca187ad..b61f41c61358 100644
@@ -38,7 +38,7 @@
 #define FN(reg_name, field_name) \
 	mi->mi_shift->field_name, mi->mi_mask->field_name
 
-static void min10_set_blank(struct mem_input *mem_input, bool blank)
+void min10_set_blank(struct mem_input *mem_input, bool blank)
 {
 	struct dcn10_mem_input *mi = TO_DCN10_MEM_INPUT(mem_input);
 	uint32_t blank_en = blank ? 1 : 0;
@@ -87,7 +87,7 @@ static void min10_vready_workaround(struct mem_input *mem_input,
 	REG_WRITE(HUBPREQ_DEBUG_DB, value);
 }
 
-static void min10_program_tiling(
+void min10_program_tiling(
 	struct dcn10_mem_input *mi,
 	const union dc_tiling_info *info,
 	const enum surface_pixel_format pixel_format)
@@ -107,7 +107,7 @@ static void min10_program_tiling(
 			PIPE_ALIGNED, info->gfx9.pipe_aligned);
 }
 
-static void min10_program_size_and_rotation(
+void min10_program_size_and_rotation(
 	struct dcn10_mem_input *mi,
 	enum dc_rotation_angle rotation,
 	enum surface_pixel_format format,
@@ -169,7 +169,7 @@ static void min10_program_size_and_rotation(
 				H_MIRROR_EN, mirror);
 }
 
-static void min10_program_pixel_format(
+void min10_program_pixel_format(
 	struct dcn10_mem_input *mi,
 	enum surface_pixel_format format)
 {
@@ -245,7 +245,7 @@ static void min10_program_pixel_format(
 	/* don't see the need of program the xbar in DCN 1.0 */
 }
 
-static bool min10_program_surface_flip_and_addr(
+bool min10_program_surface_flip_and_addr(
 	struct mem_input *mem_input,
 	const struct dc_plane_address *address,
 	bool flip_immediate)
@@ -395,7 +395,7 @@ static bool min10_program_surface_flip_and_addr(
 	return true;
 }
 
-static void min10_dcc_control(struct mem_input *mem_input, bool enable,
+void min10_dcc_control(struct mem_input *mem_input, bool enable,
 		bool independent_64b_blks)
 {
 	uint32_t dcc_en = enable ? 1 : 0;
@@ -425,7 +425,7 @@ static void min10_program_surface_config(
 	min10_program_pixel_format(mi, format);
 }
 
-static void min10_program_requestor(
+void min10_program_requestor(
 		struct mem_input *mem_input,
 		struct _vcs_dpi_display_rq_regs_st *rq_regs)
 {
@@ -459,7 +459,7 @@ static void min10_program_requestor(
 }
 
 
-static void min10_program_deadline(
+void min10_program_deadline(
 		struct mem_input *mem_input,
 		struct _vcs_dpi_display_dlg_regs_st *dlg_attr,
 		struct _vcs_dpi_display_ttu_regs_st *ttu_attr)
@@ -595,7 +595,7 @@ static void min10_setup(
 	min10_vready_workaround(mem_input, pipe_dest);
 }
 
-static void min10_program_display_marks(
+void min10_program_display_marks(
 	struct mem_input *mem_input,
 	struct dce_watermarks nbp,
 	struct dce_watermarks stutter,
@@ -607,7 +607,7 @@ static void min10_program_display_marks(
 	 */
 }
 
-static bool min10_is_flip_pending(struct mem_input *mem_input)
+bool min10_is_flip_pending(struct mem_input *mem_input)
 {
 	uint32_t flip_pending = 0;
 	struct dcn10_mem_input *mi = TO_DCN10_MEM_INPUT(mem_input);
@@ -696,7 +696,7 @@ static void min10_set_vm_context0_settings(struct mem_input *mem_input,
 			SYSTEM_ACCESS_MODE, 3);
 }
 
-static void min_set_viewport(
+void min_set_viewport(
 	struct mem_input *mem_input,
 	const struct rect *viewport,
 	const struct rect *viewport_c)
@@ -845,7 +845,7 @@ static enum cursor_lines_per_chunk ippn10_get_lines_per_chunk(
 	return line_per_chunk;
 }
 
-static void ippn10_cursor_set_attributes(
+void ippn10_cursor_set_attributes(
 		struct mem_input *mem_input,
 		const struct dc_cursor_attributes *attr)
 {
@@ -873,7 +873,7 @@ static void ippn10_cursor_set_attributes(
 			attr->color_format);
 }
 
-static void ippn10_cursor_set_position(
+void ippn10_cursor_set_position(
 		struct mem_input *mem_input,
 		const struct dc_cursor_position *pos,
 		const struct dc_cursor_mi_param *param)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.h +60 -0

72 lines not on this page · raw message

Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#51 lore raw

PATCH 50/103 drm/amd/display: Restructuring and cleaning up DML

In reply to a message not in the archive

From: Dmytro Laktyushkin <[email protected]>

---
 .../gpu/drm/amd/display/dc/calcs/dcn_calc_math.c   |   16 +
 drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c   |  103 +-
 drivers/gpu/drm/amd/display/dc/dc.h                |    1 -
 .../gpu/drm/amd/display/dc/dcn10/dcn10_resource.c  |    5 +-
 drivers/gpu/drm/amd/display/dc/dml/Makefile        |    8 +-
 drivers/gpu/drm/amd/display/dc/dml/dc_features.h   |    2 +
 .../drm/amd/display/dc/dml/display_mode_enums.h    |   52 +-
 .../gpu/drm/amd/display/dc/dml/display_mode_lib.c  |    7 +-
 .../gpu/drm/amd/display/dc/dml/display_mode_lib.h  |    8 +-
 .../drm/amd/display/dc/dml/display_mode_structs.h  |  898 +++---
 .../drm/amd/display/dc/dml/display_mode_support.c  | 2326 --------------
 .../drm/amd/display/dc/dml/display_mode_support.h  |  199 --
 .../gpu/drm/amd/display/dc/dml/display_mode_vba.c  | 3263 ++++++++++++++++++++
 .../gpu/drm/amd/display/dc/dml/display_mode_vba.h  |  410 +++
 .../drm/amd/display/dc/dml/display_pipe_clocks.c   |  377 +--
 .../drm/amd/display/dc/dml/display_pipe_clocks.h   |    8 +-
 .../drm/amd/display/dc/dml/display_rq_dlg_calc.c   | 2457 ++++++---------
 .../drm/amd/display/dc/dml/display_rq_dlg_calc.h   |  151 +-
 .../amd/display/dc/dml/display_rq_dlg_helpers.c    |  482 +--
 .../amd/display/dc/dml/display_rq_dlg_helpers.h    |   41 +-
 .../gpu/drm/amd/display/dc/dml/display_watermark.c | 1281 --------
 .../gpu/drm/amd/display/dc/dml/display_watermark.h |   98 -
 .../amd/display/dc/dml/dml1_display_rq_dlg_calc.c  | 1903 ++++++++++++
 .../amd/display/dc/dml/dml1_display_rq_dlg_calc.h  |   67 +
 .../gpu/drm/amd/display/dc/dml/dml_common_defs.c   |   12 +-
 .../gpu/drm/amd/display/dc/dml/dml_common_defs.h   |    8 +-
 .../gpu/drm/amd/display/dc/dml/soc_bounding_box.c  |   48 +-
 .../gpu/drm/amd/display/dc/dml/soc_bounding_box.h  |    9 +-
 28 files changed, 7658 insertions(+), 6582 deletions(-)
 delete mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_mode_support.c
 delete mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_mode_support.h
 create mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c
 create mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h
 delete mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_watermark.c
 delete mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_watermark.h
 create mode 100644 drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c
 create mode 100644 drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.h
drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.c +16 -0
index a18474437990..b6abe0f3bb15 100644
@@ -27,20 +27,36 @@
 
 float dcn_bw_mod(const float arg1, const float arg2)
 {
+	if (arg1 != arg1)
+		return arg2;
+	if (arg2 != arg2)
+		return arg1;
 	return arg1 - arg1 * ((int) (arg1 / arg2));
 }
 
 float dcn_bw_min2(const float arg1, const float arg2)
 {
+	if (arg1 != arg1)
+		return arg2;
+	if (arg2 != arg2)
+		return arg1;
 	return arg1 < arg2 ? arg1 : arg2;
 }
 
 unsigned int dcn_bw_max(const unsigned int arg1, const unsigned int arg2)
 {
+	if (arg1 != arg1)
+		return arg2;
+	if (arg2 != arg2)
+		return arg1;
 	return arg1 > arg2 ? arg1 : arg2;
 }
 float dcn_bw_max2(const float arg1, const float arg2)
 {
+	if (arg1 != arg1)
+		return arg2;
+	if (arg2 != arg2)
+		return arg1;
 	return arg1 > arg2 ? arg1 : arg2;
 }
 
drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c +3 -100

127 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dc.h +0 -1
index d624554efbc7..bf83459456a8 100644
@@ -200,7 +200,6 @@ struct dc_debug {
 	bool disable_hubp_power_gate;
 	bool disable_pplib_wm_range;
 	enum wm_report_mode pplib_wm_report_mode;
-	bool use_dml_wm;
 	unsigned int min_disp_clk_khz;
 	int sr_exit_time_dpm0_ns;
 	int sr_enter_plus_exit_time_dpm0_ns;
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +1 -4
index 5d618e68a744..3000ddab2357 100644
@@ -425,8 +425,6 @@ static const struct dc_debug debug_defaults_drv = {
 		.disable_pplib_clock_request = true,
 		.disable_pplib_wm_range = false,
 		.pplib_wm_report_mode = WM_REPORT_DEFAULT,
-		.use_dml_wm = false,
-
 		.pipe_split_policy = MPC_SPLIT_AVOID_MULT_DISP,
 		.disable_dcc = DCC_ENABLE,
 		.voltage_align_fclk = true,
@@ -439,8 +437,7 @@ static const struct dc_debug debug_defaults_diags = {
 		.clock_trace = true,
 		.disable_stutter = true,
 		.disable_pplib_clock_request = true,
-		.disable_pplib_wm_range = true,
-		.use_dml_wm = false,
+		.disable_pplib_wm_range = true
 };
 
 static void dcn10_dpp_destroy(struct transform **xfm)
drivers/gpu/drm/amd/display/dc/dml/Makefile +4 -4
index 9d7791d00e97..c6d9d9530d75 100644
@@ -3,19 +3,19 @@
 # It provides the general basic services required by other DAL
 # subcomponents.
 
+CFLAGS_display_mode_vba.o := -mhard-float -msse -mpreferred-stack-boundary=4
 CFLAGS_display_mode_lib.o := -mhard-float -msse -mpreferred-stack-boundary=4
 CFLAGS_display_pipe_clocks.o := -mhard-float -msse -mpreferred-stack-boundary=4
 CFLAGS_display_rq_dlg_calc.o := -mhard-float -msse -mpreferred-stack-boundary=4
+CFLAGS_dml1_display_rq_dlg_calc.o := -mhard-float -msse -mpreferred-stack-boundary=4
 CFLAGS_display_rq_dlg_helpers.o := -mhard-float -msse -mpreferred-stack-boundary=4
-CFLAGS_display_watermark.o := -mhard-float -msse -mpreferred-stack-boundary=4
 CFLAGS_soc_bounding_box.o := -mhard-float -msse -mpreferred-stack-boundary=4
 CFLAGS_dml_common_defs.o := -mhard-float -msse -mpreferred-stack-boundary=4
-CFLAGS_display_mode_support.o := -mhard-float -msse -mpreferred-stack-boundary=4
 
 
 DML = display_mode_lib.o display_pipe_clocks.o display_rq_dlg_calc.o \
-	  display_rq_dlg_helpers.o display_watermark.o \
-	  soc_bounding_box.o dml_common_defs.o display_mode_support.o
+	  display_rq_dlg_helpers.o dml1_display_rq_dlg_calc.o \
+	  soc_bounding_box.o dml_common_defs.o display_mode_vba.o
 
 AMD_DAL_DML = $(addprefix $(AMDDALPATH)/dc/dml/,$(DML))
 
drivers/gpu/drm/amd/display/dc/dml/dc_features.h +2 -0
index 745c04cb764a..ea4cde952f4f 100644
@@ -25,9 +25,11 @@
 #ifndef __DC_FEATURES_H__
 #define __DC_FEATURES_H__
 
+// local features
 #define DC__PRESENT 1
 #define DC__PRESENT__1 1
 #define DC__NUM_DPP 4
+#define DC__VOLTAGE_STATES 7
 #define DC__NUM_DPP__4 1
 #define DC__NUM_DPP__0_PRESENT 1
 #define DC__NUM_DPP__1_PRESENT 1
drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h +24 -28

84 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.c +2 -5
index c02c5522613b..09d232806de7 100644
@@ -24,6 +24,7 @@
  */
 
 #include "display_mode_lib.h"
+#include "dc_features.h"
 
 static void set_soc_bounding_box(struct _vcs_dpi_soc_bounding_box_st *soc, enum dml_project project)
 {
@@ -128,11 +129,7 @@ static void set_ip_params(struct _vcs_dpi_ip_params_st *ip, enum dml_project pro
 
 static void set_mode_evaluation(struct _vcs_dpi_mode_evaluation_st *me, enum dml_project project)
 {
-	if (project == DML_PROJECT_RAVEN1) {
-		me->voltage_override = dm_vmin;
-	} else {
-		BREAK_TO_DEBUGGER(); /* Invalid Project Specified */
-	}
+	me->voltage_override = dm_vmin;
 }
 
 void dml_init_instance(struct display_mode_lib *lib, enum dml_project project)
drivers/gpu/drm/amd/display/dc/dml/display_mode_lib.h +4 -4
index e2e311138358..57f49d29c6b9 100644
@@ -25,12 +25,13 @@
 #ifndef __DISPLAY_MODE_LIB_H__
 #define __DISPLAY_MODE_LIB_H__
 
+
 #include "dml_common_defs.h"
 #include "soc_bounding_box.h"
-#include "display_watermark.h"
+#include "display_mode_vba.h"
 #include "display_pipe_clocks.h"
 #include "display_rq_dlg_calc.h"
-#include "display_mode_support.h"
+#include "dml1_display_rq_dlg_calc.h"
 
 enum dml_project {
 	DML_PROJECT_UNDEFINED,
@@ -42,8 +43,7 @@ struct display_mode_lib {
 	struct _vcs_dpi_soc_bounding_box_st soc;
 	struct _vcs_dpi_mode_evaluation_st me;
 	enum dml_project project;
-	struct dml_ms_internal_vars vars;
-	struct _vcs_dpi_wm_calc_pipe_params_st wm_param[DC__NUM_PIPES__MAX];
+	struct vba_vars_st vba;
 	struct dal_logger *logger;
 };
 
drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h +512 -386

917 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/display_mode_support.c deleted +0 -2326

1,735 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/display_mode_support.h deleted +0 -199

199 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c new file +3263 -0

3,263 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h new file +410 -0

410 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/display_pipe_clocks.c +58 -319

377 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/display_pipe_clocks.h +2 -6
index aed5f33bb04f..a712dc3ecf4b 100644

Lines of this file are not shown. View the raw message.

drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_calc.c +982 -1475

2,457 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_calc.h +80 -71

151 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_helpers.c +277 -205

482 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/display_rq_dlg_helpers.h +11 -30

41 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/display_watermark.c deleted +0 -1281

1,281 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/display_watermark.h deleted +0 -98

98 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c new file +1903 -0

1,903 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.h new file +67 -0

67 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/dml_common_defs.c +6 -6
index 21349a022de3..9cbd415e508d 100644

Lines of this file are not shown. View the raw message.

drivers/gpu/drm/amd/display/dc/dml/dml_common_defs.h +5 -3
index c5340d41eedb..19271e79abcc 100644

Lines of this file are not shown. View the raw message.

drivers/gpu/drm/amd/display/dc/dml/soc_bounding_box.c +22 -26

48 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/soc_bounding_box.h +4 -5
index 7bbae33f163e..7a65206a6d21 100644

Lines of this file are not shown. View the raw message.

The rest of this diff is not shown. View the raw message.

Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Part of the diff is not shown here. View the raw message.

A message that is not in the archive (sent off-list or not yet received): 20171010224112.3023-51-harry.wentland-5C7GfCeVMHo@public.gmane.org

#53 lore raw

Re: [PATCH 050/103] drm/amd/display: Restructuring and cleaning up DML

Quoted 75 lines · Harry Wentland

From: Dmytro Laktyushkin <[email protected]>

---
 .../gpu/drm/amd/display/dc/calcs/dcn_calc_math.c   |   16 +
 drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c   |  103 +-
 drivers/gpu/drm/amd/display/dc/dc.h                |    1 -
 .../gpu/drm/amd/display/dc/dcn10/dcn10_resource.c  |    5 +-
 drivers/gpu/drm/amd/display/dc/dml/Makefile        |    8 +-
 drivers/gpu/drm/amd/display/dc/dml/dc_features.h   |    2 +
 .../drm/amd/display/dc/dml/display_mode_enums.h    |   52 +-
 .../gpu/drm/amd/display/dc/dml/display_mode_lib.c  |    7 +-
 .../gpu/drm/amd/display/dc/dml/display_mode_lib.h  |    8 +-
 .../drm/amd/display/dc/dml/display_mode_structs.h  |  898 +++---
 .../drm/amd/display/dc/dml/display_mode_support.c  | 2326 --------------
 .../drm/amd/display/dc/dml/display_mode_support.h  |  199 --
 .../gpu/drm/amd/display/dc/dml/display_mode_vba.c  | 3263 ++++++++++++++++++++
 .../gpu/drm/amd/display/dc/dml/display_mode_vba.h  |  410 +++
 .../drm/amd/display/dc/dml/display_pipe_clocks.c   |  377 +--
 .../drm/amd/display/dc/dml/display_pipe_clocks.h   |    8 +-
 .../drm/amd/display/dc/dml/display_rq_dlg_calc.c   | 2457 ++++++---------
 .../drm/amd/display/dc/dml/display_rq_dlg_calc.h   |  151 +-
 .../amd/display/dc/dml/display_rq_dlg_helpers.c    |  482 +--
 .../amd/display/dc/dml/display_rq_dlg_helpers.h    |   41 +-
 .../gpu/drm/amd/display/dc/dml/display_watermark.c | 1281 --------
 .../gpu/drm/amd/display/dc/dml/display_watermark.h |   98 -
 .../amd/display/dc/dml/dml1_display_rq_dlg_calc.c  | 1903 ++++++++++++
 .../amd/display/dc/dml/dml1_display_rq_dlg_calc.h  |   67 +
 .../gpu/drm/amd/display/dc/dml/dml_common_defs.c   |   12 +-
 .../gpu/drm/amd/display/dc/dml/dml_common_defs.h   |    8 +-
 .../gpu/drm/amd/display/dc/dml/soc_bounding_box.c  |   48 +-
 .../gpu/drm/amd/display/dc/dml/soc_bounding_box.h  |    9 +-
 28 files changed, 7658 insertions(+), 6582 deletions(-)
 delete mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_mode_support.c
 delete mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_mode_support.h
 create mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c
 create mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h
 delete mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_watermark.c
 delete mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_watermark.h
 create mode 100644 drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c
 create mode 100644 drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.h
drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.c +12 -0
index a18474437990..b6abe0f3bb15 100644
@@ -27,20 +27,36 @@
 
 float dcn_bw_mod(const float arg1, const float arg2)
 {
+       if (arg1 != arg1)
+               return arg2;
+       if (arg2 != arg2)
+               return arg1;
        return arg1 - arg1 * ((int) (arg1 / arg2));
 }
 
 float dcn_bw_min2(const float arg1, const float arg2)
 {
+       if (arg1 != arg1)
+               return arg2;
+       if (arg2 != arg2)
+               return arg1;
        return arg1 < arg2 ? arg1 : arg2;
 }
 
 unsigned int dcn_bw_max(const unsigned int arg1, const unsigned int arg2)
 {
+       if (arg1 != arg1)
+               return arg2;
+       if (arg2 != arg2)
+               return arg1;
        return arg1 > arg2 ? arg1 : arg2;

drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.c line 62

Quoted 10 lines
drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.c +4 -0
 }
 float dcn_bw_max2(const float arg1, const float arg2)
 {
+       if (arg1 != arg1)
+               return arg2;
+       if (arg2 != arg2)
+               return arg1;
        return arg1 > arg2 ? arg1 : arg2;
 }
 

I got this far. Wow none of the added above code makes any sense, step up the internal review please.

Dave.
_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

A message that is not in the archive (sent off-list or not yet received): CAPM=9tyvdXzi2KX7LNc_dqZRWSS98TCKABcPs2xXomS9XZZj3w-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org

#55 lore raw

Re: [PATCH 050/103] drm/amd/display: Restructuring and cleaning up DML

Quoted 81 lines · Dave Airlie
Quoted 80 lines · Harry Wentland

From: Dmytro Laktyushkin <[email protected]>

---
 .../gpu/drm/amd/display/dc/calcs/dcn_calc_math.c   |   16 +
 drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c   |  103 +-
 drivers/gpu/drm/amd/display/dc/dc.h                |    1 -
 .../gpu/drm/amd/display/dc/dcn10/dcn10_resource.c  |    5 +-
 drivers/gpu/drm/amd/display/dc/dml/Makefile        |    8 +-
 drivers/gpu/drm/amd/display/dc/dml/dc_features.h   |    2 +
 .../drm/amd/display/dc/dml/display_mode_enums.h    |   52 +-
 .../gpu/drm/amd/display/dc/dml/display_mode_lib.c  |    7 +-
 .../gpu/drm/amd/display/dc/dml/display_mode_lib.h  |    8 +-
 .../drm/amd/display/dc/dml/display_mode_structs.h  |  898 +++---
 .../drm/amd/display/dc/dml/display_mode_support.c  | 2326 --------------
 .../drm/amd/display/dc/dml/display_mode_support.h  |  199 --
 .../gpu/drm/amd/display/dc/dml/display_mode_vba.c  | 3263 ++++++++++++++++++++
 .../gpu/drm/amd/display/dc/dml/display_mode_vba.h  |  410 +++
 .../drm/amd/display/dc/dml/display_pipe_clocks.c   |  377 +--
 .../drm/amd/display/dc/dml/display_pipe_clocks.h   |    8 +-
 .../drm/amd/display/dc/dml/display_rq_dlg_calc.c   | 2457 ++++++---------
 .../drm/amd/display/dc/dml/display_rq_dlg_calc.h   |  151 +-
 .../amd/display/dc/dml/display_rq_dlg_helpers.c    |  482 +--
 .../amd/display/dc/dml/display_rq_dlg_helpers.h    |   41 +-
 .../gpu/drm/amd/display/dc/dml/display_watermark.c | 1281 --------
 .../gpu/drm/amd/display/dc/dml/display_watermark.h |   98 -
 .../amd/display/dc/dml/dml1_display_rq_dlg_calc.c  | 1903 ++++++++++++
 .../amd/display/dc/dml/dml1_display_rq_dlg_calc.h  |   67 +
 .../gpu/drm/amd/display/dc/dml/dml_common_defs.c   |   12 +-
 .../gpu/drm/amd/display/dc/dml/dml_common_defs.h   |    8 +-
 .../gpu/drm/amd/display/dc/dml/soc_bounding_box.c  |   48 +-
 .../gpu/drm/amd/display/dc/dml/soc_bounding_box.h  |    9 +-
 28 files changed, 7658 insertions(+), 6582 deletions(-)
 delete mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_mode_support.c
 delete mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_mode_support.h
 create mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c
 create mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h
 delete mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_watermark.c
 delete mode 100644 drivers/gpu/drm/amd/display/dc/dml/display_watermark.h
 create mode 100644 drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.c
 create mode 100644 drivers/gpu/drm/amd/display/dc/dml/dml1_display_rq_dlg_calc.h
drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.c +14 -0
index a18474437990..b6abe0f3bb15 100644
@@ -27,20 +27,36 @@
 
 float dcn_bw_mod(const float arg1, const float arg2)
 {
+       if (arg1 != arg1)
+               return arg2;
+       if (arg2 != arg2)
+               return arg1;
        return arg1 - arg1 * ((int) (arg1 / arg2));
 }
 
 float dcn_bw_min2(const float arg1, const float arg2)
 {
+       if (arg1 != arg1)
+               return arg2;
+       if (arg2 != arg2)
+               return arg1;
        return arg1 < arg2 ? arg1 : arg2;
 }
 
 unsigned int dcn_bw_max(const unsigned int arg1, const unsigned int arg2)
 {
+       if (arg1 != arg1)
+               return arg2;
+       if (arg2 != arg2)
+               return arg1;
        return arg1 > arg2 ? arg1 : arg2;
 }
 float dcn_bw_max2(const float arg1, const float arg2)
 {
+       if (arg1 != arg1)
+               return arg2;
Quoted 9 lines

drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.c line 62

Quoted 5 lines
drivers/gpu/drm/amd/display/dc/calcs/dcn_calc_math.c +2 -0
+       if (arg2 != arg2)
+               return arg1;
        return arg1 > arg2 ? arg1 : arg2;
 }
 

I got this far. Wow none of the added above code makes any sense, step up the internal review please.

These are NaN checks for floating point math and is the recommended way to check for those. Is there a different recommended way in the kernel?

Unfortunately this is HW magic and we'll need to find a way to deal with it.

Harry

Quoted 2 lines

Dave.

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

#56 lore raw

PATCH 51/103 drm/amd/display: add mode support check to dml vba code

In reply to a message not in the archive

From: Dmytro Laktyushkin <[email protected]>

---
 .../drm/amd/display/dc/dml/display_mode_enums.h    |   10 +-
 .../drm/amd/display/dc/dml/display_mode_structs.h  |    5 +
 .../gpu/drm/amd/display/dc/dml/display_mode_vba.c  | 5701 +++++++++++++++-----
 .../gpu/drm/amd/display/dc/dml/display_mode_vba.h  |  456 +-
 .../gpu/drm/amd/display/dc/dml/dml_common_defs.c   |   12 +
 .../gpu/drm/amd/display/dc/dml/dml_common_defs.h   |    3 +
 6 files changed, 4604 insertions(+), 1583 deletions(-)
drivers/gpu/drm/amd/display/dc/dml/display_mode_enums.h +7 -3
index eebaeb1fb61c..b1ad3553f900 100644
@@ -26,7 +26,7 @@
 #define __DISPLAY_MODE_ENUMS_H__
 
 enum output_encoder_class {
-	dm_dp = 0, dm_hdmi = 1, dm_wb = 2
+	dm_dp = 0, dm_hdmi = 1, dm_wb = 2, dm_edp
 };
 enum output_format_class {
 	dm_444 = 0, dm_420 = 1, dm_n422, dm_s422
@@ -39,7 +39,9 @@ enum source_format_class {
 	dm_420_10 = 4,
 	dm_422_8 = 5,
 	dm_422_10 = 6,
-	dm_444_8 = 7
+	dm_444_8 = 7,
+	dm_mono_8,
+	dm_mono_16
 };
 enum output_bpc_class {
 	dm_out_6 = 0, dm_out_8 = 1, dm_out_10 = 2, dm_out_12 = 3, dm_out_16 = 4
@@ -79,7 +81,9 @@ enum dm_swizzle_mode {
 	dm_sw_SPARE_15 = 28,
 	dm_sw_var_s_x = 29,
 	dm_sw_var_d_x = 30,
-	dm_sw_64kb_r_x
+	dm_sw_64kb_r_x,
+	dm_sw_gfx7_2d_thin_lvp,
+	dm_sw_gfx7_2d_thin_gl
 };
 enum lb_depth {
 	dm_lb_10 = 0, dm_lb_8 = 1, dm_lb_6 = 2, dm_lb_12 = 3, dm_lb_16
drivers/gpu/drm/amd/display/dc/dml/display_mode_structs.h +5 -0
index e5b7f7042e0b..27e20460c71b 100644
@@ -66,6 +66,7 @@ struct _vcs_dpi_mode_evaluation_st {
 
 struct _vcs_dpi_voltage_scaling_st {
 	int state;
+	double dscclk_mhz;
 	double dcfclk_mhz;
 	double socclk_mhz;
 	double dram_speed_mhz;
@@ -131,6 +132,9 @@ struct	_vcs_dpi_ip_params_st	{
 	unsigned int	writeback_chroma_line_buffer_width_pixels;
 	unsigned int	max_page_table_levels;
 	unsigned int	max_num_dpp;
+	unsigned int	max_num_otg;
+	unsigned int	cursor_chunk_size;
+	unsigned int	cursor_buffer_size;
 	unsigned int	max_num_wb;
 	unsigned int	max_dchub_pscl_bw_pix_per_clk;
 	unsigned int	max_pscl_lb_bw_pix_per_clk;
@@ -224,6 +228,7 @@ struct writeback_st {
 };
 
 struct	_vcs_dpi_display_output_params_st	{
+	int	dp_lanes;
 	int	output_bpp;
 	int	dsc_enable;
 	int	wb_enable;
drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c +4265 -1436

2,949 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h +312 -144

456 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dml/dml_common_defs.c +12 -0
index 9cbd415e508d..2af6d150cc46 100644

Lines of this file are not shown. View the raw message.

drivers/gpu/drm/amd/display/dc/dml/dml_common_defs.h +3 -0
index 19271e79abcc..ba01bbb007fc 100644

Lines of this file are not shown. View the raw message.

The rest of this diff is not shown. View the raw message.

Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Part of the diff is not shown here. View the raw message.

#57 lore raw

PATCH 52/103 drm/amd/display: extract global sync params from vba

In reply to a message not in the archive

From: Dmytro Laktyushkin <[email protected]>

---
 .../gpu/drm/amd/display/dc/dml/display_mode_vba.c  | 32 +++++++++++++++-------
 .../gpu/drm/amd/display/dc/dml/display_mode_vba.h  |  3 ++
 2 files changed, 25 insertions(+), 10 deletions(-)
drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c +22 -10
index e1a3182d58b3..054b5067fc17 100644
@@ -88,7 +88,10 @@ static bool CalculatePrefetchSchedule(
 		double *VRatioPrefetchC,
 		double *RequiredPrefetchPixDataBW,
 		unsigned int *VStartupRequiredWhenNotEnoughTimeForDynamicMetadata,
-		double *Tno_bw);
+		double *Tno_bw,
+		unsigned int *VUpdateOffsetPix,
+		unsigned int *VUpdateWidthPix,
+		unsigned int *VReadyOffsetPix);
 static double RoundToDFSGranularityUp(double Clock, double VCOSpeed);
 static double RoundToDFSGranularityDown(double Clock, double VCOSpeed);
 static double CalculatePrefetchSourceLines(
@@ -264,7 +267,7 @@ unsigned int dml_get_voltage_level(
 	memcpy(mode_lib->vba.cache_pipes, pipes, sizeof(*pipes) * num_pipes);
 	mode_lib->vba.cache_num_pipes = num_pipes;
 
-	if (need_recalculate)
+	if (need_recalculate && pipes[0].clks_cfg.dppclk_mhz != 0)
 		recalculate(mode_lib);
 	else {
 		fetch_socbb_params(mode_lib);
@@ -1066,10 +1069,13 @@ static bool CalculatePrefetchSchedule(
 		double *VRatioPrefetchC,
 		double *RequiredPrefetchPixDataBW,
 		unsigned int *VStartupRequiredWhenNotEnoughTimeForDynamicMetadata,
-		double *Tno_bw)
+		double *Tno_bw,
+		unsigned int *VUpdateOffsetPix,
+		unsigned int *VUpdateWidthPix,
+		unsigned int *VReadyOffsetPix)
 {
 	bool MyError = false;
-	unsigned int DPPCycles, DISPCLKCycles, VUpdateOffsetPix, VUpdateWidthPix, VReadyOffsetPix;
+	unsigned int DPPCycles, DISPCLKCycles;
 	double DSTTotalPixelsAfterScaler, TotalRepeaterDelayTime;
 	double Tdm, LineTime, Tsetup;
 	double dst_y_prefetch_equ;
@@ -1110,17 +1116,17 @@ static bool CalculatePrefetchSchedule(
 	*DSTYAfterScaler = dml_floor(DSTTotalPixelsAfterScaler / HTotal, 1);
 	*DSTXAfterScaler = DSTTotalPixelsAfterScaler - ((double) (*DSTYAfterScaler * HTotal));
 
-	VUpdateOffsetPix = dml_ceil(HTotal / 4.0, 1);
+	*VUpdateOffsetPix = dml_ceil(HTotal / 4.0, 1);
 	TotalRepeaterDelayTime = MaxInterDCNTileRepeaters * (2.0 / DPPCLK + 3.0 / DISPCLK);
-	VUpdateWidthPix = (14.0 / DCFClkDeepSleep + 12.0 / DPPCLK + TotalRepeaterDelayTime)
+	*VUpdateWidthPix = (14.0 / DCFClkDeepSleep + 12.0 / DPPCLK + TotalRepeaterDelayTime)
 			* PixelClock;
 
-	VReadyOffsetPix = dml_max(
+	*VReadyOffsetPix = dml_max(
 			150.0 / DPPCLK,
 			TotalRepeaterDelayTime + 20.0 / DCFClkDeepSleep + 10.0 / DPPCLK)
 			* PixelClock;
 
-	Tsetup = (double) (VUpdateOffsetPix + VUpdateWidthPix + VReadyOffsetPix) / PixelClock;
+	Tsetup = (double) (*VUpdateOffsetPix + *VUpdateWidthPix + *VReadyOffsetPix) / PixelClock;
 
 	LineTime = (double) HTotal / PixelClock;
 
@@ -2710,7 +2716,10 @@ static void DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerforman
 							&mode_lib->vba.VRatioPrefetchC[k],
 							&mode_lib->vba.RequiredPrefetchPixDataBW[k],
 							&mode_lib->vba.VStartupRequiredWhenNotEnoughTimeForDynamicMetadata,
-							&mode_lib->vba.Tno_bw[k]);
+							&mode_lib->vba.Tno_bw[k],
+							&mode_lib->vba.VUpdateOffsetPix[k],
+							&mode_lib->vba.VUpdateWidthPix[k],
+							&mode_lib->vba.VReadyOffsetPix[k]);
 			if (mode_lib->vba.BlendingAndTiming[k] == k) {
 				mode_lib->vba.VStartup[k] = dml_min(
 						mode_lib->vba.VStartupLines,
@@ -5805,7 +5814,10 @@ static void ModeSupportAndSystemConfigurationFull(struct display_mode_lib *mode_
 							&mode_lib->vba.VRatioPreC[i][k],
 							&mode_lib->vba.RequiredPrefetchPixelDataBW[i][k],
 							&mode_lib->vba.VStartupRequiredWhenNotEnoughTimeForDynamicMetadata,
-							&mode_lib->vba.Tno_bw[k]);
+							&mode_lib->vba.Tno_bw[k],
+							&mode_lib->vba.VUpdateOffsetPix[k],
+							&mode_lib->vba.VUpdateWidthPix[k],
+							&mode_lib->vba.VReadyOffsetPix[k]);
 		}
 		for (k = 0; k <= mode_lib->vba.NumberOfActivePlanes - 1; k++) {
 			mode_lib->vba.cursor_bw[k] = mode_lib->vba.NumberOfCursors[k]
drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h +3 -0
index 346ffea843c9..5758c3fd39ba 100644
@@ -383,6 +383,9 @@ struct vba_vars_st {
 	double DPPCLK_calculated[DC__NUM_DPP__MAX];
 
 	unsigned int VStartup[DC__NUM_DPP__MAX];
+	unsigned int VUpdateOffsetPix[DC__NUM_DPP__MAX];
+	unsigned int VUpdateWidthPix[DC__NUM_DPP__MAX];
+	unsigned int VReadyOffsetPix[DC__NUM_DPP__MAX];
 	unsigned int VStartupRequiredWhenNotEnoughTimeForDynamicMetadata;
 
 	double ImmediateFlipBW;
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#58 lore raw

PATCH 53/103 drm/amd/display: change dml vba cursor count define to correct one

In reply to a message not in the archive

From: Dmytro Laktyushkin <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c | 2 +-
 drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.c +1 -1
index 054b5067fc17..a11b37c50d96 100644
@@ -633,7 +633,7 @@ static void fetch_pipe_params(struct display_mode_lib *mode_lib)
 		mode_lib->vba.NumberOfCursors[mode_lib->vba.NumberOfActivePlanes] = 0;
 		// The DML spreadsheet assumes that the two cursors utilize the same amount of bandwidth. We'll
 		// calculate things a little more accurately
-		for (k = 0; k < DC__NUM_CURSOR; ++k) {
+		for (k = 0; k < DC__NUM_CURSOR__MAX; ++k) {
 			switch (k) {
 			case 0:
 				mode_lib->vba.CursorBPP[mode_lib->vba.NumberOfActivePlanes][0] =
drivers/gpu/drm/amd/display/dc/dml/display_mode_vba.h +2 -2
index 5758c3fd39ba..63ad4a7bfeb5 100644
@@ -332,8 +332,8 @@ struct vba_vars_st {
 	unsigned int BlendingAndTiming[DC__NUM_DPP__MAX];
 	bool SynchronizedVBlank;
 	unsigned int NumberOfCursors[DC__NUM_DPP__MAX];
-	unsigned int CursorWidth[DC__NUM_DPP__MAX][DC__NUM_CURSOR];
-	unsigned int CursorBPP[DC__NUM_DPP__MAX][DC__NUM_CURSOR];
+	unsigned int CursorWidth[DC__NUM_DPP__MAX][DC__NUM_CURSOR__MAX];
+	unsigned int CursorBPP[DC__NUM_DPP__MAX][DC__NUM_CURSOR__MAX];
 	bool XFCEnabled[DC__NUM_DPP__MAX];
 	bool ScalerEnabled[DC__NUM_DPP__MAX];
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#59 lore raw

PATCH 54/103 drm/amd/display: Enable double buffer as per vertical interrupt enabled.

In reply to a message not in the archive

From: Yongqiang Sun <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_timing_generator.c | 11 +++++++++++
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_timing_generator.h |  4 ++++
 2 files changed, 15 insertions(+)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_timing_generator.c +11 -0
index 7cd10cbc5497..8dbc82ff9b3a 100644
@@ -299,6 +299,17 @@ static void tgn10_program_timing(
 static void tgn10_unblank_crtc(struct timing_generator *tg)
 {
 	struct dcn10_timing_generator *tgn10 = DCN10TG_FROM_TG(tg);
+	uint32_t vertical_interrupt_enable = 0;
+
+	REG_GET(OTG_VERTICAL_INTERRUPT2_CONTROL,
+			OTG_VERTICAL_INTERRUPT2_INT_ENABLE, &vertical_interrupt_enable);
+
+	/* temporary work around for vertical interrupt, once vertical interrupt enabled,
+	 * this check will be removed.
+	 */
+	if (vertical_interrupt_enable)
+		REG_UPDATE(OTG_DOUBLE_BUFFER_CONTROL,
+				OTG_BLANK_DATA_DOUBLE_BUFFER_EN, 1);
 
 	REG_UPDATE_2(OTG_BLANK_CONTROL,
 			OTG_BLANK_DATA_EN, 0,
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_timing_generator.h +4 -0
index 0826d73b9809..7d4818d7aa31 100644
@@ -65,6 +65,7 @@
 	SRI(OTG_NOM_VERT_POSITION, OTG, inst),\
 	SRI(OTG_BLACK_COLOR, OTG, inst),\
 	SRI(OTG_CLOCK_CONTROL, OTG, inst),\
+	SRI(OTG_VERTICAL_INTERRUPT2_CONTROL, OTG, inst),\
 	SRI(OTG_VERTICAL_INTERRUPT2_POSITION, OTG, inst),\
 	SRI(OPTC_INPUT_CLOCK_CONTROL, ODM, inst),\
 	SRI(OPTC_DATA_SOURCE_SELECT, ODM, inst),\
@@ -117,6 +118,7 @@ struct dcn_tg_registers {
 	uint32_t OTG_TEST_PATTERN_CONTROL;
 	uint32_t OTG_TEST_PATTERN_COLOR;
 	uint32_t OTG_CLOCK_CONTROL;
+	uint32_t OTG_VERTICAL_INTERRUPT2_CONTROL;
 	uint32_t OTG_VERTICAL_INTERRUPT2_POSITION;
 	uint32_t OPTC_INPUT_CLOCK_CONTROL;
 	uint32_t OPTC_DATA_SOURCE_SELECT;
@@ -196,6 +198,7 @@ struct dcn_tg_registers {
 	SF(OTG0_OTG_CLOCK_CONTROL, OTG_CLOCK_EN, mask_sh),\
 	SF(OTG0_OTG_CLOCK_CONTROL, OTG_CLOCK_ON, mask_sh),\
 	SF(OTG0_OTG_CLOCK_CONTROL, OTG_CLOCK_GATE_DIS, mask_sh),\
+	SF(OTG0_OTG_VERTICAL_INTERRUPT2_CONTROL, OTG_VERTICAL_INTERRUPT2_INT_ENABLE, mask_sh),\
 	SF(OTG0_OTG_VERTICAL_INTERRUPT2_POSITION, OTG_VERTICAL_INTERRUPT2_LINE_START, mask_sh),\
 	SF(ODM0_OPTC_INPUT_CLOCK_CONTROL, OPTC_INPUT_CLK_EN, mask_sh),\
 	SF(ODM0_OPTC_INPUT_CLOCK_CONTROL, OPTC_INPUT_CLK_ON, mask_sh),\
@@ -302,6 +305,7 @@ struct dcn_tg_registers {
 	type OTG_CLOCK_EN;\
 	type OTG_CLOCK_ON;\
 	type OTG_CLOCK_GATE_DIS;\
+	type OTG_VERTICAL_INTERRUPT2_INT_ENABLE;\
 	type OTG_VERTICAL_INTERRUPT2_LINE_START;\
 	type OPTC_INPUT_CLK_EN;\
 	type OPTC_INPUT_CLK_ON;\
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#60 lore raw

PATCH 55/103 drm/amd/display: wrong VM setting cause invalid DF request from DCN

In reply to a message not in the archive

From: Tony Cheng <[email protected]>

fix typo in register field. we are lucky the shift/mask is the same, no behavior change

add globals to experiment with using different VM settings

---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.c | 11 +++++++----
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.h |  2 ++
 2 files changed, 9 insertions(+), 4 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.c +7 -4
index b61f41c61358..c09e65a0031a 100644
@@ -632,6 +632,9 @@ bool min10_is_flip_pending(struct mem_input *mem_input)
 	return false;
 }
 
+uint32_t aperture_default_system = 1;
+uint32_t context0_default_system; /* = 0;*/
+
 static void min10_set_vm_system_aperture_settings(struct mem_input *mem_input,
 		struct vm_system_aperture_param *apt)
 {
@@ -645,7 +648,7 @@ static void min10_set_vm_system_aperture_settings(struct mem_input *mem_input,
 	mc_vm_apt_high.quad_part = apt->sys_high.quad_part >> 12;
 
 	REG_SET_2(DCN_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB, 0,
-		MC_VM_SYSTEM_APERTURE_DEFAULT_SYSTEM, 1, /* 1 = system physical memory */
+		MC_VM_SYSTEM_APERTURE_DEFAULT_SYSTEM, aperture_default_system, /* 1 = system physical memory */
 		MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_MSB, mc_vm_apt_default.high_part);
 	REG_SET(DCN_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB, 0,
 		MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR_LSB, mc_vm_apt_default.low_part);
@@ -684,9 +687,9 @@ static void min10_set_vm_context0_settings(struct mem_input *mem_input,
 			VM_CONTEXT0_PAGE_TABLE_END_ADDR_LSB, vm0->pte_end.low_part);
 
 	/* fault handling */
-	REG_SET(DCN_VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR_MSB, 0,
-			VM_CONTEXT0_PAGE_TABLE_END_ADDR_MSB, vm0->fault_default.high_part);
-	/* VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_SYSTEM, 0 */
+	REG_SET_2(DCN_VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR_MSB, 0,
+			VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR_MSB, vm0->fault_default.high_part,
+			VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_SYSTEM, context0_default_system);
 	REG_SET(DCN_VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR_LSB, 0,
 			VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR_LSB, vm0->fault_default.low_part);
 
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.h +2 -0
index fdeb0e87463e..15231e86e6c6 100644
@@ -369,6 +369,7 @@ struct dcn_mi_registers {
 	MI_SF(HUBPREQ0_DCN_VM_CONTEXT0_PAGE_TABLE_END_ADDR_MSB, VM_CONTEXT0_PAGE_TABLE_END_ADDR_MSB, mask_sh),\
 	MI_SF(HUBPREQ0_DCN_VM_CONTEXT0_PAGE_TABLE_END_ADDR_LSB, VM_CONTEXT0_PAGE_TABLE_END_ADDR_LSB, mask_sh),\
 	MI_SF(HUBPREQ0_DCN_VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR_MSB, VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR_MSB, mask_sh),\
+	MI_SF(HUBPREQ0_DCN_VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR_MSB, VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_SYSTEM, mask_sh),\
 	MI_SF(HUBPREQ0_DCN_VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR_LSB, VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR_LSB, mask_sh),\
 	MI_SF(HUBPREQ0_DCN_VM_SYSTEM_APERTURE_LOW_ADDR_MSB, MC_VM_SYSTEM_APERTURE_LOW_ADDR_MSB, mask_sh),\
 	MI_SF(HUBPREQ0_DCN_VM_SYSTEM_APERTURE_LOW_ADDR_LSB, MC_VM_SYSTEM_APERTURE_LOW_ADDR_LSB, mask_sh),\
@@ -522,6 +523,7 @@ struct dcn_mi_registers {
 	type VM_CONTEXT0_PAGE_TABLE_END_ADDR_MSB;\
 	type VM_CONTEXT0_PAGE_TABLE_END_ADDR_LSB;\
 	type VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR_MSB;\
+	type VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_SYSTEM;\
 	type VM_CONTEXT0_PROTECTION_FAULT_DEFAULT_ADDR_LSB;\
 	type ENABLE_L1_TLB;\
 	type SYSTEM_ACCESS_MODE;\
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#61 lore raw

PATCH 56/103 drm/amd/display: Fix hotplug after s3 resume.

In reply to a message not in the archive

From: Roman Li <[email protected]>

On S3 resume the HPD6 irq source was not reenabled due to loop boundary bug.

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c +2 -2
index 22d41d35249e..c438ed09f52f 100644
@@ -535,7 +535,7 @@ int amdgpu_dm_irq_resume_early(struct amdgpu_device *adev)
 	DRM_DEBUG_KMS("DM_IRQ: early resume\n");
 
 	/* re-enable short pulse interrupts HW interrupt */
-	for (src = DC_IRQ_SOURCE_HPD1RX; src < DC_IRQ_SOURCE_HPD6RX + 1; src++) {
+	for (src = DC_IRQ_SOURCE_HPD1RX; src <= DC_IRQ_SOURCE_HPD6RX + 1; src++) {
 		hnd_list_l = &adev->dm.irq_handler_list_low_tab[src].head;
 		hnd_list_h = &adev->dm.irq_handler_list_high_tab[src];
 		if (!list_empty(hnd_list_l) || !list_empty(hnd_list_h))
@@ -561,7 +561,7 @@ int amdgpu_dm_irq_resume_late(struct amdgpu_device *adev)
 	 * Renable HW interrupt  for HPD and only since FLIP and VBLANK
 	 * will be enabled from manage_dm_interrupts on enable CRTC.
 	 */
-	for (src = DC_IRQ_SOURCE_HPD1; src < DC_IRQ_SOURCE_HPD6; src++) {
+	for (src = DC_IRQ_SOURCE_HPD1; src <= DC_IRQ_SOURCE_HPD6; src++) {
 		hnd_list_l = &adev->dm.irq_handler_list_low_tab[src].head;
 		hnd_list_h = &adev->dm.irq_handler_list_high_tab[src];
 		if (!list_empty(hnd_list_l) || !list_empty(hnd_list_h))
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#62 lore raw

PATCH 57/103 drm/amd/display: make FBC configurable option

In reply to a message not in the archive

From: Shirish S <[email protected]>

Currently FBC is guarded with ENABLE_FBC macro, which needs to be manually enabled in Makefile.

This patch moves it to Kconfig so that there wont be any need to additional patch to be carried for enabling or disabling on every SoC.

---
 drivers/gpu/drm/amd/display/Kconfig                      | 10 ++++++++++
 drivers/gpu/drm/amd/display/dc/core/dc.c                 |  2 +-
 drivers/gpu/drm/amd/display/dc/dc.h                      |  4 ++--
 drivers/gpu/drm/amd/display/dc/dc_types.h                |  2 +-
 .../gpu/drm/amd/display/dc/dce110/dce110_compressor.c    |  2 +-
 .../gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c  | 16 ++++++++--------
 drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c  |  4 ++--
 7 files changed, 25 insertions(+), 15 deletions(-)
drivers/gpu/drm/amd/display/Kconfig +10 -0
index 21b6b5b4c3ef..7aed462328ac 100644
@@ -9,6 +9,16 @@ config DRM_AMD_DC
 
           Will be deprecated when the DC component is upstream.
 
+config DRM_AMD_DC_FBC
+	bool "AMD FBC - Enable Frame Buffer Compression"
+	depends on DRM_AMD_DC
+	help
+	  Choose this option if you want to use frame buffer compression
+	  support.
+	  This is a power optimisation feature, check its availability
+	  on your hardware before enabling this option.
+
+
 config DRM_AMD_DC_DCN1_0
 	bool "DCN 1.0 Raven family"
 	depends on DRM_AMD_DC && X86
drivers/gpu/drm/amd/display/dc/core/dc.c +1 -1
index 3439bc46a4a2..a5c72dbc2dcc 100644
@@ -507,7 +507,7 @@ static bool construct(struct dc *dc,
 
 	dc_version = resource_parse_asic_id(init_params->asic_id);
 	dc->ctx->dce_version = dc_version;
-#ifdef ENABLE_FBC
+#if defined(CONFIG_DRM_AMD_DC_FBC)
 	dc->ctx->fbc_gpu_addr = init_params->fbc_gpu_addr;
 #endif
 	/* Resource should construct all asic specific resources.
drivers/gpu/drm/amd/display/dc/dc.h +2 -2
index bf83459456a8..adadfad2bd51 100644
@@ -258,7 +258,7 @@ struct dc {
 	struct dm_pp_display_configuration prev_display_config;
 
 	/* FBC compressor */
-#ifdef ENABLE_FBC
+#if defined(CONFIG_DRM_AMD_DC_FBC)
 	struct compressor *fbc_compressor;
 #endif
 };
@@ -293,7 +293,7 @@ struct dc_init_data {
 
 	struct dc_config flags;
 	uint32_t log_mask;
-#ifdef ENABLE_FBC
+#if defined(CONFIG_DRM_AMD_DC_FBC)
 	uint64_t fbc_gpu_addr;
 #endif
 };
drivers/gpu/drm/amd/display/dc/dc_types.h +1 -1
index 4bd74fc6c782..a8698e399111 100644
@@ -92,7 +92,7 @@ struct dc_context {
 	bool created_bios;
 	struct gpio_service *gpio_service;
 	struct i2caux *i2caux;
-#ifdef ENABLE_FBC
+#if defined(CONFIG_DRM_AMD_DC_FBC)
 	uint64_t fbc_gpu_addr;
 #endif
 };
drivers/gpu/drm/amd/display/dc/dce110/dce110_compressor.c +1 -1
index 3872febb4f6b..6923662413cd 100644
@@ -514,7 +514,7 @@ void dce110_compressor_construct(struct dce110_compressor *compressor,
 	compressor->base.lpt_channels_num = 0;
 	compressor->base.attached_inst = 0;
 	compressor->base.is_enabled = false;
-#ifdef ENABLE_FBC
+#if defined(CONFIG_DRM_AMD_DC_FBC)
 	compressor->base.funcs = &dce110_compressor_funcs;
 
 #endif
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +8 -8
index a891e387ed7b..247e81806c4a 100644
@@ -34,7 +34,7 @@
 #include "dce/dce_hwseq.h"
 #include "gpio_service_interface.h"
 
-#ifdef ENABLE_FBC
+#if defined(CONFIG_DRM_AMD_DC_FBC)
 #include "dce110_compressor.h"
 #endif
 
@@ -1445,7 +1445,7 @@ static void power_down_all_hw_blocks(struct dc *dc)
 
 	power_down_clock_sources(dc);
 
-#ifdef ENABLE_FBC
+#if defined(CONFIG_DRM_AMD_DC_FBC)
 	if (dc->fbc_compressor)
 		dc->fbc_compressor->funcs->disable_fbc(dc->fbc_compressor);
 #endif
@@ -1618,7 +1618,7 @@ static void set_static_screen_control(struct pipe_ctx **pipe_ctx,
 	if (events->cursor_update)
 		value |= 0x2;
 
-#ifdef ENABLE_FBC
+#if defined(CONFIG_DRM_AMD_DC_FBC)
 	value |= 0x84;
 #endif
 
@@ -1748,7 +1748,7 @@ static void apply_min_clocks(
 	}
 }
 
-#ifdef ENABLE_FBC
+#if defined(CONFIG_DRM_AMD_DC_FBC)
 
 /*
  *  Check if FBC can be enabled
@@ -1940,7 +1940,7 @@ enum dc_status dce110_apply_ctx_to_hw(
 
 	set_safe_displaymarks(&context->res_ctx, dc->res_pool);
 
-#ifdef ENABLE_FBC
+#if defined(CONFIG_DRM_AMD_DC_FBC)
 	if (dc->fbc_compressor)
 		dc->fbc_compressor->funcs->disable_fbc(dc->fbc_compressor);
 #endif
@@ -2124,7 +2124,7 @@ enum dc_status dce110_apply_ctx_to_hw(
 
 	dcb->funcs->set_scratch_critical_state(dcb, false);
 
-#ifdef ENABLE_FBC
+#if defined(CONFIG_DRM_AMD_DC_FBC)
 	if (dc->fbc_compressor)
 		enable_fbc(dc, context);
 
@@ -2533,7 +2533,7 @@ static void init_hw(struct dc *dc)
 		abm->funcs->init_backlight(abm);
 		abm->funcs->abm_init(abm);
 	}
-#ifdef ENABLE_FBC
+#if defined(CONFIG_DRM_AMD_DC_FBC)
 	if (dc->fbc_compressor)
 		dc->fbc_compressor->funcs->power_up_fbc(dc->fbc_compressor);
 #endif
@@ -2781,7 +2781,7 @@ static void dce110_program_front_end_for_pipe(
 
 	program_scaler(dc, pipe_ctx);
 
-#ifdef ENABLE_FBC
+#if defined(CONFIG_DRM_AMD_DC_FBC)
 	if (dc->fbc_compressor && old_pipe->stream) {
 		if (plane_state->tiling_info.gfx8.array_mode == DC_ARRAY_LINEAR_GENERAL)
 			dc->fbc_compressor->funcs->disable_fbc(dc->fbc_compressor);
drivers/gpu/drm/amd/display/dc/dce110/dce110_resource.c +2 -2
index 8e4ec61d59a4..db96d2b47ff1 100644
@@ -52,7 +52,7 @@
 #include "dce/dce_abm.h"
 #include "dce/dce_dmcu.h"
 
-#ifdef ENABLE_FBC
+#if defined(CONFIG_DRM_AMD_DC_FBC)
 #include "dce110/dce110_compressor.h"
 #endif
 
@@ -1279,7 +1279,7 @@ static bool construct(
 		}
 	}
 
-#ifdef ENABLE_FBC
+#if defined(CONFIG_DRM_AMD_DC_FBC)
 	dc->fbc_compressor = dce110_compressor_create(ctx);
 
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#63 lore raw

PATCH 58/103 drm/amd/display: Program front end first when set mode.

In reply to a message not in the archive

From: Yongqiang Sun <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/core/dc.c           | 27 ++++++++++++++--------
 .../drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c  |  5 ++--
 2 files changed, 20 insertions(+), 12 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +17 -10
index a5c72dbc2dcc..f9bac11d3950 100644
@@ -763,16 +763,6 @@ static bool dc_commit_state_no_check(struct dc *dc, struct dc_state *context)
 	if (!dcb->funcs->is_accelerated_mode(dcb))
 		dc->hwss.enable_accelerated_mode(dc);
 
-	dc->hwss.ready_shared_resources(dc, context);
-
-	for (i = 0; i < dc->res_pool->pipe_count; i++) {
-		pipe = &context->res_ctx.pipe_ctx[i];
-		dc->hwss.wait_for_mpcc_disconnect(dc, dc->res_pool, pipe);
-	}
-	result = dc->hwss.apply_ctx_to_hw(dc, context);
-
-	program_timing_sync(dc, context);
-
 	for (i = 0; i < context->stream_count; i++) {
 		const struct dc_sink *sink = context->streams[i]->sink;
 
@@ -798,6 +788,13 @@ static bool dc_commit_state_no_check(struct dc *dc, struct dc_state *context)
 			}
 		}
 
+		for (j = 0; j < MAX_PIPES; j++) {
+			pipe = &context->res_ctx.pipe_ctx[j];
+
+			if (!pipe->top_pipe && pipe->stream == context->streams[i])
+				dc->hwss.pipe_control_lock(dc, pipe, false);
+		}
+
 		CONN_MSG_MODE(sink->link, "{%dx%d, %dx%d@%dKhz}",
 				context->streams[i]->timing.h_addressable,
 				context->streams[i]->timing.v_addressable,
@@ -806,6 +803,16 @@ static bool dc_commit_state_no_check(struct dc *dc, struct dc_state *context)
 				context->streams[i]->timing.pix_clk_khz);
 	}
 
+	dc->hwss.ready_shared_resources(dc, context);
+
+	for (i = 0; i < dc->res_pool->pipe_count; i++) {
+		pipe = &context->res_ctx.pipe_ctx[i];
+		dc->hwss.wait_for_mpcc_disconnect(dc, dc->res_pool, pipe);
+	}
+	result = dc->hwss.apply_ctx_to_hw(dc, context);
+
+	program_timing_sync(dc, context);
+
 	dc_enable_stereo(dc, context, dc_streams, context->stream_count);
 
 	dc_release_state(dc->current_state);
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +3 -2
index 86cfab33feea..84d9d202d3b5 100644
@@ -2283,8 +2283,6 @@ static void update_dchubp_dpp(
 	 */
 	REG_UPDATE(DCHUBP_CNTL[pipe_ctx->pipe_idx], HUBP_VTG_SEL, pipe_ctx->stream_res.tg->inst);
 
-	dc->hwss.update_plane_addr(dc, pipe_ctx);
-
 	mi->funcs->mem_input_setup(
 		mi,
 		&pipe_ctx->dlg_regs,
@@ -2353,6 +2351,8 @@ static void update_dchubp_dpp(
 		&plane_state->dcc,
 		plane_state->horizontal_mirror);
 
+	dc->hwss.update_plane_addr(dc, pipe_ctx);
+
 	if (is_pipe_tree_visible(pipe_ctx))
 		mi->funcs->set_blank(mi, false);
 }
@@ -2556,6 +2556,7 @@ static void dcn10_apply_ctx_for_surface(
 			old_pipe_ctx->top_pipe = NULL;
 			old_pipe_ctx->bottom_pipe = NULL;
 			old_pipe_ctx->plane_state = NULL;
+			old_pipe_ctx->stream = NULL;
 
 			dm_logger_write(dc->ctx->logger, LOG_DC,
 					"Reset mpcc for pipe %d\n",
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#64 lore raw

PATCH 59/103 drm/amd/display: audio dynamic resource acquired related

In reply to a message not in the archive

From: Charlene Liu <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/core/dc_link.c      |  4 ++--
 drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c |  2 +-
 drivers/gpu/drm/amd/display/dc/core/dc_resource.c  | 24 ++++++++++++++++------
 drivers/gpu/drm/amd/display/dc/dc.h                |  1 +
 .../amd/display/dc/dce110/dce110_hw_sequencer.c    | 12 +++++++++--
 .../amd/display/dc/dce110/dce110_hw_sequencer.h    |  2 +-
 .../drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c  |  2 +-
 drivers/gpu/drm/amd/display/dc/inc/core_types.h    |  7 ++++++-
 drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h  |  3 ++-
 drivers/gpu/drm/amd/display/dc/inc/resource.h      |  5 +++++
 10 files changed, 47 insertions(+), 15 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +2 -2
index e414e4770789..2d59f77fa9a2 100644
@@ -2323,14 +2323,14 @@ void core_link_enable_stream(
 		allocate_mst_payload(pipe_ctx);
 }
 
-void core_link_disable_stream(struct pipe_ctx *pipe_ctx)
+void core_link_disable_stream(struct pipe_ctx *pipe_ctx, int option)
 {
 	struct dc  *core_dc = pipe_ctx->stream->ctx->dc;
 
 	if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST)
 		deallocate_mst_payload(pipe_ctx);
 
-	core_dc->hwss.disable_stream(pipe_ctx);
+	core_dc->hwss.disable_stream(pipe_ctx, option);
 
 	disable_link(pipe_ctx->stream->sink->link, pipe_ctx->stream->signal);
 }
drivers/gpu/drm/amd/display/dc/core/dc_link_hwss.c +1 -1
index 099f3ea7d7e9..7b0d408f8540 100644
@@ -282,7 +282,7 @@ void dp_retrain_link_dp_test(struct dc_link *link,
 
 			dp_receiver_power_ctrl(link, false);
 
-			link->dc->hwss.disable_stream(&pipes[i]);
+			link->dc->hwss.disable_stream(&pipes[i], KEEP_ACQUIRED_RESOURCE);
 
 			link->link_enc->funcs->disable_output(
 					link->link_enc,
drivers/gpu/drm/amd/display/dc/core/dc_resource.c +18 -6
index bbfec7cb2ad1..68c613229a10 100644
@@ -242,7 +242,10 @@ bool resource_construct(
 			pool->stream_enc_count++;
 		}
 	}
-
+	dc->caps.dynamic_audio = false;
+	if (pool->audio_count < pool->stream_enc_count) {
+		dc->caps.dynamic_audio = true;
+	}
 	for (i = 0; i < num_virtual_links; i++) {
 		pool->stream_enc[pool->stream_enc_count] =
 			virtual_stream_encoder_create(
@@ -1330,7 +1333,7 @@ static void update_stream_engine_usage(
 }
 
 /* TODO: release audio object */
-static void update_audio_usage(
+void update_audio_usage(
 		struct resource_context *res_ctx,
 		const struct resource_pool *pool,
 		struct audio *audio,
@@ -1414,12 +1417,21 @@ static struct audio *find_first_free_audio(
 		const struct resource_pool *pool)
 {
 	int i;
-	for (i = 0; i < pool->audio_count; i++) {
-		if (res_ctx->is_audio_acquired[i] == false) {
-			return pool->audios[i];
+	if (pool->audio_count >=  pool->stream_enc_count) {
+		for (i = 0; i < pool->audio_count; i++) {
+			if ((res_ctx->is_audio_acquired[i] == false) && (res_ctx->is_stream_enc_acquired[i] == true)) {
+				/*we have enough audio endpoint, no need to do dynamic distribution*/
+				return pool->audios[i];
+			}
 		}
-	}
+	} else { /*first come first serve*/
+		for (i = 0; i < pool->audio_count; i++) {
+			if (res_ctx->is_audio_acquired[i] == false) {
 
+				return pool->audios[i];
+			}
+		}
+	}
 	return 0;
 }
 
drivers/gpu/drm/amd/display/dc/dc.h +1 -0
index adadfad2bd51..2517fb821fff 100644
@@ -58,6 +58,7 @@ struct dc_caps {
 	uint32_t i2c_speed_in_khz;
 	unsigned int max_cursor_size;
 	bool dcc_const_color;
+	bool dynamic_audio;
 };
 
 struct dc_dcc_surface_param {
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +10 -2
index 247e81806c4a..453d2ffd5c4a 100644
@@ -985,10 +985,11 @@ void hwss_edp_backlight_control(
 	link_transmitter_control(link->dc->ctx->dc_bios, &cntl);
 }
 
-void dce110_disable_stream(struct pipe_ctx *pipe_ctx)
+void dce110_disable_stream(struct pipe_ctx *pipe_ctx, int option)
 {
 	struct dc_stream_state *stream = pipe_ctx->stream;
 	struct dc_link *link = stream->sink->link;
+	struct dc *dc = pipe_ctx->stream->ctx->dc;
 
 	if (pipe_ctx->stream_res.audio) {
 		pipe_ctx->stream_res.audio->funcs->az_disable(pipe_ctx->stream_res.audio);
@@ -999,6 +1000,13 @@ void dce110_disable_stream(struct pipe_ctx *pipe_ctx)
 		else
 			pipe_ctx->stream_res.stream_enc->funcs->hdmi_audio_disable(
 					pipe_ctx->stream_res.stream_enc);
+		/*don't free audio if it is from retrain or internal disable stream*/
+		if (option == FREE_ACQUIRED_RESOURCE && dc->caps.dynamic_audio == true) {
+			/*we have to dynamic arbitrate the audio endpoints*/
+			pipe_ctx->stream_res.audio = NULL;
+			/*we free the resource, need reset is_audio_acquired*/
+			update_audio_usage(&dc->current_state->res_ctx, dc->res_pool, pipe_ctx->stream_res.audio, false);
+		}
 
 		/* TODO: notify audio driver for if audio modes list changed
 		 * add audio mode list change flag */
@@ -1871,7 +1879,7 @@ static void dce110_reset_hw_ctx_wrap(
 				pipe_need_reprogram(pipe_ctx_old, pipe_ctx)) {
 			struct clock_source *old_clk = pipe_ctx_old->clock_source;
 
-			core_link_disable_stream(pipe_ctx_old);
+			core_link_disable_stream(pipe_ctx_old, FREE_ACQUIRED_RESOURCE);
 			pipe_ctx_old->stream_res.tg->funcs->set_blank(pipe_ctx_old->stream_res.tg, true);
 			if (!hwss_wait_for_blank_complete(pipe_ctx_old->stream_res.tg)) {
 				dm_error("DC: failed to blank crtc!\n");
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.h +1 -1
index a1e964af60ac..4d72bb99be93 100644
@@ -47,7 +47,7 @@ void dce110_set_displaymarks(
 
 void dce110_enable_stream(struct pipe_ctx *pipe_ctx);
 
-void dce110_disable_stream(struct pipe_ctx *pipe_ctx);
+void dce110_disable_stream(struct pipe_ctx *pipe_ctx, int option);
 
 void dce110_unblank_stream(struct pipe_ctx *pipe_ctx,
 		struct dc_link_settings *link_settings);
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +1 -1
index 84d9d202d3b5..072c48188b79 100644
@@ -1025,7 +1025,7 @@ static void reset_back_end_for_pipe(
 	 * which is used by otg. Move disable_link after disable_crtc
 	 */
 	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment))
-		core_link_disable_stream(pipe_ctx);
+		core_link_disable_stream(pipe_ctx, FREE_ACQUIRED_RESOURCE);
 
 	/* by upper caller loop, parent pipe: pipe0, will be reset last.
 	 * back end share by all pipes and will be disable only when disable
drivers/gpu/drm/amd/display/dc/inc/core_types.h +6 -1
index ff23f268fe02..5230dddc8617 100644
@@ -58,6 +58,11 @@ struct link_init_data {
 				TODO: remove it when DC is complete. */
 };
 
+enum {
+	FREE_ACQUIRED_RESOURCE = 0,
+	KEEP_ACQUIRED_RESOURCE = 1,
+};
+
 struct dc_link *link_create(const struct link_init_data *init_params);
 void link_destroy(struct dc_link **link);
 
@@ -72,7 +77,7 @@ void core_link_enable_stream(
 		struct dc_state *state,
 		struct pipe_ctx *pipe_ctx);
 
-void core_link_disable_stream(struct pipe_ctx *pipe_ctx);
+void core_link_disable_stream(struct pipe_ctx *pipe_ctx, int option);
 
 void core_link_set_avmute(struct pipe_ctx *pipe_ctx, bool enable);
 /********** DAL Core*********************/
drivers/gpu/drm/amd/display/dc/inc/hw_sequencer.h +2 -1
index bf3ab5d7398e..8734689a9245 100644
@@ -134,7 +134,8 @@ struct hw_sequencer_funcs {
 
 	void (*enable_stream)(struct pipe_ctx *pipe_ctx);
 
-	void (*disable_stream)(struct pipe_ctx *pipe_ctx);
+	void (*disable_stream)(struct pipe_ctx *pipe_ctx,
+			int option);
 
 	void (*unblank_stream)(struct pipe_ctx *pipe_ctx,
 			struct dc_link_settings *link_settings);
drivers/gpu/drm/amd/display/dc/inc/resource.h +5 -0
index 614bb691ab59..5467332faf7b 100644
@@ -164,4 +164,9 @@ bool pipe_need_reprogram(
 void resource_build_bit_depth_reduction_params(struct dc_stream_state *stream,
 		struct bit_depth_reduction_params *fmt_bit_depth);
 
+void update_audio_usage(
+		struct resource_context *res_ctx,
+		const struct resource_pool *pool,
+		struct audio *audio,
+		bool acquired);
 #endif /* DRIVERS_GPU_DRM_AMD_DC_DEV_DC_INC_RESOURCE_H_ */
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#65 lore raw

PATCH 60/103 drm/amd/display: renaming mem input to hubp

In reply to a message not in the archive

From: Yue Hin Lau <[email protected]>

---
 .../gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.c | 76 +++++++++++-----------
 .../gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.h | 24 +++----
 2 files changed, 50 insertions(+), 50 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.c +38 -38
index c09e65a0031a..46086bec9982 100644
@@ -38,7 +38,7 @@
 #define FN(reg_name, field_name) \
 	mi->mi_shift->field_name, mi->mi_mask->field_name
 
-void min10_set_blank(struct mem_input *mem_input, bool blank)
+void hubp1_set_blank(struct mem_input *mem_input, bool blank)
 {
 	struct dcn10_mem_input *mi = TO_DCN10_MEM_INPUT(mem_input);
 	uint32_t blank_en = blank ? 1 : 0;
@@ -56,7 +56,7 @@ void min10_set_blank(struct mem_input *mem_input, bool blank)
 	}
 }
 
-static void min10_set_hubp_blank_en(struct mem_input *mem_input, bool blank)
+static void hubp1_set_hubp_blank_en(struct mem_input *mem_input, bool blank)
 {
 	struct dcn10_mem_input *mi = TO_DCN10_MEM_INPUT(mem_input);
 	uint32_t blank_en = blank ? 1 : 0;
@@ -64,7 +64,7 @@ static void min10_set_hubp_blank_en(struct mem_input *mem_input, bool blank)
 	REG_UPDATE(DCHUBP_CNTL, HUBP_BLANK_EN, blank_en);
 }
 
-static void min10_vready_workaround(struct mem_input *mem_input,
+static void hubp1_vready_workaround(struct mem_input *mem_input,
 		struct _vcs_dpi_display_pipe_dest_params_st *pipe_dest)
 {
 	uint32_t value = 0;
@@ -87,7 +87,7 @@ static void min10_vready_workaround(struct mem_input *mem_input,
 	REG_WRITE(HUBPREQ_DEBUG_DB, value);
 }
 
-void min10_program_tiling(
+void hubp1_program_tiling(
 	struct dcn10_mem_input *mi,
 	const union dc_tiling_info *info,
 	const enum surface_pixel_format pixel_format)
@@ -107,7 +107,7 @@ void min10_program_tiling(
 			PIPE_ALIGNED, info->gfx9.pipe_aligned);
 }
 
-void min10_program_size_and_rotation(
+void hubp1_program_size_and_rotation(
 	struct dcn10_mem_input *mi,
 	enum dc_rotation_angle rotation,
 	enum surface_pixel_format format,
@@ -169,7 +169,7 @@ void min10_program_size_and_rotation(
 				H_MIRROR_EN, mirror);
 }
 
-void min10_program_pixel_format(
+void hubp1_program_pixel_format(
 	struct dcn10_mem_input *mi,
 	enum surface_pixel_format format)
 {
@@ -245,7 +245,7 @@ void min10_program_pixel_format(
 	/* don't see the need of program the xbar in DCN 1.0 */
 }
 
-bool min10_program_surface_flip_and_addr(
+bool hubp1_program_surface_flip_and_addr(
 	struct mem_input *mem_input,
 	const struct dc_plane_address *address,
 	bool flip_immediate)
@@ -395,7 +395,7 @@ bool min10_program_surface_flip_and_addr(
 	return true;
 }
 
-void min10_dcc_control(struct mem_input *mem_input, bool enable,
+void hubp1_dcc_control(struct mem_input *mem_input, bool enable,
 		bool independent_64b_blks)
 {
 	uint32_t dcc_en = enable ? 1 : 0;
@@ -407,7 +407,7 @@ void min10_dcc_control(struct mem_input *mem_input, bool enable,
 			PRIMARY_SURFACE_DCC_IND_64B_BLK, dcc_ind_64b_blk);
 }
 
-static void min10_program_surface_config(
+static void hubp1_program_surface_config(
 	struct mem_input *mem_input,
 	enum surface_pixel_format format,
 	union dc_tiling_info *tiling_info,
@@ -418,14 +418,14 @@ static void min10_program_surface_config(
 {
 	struct dcn10_mem_input *mi = TO_DCN10_MEM_INPUT(mem_input);
 
-	min10_dcc_control(mem_input, dcc->enable, dcc->grph.independent_64b_blks);
-	min10_program_tiling(mi, tiling_info, format);
-	min10_program_size_and_rotation(
+	hubp1_dcc_control(mem_input, dcc->enable, dcc->grph.independent_64b_blks);
+	hubp1_program_tiling(mi, tiling_info, format);
+	hubp1_program_size_and_rotation(
 		mi, rotation, format, plane_size, dcc, horizontal_mirror);
-	min10_program_pixel_format(mi, format);
+	hubp1_program_pixel_format(mi, format);
 }
 
-void min10_program_requestor(
+void hubp1_program_requestor(
 		struct mem_input *mem_input,
 		struct _vcs_dpi_display_rq_regs_st *rq_regs)
 {
@@ -459,7 +459,7 @@ void min10_program_requestor(
 }
 
 
-void min10_program_deadline(
+void hubp1_program_deadline(
 		struct mem_input *mem_input,
 		struct _vcs_dpi_display_dlg_regs_st *dlg_attr,
 		struct _vcs_dpi_display_ttu_regs_st *ttu_attr)
@@ -580,7 +580,7 @@ void min10_program_deadline(
 		ttu_attr->refcyc_per_req_delivery_pre_c);
 }
 
-static void min10_setup(
+static void hubp1_setup(
 		struct mem_input *mem_input,
 		struct _vcs_dpi_display_dlg_regs_st *dlg_attr,
 		struct _vcs_dpi_display_ttu_regs_st *ttu_attr,
@@ -590,12 +590,12 @@ static void min10_setup(
 	/* otg is locked when this func is called. Register are double buffered.
 	 * disable the requestors is not needed
 	 */
-	min10_program_requestor(mem_input, rq_regs);
-	min10_program_deadline(mem_input, dlg_attr, ttu_attr);
-	min10_vready_workaround(mem_input, pipe_dest);
+	hubp1_program_requestor(mem_input, rq_regs);
+	hubp1_program_deadline(mem_input, dlg_attr, ttu_attr);
+	hubp1_vready_workaround(mem_input, pipe_dest);
 }
 
-void min10_program_display_marks(
+void hubp1_program_display_marks(
 	struct mem_input *mem_input,
 	struct dce_watermarks nbp,
 	struct dce_watermarks stutter,
@@ -607,7 +607,7 @@ void min10_program_display_marks(
 	 */
 }
 
-bool min10_is_flip_pending(struct mem_input *mem_input)
+bool hubp1_is_flip_pending(struct mem_input *mem_input)
 {
 	uint32_t flip_pending = 0;
 	struct dcn10_mem_input *mi = TO_DCN10_MEM_INPUT(mem_input);
@@ -635,7 +635,7 @@ bool min10_is_flip_pending(struct mem_input *mem_input)
 uint32_t aperture_default_system = 1;
 uint32_t context0_default_system; /* = 0;*/
 
-static void min10_set_vm_system_aperture_settings(struct mem_input *mem_input,
+static void hubp1_set_vm_system_aperture_settings(struct mem_input *mem_input,
 		struct vm_system_aperture_param *apt)
 {
 	struct dcn10_mem_input *mi = TO_DCN10_MEM_INPUT(mem_input);
@@ -664,7 +664,7 @@ static void min10_set_vm_system_aperture_settings(struct mem_input *mem_input,
 			MC_VM_SYSTEM_APERTURE_HIGH_ADDR_LSB, mc_vm_apt_high.low_part);
 }
 
-static void min10_set_vm_context0_settings(struct mem_input *mem_input,
+static void hubp1_set_vm_context0_settings(struct mem_input *mem_input,
 		const struct vm_context0_param *vm0)
 {
 	struct dcn10_mem_input *mi = TO_DCN10_MEM_INPUT(mem_input);
@@ -848,7 +848,7 @@ static enum cursor_lines_per_chunk ippn10_get_lines_per_chunk(
 	return line_per_chunk;
 }
 
-void ippn10_cursor_set_attributes(
+void hubp1_cursor_set_attributes(
 		struct mem_input *mem_input,
 		const struct dc_cursor_attributes *attr)
 {
@@ -876,7 +876,7 @@ void ippn10_cursor_set_attributes(
 			attr->color_format);
 }
 
-void ippn10_cursor_set_position(
+void hubp1_cursor_set_position(
 		struct mem_input *mem_input,
 		const struct dc_cursor_position *pos,
 		const struct dc_cursor_mi_param *param)
@@ -913,7 +913,7 @@ void ippn10_cursor_set_position(
 		cur_en = 0;  /* not visible beyond left edge*/
 
 	if (cur_en && REG_READ(CURSOR_SURFACE_ADDRESS) == 0)
-		ippn10_cursor_set_attributes(mem_input, &mem_input->curs_attr);
+		hubp1_cursor_set_attributes(mem_input, &mem_input->curs_attr);
 	REG_UPDATE(CURSOR_CONTROL,
 			CURSOR_ENABLE, cur_en);
 
@@ -931,21 +931,21 @@ void ippn10_cursor_set_position(
 }
 
 static struct mem_input_funcs dcn10_mem_input_funcs = {
-	.mem_input_program_display_marks = min10_program_display_marks,
+	.mem_input_program_display_marks = hubp1_program_display_marks,
 	.mem_input_program_surface_flip_and_addr =
-			min10_program_surface_flip_and_addr,
+			hubp1_program_surface_flip_and_addr,
 	.mem_input_program_surface_config =
-			min10_program_surface_config,
-	.mem_input_is_flip_pending = min10_is_flip_pending,
-	.mem_input_setup = min10_setup,
-	.mem_input_set_vm_system_aperture_settings = min10_set_vm_system_aperture_settings,
-	.mem_input_set_vm_context0_settings = min10_set_vm_context0_settings,
-	.set_blank = min10_set_blank,
-	.dcc_control = min10_dcc_control,
+			hubp1_program_surface_config,
+	.mem_input_is_flip_pending = hubp1_is_flip_pending,
+	.mem_input_setup = hubp1_setup,
+	.mem_input_set_vm_system_aperture_settings = hubp1_set_vm_system_aperture_settings,
+	.mem_input_set_vm_context0_settings = hubp1_set_vm_context0_settings,
+	.set_blank = hubp1_set_blank,
+	.dcc_control = hubp1_dcc_control,
 	.mem_program_viewport = min_set_viewport,
-	.set_hubp_blank_en = min10_set_hubp_blank_en,
-	.set_cursor_attributes	= ippn10_cursor_set_attributes,
-	.set_cursor_position	= ippn10_cursor_set_position,
+	.set_hubp_blank_en = hubp1_set_hubp_blank_en,
+	.set_cursor_attributes	= hubp1_cursor_set_attributes,
+	.set_cursor_position	= hubp1_cursor_set_position,
 };
 
 /*****************************************/
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.h +12 -12
index 15231e86e6c6..653875285b70 100644
@@ -591,20 +591,20 @@ struct dcn10_mem_input {
 	const struct dcn_mi_mask *mi_mask;
 };
 
-void min10_program_deadline(
+void hubp1_program_deadline(
 		struct mem_input *mem_input,
 		struct _vcs_dpi_display_dlg_regs_st *dlg_attr,
 		struct _vcs_dpi_display_ttu_regs_st *ttu_attr);
 
-void min10_program_requestor(
+void hubp1_program_requestor(
 		struct mem_input *mem_input,
 		struct _vcs_dpi_display_rq_regs_st *rq_regs);
 
-void min10_program_pixel_format(
+void hubp1_program_pixel_format(
 	struct dcn10_mem_input *mi,
 	enum surface_pixel_format format);
 
-void min10_program_size_and_rotation(
+void hubp1_program_size_and_rotation(
 	struct dcn10_mem_input *mi,
 	enum dc_rotation_angle rotation,
 	enum surface_pixel_format format,
@@ -612,39 +612,39 @@ void min10_program_size_and_rotation(
 	struct dc_plane_dcc_param *dcc,
 	bool horizontal_mirror);
 
-void min10_program_tiling(
+void hubp1_program_tiling(
 	struct dcn10_mem_input *mi,
 	const union dc_tiling_info *info,
 	const enum surface_pixel_format pixel_format);
 
-void min10_dcc_control(struct mem_input *mem_input,
+void hubp1_dcc_control(struct mem_input *mem_input,
 		bool enable,
 		bool independent_64b_blks);
 
-void min10_program_display_marks(
+void hubp1_program_display_marks(
 	struct mem_input *mem_input,
 	struct dce_watermarks nbp,
 	struct dce_watermarks stutter,
 	struct dce_watermarks urgent,
 	uint32_t total_dest_line_time_ns);
 
-bool min10_program_surface_flip_and_addr(
+bool hubp1_program_surface_flip_and_addr(
 	struct mem_input *mem_input,
 	const struct dc_plane_address *address,
 	bool flip_immediate);
 
-bool min10_is_flip_pending(struct mem_input *mem_input);
+bool hubp1_is_flip_pending(struct mem_input *mem_input);
 
-void ippn10_cursor_set_attributes(
+void hubp1_cursor_set_attributes(
 		struct mem_input *mem_input,
 		const struct dc_cursor_attributes *attr);
 
-void ippn10_cursor_set_position(
+void hubp1_cursor_set_position(
 		struct mem_input *mem_input,
 		const struct dc_cursor_position *pos,
 		const struct dc_cursor_mi_param *param);
 
-void min10_set_blank(struct mem_input *mem_input, bool blank);
+void hubp1_set_blank(struct mem_input *mem_input, bool blank);
 
 void min_set_viewport(struct mem_input *mem_input,
 		const struct rect *viewport,
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#66 lore raw

PATCH 61/103 drm/amd/display: arbitration find the matching dig-az first.

In reply to a message not in the archive

From: Charlene Liu <[email protected]>

[Description]
this change is in branch already.
without this change, after resume, az_inst might swapped.

---
 drivers/gpu/drm/amd/display/dc/core/dc_resource.c | 20 ++++++++------------
 1 file changed, 8 insertions(+), 12 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_resource.c +8 -12
index 68c613229a10..c31dccdc3d2e 100644
@@ -1417,19 +1417,15 @@ static struct audio *find_first_free_audio(
 		const struct resource_pool *pool)
 {
 	int i;
-	if (pool->audio_count >=  pool->stream_enc_count) {
-		for (i = 0; i < pool->audio_count; i++) {
-			if ((res_ctx->is_audio_acquired[i] == false) && (res_ctx->is_stream_enc_acquired[i] == true)) {
-				/*we have enough audio endpoint, no need to do dynamic distribution*/
-				return pool->audios[i];
-			}
+	for (i = 0; i < pool->audio_count; i++) {
+		if ((res_ctx->is_audio_acquired[i] == false) && (res_ctx->is_stream_enc_acquired[i] == true)) {
+			return pool->audios[i];
 		}
-	} else { /*first come first serve*/
-		for (i = 0; i < pool->audio_count; i++) {
-			if (res_ctx->is_audio_acquired[i] == false) {
-
-				return pool->audios[i];
-			}
+	}
+	/*not found the matching one, first come first serve*/
+	for (i = 0; i < pool->audio_count; i++) {
+		if (res_ctx->is_audio_acquired[i] == false) {
+			return pool->audios[i];
 		}
 	}
 	return 0;
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#67 lore raw

PATCH 62/103 drm/amd/display: Making hubp1_program_surface_config public

In reply to a message not in the archive

From: Yue Hin Lau <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.c | 3 ++-
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.h | 9 +++++++++
 2 files changed, 11 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.c +2 -1
index 46086bec9982..2336da54000c 100644
@@ -407,7 +407,7 @@ void hubp1_dcc_control(struct mem_input *mem_input, bool enable,
 			PRIMARY_SURFACE_DCC_IND_64B_BLK, dcc_ind_64b_blk);
 }
 
-static void hubp1_program_surface_config(
+void hubp1_program_surface_config(
 	struct mem_input *mem_input,
 	enum surface_pixel_format format,
 	union dc_tiling_info *tiling_info,
@@ -970,3 +970,4 @@ void dcn10_mem_input_construct(
 	mi->base.mpcc_id = 0xf;
 }
 
+
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.h +9 -0
index 653875285b70..5c5eed5c95d5 100644
@@ -591,6 +591,15 @@ struct dcn10_mem_input {
 	const struct dcn_mi_mask *mi_mask;
 };
 
+void hubp1_program_surface_config(
+	struct mem_input *mem_input,
+	enum surface_pixel_format format,
+	union dc_tiling_info *tiling_info,
+	union plane_size *plane_size,
+	enum dc_rotation_angle rotation,
+	struct dc_plane_dcc_param *dcc,
+	bool horizontal_mirror);
+
 void hubp1_program_deadline(
 		struct mem_input *mem_input,
 		struct _vcs_dpi_display_dlg_regs_st *dlg_attr,
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#68 lore raw

PATCH 63/103 drm/amd/display: Fix irq enable/disable on resume.

In reply to a message not in the archive

From: Roman Li <[email protected]>

Fixing loop boundaries in irq enable/disable on resume.

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_irq.c +2 -2
index c438ed09f52f..544382d0a633 100644
@@ -508,7 +508,7 @@ int amdgpu_dm_irq_suspend(
 	 * Disable HW interrupt  for HPD and HPDRX only since FLIP and VBLANK
 	 * will be disabled from manage_dm_interrupts on disable CRTC.
 	 */
-	for (src = DC_IRQ_SOURCE_HPD1; src < DC_IRQ_SOURCE_HPD6RX; src++) {
+	for (src = DC_IRQ_SOURCE_HPD1; src <= DC_IRQ_SOURCE_HPD6RX; src++) {
 		hnd_list_l = &adev->dm.irq_handler_list_low_tab[src].head;
 		hnd_list_h = &adev->dm.irq_handler_list_high_tab[src];
 		if (!list_empty(hnd_list_l) || !list_empty(hnd_list_h))
@@ -535,7 +535,7 @@ int amdgpu_dm_irq_resume_early(struct amdgpu_device *adev)
 	DRM_DEBUG_KMS("DM_IRQ: early resume\n");
 
 	/* re-enable short pulse interrupts HW interrupt */
-	for (src = DC_IRQ_SOURCE_HPD1RX; src <= DC_IRQ_SOURCE_HPD6RX + 1; src++) {
+	for (src = DC_IRQ_SOURCE_HPD1RX; src <= DC_IRQ_SOURCE_HPD6RX; src++) {
 		hnd_list_l = &adev->dm.irq_handler_list_low_tab[src].head;
 		hnd_list_h = &adev->dm.irq_handler_list_high_tab[src];
 		if (!list_empty(hnd_list_l) || !list_empty(hnd_list_h))
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#69 lore raw

PATCH 64/103 drm/amd/display: Do not access fbdev helper members

In reply to a message not in the archive

From: Jerry Zuo <[email protected]>

Original code is no longer needed and tested without loop.

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c | 13 +------------
 1 file changed, 1 insertion(+), 12 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c +1 -12
index 11a10c1c0adb..aafa53994c14 100644
@@ -395,21 +395,10 @@ static void dm_dp_mst_register_connector(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
 	struct amdgpu_device *adev = dev->dev_private;
-	int i;
 
 	drm_modeset_lock_all(dev);
-	if (adev->mode_info.rfbdev) {
-		/*Do not add if already registered in past*/
-		for (i = 0; i < adev->mode_info.rfbdev->helper.connector_count; i++) {
-			if (adev->mode_info.rfbdev->helper.connector_info[i]->connector
-					== connector) {
-				drm_modeset_unlock_all(dev);
-				return;
-			}
-		}
-
+	if (adev->mode_info.rfbdev)
 		drm_fb_helper_add_one_connector(&adev->mode_info.rfbdev->helper, connector);
-	}
 	else
 		DRM_ERROR("adev->mode_info.rfbdev is NULL\n");
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#70 lore raw

PATCH 65/103 drm/amd/display: Remove drm_modeset_lock in MST code

In reply to a message not in the archive

From: Jerry Zuo <[email protected]>

Locks are no longer needed since the drm framework takes care of correct locking.

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c | 12 +++++-------
 1 file changed, 5 insertions(+), 7 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c +5 -7
index aafa53994c14..d86f03900b0c 100644
@@ -284,9 +284,10 @@ static struct drm_connector *dm_dp_add_mst_connector(struct drm_dp_mst_topology_
 	struct amdgpu_device *adev = dev->dev_private;
 	struct amdgpu_dm_connector *aconnector;
 	struct drm_connector *connector;
+	struct drm_connector_list_iter conn_iter;
 
-	drm_modeset_lock(&dev->mode_config.connection_mutex, NULL);
-	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
+	drm_connector_list_iter_begin(dev, &conn_iter);
+	drm_for_each_connector_iter(connector, &conn_iter) {
 		aconnector = to_amdgpu_dm_connector(connector);
 		if (aconnector->mst_port == master
 				&& !aconnector->port) {
@@ -296,11 +297,11 @@ static struct drm_connector *dm_dp_add_mst_connector(struct drm_dp_mst_topology_
 			aconnector->port = port;
 			drm_mode_connector_set_path_property(connector, pathprop);
 
-			drm_modeset_unlock(&dev->mode_config.connection_mutex);
+			drm_connector_list_iter_end(&conn_iter);
 			return &aconnector->base;
 		}
 	}
-	drm_modeset_unlock(&dev->mode_config.connection_mutex);
+	drm_connector_list_iter_end(&conn_iter);
 
 	aconnector = kzalloc(sizeof(*aconnector), GFP_KERNEL);
 	if (!aconnector)
@@ -396,14 +397,11 @@ static void dm_dp_mst_register_connector(struct drm_connector *connector)
 	struct drm_device *dev = connector->dev;
 	struct amdgpu_device *adev = dev->dev_private;
 
-	drm_modeset_lock_all(dev);
 	if (adev->mode_info.rfbdev)
 		drm_fb_helper_add_one_connector(&adev->mode_info.rfbdev->helper, connector);
 	else
 		DRM_ERROR("adev->mode_info.rfbdev is NULL\n");
 
-	drm_modeset_unlock_all(dev);
-
 	drm_connector_register(connector);
 
 }
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#71 lore raw

PATCH 66/103 drm/amd/display: clean up dcn10 dpp after HW review

In reply to a message not in the archive

From: Eric Bernstein <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dc_hw_types.h       |  2 +
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.h   | 20 +++++
 .../gpu/drm/amd/display/dc/dcn10/dcn10_dpp_dscl.c  | 85 +++++++++++-----------
 3 files changed, 65 insertions(+), 42 deletions(-)
drivers/gpu/drm/amd/display/dc/dc_hw_types.h +2 -0
index 4ab109314e4b..1a9f57fb0838 100644
@@ -204,6 +204,8 @@ enum surface_pixel_format {
 	/*grow 444 video here if necessary */
 };
 
+
+
 /* Pixel format */
 enum pixel_format {
 	/*graph*/
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.h +20 -0
index 889f91cf47cf..153aa3c32762 100644
@@ -1265,6 +1265,21 @@ struct dcn10_dpp {
 };
 
 
+enum lb_memory_config {
+	/* Enable all 3 pieces of memory */
+	LB_MEMORY_CONFIG_0 = 0,
+
+	/* Enable only the first piece of memory */
+	LB_MEMORY_CONFIG_1 = 1,
+
+	/* Enable only the second piece of memory */
+	LB_MEMORY_CONFIG_2 = 2,
+
+	/* Only applicable in 4:2:0 mode, enable all 3 pieces of memory and the
+	 * last piece of chroma memory used for the luma storage
+	 */
+	LB_MEMORY_CONFIG_3 = 3
+};
 
 enum dcn10_input_csc_select {
 	INPUT_CSC_SELECT_BYPASS = 0,
@@ -1272,6 +1287,11 @@ enum dcn10_input_csc_select {
 	INPUT_CSC_SELECT_COMA
 };
 
+bool is_lb_conf_valid(
+		int ceil_vratio,
+		int num_partitions,
+		int vtaps);
+
 void ippn10_degamma_ram_select(
 		struct transform *xfm_base,
 							bool use_ram_a);
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_dscl.c +43 -42

175 lines not on this page · raw message

Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#72 lore raw

PATCH 67/103 drm/amd/display: Add stereo_support debug flag

In reply to a message not in the archive

From: Charlene Liu <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dc.h | 1 +
 1 file changed, 1 insertion(+)
drivers/gpu/drm/amd/display/dc/dc.h +1 -0
index 2517fb821fff..172050ad845d 100644
@@ -217,6 +217,7 @@ struct dc_debug {
 	bool force_abm_enable;
 	bool disable_hbup_pg;
 	bool disable_dpp_pg;
+	bool stereo_support;
 };
 struct dc_state;
 struct resource_pool;
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#73 lore raw

PATCH 68/103 drm/amd/display: Fixed extend to second screen mode hang

In reply to a message not in the archive

From: Yongqiang Sun <[email protected]>

1. Fixed acquire free split pipe bug.
2. Change return value for dc_add_stream_to_ctx from bool to enum.
4. Remove redundant apply_ctx_for_surface calling
5. Unlock pipe after back end programming.

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c |  4 +-
 drivers/gpu/drm/amd/display/dc/core/dc.c          | 48 ++++++++++++-----------
 drivers/gpu/drm/amd/display/dc/core/dc_resource.c |  9 ++---
 drivers/gpu/drm/amd/display/dc/dc.h               |  2 +-
 4 files changed, 32 insertions(+), 31 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +2 -2
index 8f9dcc697aad..1dc110042956 100644
@@ -4661,10 +4661,10 @@ static int dm_update_crtcs_state(
 				DRM_DEBUG_DRIVER("Enabling DRM crtc: %d\n",
 							crtc->base.id);
 
-				if (!dc_add_stream_to_ctx(
+				if (dc_add_stream_to_ctx(
 						dc,
 						dm_state->context,
-						new_acrtc_state->stream)) {
+						new_acrtc_state->stream) != DC_OK) {
 					ret = -EINVAL;
 					goto fail;
 				}
drivers/gpu/drm/amd/display/dc/core/dc.c +25 -23

65 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/core/dc_resource.c +4 -5
index c31dccdc3d2e..c60f53069708 100644
@@ -1446,7 +1446,7 @@ bool resource_is_stream_unchanged(
 	return false;
 }
 
-bool dc_add_stream_to_ctx(
+enum dc_status dc_add_stream_to_ctx(
 		struct dc *dc,
 		struct dc_state *new_ctx,
 		struct dc_stream_state *stream)
@@ -1467,7 +1467,7 @@ bool dc_add_stream_to_ctx(
 	if (res != DC_OK)
 		DC_ERROR("Adding stream %p to context failed with err %d!\n", stream, res);
 
-	return res == DC_OK;
+	return res;
 }
 
 bool dc_remove_stream_from_ctx(
@@ -1640,10 +1640,9 @@ enum dc_status resource_map_pool_resources(
 	/* acquire new resources */
 	pipe_idx = acquire_first_free_pipe(&context->res_ctx, pool, stream);
 
-#if defined(CONFIG_DRM_AMD_DC_DCN1_0)
 	if (pipe_idx < 0)
-		acquire_first_split_pipe(&context->res_ctx, pool, stream);
-#endif
+		pipe_idx = acquire_first_split_pipe(&context->res_ctx, pool, stream);
+
 	if (pipe_idx < 0)
 		return DC_NO_CONTROLLER_RESOURCE;
 
drivers/gpu/drm/amd/display/dc/dc.h +1 -1
index 172050ad845d..eeb8ee5acdc1 100644
@@ -631,7 +631,7 @@ bool dc_stream_get_scanoutpos(const struct dc_stream_state *stream,
 				  uint32_t *h_position,
 				  uint32_t *v_position);
 
-bool dc_add_stream_to_ctx(
+enum dc_status dc_add_stream_to_ctx(
 			struct dc *dc,
 		struct dc_state *new_ctx,
 		struct dc_stream_state *stream);
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#74 lore raw

PATCH 69/103 drm/amd/display: wa to program split pipedegamma regs after mode set.

In reply to a message not in the archive

From: Yongqiang Sun <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/core/dc.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +3 -1
index 33a63c7d7a8e..09bab93815a3 100644
@@ -1282,7 +1282,9 @@ static void commit_planes_for_stream(struct dc *dc,
 			if (update_type == UPDATE_TYPE_FAST)
 				continue;
 
-			if (srf_updates[i].in_transfer_func)
+			/* work around to program degamma regs for split pipe after set mode. */
+			if (srf_updates[i].in_transfer_func || (pipe_ctx->top_pipe &&
+					pipe_ctx->top_pipe->plane_state == pipe_ctx->plane_state))
 				dc->hwss.set_input_transfer_func(
 						pipe_ctx, pipe_ctx->plane_state);
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#75 lore raw

PATCH 70/103 drm/amd/display: Add debug flag for VSR support

In reply to a message not in the archive

From: Charlene Liu <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dc.h                           | 1 +
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c         | 1 +
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_timing_generator.c | 1 +
 3 files changed, 3 insertions(+)
drivers/gpu/drm/amd/display/dc/dc.h +1 -0
index eeb8ee5acdc1..defc44393b38 100644
@@ -218,6 +218,7 @@ struct dc_debug {
 	bool disable_hbup_pg;
 	bool disable_dpp_pg;
 	bool stereo_support;
+	bool vsr_support;
 };
 struct dc_state;
 struct resource_pool;
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +1 -0
index 3000ddab2357..1b108ae2e656 100644
@@ -428,6 +428,7 @@ static const struct dc_debug debug_defaults_drv = {
 		.pipe_split_policy = MPC_SPLIT_AVOID_MULT_DISP,
 		.disable_dcc = DCC_ENABLE,
 		.voltage_align_fclk = true,
+		.vsr_support = true,
 };
 
 static const struct dc_debug debug_defaults_diags = {
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_timing_generator.c +1 -0
index 8dbc82ff9b3a..178dadda74f9 100644
@@ -25,6 +25,7 @@
 
 #include "reg_helper.h"
 #include "dcn10_timing_generator.h"
+#include "dc.h"
 
 #define REG(reg)\
 	tgn10->tg_regs->reg
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#76 lore raw

PATCH 71/103 drm/amd/display: fix re-enabling stutter for raven

In reply to a message not in the archive

From: Bhawanpreet Lakha <[email protected]>

We were overwriting the whole register which was re-enabling stutter for raven. Now we are reading the register then setting the values only for pstate.

---
 drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h            | 4 ++++
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c | 9 ++++++---
 2 files changed, 10 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h +4 -0
index 0a058e0c3fec..9a04b8758888 100644
@@ -446,6 +446,8 @@ struct dce_hwseq_registers {
 	HWS_SF(, DCHUBBUB_ARB_WATERMARK_CHANGE_CNTL, DCHUBBUB_ARB_WATERMARK_CHANGE_DONE_INTERRUPT_DISABLE, mask_sh), \
 	HWS_SF(, DCHUBBUB_ARB_DRAM_STATE_CNTL, DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_VALUE, mask_sh), \
 	HWS_SF(, DCHUBBUB_ARB_DRAM_STATE_CNTL, DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE, mask_sh), \
+	HWS_SF(, DCHUBBUB_ARB_DRAM_STATE_CNTL, DCHUBBUB_ARB_ALLOW_PSTATE_CHANGE_FORCE_VALUE, mask_sh), \
+	HWS_SF(, DCHUBBUB_ARB_DRAM_STATE_CNTL, DCHUBBUB_ARB_ALLOW_PSTATE_CHANGE_FORCE_ENABLE, mask_sh), \
 	HWS_SF(, DCHUBBUB_ARB_SAT_LEVEL, DCHUBBUB_ARB_SAT_LEVEL, mask_sh), \
 	HWS_SF(, DCHUBBUB_ARB_DF_REQ_OUTSTAND, DCHUBBUB_ARB_MIN_REQ_OUTSTAND, mask_sh), \
 	HWS_SF(, DCFCLK_CNTL, DCFCLK_GATE_DIS, mask_sh)
@@ -554,6 +556,8 @@ struct dce_hwseq_registers {
 	type DCHUBBUB_ARB_WATERMARK_CHANGE_DONE_INTERRUPT_DISABLE;\
 	type DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_VALUE;\
 	type DCHUBBUB_ARB_ALLOW_SELF_REFRESH_FORCE_ENABLE;\
+	type DCHUBBUB_ARB_ALLOW_PSTATE_CHANGE_FORCE_VALUE;\
+	type DCHUBBUB_ARB_ALLOW_PSTATE_CHANGE_FORCE_ENABLE;\
 	type DCHUBBUB_ARB_SAT_LEVEL;\
 	type DCHUBBUB_ARB_MIN_REQ_OUTSTAND;\
 	type OPP_PIPE_CLOCK_EN;\
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +6 -3
index 072c48188b79..db774b3c3721 100644
@@ -253,7 +253,6 @@ static void verify_allow_pstate_change_high(
 
 	unsigned int debug_index = 0x7;
 	unsigned int debug_data;
-	unsigned int force_allow_pstate = 0x30;
 	unsigned int i;
 
 	if (forced_pstate_allow) {
@@ -261,7 +260,9 @@ static void verify_allow_pstate_change_high(
 		 * we verify_allow_pstate_change_high.  so disable force
 		 * here so we can check status
 		 */
-		REG_WRITE(DCHUBBUB_ARB_DRAM_STATE_CNTL, 0);
+		REG_UPDATE_2(DCHUBBUB_ARB_DRAM_STATE_CNTL,
+			     DCHUBBUB_ARB_ALLOW_PSTATE_CHANGE_FORCE_VALUE, 0,
+			     DCHUBBUB_ARB_ALLOW_PSTATE_CHANGE_FORCE_ENABLE, 0);
 		forced_pstate_allow = false;
 	}
 
@@ -304,7 +305,9 @@ static void verify_allow_pstate_change_high(
 	/* force pstate allow to prevent system hang
 	 * and break to debugger to investigate
 	 */
-	REG_WRITE(DCHUBBUB_ARB_DRAM_STATE_CNTL, force_allow_pstate);
+	REG_UPDATE_2(DCHUBBUB_ARB_DRAM_STATE_CNTL,
+		     DCHUBBUB_ARB_ALLOW_PSTATE_CHANGE_FORCE_VALUE, 1,
+		     DCHUBBUB_ARB_ALLOW_PSTATE_CHANGE_FORCE_ENABLE, 1);
 	forced_pstate_allow = true;
 
 	if (should_log_hw_state) {
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#77 lore raw

PATCH 72/103 drm/amd/display: reject 3D mode based on static debug flag.

In reply to a message not in the archive

From: Charlene Liu <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dc.h                           | 2 +-
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c         | 1 +
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_timing_generator.c | 3 +++
 3 files changed, 5 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dc.h +1 -1
index defc44393b38..4c47a489c87a 100644
@@ -217,7 +217,7 @@ struct dc_debug {
 	bool force_abm_enable;
 	bool disable_hbup_pg;
 	bool disable_dpp_pg;
-	bool stereo_support;
+	bool disable_stereo_support;
 	bool vsr_support;
 };
 struct dc_state;
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +1 -0
index 1b108ae2e656..f168335e95bf 100644
@@ -428,6 +428,7 @@ static const struct dc_debug debug_defaults_drv = {
 		.pipe_split_policy = MPC_SPLIT_AVOID_MULT_DISP,
 		.disable_dcc = DCC_ENABLE,
 		.voltage_align_fclk = true,
+		.disable_stereo_support = true,
 		.vsr_support = true,
 };
 
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_timing_generator.c +3 -0
index 178dadda74f9..1994865d4351 100644
@@ -499,6 +499,9 @@ static bool tgn10_validate_timing(
 		timing->timing_3d_format != TIMING_3D_FORMAT_INBAND_FA)
 		return false;
 
+	if (timing->timing_3d_format != TIMING_3D_FORMAT_NONE &&
+		tg->ctx->dc->debug.disable_stereo_support)
+		return false;
 	/* Temporarily blocking interlacing mode until it's supported */
 	if (timing->flags.INTERLACE == 1)
 		return false;
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#78 lore raw

PATCH 73/103 drm/amd/display: increase pstate wait timeout

In reply to a message not in the archive

From: Bhawanpreet Lakha <[email protected]>

For linux it takes longer than 40us so increasing it to 200us. Also added debug prints regarding the change

---
 .../drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c    | 20 +++++++++++++++++---
 1 file changed, 17 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +17 -3
index db774b3c3721..012ba7c9ca8d 100644
@@ -245,8 +245,13 @@ static void verify_allow_pstate_change_high(
 {
 	/* pstate latency is ~20us so if we wait over 40us and pstate allow
 	 * still not asserted, we are probably stuck and going to hang
+	 *
+	 * TODO: Figure out why it takes ~100us on linux
+	 * pstate takes around ~100us on linux. Unknown currently as to
+	 * why it takes that long on linux
 	 */
-	static unsigned int pstate_wait_timeout_us = 40;
+	static unsigned int pstate_wait_timeout_us = 200;
+	static unsigned int pstate_wait_expected_timeout_us = 40;
 	static unsigned int max_sampled_pstate_wait_us; /* data collection */
 	static bool forced_pstate_allow; /* help with revert wa */
 	static bool should_log_hw_state; /* prevent hw state log by default */
@@ -293,9 +298,15 @@ static void verify_allow_pstate_change_high(
 	for (i = 0; i < pstate_wait_timeout_us; i++) {
 		debug_data = REG_READ(DCHUBBUB_TEST_DEBUG_DATA);
 
-		if (debug_data & (1 << 30))
-			return;
+		if (debug_data & (1 << 30)) {
+
+			if (i > pstate_wait_expected_timeout_us)
+				dm_logger_write(hws->ctx->logger, LOG_WARNING,
+						"pstate took longer than expected ~%dus",
+						i);
 
+			return;
+		}
 		if (max_sampled_pstate_wait_us < i)
 			max_sampled_pstate_wait_us = i;
 
@@ -314,6 +325,9 @@ static void verify_allow_pstate_change_high(
 		dcn10_log_hw_state(hws->ctx->dc);
 	}
 
+	dm_logger_write(hws->ctx->logger, LOG_WARNING,
+			"pstate TEST_DEBUG_DATA: 0x%X",
+			debug_data);
 	BREAK_TO_DEBUGGER();
 }
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#79 lore raw

PATCH 74/103 drm/amd/display: block video planes >4k on dcn10

In reply to a message not in the archive

From: Dmytro Laktyushkin <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c | 9 +++++++++
 1 file changed, 9 insertions(+)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +9 -0
index f168335e95bf..4dd7c3af51e6 100644
@@ -1163,6 +1163,14 @@ static void dcn10_destroy_resource_pool(struct resource_pool **pool)
 	*pool = NULL;
 }
 
+enum dc_status dcn10_validate_plane(const struct dc_plane_state *plane_state)
+{
+	if (plane_state->format >= SURFACE_PIXEL_FORMAT_VIDEO_BEGIN
+			&& plane_state->src_rect.width > 3840)
+		return DC_FAIL_SURFACE_VALIDATE;
+
+	return DC_OK;
+}
 
 static struct dc_cap_funcs cap_funcs = {
 	.get_dcc_compression_cap = get_dcc_compression_cap
@@ -1174,6 +1182,7 @@ static struct resource_funcs dcn10_res_pool_funcs = {
 	.validate_guaranteed = dcn10_validate_guaranteed,
 	.validate_bandwidth = dcn_validate_bandwidth,
 	.acquire_idle_pipe_for_layer = dcn10_acquire_idle_pipe_for_layer,
+	.validate_plane = dcn10_validate_plane,
 	.add_stream_to_ctx = dcn10_add_stream_to_ctx
 };
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#80 lore raw

PATCH 75/103 drm/amd/display: add max_video_width cap to dc

In reply to a message not in the archive

From: Dmytro Laktyushkin <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/core/dc_resource.c       | 6 ++----
 drivers/gpu/drm/amd/display/dc/dc.h                     | 1 +
 drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c | 2 +-
 drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.h | 2 +-
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c   | 6 ++++--
 drivers/gpu/drm/amd/display/dc/inc/core_types.h         | 2 +-
 6 files changed, 10 insertions(+), 9 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_resource.c +2 -4
index c60f53069708..7d750369b78b 100644
@@ -2750,11 +2750,9 @@ bool dc_validate_stream(struct dc *dc, struct dc_stream_state *stream)
 
 bool dc_validate_plane(struct dc *dc, const struct dc_plane_state *plane_state)
 {
-	struct dc *core_dc = dc;
-
 	/* TODO For now validates pixel format only */
-	if (core_dc->res_pool->funcs->validate_plane)
-		return core_dc->res_pool->funcs->validate_plane(plane_state) == DC_OK;
+	if (dc->res_pool->funcs->validate_plane)
+		return dc->res_pool->funcs->validate_plane(plane_state, &dc->caps) == DC_OK;
 
 	return true;
 }
drivers/gpu/drm/amd/display/dc/dc.h +1 -0
index 4c47a489c87a..5a378cc5daaf 100644
@@ -57,6 +57,7 @@ struct dc_caps {
 	uint32_t max_downscale_ratio;
 	uint32_t i2c_speed_in_khz;
 	unsigned int max_cursor_size;
+	unsigned int max_video_width;
 	bool dcc_const_color;
 	bool dynamic_audio;
 };
drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.c +1 -1
index b1cf591b3408..90911258bdb3 100644
@@ -743,7 +743,7 @@ static void dce100_destroy_resource_pool(struct resource_pool **pool)
 	*pool = NULL;
 }
 
-enum dc_status dce100_validate_plane(const struct dc_plane_state *plane_state)
+enum dc_status dce100_validate_plane(const struct dc_plane_state *plane_state, struct dc_caps *caps)
 {
 
 	if (plane_state->format < SURFACE_PIXEL_FORMAT_VIDEO_BEGIN)
drivers/gpu/drm/amd/display/dc/dce100/dce100_resource.h +1 -1
index f52cae24ee23..de8fdf438f9b 100644
@@ -16,7 +16,7 @@ struct resource_pool *dce100_create_resource_pool(
 	uint8_t num_virtual_links,
 	struct dc *dc);
 
-enum dc_status dce100_validate_plane(const struct dc_plane_state *plane_state);
+enum dc_status dce100_validate_plane(const struct dc_plane_state *plane_state, struct dc_caps *caps);
 
 enum dc_status dce100_add_stream_to_ctx(
 		struct dc *dc,
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +4 -2
index 4dd7c3af51e6..2caad73257e0 100644
@@ -1163,10 +1163,11 @@ static void dcn10_destroy_resource_pool(struct resource_pool **pool)
 	*pool = NULL;
 }
 
-enum dc_status dcn10_validate_plane(const struct dc_plane_state *plane_state)
+static enum dc_status dcn10_validate_plane(const struct dc_plane_state *plane_state, struct dc_caps *caps)
 {
 	if (plane_state->format >= SURFACE_PIXEL_FORMAT_VIDEO_BEGIN
-			&& plane_state->src_rect.width > 3840)
+			&& caps->max_video_width != 0
+			&& plane_state->src_rect.width > caps->max_video_width)
 		return DC_FAIL_SURFACE_VALIDATE;
 
 	return DC_OK;
@@ -1222,6 +1223,7 @@ static bool construct(
 	/* max pipe num for ASIC before check pipe fuses */
 	pool->base.pipe_count = pool->base.res_cap->num_timing_generator;
 
+	dc->caps.max_video_width = 3840;
 	dc->caps.max_downscale_ratio = 200;
 	dc->caps.i2c_speed_in_khz = 100;
 	dc->caps.max_cursor_size = 256;
drivers/gpu/drm/amd/display/dc/inc/core_types.h +1 -1
index 5230dddc8617..118122d12e2b 100644
@@ -111,7 +111,7 @@ struct resource_funcs {
 			const struct resource_pool *pool,
 			struct dc_stream_state *stream);
 
-	enum dc_status (*validate_plane)(const struct dc_plane_state *plane_state);
+	enum dc_status (*validate_plane)(const struct dc_plane_state *plane_state, struct dc_caps *caps);
 
 	enum dc_status (*add_stream_to_ctx)(
 			struct dc *dc,
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#81 lore raw

PATCH 76/103 drm/amd/display: add dpms state to DC

In reply to a message not in the archive

From: Hersen Wu <[email protected]>

- avoid eDP screen flash 4 times when resume from s3
- improve s3 and boot time

---
 drivers/gpu/drm/amd/display/dc/core/dc.c           | 39 +++++++++++++++++++++-
 drivers/gpu/drm/amd/display/dc/core/dc_link.c      |  4 +++
 drivers/gpu/drm/amd/display/dc/dc.h                |  5 +++
 .../amd/display/dc/dce110/dce110_hw_sequencer.c    | 13 ++++----
 .../drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c  | 11 +++---
 5 files changed, 58 insertions(+), 14 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +38 -1
index 09bab93815a3..d963de753e63 100644
@@ -332,7 +332,16 @@ static void set_dither_option(struct dc_stream_state *stream,
 {
 	struct bit_depth_reduction_params params;
 	struct dc_link *link = stream->status.link;
-	struct pipe_ctx *pipes = link->dc->current_state->res_ctx.pipe_ctx;
+	struct pipe_ctx *pipes;
+	int i;
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		if (link->dc->current_state->res_ctx.pipe_ctx[i].stream ==
+				stream) {
+			pipes = &link->dc->current_state->res_ctx.pipe_ctx[i];
+			break;
+		}
+	}
 
 	memset(&params, 0, sizeof(params));
 	if (!stream)
@@ -349,6 +358,31 @@ static void set_dither_option(struct dc_stream_state *stream,
 		opp_program_bit_depth_reduction(pipes->stream_res.opp, &params);
 }
 
+void set_dpms(
+	struct dc *dc,
+	struct dc_stream_state *stream,
+	bool dpms_off)
+{
+	struct pipe_ctx *pipe_ctx;
+	int i;
+
+	for (i = 0; i < MAX_PIPES; i++) {
+		if (dc->current_state->res_ctx.pipe_ctx[i].stream == stream) {
+			pipe_ctx = &dc->current_state->res_ctx.pipe_ctx[i];
+			break;
+		}
+	}
+
+	if (stream->dpms_off != dpms_off) {
+		stream->dpms_off = dpms_off;
+		if (dpms_off)
+			core_link_disable_stream(pipe_ctx,
+					KEEP_ACQUIRED_RESOURCE);
+		else
+			core_link_enable_stream(dc->current_state, pipe_ctx);
+	}
+}
+
 static void allocate_dc_stream_funcs(struct dc  *dc)
 {
 	if (dc->hwss.set_drr != NULL) {
@@ -371,6 +405,9 @@ static void allocate_dc_stream_funcs(struct dc  *dc)
 	dc->stream_funcs.set_dither_option =
 			set_dither_option;
 
+	dc->stream_funcs.set_dpms =
+			set_dpms;
+
 	dc->link_funcs.set_drive_settings =
 			set_drive_settings;
 
drivers/gpu/drm/amd/display/dc/core/dc_link.c +4 -0
index 2d59f77fa9a2..72e56fbe4a42 100644
@@ -2321,6 +2321,10 @@ void core_link_enable_stream(
 
 	if (pipe_ctx->stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST)
 		allocate_mst_payload(pipe_ctx);
+
+	if (dc_is_dp_signal(pipe_ctx->stream->signal))
+		core_dc->hwss.unblank_stream(pipe_ctx,
+			&pipe_ctx->stream->sink->link->cur_link_settings);
 }
 
 void core_link_disable_stream(struct pipe_ctx *pipe_ctx, int option)
drivers/gpu/drm/amd/display/dc/dc.h +5 -0
index 5a378cc5daaf..7a5e53ffac2e 100644
@@ -133,6 +133,10 @@ struct dc_stream_state_funcs {
 
 	void (*set_dither_option)(struct dc_stream_state *stream,
 			enum dc_dither_option option);
+
+	void (*set_dpms)(struct dc *dc,
+			struct dc_stream_state *stream,
+			bool dpms_off);
 };
 
 struct link_training_settings;
@@ -566,6 +570,7 @@ struct dc_stream_state {
 
 	int phy_pix_clk;
 	enum signal_type signal;
+	bool dpms_off;
 
 	struct dc_stream_status status;
 
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +6 -7
index 453d2ffd5c4a..81ebf2b9c71d 100644
@@ -1364,12 +1364,8 @@ static enum dc_status apply_single_controller_ctx_to_hw(
 	resource_build_info_frame(pipe_ctx);
 	dce110_update_info_frame(pipe_ctx);
 	if (!pipe_ctx_old->stream) {
-		core_link_enable_stream(context, pipe_ctx);
-
-
-		if (dc_is_dp_signal(pipe_ctx->stream->signal))
-			dce110_unblank_stream(pipe_ctx,
-				&stream->sink->link->cur_link_settings);
+		if (!pipe_ctx->stream->dpms_off)
+			core_link_enable_stream(context, pipe_ctx);
 	}
 
 	pipe_ctx->plane_res.scl_data.lb_params.alpha_en = pipe_ctx->bottom_pipe != 0;
@@ -1879,7 +1875,10 @@ static void dce110_reset_hw_ctx_wrap(
 				pipe_need_reprogram(pipe_ctx_old, pipe_ctx)) {
 			struct clock_source *old_clk = pipe_ctx_old->clock_source;
 
-			core_link_disable_stream(pipe_ctx_old, FREE_ACQUIRED_RESOURCE);
+			/* disable already, no need to disable again */
+			if (!pipe_ctx->stream->dpms_off)
+				core_link_disable_stream(pipe_ctx_old, FREE_ACQUIRED_RESOURCE);
+
 			pipe_ctx_old->stream_res.tg->funcs->set_blank(pipe_ctx_old->stream_res.tg, true);
 			if (!hwss_wait_for_blank_complete(pipe_ctx_old->stream_res.tg)) {
 				dm_error("DC: failed to blank crtc!\n");
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +5 -6
index 012ba7c9ca8d..494f833208f9 100644
@@ -1037,12 +1037,11 @@ static void reset_back_end_for_pipe(
 		return;
 	}
 
-	/* TODOFPGA break core_link_disable_stream into 2 functions:
-	 * disable_stream and disable_link. disable_link will disable PHYPLL
-	 * which is used by otg. Move disable_link after disable_crtc
-	 */
-	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment))
-		core_link_disable_stream(pipe_ctx, FREE_ACQUIRED_RESOURCE);
+	if (!IS_FPGA_MAXIMUS_DC(dc->ctx->dce_environment)) {
+		/* DPMS may already disable */
+		if (!pipe_ctx->stream->dpms_off)
+			core_link_disable_stream(pipe_ctx, FREE_ACQUIRED_RESOURCE);
+	}
 
 	/* by upper caller loop, parent pipe: pipe0, will be reset last.
 	 * back end share by all pipes and will be disable only when disable
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#82 lore raw

PATCH 77/103 drm/amd/display: Fixed switching mode half screen gamma incorrect.

In reply to a message not in the archive

From: Yongqiang Sun <[email protected]>

Half screen gamma setting and cursor are incorrect when switching mode through win+p due to wrong programming gamma sequence (In case of bottom pipe, gamma and cursor are programmed before front end programmed, pipe is power gated).

change:
1. Cache curor attributes to stream
2. Move set gamma and cursor inside front end programming.

---
 drivers/gpu/drm/amd/display/dc/core/dc.c                | 15 ---------------
 drivers/gpu/drm/amd/display/dc/core/dc_stream.c         |  9 ++++++++-
 drivers/gpu/drm/amd/display/dc/dc.h                     |  4 +++-
 .../gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c | 10 ++++++++++
 .../gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c   | 17 +++++++++++++++++
 5 files changed, 38 insertions(+), 17 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +0 -15
index d963de753e63..92fab09843e4 100644
@@ -1269,10 +1269,6 @@ static void commit_planes_for_stream(struct dc *dc,
 	/* Full fe update*/
 	for (j = 0; j < dc->res_pool->pipe_count; j++) {
 		struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j];
-		struct pipe_ctx *cur_pipe_ctx = &dc->current_state->res_ctx.pipe_ctx[j];
-		bool is_new_pipe_surface = cur_pipe_ctx->plane_state != pipe_ctx->plane_state;
-		struct dc_cursor_position position = { 0 };
-
 
 		if (update_type != UPDATE_TYPE_FULL || !pipe_ctx->plane_state)
 			continue;
@@ -1283,17 +1279,6 @@ static void commit_planes_for_stream(struct dc *dc,
 			dc->hwss.apply_ctx_for_surface(
 					dc, pipe_ctx->stream, stream_status->plane_count, context);
 		}
-
-		/* TODO: this is a hack w/a for switching from mpo to pipe split */
-		dc_stream_set_cursor_position(pipe_ctx->stream, &position);
-
-		if (is_new_pipe_surface) {
-			dc->hwss.update_plane_addr(dc, pipe_ctx);
-			dc->hwss.set_input_transfer_func(
-					pipe_ctx, pipe_ctx->plane_state);
-			dc->hwss.set_output_transfer_func(
-					pipe_ctx, pipe_ctx->stream);
-		}
 	}
 
 	if (update_type > UPDATE_TYPE_FAST)
drivers/gpu/drm/amd/display/dc/core/dc_stream.c +8 -1
index c19b478bcdc7..3dd44bef56eb 100644
@@ -173,7 +173,7 @@ struct dc_stream_status *dc_stream_get_status(
  * Update the cursor attributes and set cursor surface address
  */
 bool dc_stream_set_cursor_attributes(
-	const struct dc_stream_state *stream,
+	struct dc_stream_state *stream,
 	const struct dc_cursor_attributes *attributes)
 {
 	int i;
@@ -189,6 +189,11 @@ bool dc_stream_set_cursor_attributes(
 			return false;
 	}
 
+	if (attributes->address.quad_part == 0) {
+		dm_error("DC: Cursor address is 0!\n");
+		return false;
+	}
+
 	core_dc = stream->ctx->dc;
 	res_ctx = &core_dc->current_state->res_ctx;
 
@@ -214,6 +219,8 @@ bool dc_stream_set_cursor_attributes(
 				pipe_ctx->plane_res.xfm, attributes);
 	}
 
+	stream->cursor_attributes = *attributes;
+
 	return true;
 }
 
drivers/gpu/drm/amd/display/dc/dc.h +3 -1
index 7a5e53ffac2e..c832d5abbbdf 100644
@@ -574,6 +574,8 @@ struct dc_stream_state {
 
 	struct dc_stream_status status;
 
+	struct dc_cursor_attributes cursor_attributes;
+
 	/* from stream struct */
 	struct kref refcount;
 };
@@ -1013,7 +1015,7 @@ struct dc_sink *dc_sink_create(const struct dc_sink_init_data *init_params);
  ******************************************************************************/
 /* TODO: Deprecated once we switch to dc_set_cursor_position */
 bool dc_stream_set_cursor_attributes(
-	const struct dc_stream_state *stream,
+	struct dc_stream_state *stream,
 	const struct dc_cursor_attributes *attributes);
 
 bool dc_stream_set_cursor_position(
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +10 -0
index 81ebf2b9c71d..bf7318246401 100644
@@ -2725,6 +2725,8 @@ static void dce110_program_front_end_for_pipe(
 	struct dc_plane_state *plane_state = pipe_ctx->plane_state;
 	struct xfm_grph_csc_adjustment adjust;
 	struct out_csc_color_matrix tbl_entry;
+	struct pipe_ctx *cur_pipe_ctx =
+					&dc->current_state->res_ctx.pipe_ctx[pipe_ctx->pipe_idx];
 	unsigned int i;
 
 	memset(&tbl_entry, 0, sizeof(tbl_entry));
@@ -2815,6 +2817,14 @@ static void dce110_program_front_end_for_pipe(
 				&plane_state->tiling_info,
 				plane_state->rotation);
 
+	/* Moved programming gamma from dc to hwss */
+	if (cur_pipe_ctx->plane_state != pipe_ctx->plane_state) {
+		dc->hwss.set_input_transfer_func(
+				pipe_ctx, pipe_ctx->plane_state);
+		dc->hwss.set_output_transfer_func(
+				pipe_ctx, pipe_ctx->stream);
+	}
+
 	dm_logger_write(dc->ctx->logger, LOG_SURFACE,
 			"Pipe:%d 0x%x: addr hi:0x%x, "
 			"addr low:0x%x, "
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +17 -0
index 494f833208f9..ee63155c1895 100644
@@ -2408,6 +2408,10 @@ static void program_all_pipe_in_tree(
 	}
 
 	if (pipe_ctx->plane_state != NULL) {
+		struct dc_cursor_position position = { 0 };
+		struct pipe_ctx *cur_pipe_ctx =
+				&dc->current_state->res_ctx.pipe_ctx[pipe_ctx->pipe_idx];
+
 		dcn10_power_on_fe(dc, pipe_ctx, context);
 
 		/* temporary dcn1 wa:
@@ -2422,6 +2426,19 @@ static void program_all_pipe_in_tree(
 		toggle_watermark_change_req(dc->hwseq);
 
 		update_dchubp_dpp(dc, pipe_ctx, context);
+
+		/* TODO: this is a hack w/a for switching from mpo to pipe split */
+		dc_stream_set_cursor_position(pipe_ctx->stream, &position);
+
+		dc_stream_set_cursor_attributes(pipe_ctx->stream,
+				&pipe_ctx->stream->cursor_attributes);
+
+		if (cur_pipe_ctx->plane_state != pipe_ctx->plane_state) {
+			dc->hwss.set_input_transfer_func(
+					pipe_ctx, pipe_ctx->plane_state);
+			dc->hwss.set_output_transfer_func(
+					pipe_ctx, pipe_ctx->stream);
+		}
 	}
 
 	if (dc->debug.sanity_checks) {
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#83 lore raw

PATCH 78/103 drm/amd/display: add dpms state to DC follow up

In reply to a message not in the archive

From: Hersen Wu <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/core/dc.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +2 -2
index 92fab09843e4..4cdb855ef855 100644
@@ -332,7 +332,7 @@ static void set_dither_option(struct dc_stream_state *stream,
 {
 	struct bit_depth_reduction_params params;
 	struct dc_link *link = stream->status.link;
-	struct pipe_ctx *pipes;
+	struct pipe_ctx *pipes = NULL;
 	int i;
 
 	for (i = 0; i < MAX_PIPES; i++) {
@@ -344,7 +344,7 @@ static void set_dither_option(struct dc_stream_state *stream,
 	}
 
 	memset(&params, 0, sizeof(params));
-	if (!stream)
+	if (!pipes)
 		return;
 	if (option > DITHER_OPTION_MAX)
 		return;
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#84 lore raw

PATCH 79/103 drm/amd/display: Handle case when stream not found in set_dpms

In reply to a message not in the archive

From: Eric Yang <[email protected]>

When validate with context fail to add stream to the context, we have a case where set_dpms won't be able to find the stream that need to be disabled.

---
 drivers/gpu/drm/amd/display/dc/core/dc.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +6 -1
index 4cdb855ef855..fe66c6a21bc3 100644
@@ -363,7 +363,7 @@ void set_dpms(
 	struct dc_stream_state *stream,
 	bool dpms_off)
 {
-	struct pipe_ctx *pipe_ctx;
+	struct pipe_ctx *pipe_ctx = NULL;
 	int i;
 
 	for (i = 0; i < MAX_PIPES; i++) {
@@ -373,6 +373,11 @@ void set_dpms(
 		}
 	}
 
+	if (!pipe_ctx) {
+		ASSERT(0);
+		return;
+	}
+
 	if (stream->dpms_off != dpms_off) {
 		stream->dpms_off = dpms_off;
 		if (dpms_off)
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#85 lore raw

PATCH 80/103 drm/amd/display: Ignore wrong tautological warning

In reply to a message not in the archive

Some of our HW calcs functions compares a var against itself to check for NaN. Gcc isn't fond of it and wrongfully warns about a tautological comparison. Disable this check for dcn_calcs_math.c.

---
 drivers/gpu/drm/amd/display/dc/calcs/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/calcs/Makefile +1 -1
index a095472bf4b5..41ef35995b02 100644
@@ -5,7 +5,7 @@
 
 CFLAGS_dcn_calcs.o := -mhard-float -msse -mpreferred-stack-boundary=4
 CFLAGS_dcn_calc_auto.o := -mhard-float -msse -mpreferred-stack-boundary=4
-CFLAGS_dcn_calc_math.o := -mhard-float -msse -mpreferred-stack-boundary=4
+CFLAGS_dcn_calc_math.o := -mhard-float -msse -mpreferred-stack-boundary=4 -Wno-tautological-compare
 
 BW_CALCS = dce_calcs.o bw_fixed.o custom_float.o
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#86 lore raw

PATCH 81/103 drm/amd/display: dal 3.1.04

In reply to a message not in the archive

From: Tony Cheng <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dc.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dc.h +1 -1
index c832d5abbbdf..8fa9ef1167fa 100644
@@ -38,7 +38,7 @@
 #include "inc/compressor.h"
 #include "dml/display_mode_lib.h"
 
-#define DC_VER "3.1.03"
+#define DC_VER "3.1.04"
 
 #define MAX_SURFACES 3
 #define MAX_STREAMS 6
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#87 lore raw

PATCH 82/103 drm/amd/display: group DCN watermark registers

In reply to a message not in the archive

From: Eric Bernstein <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h | 19 +++++++++++--------
 1 file changed, 11 insertions(+), 8 deletions(-)
drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h +11 -8
index 9a04b8758888..52506155e361 100644
@@ -163,23 +163,15 @@
 	SR(REFCLK_CNTL), \
 	SR(DCHUBBUB_ARB_DATA_URGENCY_WATERMARK_A),\
 	SR(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_A),\
-	SR(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_A),\
-	SR(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_A),\
 	SR(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_A),\
 	SR(DCHUBBUB_ARB_DATA_URGENCY_WATERMARK_B),\
 	SR(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_B),\
-	SR(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_B),\
-	SR(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_B),\
 	SR(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_B),\
 	SR(DCHUBBUB_ARB_DATA_URGENCY_WATERMARK_C),\
 	SR(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_C),\
-	SR(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_C),\
-	SR(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_C),\
 	SR(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_C),\
 	SR(DCHUBBUB_ARB_DATA_URGENCY_WATERMARK_D),\
 	SR(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_D),\
-	SR(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_D),\
-	SR(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_D),\
 	SR(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_D),\
 	SR(DCHUBBUB_ARB_WATERMARK_CHANGE_CNTL),\
 	SR(DCHUBBUB_ARB_DRAM_STATE_CNTL),\
@@ -207,8 +199,19 @@
 	MMHUB_SR(MC_VM_SYSTEM_APERTURE_LOW_ADDR),\
 	MMHUB_SR(MC_VM_SYSTEM_APERTURE_HIGH_ADDR)
 
+#define HWSEQ_SR_WATERMARK_REG_LIST()\
+	SR(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_A),\
+	SR(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_A),\
+	SR(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_B),\
+	SR(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_B),\
+	SR(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_C),\
+	SR(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_C),\
+	SR(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_D),\
+	SR(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_D)
+
 #define HWSEQ_DCN1_REG_LIST()\
 	HWSEQ_DCN_REG_LIST(), \
+	HWSEQ_SR_WATERMARK_REG_LIST(), \
 	HWSEQ_PIXEL_RATE_REG_LIST(OTG), \
 	HWSEQ_PHYPLL_REG_LIST(OTG), \
 	SR(DCHUBBUB_SDPIF_FB_TOP),\
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#88 lore raw

PATCH 83/103 drm/amd/display: rename struct mem_input to hubp for dcn

In reply to a message not in the archive

From: Yue Hin Lau <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c   |   1 +
 drivers/gpu/drm/amd/display/dc/core/dc.c           |   1 +
 drivers/gpu/drm/amd/display/dc/core/dc_resource.c  |   3 +-
 drivers/gpu/drm/amd/display/dc/core/dc_stream.c    |  22 +--
 drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c |   3 -
 .../amd/display/dc/dce110/dce110_hw_sequencer.c    |  19 +--
 .../drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c  |  88 ++++++------
 .../gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.c | 157 ++++++++++-----------
 .../gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.h |  49 +++----
 .../gpu/drm/amd/display/dc/dcn10/dcn10_resource.c  |  26 ++--
 drivers/gpu/drm/amd/display/dc/inc/core_types.h    |   4 +-
 drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h       | 105 ++++++++++++++
 drivers/gpu/drm/amd/display/dc/inc/hw/mem_input.h  |   3 -
 13 files changed, 286 insertions(+), 195 deletions(-)
 create mode 100644 drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h
drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c +1 -0
index fd1db8dc2f35..c37f130b0170 100644
@@ -485,6 +485,7 @@ static void split_stream_across_pipes(
 
 	secondary_pipe->pipe_idx = pipe_idx;
 	secondary_pipe->plane_res.mi = pool->mis[secondary_pipe->pipe_idx];
+	secondary_pipe->plane_res.hubp = pool->hubps[secondary_pipe->pipe_idx];
 	secondary_pipe->plane_res.ipp = pool->ipps[secondary_pipe->pipe_idx];
 	secondary_pipe->plane_res.xfm = pool->transforms[secondary_pipe->pipe_idx];
 	if (primary_pipe->bottom_pipe) {
drivers/gpu/drm/amd/display/dc/core/dc.c +1 -0
index fe66c6a21bc3..4df5bf750257 100644
@@ -47,6 +47,7 @@
 #include "dc_link_ddc.h"
 #include "dm_helpers.h"
 #include "mem_input.h"
+#include "hubp.h"
 
 
 /*******************************************************************************
drivers/gpu/drm/amd/display/dc/core/dc_resource.c +2 -1
index 7d750369b78b..4fb4f0068f1b 100644
@@ -1024,7 +1024,7 @@ static int acquire_first_split_pipe(
 
 			memset(pipe_ctx, 0, sizeof(*pipe_ctx));
 			pipe_ctx->stream_res.tg = pool->timing_generators[i];
-			pipe_ctx->plane_res.mi = pool->mis[i];
+			pipe_ctx->plane_res.hubp = pool->hubps[i];
 			pipe_ctx->plane_res.ipp = pool->ipps[i];
 			pipe_ctx->plane_res.xfm = pool->transforms[i];
 			pipe_ctx->stream_res.opp = pool->opps[i];
@@ -1359,6 +1359,7 @@ static int acquire_first_free_pipe(
 
 			pipe_ctx->stream_res.tg = pool->timing_generators[i];
 			pipe_ctx->plane_res.mi = pool->mis[i];
+			pipe_ctx->plane_res.hubp = pool->hubps[i];
 			pipe_ctx->plane_res.ipp = pool->ipps[i];
 			pipe_ctx->plane_res.xfm = pool->transforms[i];
 			pipe_ctx->stream_res.opp = pool->opps[i];
drivers/gpu/drm/amd/display/dc/core/dc_stream.c +14 -8
index 3dd44bef56eb..411678736ed4 100644
@@ -200,7 +200,7 @@ bool dc_stream_set_cursor_attributes(
 	for (i = 0; i < MAX_PIPES; i++) {
 		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
 
-		if (pipe_ctx->stream != stream || !pipe_ctx->plane_res.mi || !pipe_ctx->plane_res.xfm)
+		if (pipe_ctx->stream != stream || !pipe_ctx->plane_res.xfm)
 			continue;
 		if (pipe_ctx->top_pipe && pipe_ctx->plane_state != pipe_ctx->top_pipe->plane_state)
 			continue;
@@ -210,10 +210,15 @@ bool dc_stream_set_cursor_attributes(
 			pipe_ctx->plane_res.ipp->funcs->ipp_cursor_set_attributes(
 						pipe_ctx->plane_res.ipp, attributes);
 
-		if (pipe_ctx->plane_res.mi->funcs->set_cursor_attributes != NULL)
+		if (pipe_ctx->plane_res.hubp != NULL)
+			pipe_ctx->plane_res.hubp->funcs->set_cursor_attributes(
+					pipe_ctx->plane_res.hubp, attributes);
+
+		if (pipe_ctx->plane_res.mi != NULL)
 			pipe_ctx->plane_res.mi->funcs->set_cursor_attributes(
 					pipe_ctx->plane_res.mi, attributes);
 
+
 		if (pipe_ctx->plane_res.xfm->funcs->set_cursor_attributes != NULL)
 			pipe_ctx->plane_res.xfm->funcs->set_cursor_attributes(
 				pipe_ctx->plane_res.xfm, attributes);
@@ -248,7 +253,7 @@ bool dc_stream_set_cursor_position(
 	for (i = 0; i < MAX_PIPES; i++) {
 		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
 		struct input_pixel_processor *ipp = pipe_ctx->plane_res.ipp;
-		struct mem_input *mi = pipe_ctx->plane_res.mi;
+		struct hubp *hubp = pipe_ctx->plane_res.hubp;
 		struct transform *xfm = pipe_ctx->plane_res.xfm;
 		struct dc_cursor_position pos_cpy = *position;
 		struct dc_cursor_mi_param param = {
@@ -260,7 +265,7 @@ bool dc_stream_set_cursor_position(
 		};
 
 		if (pipe_ctx->stream != stream ||
-				!pipe_ctx->plane_res.mi ||
+				(!pipe_ctx->plane_res.mi  && !pipe_ctx->plane_res.hubp) ||
 				!pipe_ctx->plane_state ||
 				!pipe_ctx->plane_res.xfm)
 			continue;
@@ -276,11 +281,12 @@ bool dc_stream_set_cursor_position(
 		if (ipp->funcs->ipp_cursor_set_position != NULL)
 			ipp->funcs->ipp_cursor_set_position(ipp, &pos_cpy, &param);
 
-		if (mi->funcs->set_cursor_attributes != NULL)
-			mi->funcs->set_cursor_position(mi, &pos_cpy, &param);
+		if (hubp != NULL)
+			hubp->funcs->set_cursor_position(hubp, &pos_cpy, &param);
+
+		if (xfm != NULL)
+			xfm->funcs->set_cursor_position(xfm, &pos_cpy, &param, hubp->curs_attr.width);
 
-		if (xfm->funcs->set_cursor_attributes != NULL)
-			xfm->funcs->set_cursor_position(xfm, &pos_cpy, &param, mi->curs_attr.width);
 	}
 
 	return true;
drivers/gpu/drm/amd/display/dc/dce/dce_mem_input.c +0 -3
index 83d9caa4f438..ed301c3b2736 100644
@@ -685,9 +685,6 @@ void dce_mem_input_construct(
 	dce_mi->regs = regs;
 	dce_mi->shifts = mi_shift;
 	dce_mi->masks = mi_mask;
-
-	dce_mi->base.mpcc_id = 0xf;
-	dce_mi->base.opp_id = 0xf;
 }
 
 void dce112_mem_input_construct(
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +10 -9
index bf7318246401..867a0eb9f2e0 100644
@@ -1378,15 +1378,14 @@ static enum dc_status apply_single_controller_ctx_to_hw(
 	}
 
 	/* mst support - use total stream count */
-#if defined(CONFIG_DRM_AMD_DC_DCN1_0)
-	if (pipe_ctx->plane_res.mi->funcs->allocate_mem_input != NULL)
-#endif
+	if (pipe_ctx->plane_res.mi != NULL) {
 		pipe_ctx->plane_res.mi->funcs->allocate_mem_input(
-					pipe_ctx->plane_res.mi,
-					stream->timing.h_total,
-					stream->timing.v_total,
-					stream->timing.pix_clk_khz,
-					context->stream_count);
+				pipe_ctx->plane_res.mi,
+				stream->timing.h_total,
+				stream->timing.v_total,
+				stream->timing.pix_clk_khz,
+				context->stream_count);
+	}
 
 	pipe_ctx->stream->sink->link->psr_enabled = false;
 
@@ -1556,7 +1555,7 @@ static void set_safe_displaymarks(
 		SAFE_NBP_MARK, SAFE_NBP_MARK, SAFE_NBP_MARK, SAFE_NBP_MARK };
 
 	for (i = 0; i < MAX_PIPES; i++) {
-		if (res_ctx->pipe_ctx[i].stream == NULL)
+		if (res_ctx->pipe_ctx[i].stream == NULL || res_ctx->pipe_ctx[i].plane_res.mi == NULL)
 			continue;
 
 		res_ctx->pipe_ctx[i].plane_res.mi->funcs->mem_input_program_display_marks(
@@ -1565,6 +1564,7 @@ static void set_safe_displaymarks(
 				max_marks,
 				max_marks,
 				MAX_WATERMARK);
+
 		if (i == underlay_idx)
 			res_ctx->pipe_ctx[i].plane_res.mi->funcs->mem_input_program_chroma_display_marks(
 				res_ctx->pipe_ctx[i].plane_res.mi,
@@ -1572,6 +1572,7 @@ static void set_safe_displaymarks(
 				max_marks,
 				max_marks,
 				MAX_WATERMARK);
+
 	}
 }
 
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +44 -44

266 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.c +72 -85

381 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.h +21 -28

118 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +13 -13
index 2caad73257e0..93edbba762e8 100644
@@ -704,9 +704,9 @@ static void destruct(struct dcn10_resource_pool *pool)
 		if (pool->base.ipps[i] != NULL)
 			pool->base.ipps[i]->funcs->ipp_destroy(&pool->base.ipps[i]);
 
-		if (pool->base.mis[i] != NULL) {
-			kfree(TO_DCN10_MEM_INPUT(pool->base.mis[i]));
-			pool->base.mis[i] = NULL;
+		if (pool->base.hubps[i] != NULL) {
+			kfree(TO_DCN10_HUBP(pool->base.hubps[i]));
+			pool->base.hubps[i] = NULL;
 		}
 
 		if (pool->base.irqs != NULL) {
@@ -753,19 +753,19 @@ static void destruct(struct dcn10_resource_pool *pool)
 	kfree(pool->base.pp_smu);
 }
 
-static struct mem_input *dcn10_mem_input_create(
+static struct hubp *dcn10_hubp_create(
 	struct dc_context *ctx,
 	uint32_t inst)
 {
-	struct dcn10_mem_input *mem_inputn10 =
-		kzalloc(sizeof(struct dcn10_mem_input), GFP_KERNEL);
+	struct dcn10_hubp *hubp1 =
+		kzalloc(sizeof(struct dcn10_hubp), GFP_KERNEL);
 
-	if (!mem_inputn10)
+	if (!hubp1)
 		return NULL;
 
-	dcn10_mem_input_construct(mem_inputn10, ctx, inst,
-				  &mi_regs[inst], &mi_shift, &mi_mask);
-	return &mem_inputn10->base;
+	dcn10_hubp_construct(hubp1, ctx, inst,
+			     &mi_regs[inst], &mi_shift, &mi_mask);
+	return &hubp1->base;
 }
 
 static void get_pixel_clock_parameters(
@@ -918,7 +918,7 @@ static struct pipe_ctx *dcn10_acquire_idle_pipe_for_layer(
 	idle_pipe->stream_res.tg = head_pipe->stream_res.tg;
 	idle_pipe->stream_res.opp = head_pipe->stream_res.opp;
 
-	idle_pipe->plane_res.mi = pool->mis[idle_pipe->pipe_idx];
+	idle_pipe->plane_res.hubp = pool->hubps[idle_pipe->pipe_idx];
 	idle_pipe->plane_res.ipp = pool->ipps[idle_pipe->pipe_idx];
 	idle_pipe->plane_res.xfm = pool->transforms[idle_pipe->pipe_idx];
 
@@ -1357,8 +1357,8 @@ static bool construct(
 		if ((pipe_fuses & (1 << i)) != 0)
 			continue;
 
-		pool->base.mis[j] = dcn10_mem_input_create(ctx, i);
-		if (pool->base.mis[j] == NULL) {
+		pool->base.hubps[j] = dcn10_hubp_create(ctx, i);
+		if (pool->base.hubps[j] == NULL) {
 			BREAK_TO_DEBUGGER();
 			dm_error(
 				"DC: failed to create memory input!\n");
drivers/gpu/drm/amd/display/dc/inc/core_types.h +3 -1
index 118122d12e2b..45518c05e846 100644
@@ -32,6 +32,7 @@
 #include "ddc_service_types.h"
 #include "dc_bios_types.h"
 #include "mem_input.h"
+#include "hubp.h"
 #if defined(CONFIG_DRM_AMD_DC_DCN1_0)
 #include "mpc.h"
 #endif
@@ -129,6 +130,7 @@ struct audio_support{
 
 struct resource_pool {
 	struct mem_input *mis[MAX_PIPES];
+	struct hubp *hubps[MAX_PIPES];
 	struct input_pixel_processor *ipps[MAX_PIPES];
 	struct transform *transforms[MAX_PIPES];
 	struct output_pixel_processor *opps[MAX_PIPES];
@@ -178,7 +180,7 @@ struct stream_resource {
 
 struct plane_resource {
 	struct scaler_data scl_data;
-
+	struct hubp *hubp;
 	struct mem_input *mi;
 	struct input_pixel_processor *ipp;
 	struct transform *xfm;
drivers/gpu/drm/amd/display/dc/inc/hw/hubp.h new file +105 -0

105 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/inc/hw/mem_input.h +0 -3
index 5c8e45ba8f45..3e1e7e6a8792 100644
@@ -69,10 +69,7 @@ struct mem_input {
 	struct dc_plane_address request_address;
 	struct dc_plane_address current_address;
 	int inst;
-	int opp_id;
-	int mpcc_id;
 	struct stutter_modes stutter_mode;
-	struct dc_cursor_attributes curs_attr;
 };
 
 struct vm_system_aperture_param {
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#89 lore raw

PATCH 84/103 drm/amd/display: Add DPP capabilities

In reply to a message not in the archive

From: Eric Bernstein <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c   |  5 +++
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.h   | 23 +++--------
 .../gpu/drm/amd/display/dc/dcn10/dcn10_dpp_dscl.c  | 42 +++++++-------------
 drivers/gpu/drm/amd/display/dc/inc/hw/transform.h  | 46 ++++++++++++++++++++++
 4 files changed, 72 insertions(+), 44 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c +5 -0
index 05df3b222945..fb3c78858ebe 100644
@@ -444,6 +444,10 @@ static const struct transform_funcs dcn10_dpp_funcs = {
 		.set_cursor_position = dcn10_set_cursor_position,
 };
 
+static struct dpp_caps dcn10_dpp_cap = {
+	.dscl_data_proc_format = DSCL_DATA_PRCESSING_FIXED_FORMAT,
+	.dscl_calc_lb_num_partitions = dscl1_calc_lb_num_partitions,
+};
 
 /*****************************************/
 /* Constructor, Destructor               */
@@ -461,6 +465,7 @@ void dcn10_dpp_construct(
 
 	xfm->base.inst = inst;
 	xfm->base.funcs = &dcn10_dpp_funcs;
+	xfm->base.caps = &dcn10_dpp_cap;
 
 	xfm->tf_regs = tf_regs;
 	xfm->tf_shift = tf_shift;
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.h +6 -17
index 153aa3c32762..4cf18a500dcd 100644
@@ -1264,23 +1264,6 @@ struct dcn10_dpp {
 	bool is_write_to_ram_a_safe;
 };
 
-
-enum lb_memory_config {
-	/* Enable all 3 pieces of memory */
-	LB_MEMORY_CONFIG_0 = 0,
-
-	/* Enable only the first piece of memory */
-	LB_MEMORY_CONFIG_1 = 1,
-
-	/* Enable only the second piece of memory */
-	LB_MEMORY_CONFIG_2 = 2,
-
-	/* Only applicable in 4:2:0 mode, enable all 3 pieces of memory and the
-	 * last piece of chroma memory used for the luma storage
-	 */
-	LB_MEMORY_CONFIG_3 = 3
-};
-
 enum dcn10_input_csc_select {
 	INPUT_CSC_SELECT_BYPASS = 0,
 	INPUT_CSC_SELECT_ICSC,
@@ -1292,6 +1275,12 @@ bool is_lb_conf_valid(
 		int num_partitions,
 		int vtaps);
 
+void dscl1_calc_lb_num_partitions(
+		const struct scaler_data *scl_data,
+		enum lb_memory_config lb_config,
+		int *num_part_y,
+		int *num_part_c);
+
 void ippn10_degamma_ram_select(
 		struct transform *xfm_base,
 							bool use_ram_a);
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_dscl.c +15 -27

105 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/inc/hw/transform.h +46 -0

62 lines not on this page · raw message

Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#90 lore raw

PATCH 85/103 drm/amd/display: Add missing newlines in pstate wait debug messages

In reply to a message not in the archive

From: Andrew Jiang <[email protected]>

These were spamming the debugger logs.

---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +2 -2
index c34dbcd1fba4..059da7882a68 100644
@@ -302,7 +302,7 @@ static void verify_allow_pstate_change_high(
 
 			if (i > pstate_wait_expected_timeout_us)
 				dm_logger_write(hws->ctx->logger, LOG_WARNING,
-						"pstate took longer than expected ~%dus",
+						"pstate took longer than expected ~%dus\n",
 						i);
 
 			return;
@@ -326,7 +326,7 @@ static void verify_allow_pstate_change_high(
 	}
 
 	dm_logger_write(hws->ctx->logger, LOG_WARNING,
-			"pstate TEST_DEBUG_DATA: 0x%X",
+			"pstate TEST_DEBUG_DATA: 0x%X\n",
 			debug_data);
 	BREAK_TO_DEBUGGER();
 }
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#91 lore raw

PATCH 86/103 drm/amd/display: Fix up plane_states add/remove logic

In reply to a message not in the archive

From: Andrew Jiang <[email protected]>

Our plane_states array trimming logic was faulty, we should be starting to shuffle from the plane that was just released, not from the very beginning of the array.

Also get rid of a leftover line that was setting the plane state at the stream index to null, leading to issues. Also move the dc_plane_state_retain call to where we reference plane_state, in case we do hit the error case where we can't get a free pipe.

---
 drivers/gpu/drm/amd/display/dc/core/dc_resource.c | 13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_resource.c +5 -8
index 4fb4f0068f1b..c80bde94472f 100644
@@ -1073,9 +1073,6 @@ bool dc_add_plane_to_context(
 		return false;
 	}
 
-	/* retain new surfaces */
-	dc_plane_state_retain(plane_state);
-
 	free_pipe = acquire_free_pipe_for_stream(context, pool, stream);
 
 #if defined(CONFIG_DRM_AMD_DC_DCN1_0)
@@ -1085,11 +1082,11 @@ bool dc_add_plane_to_context(
 			free_pipe = &context->res_ctx.pipe_ctx[pipe_idx];
 	}
 #endif
-	if (!free_pipe) {
-		stream_status->plane_states[i] = NULL;
+	if (!free_pipe)
 		return false;
-	}
 
+	/* retain new surfaces */
+	dc_plane_state_retain(plane_state);
 	free_pipe->plane_state = plane_state;
 
 	if (head_pipe != free_pipe) {
@@ -1181,8 +1178,8 @@ bool dc_remove_plane_from_context(
 
 	stream_status->plane_count--;
 
-	/* Trim back arrays */
-	for (i = 0; i < stream_status->plane_count; i++)
+	/* Start at the plane we've just released, and move all the planes one index forward to "trim" the array */
+	for (; i < stream_status->plane_count; i++)
 		stream_status->plane_states[i] = stream_status->plane_states[i + 1];
 
 	stream_status->plane_states[stream_status->plane_count] = NULL;
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#92 lore raw

PATCH 87/103 drm/amd/display: dal 3.1.05

In reply to a message not in the archive

From: Tony Cheng <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dc.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dc.h +1 -1
index 8fa9ef1167fa..6e9b21928d4f 100644
@@ -38,7 +38,7 @@
 #include "inc/compressor.h"
 #include "dml/display_mode_lib.h"
 
-#define DC_VER "3.1.04"
+#define DC_VER "3.1.05"
 
 #define MAX_SURFACES 3
 #define MAX_STREAMS 6
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#93 lore raw

PATCH 88/103 drm/amd/display: renaming filename for hubp

In reply to a message not in the archive

From: Yue Hin Lau <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dcn10/Makefile                           | 2 +-
 .../gpu/drm/amd/display/dc/dcn10/{dcn10_mem_input.c => dcn10_hubp.c}    | 2 +-
 .../gpu/drm/amd/display/dc/dcn10/{dcn10_mem_input.h => dcn10_hubp.h}    | 0
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c               | 2 +-
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c                   | 2 +-
 5 files changed, 4 insertions(+), 4 deletions(-)
 rename drivers/gpu/drm/amd/display/dc/dcn10/{dcn10_mem_input.c => dcn10_hubp.c} (99%)
 rename drivers/gpu/drm/amd/display/dc/dcn10/{dcn10_mem_input.h => dcn10_hubp.h} (100%)
drivers/gpu/drm/amd/display/dc/dcn10/Makefile +1 -1
index 6b19106f4679..ebeb88283a14 100644
@@ -3,7 +3,7 @@
 
 DCN10 = dcn10_resource.o dcn10_ipp.o dcn10_hw_sequencer.o \
 		dcn10_dpp.o dcn10_opp.o dcn10_timing_generator.o \
-		dcn10_mem_input.o dcn10_mpc.o \
+		dcn10_hubp.o dcn10_mpc.o \
 		dcn10_dpp_dscl.o dcn10_dpp_cm.o dcn10_cm_common.o
 
 AMD_DAL_DCN10 = $(addprefix $(AMDDALPATH)/dc/dcn10/,$(DCN10))
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c renamed from drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.c +1 -1
similarity index 99%
rename from drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.c
rename to drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.c
index 9ee42d9840fb..b13dee64e0ce 100644
@@ -24,9 +24,9 @@
  */
 #include "dm_services.h"
 #include "dce_calcs.h"
-#include "dcn10_mem_input.h"
 #include "reg_helper.h"
 #include "basics/conversion.h"
+#include "dcn10_hubp.h"
 
 #define REG(reg)\
 	hubp1->mi_regs->reg
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h renamed from drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.h +0 -0
similarity index 100%
rename from drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mem_input.h
rename to drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubp.h
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +1 -1
index 059da7882a68..6b866b169985 100644
@@ -31,7 +31,6 @@
 #include "dce110/dce110_hw_sequencer.h"
 #include "dce/dce_hwseq.h"
 #include "abm.h"
-#include "dcn10/dcn10_mem_input.h"
 #include "dcn10/dcn10_timing_generator.h"
 #include "dcn10/dcn10_dpp.h"
 #include "dcn10/dcn10_mpc.h"
@@ -41,6 +40,7 @@
 #include "mpc.h"
 #include "reg_helper.h"
 #include "custom_float.h"
+#include "dcn10_hubp.h"
 
 #define CTX \
 	hws->ctx
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +1 -1
index 93edbba762e8..1da5105ed30a 100644
@@ -42,12 +42,12 @@
 #include "dce/dce_stream_encoder.h"
 #include "dce/dce_clocks.h"
 #include "dce/dce_clock_source.h"
-#include "dcn10/dcn10_mem_input.h"
 #include "dce/dce_audio.h"
 #include "dce/dce_hwseq.h"
 #include "../virtual/virtual_stream_encoder.h"
 #include "dce110/dce110_resource.h"
 #include "dce112/dce112_resource.h"
+#include "dcn10_hubp.h"
 
 #include "vega10/soc15ip.h"
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#94 lore raw

PATCH 89/103 drm/amd/display: Pull transfer func defines into DC

In reply to a message not in the archive

From: SivapiriyanKumarasamy <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dc.h | 38 +++++++++++++++++++++++++++++++++++++
 1 file changed, 38 insertions(+)
drivers/gpu/drm/amd/display/dc/dc.h +38 -0
index 6e9b21928d4f..c7b31adc736e 100644
@@ -318,6 +318,38 @@ enum {
 	TRANSFER_FUNC_POINTS = 1025
 };
 
+// Moved here from color module for linux
+enum color_transfer_func {
+	transfer_func_unknown,
+	transfer_func_srgb,
+	transfer_func_bt709,
+	transfer_func_pq2084,
+	transfer_func_pq2084_interim,
+	transfer_func_linear_0_1,
+	transfer_func_linear_0_125,
+	transfer_func_dolbyvision,
+	transfer_func_gamma_22,
+	transfer_func_gamma_26
+};
+
+enum color_color_space {
+	color_space_unsupported,
+	color_space_srgb,
+	color_space_bt601,
+	color_space_bt709,
+	color_space_xv_ycc_bt601,
+	color_space_xv_ycc_bt709,
+	color_space_xr_rgb,
+	color_space_bt2020,
+	color_space_adobe,
+	color_space_dci_p3,
+	color_space_sc_rgb_ms_ref,
+	color_space_display_native,
+	color_space_app_ctrl,
+	color_space_dolby_vision,
+	color_space_custom_coordinates
+};
+
 struct dc_hdr_static_metadata {
 	/* display chromaticities and white point in units of 0.00001 */
 	unsigned int chromaticity_green_x;
@@ -398,6 +430,12 @@ struct dc_plane_state {
 	struct dc_gamma *gamma_correction;
 	struct dc_transfer_func *in_transfer_func;
 
+	// sourceContentAttribute cache
+	bool is_source_input_valid;
+	struct dc_hdr_static_metadata source_input_mastering_info;
+	enum color_color_space source_input_color_space;
+	enum color_transfer_func source_input_tf;
+
 	enum dc_color_space color_space;
 	enum surface_pixel_format format;
 	enum dc_rotation_angle rotation;
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#95 lore raw

PATCH 90/103 drm/amd/display: Fix underscan not using proper scaling

In reply to a message not in the archive

From: Bhawanpreet Lakha <[email protected]>

fix underscan not being set correctly

---
 drivers/gpu/drm/amd/display/dc/core/dc.c | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +11 -2
index 4df5bf750257..12aeb6d83f42 100644
@@ -1362,7 +1362,7 @@ void dc_commit_updates_for_stream(struct dc *dc,
 	enum surface_update_type update_type;
 	struct dc_state *context;
 	struct dc_context *dc_ctx = dc->ctx;
-	int i;
+	int i, j;
 
 	stream_status = dc_stream_get_status(stream);
 	context = dc->current_state;
@@ -1396,8 +1396,17 @@ void dc_commit_updates_for_stream(struct dc *dc,
 		if (srf_updates[i].flip_addr)
 			surface->address = srf_updates[i].flip_addr->address;
 
-	}
+		if (update_type >= UPDATE_TYPE_MED) {
+			for (j = 0; j < dc->res_pool->pipe_count; j++) {
+				struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[j];
+
+				if (pipe_ctx->plane_state != surface)
+					continue;
 
+				resource_build_scaling_params(pipe_ctx);
+			}
+		}
+	}
 
 	commit_planes_for_stream(
 				dc,
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#96 lore raw

PATCH 91/103 drm/amd/display: Expose some MPC functions for reuse

In reply to a message not in the archive

From: Eric Bernstein <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c   | 11 +++---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.h   | 45 ++++++++++++++++------
 .../gpu/drm/amd/display/dc/dcn10/dcn10_resource.c  |  6 ++-
 drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h        |  7 ++--
 4 files changed, 48 insertions(+), 21 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c +5 -6
index 334f072cea1d..76573e1f5b01 100644
@@ -65,7 +65,7 @@ static void mpc10_set_bg_color(
 			MPCC_BG_B_CB, bg_b_cb);
 }
 
-static void mpc10_assert_idle_mpcc(struct mpc *mpc, int id)
+void mpc10_assert_idle_mpcc(struct mpc *mpc, int id)
 {
 	struct dcn10_mpc *mpc10 = TO_DCN10_MPC(mpc);
 
@@ -116,7 +116,7 @@ static void mpc10_assert_mpcc_idle_before_connect(struct dcn10_mpc *mpc10, int i
 	}
 }
 
-static void mpc10_mpcc_remove(
+void mpc10_mpcc_remove(
 		struct mpc *mpc,
 		struct mpc_tree_cfg *tree_cfg,
 		int opp_id,
@@ -265,7 +265,7 @@ static void mpc10_add_to_tree_cfg(
 	tree_cfg->num_pipes++;
 }
 
-static int mpc10_mpcc_add(struct mpc *mpc, struct mpcc_cfg *cfg)
+int mpc10_mpcc_add(struct mpc *mpc, struct mpcc_cfg *cfg)
 {
 	struct dcn10_mpc *mpc10 = TO_DCN10_MPC(mpc);
 	int mpcc_id, z_idx;
@@ -313,7 +313,7 @@ static int mpc10_mpcc_add(struct mpc *mpc, struct mpcc_cfg *cfg)
 	return mpcc_id;
 }
 
-static void mpc10_update_blend_mode(
+void mpc10_update_blend_mode(
 		struct mpc *mpc,
 		struct mpcc_cfg *cfg)
 {
@@ -339,8 +339,7 @@ const struct mpc_funcs dcn10_mpc_funcs = {
 		.add = mpc10_mpcc_add,
 		.remove = mpc10_mpcc_remove,
 		.wait_for_idle = mpc10_assert_idle_mpcc,
-		.set_denorm = NULL,
-		.update_blend_mode = mpc10_update_blend_mode
+		.update_blend_mode = mpc10_update_blend_mode,
 };
 
 void dcn10_mpc_construct(struct dcn10_mpc *mpc10,
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.h +34 -11

68 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +5 -1
index 1da5105ed30a..4764590f9097 100644
@@ -329,7 +329,11 @@ static const struct dcn_mpc_registers mpc_regs = {
 		MPC_COMMON_REG_LIST_DCN1_0(0),
 		MPC_COMMON_REG_LIST_DCN1_0(1),
 		MPC_COMMON_REG_LIST_DCN1_0(2),
-		MPC_COMMON_REG_LIST_DCN1_0(3)
+		MPC_COMMON_REG_LIST_DCN1_0(3),
+		MPC_OUT_MUX_COMMON_REG_LIST_DCN1_0(0),
+		MPC_OUT_MUX_COMMON_REG_LIST_DCN1_0(1),
+		MPC_OUT_MUX_COMMON_REG_LIST_DCN1_0(2),
+		MPC_OUT_MUX_COMMON_REG_LIST_DCN1_0(3)
 };
 
 static const struct dcn_mpc_shift mpc_shift = {
drivers/gpu/drm/amd/display/dc/inc/hw/mpc.h +4 -3
index fb590f5353a0..d4188b2c0626 100644
@@ -46,15 +46,16 @@ struct mpc {
 
 struct mpc_funcs {
 	int (*add)(struct mpc *mpc, struct mpcc_cfg *cfg);
+
 	void (*remove)(struct mpc *mpc,
 			struct mpc_tree_cfg *tree_cfg,
 			int opp_id,
 			int mpcc_inst);
+
 	void (*wait_for_idle)(struct mpc *mpc, int id);
-	void (*set_denorm)(struct mpc *mpc,
-			int opp_id,
-			enum dc_color_depth output_depth);
+
 	void (*update_blend_mode)(struct mpc *mpc, struct mpcc_cfg *cfg);
+
 };
 
 #endif
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#97 lore raw

PATCH 92/103 drm/amd/display: change when to set tap_hardcode_coeff

In reply to a message not in the archive

From: Bhawanpreet Lakha <[email protected]>

Set hardcode_coeff only when BOTH chroma and luma taps are more than 1

without this underscan with h or v set to 0 darkens the screen (either h or v set to 0, not both)

---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_dscl.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_dscl.c +2 -2
index 830dbd5d4144..eda6c8010fb8 100644
@@ -306,10 +306,10 @@ static void dpp_set_scl_filter(
 
 	h_2tap_hardcode_coef_en = scl_data->taps.h_taps < 3
 					&& scl_data->taps.h_taps_c < 3
-		&& (scl_data->taps.h_taps > 1 || scl_data->taps.h_taps_c > 1);
+		&& (scl_data->taps.h_taps > 1 && scl_data->taps.h_taps_c > 1);
 	v_2tap_hardcode_coef_en = scl_data->taps.v_taps < 3
 					&& scl_data->taps.v_taps_c < 3
-		&& (scl_data->taps.v_taps > 1 || scl_data->taps.v_taps_c > 1);
+		&& (scl_data->taps.v_taps > 1 && scl_data->taps.v_taps_c > 1);
 
 	h_2tap_sharp_en = h_2tap_hardcode_coef_en && h_2tap_sharp_factor != 0;
 	v_2tap_sharp_en = v_2tap_hardcode_coef_en && v_2tap_sharp_factor != 0;
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#98 lore raw

PATCH 93/103 drm/amd/display: [RV] bug in cm programming

In reply to a message not in the archive

From: Vitaly Prosyak <[email protected]>

When surface bigger then 10 bpc the output pixel does not match to the required value.Update CRC's.

---
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.h    | 6 +++++-
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm.c | 2 +-
 2 files changed, 6 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.h +5 -1
index 4cf18a500dcd..c2e3ff883537 100644
@@ -164,7 +164,8 @@
 	SRI(CM_IGAM_LUT_RW_CONTROL, CM, id), \
 	SRI(CM_IGAM_LUT_RW_INDEX, CM, id), \
 	SRI(CM_IGAM_LUT_SEQ_COLOR, CM, id), \
-	SRI(CURSOR_CONTROL, CURSOR, id)
+	SRI(CURSOR_CONTROL, CURSOR, id), \
+	SRI(CM_CMOUT_CONTROL, CM, id)
 
 
 #define TF_REG_LIST_SH_MASK_DCN(mask_sh)\
@@ -401,6 +402,7 @@
 	TF_SF(CM0_CM_CONTROL, CM_BYPASS_EN, mask_sh), \
 	TF_SF(CM0_CM_IGAM_LUT_SEQ_COLOR, CM_IGAM_LUT_SEQ_COLOR, mask_sh), \
 	TF_SF(CNVC_CFG0_FORMAT_CONTROL, OUTPUT_FP, mask_sh), \
+	TF_SF(CM0_CM_CMOUT_CONTROL, CM_CMOUT_ROUND_TRUNC_MODE, mask_sh), \
 	TF_SF(CURSOR0_CURSOR_CONTROL, CURSOR_MODE, mask_sh), \
 	TF_SF(CURSOR0_CURSOR_CONTROL, CURSOR_PITCH, mask_sh), \
 	TF_SF(CURSOR0_CURSOR_CONTROL, CURSOR_LINES_PER_CHUNK, mask_sh), \
@@ -549,6 +551,7 @@
 	type CM_RGAM_RAMA_EXP_REGION33_LUT_OFFSET; \
 	type CM_RGAM_RAMA_EXP_REGION33_NUM_SEGMENTS; \
 	type CM_RGAM_LUT_MODE; \
+	type CM_CMOUT_ROUND_TRUNC_MODE; \
 	type OBUF_BYPASS; \
 	type OBUF_H_2X_UPSCALE_EN; \
 	type CM_BLNDGAM_LUT_MODE; \
@@ -1081,6 +1084,7 @@ struct dcn_dpp_registers {
 	uint32_t CM_RGAM_RAMA_REGION_0_1;
 	uint32_t CM_RGAM_RAMA_REGION_32_33;
 	uint32_t CM_RGAM_CONTROL;
+	uint32_t CM_CMOUT_CONTROL;
 	uint32_t OBUF_CONTROL;
 	uint32_t CM_BLNDGAM_LUT_WRITE_EN_MASK;
 	uint32_t CM_BLNDGAM_CONTROL;
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm.c +1 -1
index 61e2a8919e9c..f88fd2e2e0aa 100644
@@ -602,7 +602,7 @@ static void ippn10_enable_cm_block(
 		struct transform *xfm_base)
 {
 	struct dcn10_dpp *xfm = TO_DCN10_DPP(xfm_base);
-
+	REG_UPDATE(CM_CMOUT_CONTROL, CM_CMOUT_ROUND_TRUNC_MODE, 8);
 	REG_UPDATE(CM_CONTROL, CM_BYPASS_EN, 0);
 }
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#99 lore raw

PATCH 94/103 drm/amd/display: Fix a logic defect in cursor move

In reply to a message not in the archive

From: Jerry Zuo <[email protected]>

Regression caused by: Ib98354194d7

Need to check crtc->stream before updating cursor attributes and position.

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +5 -4
index 1dc110042956..ff9a35ecf040 100644
@@ -3899,14 +3899,15 @@ static void handle_cursor_update(
 
 	attributes.pitch = attributes.width;
 
-	if (!dc_stream_set_cursor_attributes(crtc_state->stream,
-					     &attributes))
-		DRM_ERROR("DC failed to set cursor attributes\n");
+	if (crtc_state->stream) {
+		if (!dc_stream_set_cursor_attributes(crtc_state->stream,
+							 &attributes))
+			DRM_ERROR("DC failed to set cursor attributes\n");
 
-	if (crtc_state->stream)
 		if (!dc_stream_set_cursor_position(crtc_state->stream,
 						   &position))
 			DRM_ERROR("DC failed to set cursor position\n");
+	}
 }
 
 static void prepare_flip_isr(struct amdgpu_crtc *acrtc)
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#100 lore raw

PATCH 95/103 drm/amd/display: fix null dereference in reset hw ctx

In reply to a message not in the archive

From: Eric Yang <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c +1 -1
index 867a0eb9f2e0..5e5763c0ee94 100644
@@ -1877,7 +1877,7 @@ static void dce110_reset_hw_ctx_wrap(
 			struct clock_source *old_clk = pipe_ctx_old->clock_source;
 
 			/* disable already, no need to disable again */
-			if (!pipe_ctx->stream->dpms_off)
+			if (pipe_ctx->stream && !pipe_ctx->stream->dpms_off)
 				core_link_disable_stream(pipe_ctx_old, FREE_ACQUIRED_RESOURCE);
 
 			pipe_ctx_old->stream_res.tg->funcs->set_blank(pipe_ctx_old->stream_res.tg, true);
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#101 lore raw

PATCH 96/103 drm/amd/display: add null check for cursor functions

In reply to a message not in the archive

From: Yue Hin Lau <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/core/dc_stream.c | 14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_stream.c +10 -4
index 411678736ed4..10421ea31373 100644
@@ -210,11 +210,13 @@ bool dc_stream_set_cursor_attributes(
 			pipe_ctx->plane_res.ipp->funcs->ipp_cursor_set_attributes(
 						pipe_ctx->plane_res.ipp, attributes);
 
-		if (pipe_ctx->plane_res.hubp != NULL)
+		if (pipe_ctx->plane_res.hubp != NULL &&
+				pipe_ctx->plane_res.hubp->funcs->set_cursor_attributes != NULL)
 			pipe_ctx->plane_res.hubp->funcs->set_cursor_attributes(
 					pipe_ctx->plane_res.hubp, attributes);
 
-		if (pipe_ctx->plane_res.mi != NULL)
+		if (pipe_ctx->plane_res.mi != NULL &&
+				pipe_ctx->plane_res.mi->funcs->set_cursor_attributes != NULL)
 			pipe_ctx->plane_res.mi->funcs->set_cursor_attributes(
 					pipe_ctx->plane_res.mi, attributes);
 
@@ -253,6 +255,7 @@ bool dc_stream_set_cursor_position(
 	for (i = 0; i < MAX_PIPES; i++) {
 		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
 		struct input_pixel_processor *ipp = pipe_ctx->plane_res.ipp;
+		struct mem_input *mi = pipe_ctx->plane_res.mi;
 		struct hubp *hubp = pipe_ctx->plane_res.hubp;
 		struct transform *xfm = pipe_ctx->plane_res.xfm;
 		struct dc_cursor_position pos_cpy = *position;
@@ -281,10 +284,13 @@ bool dc_stream_set_cursor_position(
 		if (ipp->funcs->ipp_cursor_set_position != NULL)
 			ipp->funcs->ipp_cursor_set_position(ipp, &pos_cpy, &param);
 
-		if (hubp != NULL)
+		if (mi != NULL && mi->funcs->set_cursor_position != NULL)
+			mi->funcs->set_cursor_position(mi, &pos_cpy, &param);
+
+		if (hubp != NULL && hubp->funcs->set_cursor_position != NULL)
 			hubp->funcs->set_cursor_position(hubp, &pos_cpy, &param);
 
-		if (xfm != NULL)
+		if (xfm != NULL && xfm->funcs->set_cursor_position != NULL)
 			xfm->funcs->set_cursor_position(xfm, &pos_cpy, &param, hubp->curs_attr.width);
 
 	}
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#102 lore raw

PATCH 97/103 drm/amd/display: rename transform to dpp for dcn

In reply to a message not in the archive

From: Yue Hin Lau <[email protected]>

---
 drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c   |   1 +
 drivers/gpu/drm/amd/display/dc/core/dc.c           |   1 +
 drivers/gpu/drm/amd/display/dc/core/dc_resource.c  |  16 +-
 drivers/gpu/drm/amd/display/dc/core/dc_stream.c    |  16 +-
 drivers/gpu/drm/amd/display/dc/core/dc_surface.c   |   1 +
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c   | 144 ++++++------
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.h   |  94 ++++----
 .../gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm.c    | 261 +++++++++++----------
 .../gpu/drm/amd/display/dc/dcn10/dcn10_dpp_dscl.c  | 104 ++++----
 .../drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c  |  59 ++---
 .../gpu/drm/amd/display/dc/dcn10/dcn10_resource.c  |  22 +-
 drivers/gpu/drm/amd/display/dc/inc/core_types.h    |   3 +
 drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h        | 134 +++++++++++
 13 files changed, 509 insertions(+), 347 deletions(-)
 create mode 100644 drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h
drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c +1 -0
index c37f130b0170..cf0c1459f546 100644
@@ -488,6 +488,7 @@ static void split_stream_across_pipes(
 	secondary_pipe->plane_res.hubp = pool->hubps[secondary_pipe->pipe_idx];
 	secondary_pipe->plane_res.ipp = pool->ipps[secondary_pipe->pipe_idx];
 	secondary_pipe->plane_res.xfm = pool->transforms[secondary_pipe->pipe_idx];
+	secondary_pipe->plane_res.dpp = pool->dpps[secondary_pipe->pipe_idx];
 	if (primary_pipe->bottom_pipe) {
 		ASSERT(primary_pipe->bottom_pipe != secondary_pipe);
 		secondary_pipe->bottom_pipe = primary_pipe->bottom_pipe;
drivers/gpu/drm/amd/display/dc/core/dc.c +1 -0
index 12aeb6d83f42..80ddccc2570a 100644
@@ -38,6 +38,7 @@
 #include "bios_parser_interface.h"
 #include "include/irq_service_interface.h"
 #include "transform.h"
+#include "dpp.h"
 #include "timing_generator.h"
 #include "virtual/virtual_link_encoder.h"
 
drivers/gpu/drm/amd/display/dc/core/dc_resource.c +13 -3
index c80bde94472f..303d7ec684d9 100644
@@ -31,6 +31,7 @@
 #include "opp.h"
 #include "timing_generator.h"
 #include "transform.h"
+#include "dpp.h"
 #include "core_types.h"
 #include "set_mode_types.h"
 #include "virtual/virtual_stream_encoder.h"
@@ -853,8 +854,13 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx)
 	pipe_ctx->plane_res.scl_data.v_active = timing->v_addressable;
 
 	/* Taps calculations */
-	res = pipe_ctx->plane_res.xfm->funcs->transform_get_optimal_number_of_taps(
-		pipe_ctx->plane_res.xfm, &pipe_ctx->plane_res.scl_data, &plane_state->scaling_quality);
+	if (pipe_ctx->plane_res.xfm != NULL)
+		res = pipe_ctx->plane_res.xfm->funcs->transform_get_optimal_number_of_taps(
+				pipe_ctx->plane_res.xfm, &pipe_ctx->plane_res.scl_data, &plane_state->scaling_quality);
+
+	if (pipe_ctx->plane_res.dpp != NULL)
+		res = pipe_ctx->plane_res.dpp->funcs->dpp_get_optimal_number_of_taps(
+				pipe_ctx->plane_res.dpp, &pipe_ctx->plane_res.scl_data, &plane_state->scaling_quality);
 
 	if (!res) {
 		/* Try 24 bpp linebuffer */
@@ -862,6 +868,9 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx)
 
 		res = pipe_ctx->plane_res.xfm->funcs->transform_get_optimal_number_of_taps(
 			pipe_ctx->plane_res.xfm, &pipe_ctx->plane_res.scl_data, &plane_state->scaling_quality);
+
+		res = pipe_ctx->plane_res.dpp->funcs->dpp_get_optimal_number_of_taps(
+			pipe_ctx->plane_res.dpp, &pipe_ctx->plane_res.scl_data, &plane_state->scaling_quality);
 	}
 
 	if (res)
@@ -1026,7 +1035,7 @@ static int acquire_first_split_pipe(
 			pipe_ctx->stream_res.tg = pool->timing_generators[i];
 			pipe_ctx->plane_res.hubp = pool->hubps[i];
 			pipe_ctx->plane_res.ipp = pool->ipps[i];
-			pipe_ctx->plane_res.xfm = pool->transforms[i];
+			pipe_ctx->plane_res.dpp = pool->dpps[i];
 			pipe_ctx->stream_res.opp = pool->opps[i];
 			pipe_ctx->pipe_idx = i;
 
@@ -1359,6 +1368,7 @@ static int acquire_first_free_pipe(
 			pipe_ctx->plane_res.hubp = pool->hubps[i];
 			pipe_ctx->plane_res.ipp = pool->ipps[i];
 			pipe_ctx->plane_res.xfm = pool->transforms[i];
+			pipe_ctx->plane_res.dpp = pool->dpps[i];
 			pipe_ctx->stream_res.opp = pool->opps[i];
 			pipe_ctx->pipe_idx = i;
 
drivers/gpu/drm/amd/display/dc/core/dc_stream.c +13 -3
index 10421ea31373..97f040a25820 100644
@@ -200,7 +200,7 @@ bool dc_stream_set_cursor_attributes(
 	for (i = 0; i < MAX_PIPES; i++) {
 		struct pipe_ctx *pipe_ctx = &res_ctx->pipe_ctx[i];
 
-		if (pipe_ctx->stream != stream || !pipe_ctx->plane_res.xfm)
+		if (pipe_ctx->stream != stream || (!pipe_ctx->plane_res.xfm && !pipe_ctx->plane_res.dpp))
 			continue;
 		if (pipe_ctx->top_pipe && pipe_ctx->plane_state != pipe_ctx->top_pipe->plane_state)
 			continue;
@@ -221,9 +221,15 @@ bool dc_stream_set_cursor_attributes(
 					pipe_ctx->plane_res.mi, attributes);
 
 
-		if (pipe_ctx->plane_res.xfm->funcs->set_cursor_attributes != NULL)
+		if (pipe_ctx->plane_res.xfm != NULL &&
+				pipe_ctx->plane_res.xfm->funcs->set_cursor_attributes != NULL)
 			pipe_ctx->plane_res.xfm->funcs->set_cursor_attributes(
 				pipe_ctx->plane_res.xfm, attributes);
+
+		if (pipe_ctx->plane_res.dpp != NULL &&
+				pipe_ctx->plane_res.dpp->funcs->set_cursor_attributes != NULL)
+			pipe_ctx->plane_res.dpp->funcs->set_cursor_attributes(
+				pipe_ctx->plane_res.dpp, attributes);
 	}
 
 	stream->cursor_attributes = *attributes;
@@ -258,6 +264,7 @@ bool dc_stream_set_cursor_position(
 		struct mem_input *mi = pipe_ctx->plane_res.mi;
 		struct hubp *hubp = pipe_ctx->plane_res.hubp;
 		struct transform *xfm = pipe_ctx->plane_res.xfm;
+		struct dpp *dpp = pipe_ctx->plane_res.dpp;
 		struct dc_cursor_position pos_cpy = *position;
 		struct dc_cursor_mi_param param = {
 			.pixel_clk_khz = stream->timing.pix_clk_khz,
@@ -270,7 +277,7 @@ bool dc_stream_set_cursor_position(
 		if (pipe_ctx->stream != stream ||
 				(!pipe_ctx->plane_res.mi  && !pipe_ctx->plane_res.hubp) ||
 				!pipe_ctx->plane_state ||
-				!pipe_ctx->plane_res.xfm)
+				(!pipe_ctx->plane_res.xfm && !pipe_ctx->plane_res.dpp))
 			continue;
 
 		if (pipe_ctx->plane_state->address.type
@@ -293,6 +300,9 @@ bool dc_stream_set_cursor_position(
 		if (xfm != NULL && xfm->funcs->set_cursor_position != NULL)
 			xfm->funcs->set_cursor_position(xfm, &pos_cpy, &param, hubp->curs_attr.width);
 
+		if (dpp != NULL && dpp->funcs->set_cursor_position != NULL)
+			dpp->funcs->set_cursor_position(dpp, &pos_cpy, &param, hubp->curs_attr.width);
+
 	}
 
 	return true;
drivers/gpu/drm/amd/display/dc/core/dc_surface.c +1 -0
index d43783a45ab6..5aa2270f36fd 100644
@@ -30,6 +30,7 @@
 /* DC core (private) */
 #include "core_types.h"
 #include "transform.h"
+#include "dpp.h"
 
 /*******************************************************************************
  * Private functions
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.c +72 -72

283 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp.h +47 -47

163 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_cm.c +131 -130

562 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_dpp_dscl.c +52 -52

297 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +30 -29

177 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +11 -11
index 4764590f9097..4dbaf9a2f275 100644
@@ -446,13 +446,13 @@ static const struct dc_debug debug_defaults_diags = {
 		.disable_pplib_wm_range = true
 };
 
-static void dcn10_dpp_destroy(struct transform **xfm)
+static void dcn10_dpp_destroy(struct dpp **dpp)
 {
-	kfree(TO_DCN10_DPP(*xfm));
-	*xfm = NULL;
+	kfree(TO_DCN10_DPP(*dpp));
+	*dpp = NULL;
 }
 
-static struct transform *dcn10_dpp_create(
+static struct dpp *dcn10_dpp_create(
 	struct dc_context *ctx,
 	uint32_t inst)
 {
@@ -462,8 +462,8 @@ static struct transform *dcn10_dpp_create(
 	if (!dpp)
 		return NULL;
 
-	dcn10_dpp_construct(dpp, ctx, inst,
-			    &tf_regs[inst], &tf_shift, &tf_mask);
+	dpp1_construct(dpp, ctx, inst,
+		       &tf_regs[inst], &tf_shift, &tf_mask);
 	return &dpp->base;
 }
 
@@ -702,8 +702,8 @@ static void destruct(struct dcn10_resource_pool *pool)
 		if (pool->base.opps[i] != NULL)
 			pool->base.opps[i]->funcs->opp_destroy(&pool->base.opps[i]);
 
-		if (pool->base.transforms[i] != NULL)
-			dcn10_dpp_destroy(&pool->base.transforms[i]);
+		if (pool->base.dpps[i] != NULL)
+			dcn10_dpp_destroy(&pool->base.dpps[i]);
 
 		if (pool->base.ipps[i] != NULL)
 			pool->base.ipps[i]->funcs->ipp_destroy(&pool->base.ipps[i]);
@@ -924,7 +924,7 @@ static struct pipe_ctx *dcn10_acquire_idle_pipe_for_layer(
 
 	idle_pipe->plane_res.hubp = pool->hubps[idle_pipe->pipe_idx];
 	idle_pipe->plane_res.ipp = pool->ipps[idle_pipe->pipe_idx];
-	idle_pipe->plane_res.xfm = pool->transforms[idle_pipe->pipe_idx];
+	idle_pipe->plane_res.dpp = pool->dpps[idle_pipe->pipe_idx];
 
 	return idle_pipe;
 }
@@ -1377,8 +1377,8 @@ static bool construct(
 			goto ipp_create_fail;
 		}
 
-		pool->base.transforms[j] = dcn10_dpp_create(ctx, i);
-		if (pool->base.transforms[j] == NULL) {
+		pool->base.dpps[j] = dcn10_dpp_create(ctx, i);
+		if (pool->base.dpps[j] == NULL) {
 			BREAK_TO_DEBUGGER();
 			dm_error(
 				"DC: failed to create dpp!\n");
drivers/gpu/drm/amd/display/dc/inc/core_types.h +3 -0
index 45518c05e846..b69f321e2ab6 100644
@@ -84,6 +84,7 @@ void core_link_set_avmute(struct pipe_ctx *pipe_ctx, bool enable);
 /********** DAL Core*********************/
 #include "display_clock.h"
 #include "transform.h"
+#include "dpp.h"
 
 struct resource_pool;
 struct dc_state;
@@ -133,6 +134,7 @@ struct resource_pool {
 	struct hubp *hubps[MAX_PIPES];
 	struct input_pixel_processor *ipps[MAX_PIPES];
 	struct transform *transforms[MAX_PIPES];
+	struct dpp *dpps[MAX_PIPES];
 	struct output_pixel_processor *opps[MAX_PIPES];
 	struct timing_generator *timing_generators[MAX_PIPES];
 	struct stream_encoder *stream_enc[MAX_PIPES * 2];
@@ -184,6 +186,7 @@ struct plane_resource {
 	struct mem_input *mi;
 	struct input_pixel_processor *ipp;
 	struct transform *xfm;
+	struct dpp *dpp;
 };
 
 struct pipe_ctx {
drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h new file +134 -0

134 lines not on this page · raw message

Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#103 lore raw

PATCH 98/103 drm/amd/display: Do not set bandwidth on page flips.

In reply to a message not in the archive

From: "Leo (Sunpeng) Li" <[email protected]>

Which removes a lockdep warning for a possible deadlock situation:

While holding the drm event mutex (hard irq safe), dc_post_update_surfaces_to_stream eventually acquires the atom context lock, which is hard irq unsafe. We should only be calling it on full updates anyways.

Also remove a redundant call to dc_post_update_surfaces_to_stream, dc_commit_updates_for_stream already calls it.

---
 drivers/gpu/drm/amd/display/dc/core/dc.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +2 -3
index 80ddccc2570a..9d750fa29a87 100644
@@ -980,8 +980,6 @@ bool dc_commit_planes_to_stream(
 			new_plane_count,
 			dc_stream, stream_update, plane_states, state);
 
-	dc_post_update_surfaces_to_stream(dc);
-
 	kfree(stream_update);
 	return true;
 }
@@ -1418,7 +1416,8 @@ void dc_commit_updates_for_stream(struct dc *dc,
 				update_type,
 				context);
 
-	dc_post_update_surfaces_to_stream(dc);
+	if (update_type >= UPDATE_TYPE_FULL)
+		dc_post_update_surfaces_to_stream(dc);
 
 	if (dc->current_state != context) {
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#104 lore raw

PATCH 99/103 drm/amd/display: Remove a false-positive error message

In reply to a message not in the archive

From: Jerry Zuo <[email protected]>

To avoid confusion, need to suppress the error message when get -ERESTARTSYS error code. It is normal when getting interrupted by signals in the process of a wait for the buffer to become unreserved. Only propagate to user-mode for further action, no need to pop up error message.

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +3 -1
index ff9a35ecf040..351e398cb445 100644
@@ -1779,7 +1779,9 @@ static int get_fb_info(
 	int r = amdgpu_bo_reserve(rbo, false);
 
 	if (unlikely(r)) {
-		DRM_ERROR("Unable to reserve buffer\n");
+		// Don't show error msg. when return -ERESTARTSYS
+		if (r != -ERESTARTSYS)
+			DRM_ERROR("Unable to reserve buffer: %d\n", r);
 		return r;
 	}
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#105 lore raw

PATCH 100/103 drm/amd/display: add performance trace macro to dc

In reply to a message not in the archive

From: Dmytro Laktyushkin <[email protected]>

---
 .../drm/amd/display/amdgpu_dm/amdgpu_dm_services.c |  6 ++
 drivers/gpu/drm/amd/display/dc/basics/logger.c     | 78 ++++++----------------
 drivers/gpu/drm/amd/display/dc/basics/logger.h     | 37 ----------
 drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c   |  5 +-
 drivers/gpu/drm/amd/display/dc/dc.h                |  1 +
 .../gpu/drm/amd/display/dc/dce/dce_link_encoder.c  |  5 --
 .../drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c  | 40 +++++------
 .../gpu/drm/amd/display/dc/dcn10/dcn10_resource.c  |  1 +
 drivers/gpu/drm/amd/display/dc/dm_services.h       |  2 +
 .../gpu/drm/amd/display/include/logger_interface.h | 28 ++++++++
 drivers/gpu/drm/amd/display/include/logger_types.h | 36 +++++++++-
 11 files changed, 118 insertions(+), 121 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_services.c +6 -0
index 56bc7ccda414..10f9cc95a25f 100644
@@ -76,6 +76,12 @@ void dal_isr_release_lock(struct dc_context *ctx)
  * End-of-IRQ Interfaces.
  *****************************************************************************/
 
+unsigned long long dm_get_timestamp(struct dc_context *ctx)
+{
+	/* TODO: return actual timestamp */
+	return 0;
+}
+
 bool dm_write_persistent_data(struct dc_context *ctx,
 		const struct dc_sink *sink,
 		const char *module_name,
drivers/gpu/drm/amd/display/dc/basics/logger.c +22 -56

125 lines not on this page · raw message

drivers/gpu/drm/amd/display/dc/basics/logger.h +0 -37
index 2f7a5df4c811..09722f0f8aa3 100644
@@ -26,42 +26,5 @@
 #ifndef __DAL_LOGGER_H__
 #define __DAL_LOGGER_H__
 
-/* Structure for keeping track of offsets, buffer, etc */
-
-#define DAL_LOGGER_BUFFER_MAX_SIZE 2048
-
-/*Connectivity log needs to output EDID, which needs at lease 256x3 bytes,
- * change log line size to 896 to meet the request.
- */
-#define LOG_MAX_LINE_SIZE 896
-
-#include "include/logger_types.h"
-
-struct dal_logger {
-
-	/* How far into the circular buffer has been read by dsat
-	 * Read offset should never cross write offset. Write \0's to
-	 * read data just to be sure?
-	 */
-	uint32_t buffer_read_offset;
-
-	/* How far into the circular buffer we have written
-	 * Write offset should never cross read offset
-	 */
-	uint32_t buffer_write_offset;
-
-	uint32_t write_wrap_count;
-	uint32_t read_wrap_count;
-
-	uint32_t open_count;
-
-	char *log_buffer;	/* Pointer to malloc'ed buffer */
-	uint32_t log_buffer_size; /* Size of circular buffer */
-
-	uint32_t mask; /*array of masks for major elements*/
-
-	union logger_flags flags;
-	struct dc_context *ctx;
-};
 
 #endif /* __DAL_LOGGER_H__ */
drivers/gpu/drm/amd/display/dc/calcs/dcn_calcs.c +4 -1
index cf0c1459f546..e1515230c661 100644
@@ -682,6 +682,7 @@ bool dcn_validate_bandwidth(
 	bool bw_limit_pass;
 	float bw_limit;
 
+	PERFORMANCE_TRACE_START();
 	if (dcn_bw_apply_registry_override(dc))
 		dcn_bw_sync_calcs_and_dml(dc);
 
@@ -1089,6 +1090,8 @@ bool dcn_validate_bandwidth(
 
 	kernel_fpu_end();
 
+	PERFORMANCE_TRACE_END();
+
 	if (bw_limit_pass && v->voltage_level != 5)
 		return true;
 	else
@@ -1223,7 +1226,7 @@ unsigned int dcn_find_dcfclk_suits_all(
 	else
 		dcf_clk =  dc->dcn_soc->dcfclkv_min0p65*1000;
 
-	dm_logger_write(dc->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(dc->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"\tdcf_clk for voltage = %d\n", dcf_clk);
 	return dcf_clk;
 }
drivers/gpu/drm/amd/display/dc/dc.h +1 -0
index c7b31adc736e..ace17d13683a 100644
@@ -224,6 +224,7 @@ struct dc_debug {
 	bool disable_dpp_pg;
 	bool disable_stereo_support;
 	bool vsr_support;
+	bool performance_trace;
 };
 struct dc_state;
 struct resource_pool;
drivers/gpu/drm/amd/display/dc/dce/dce_link_encoder.c +0 -5
index 0cf0fff74d44..fe88852b4774 100644
@@ -757,11 +757,6 @@ void dce110_link_encoder_construct(
 		enc110->base.preferred_engine = ENGINE_ID_UNKNOWN;
 	}
 
-	dm_logger_write(init_data->ctx->logger, LOG_I2C_AUX,
-			"Using channel: %s [%d]\n",
-			DECODE_CHANNEL_ID(init_data->channel),
-			init_data->channel);
-
 	/* Override features with DCE-specific values */
 	if (BP_RESULT_OK == bp_funcs->get_encoder_cap_info(
 			enc110->base.ctx->dc_bios, enc110->base.id,
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +20 -20
index b69884549fc5..6e2cec643f31 100644
@@ -468,7 +468,7 @@ static void program_watermarks(
 			refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_DATA_URGENCY_WATERMARK_A, prog_wm_value);
 
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"URGENCY_WATERMARK_A calculated =%d\n"
 		"HW register value = 0x%x\n",
 		watermarks->a.urgent_ns, prog_wm_value);
@@ -476,7 +476,7 @@ static void program_watermarks(
 	prog_wm_value = convert_and_clamp(watermarks->a.pte_meta_urgent_ns,
 			refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_A, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"PTE_META_URGENCY_WATERMARK_A calculated =%d\n"
 		"HW register value = 0x%x\n",
 		watermarks->a.pte_meta_urgent_ns, prog_wm_value);
@@ -487,7 +487,7 @@ static void program_watermarks(
 			refclk_mhz, 0x1fffff);
 
 	REG_WRITE(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_A, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"SR_ENTER_EXIT_WATERMARK_A calculated =%d\n"
 		"HW register value = 0x%x\n",
 		watermarks->a.cstate_pstate.cstate_enter_plus_exit_ns, prog_wm_value);
@@ -497,7 +497,7 @@ static void program_watermarks(
 			watermarks->a.cstate_pstate.cstate_exit_ns,
 			refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_A, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"SR_EXIT_WATERMARK_A calculated =%d\n"
 		"HW register value = 0x%x\n",
 		watermarks->a.cstate_pstate.cstate_exit_ns, prog_wm_value);
@@ -507,7 +507,7 @@ static void program_watermarks(
 			watermarks->a.cstate_pstate.pstate_change_ns,
 			refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_A, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"DRAM_CLK_CHANGE_WATERMARK_A calculated =%d\n"
 		"HW register value = 0x%x\n\n",
 		watermarks->a.cstate_pstate.pstate_change_ns, prog_wm_value);
@@ -517,7 +517,7 @@ static void program_watermarks(
 	prog_wm_value = convert_and_clamp(
 			watermarks->b.urgent_ns, refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_DATA_URGENCY_WATERMARK_B, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"URGENCY_WATERMARK_B calculated =%d\n"
 		"HW register value = 0x%x\n",
 		watermarks->b.urgent_ns, prog_wm_value);
@@ -527,7 +527,7 @@ static void program_watermarks(
 			watermarks->b.pte_meta_urgent_ns,
 			refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_B, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"PTE_META_URGENCY_WATERMARK_B calculated =%d\n"
 		"HW register value = 0x%x\n",
 		watermarks->b.pte_meta_urgent_ns, prog_wm_value);
@@ -537,7 +537,7 @@ static void program_watermarks(
 			watermarks->b.cstate_pstate.cstate_enter_plus_exit_ns,
 			refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_B, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"SR_ENTER_WATERMARK_B calculated =%d\n"
 		"HW register value = 0x%x\n",
 		watermarks->b.cstate_pstate.cstate_enter_plus_exit_ns, prog_wm_value);
@@ -547,7 +547,7 @@ static void program_watermarks(
 			watermarks->b.cstate_pstate.cstate_exit_ns,
 			refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_B, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"SR_EXIT_WATERMARK_B calculated =%d\n"
 		"HW register value = 0x%x\n",
 		watermarks->b.cstate_pstate.cstate_exit_ns, prog_wm_value);
@@ -556,7 +556,7 @@ static void program_watermarks(
 			watermarks->b.cstate_pstate.pstate_change_ns,
 			refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_B, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"DRAM_CLK_CHANGE_WATERMARK_B calculated =%d\n\n"
 		"HW register value = 0x%x\n",
 		watermarks->b.cstate_pstate.pstate_change_ns, prog_wm_value);
@@ -565,7 +565,7 @@ static void program_watermarks(
 	prog_wm_value = convert_and_clamp(
 			watermarks->c.urgent_ns, refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_DATA_URGENCY_WATERMARK_C, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"URGENCY_WATERMARK_C calculated =%d\n"
 		"HW register value = 0x%x\n",
 		watermarks->c.urgent_ns, prog_wm_value);
@@ -575,7 +575,7 @@ static void program_watermarks(
 			watermarks->c.pte_meta_urgent_ns,
 			refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_C, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"PTE_META_URGENCY_WATERMARK_C calculated =%d\n"
 		"HW register value = 0x%x\n",
 		watermarks->c.pte_meta_urgent_ns, prog_wm_value);
@@ -585,7 +585,7 @@ static void program_watermarks(
 			watermarks->c.cstate_pstate.cstate_enter_plus_exit_ns,
 			refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_C, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"SR_ENTER_WATERMARK_C calculated =%d\n"
 		"HW register value = 0x%x\n",
 		watermarks->c.cstate_pstate.cstate_enter_plus_exit_ns, prog_wm_value);
@@ -595,7 +595,7 @@ static void program_watermarks(
 			watermarks->c.cstate_pstate.cstate_exit_ns,
 			refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_C, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"SR_EXIT_WATERMARK_C calculated =%d\n"
 		"HW register value = 0x%x\n",
 		watermarks->c.cstate_pstate.cstate_exit_ns, prog_wm_value);
@@ -605,7 +605,7 @@ static void program_watermarks(
 			watermarks->c.cstate_pstate.pstate_change_ns,
 			refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_C, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"DRAM_CLK_CHANGE_WATERMARK_C calculated =%d\n\n"
 		"HW register value = 0x%x\n",
 		watermarks->c.cstate_pstate.pstate_change_ns, prog_wm_value);
@@ -614,7 +614,7 @@ static void program_watermarks(
 	prog_wm_value = convert_and_clamp(
 			watermarks->d.urgent_ns, refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_DATA_URGENCY_WATERMARK_D, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"URGENCY_WATERMARK_D calculated =%d\n"
 		"HW register value = 0x%x\n",
 		watermarks->d.urgent_ns, prog_wm_value);
@@ -623,7 +623,7 @@ static void program_watermarks(
 			watermarks->d.pte_meta_urgent_ns,
 			refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_PTE_META_URGENCY_WATERMARK_D, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"PTE_META_URGENCY_WATERMARK_D calculated =%d\n"
 		"HW register value = 0x%x\n",
 		watermarks->d.pte_meta_urgent_ns, prog_wm_value);
@@ -633,7 +633,7 @@ static void program_watermarks(
 			watermarks->d.cstate_pstate.cstate_enter_plus_exit_ns,
 			refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_D, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"SR_ENTER_WATERMARK_D calculated =%d\n"
 		"HW register value = 0x%x\n",
 		watermarks->d.cstate_pstate.cstate_enter_plus_exit_ns, prog_wm_value);
@@ -643,7 +643,7 @@ static void program_watermarks(
 			watermarks->d.cstate_pstate.cstate_exit_ns,
 			refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_D, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"SR_EXIT_WATERMARK_D calculated =%d\n"
 		"HW register value = 0x%x\n",
 		watermarks->d.cstate_pstate.cstate_exit_ns, prog_wm_value);
@@ -653,7 +653,7 @@ static void program_watermarks(
 			watermarks->d.cstate_pstate.pstate_change_ns,
 			refclk_mhz, 0x1fffff);
 	REG_WRITE(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_D, prog_wm_value);
-	dm_logger_write(hws->ctx->logger, LOG_HW_MARKS,
+	dm_logger_write(hws->ctx->logger, LOG_BANDWIDTH_CALCS,
 		"DRAM_CLK_CHANGE_WATERMARK_D calculated =%d\n"
 		"HW register value = 0x%x\n\n",
 		watermarks->d.cstate_pstate.pstate_change_ns, prog_wm_value);
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_resource.c +1 -0
index 4dbaf9a2f275..d47ff7e6a525 100644
@@ -434,6 +434,7 @@ static const struct dc_debug debug_defaults_drv = {
 		.voltage_align_fclk = true,
 		.disable_stereo_support = true,
 		.vsr_support = true,
+		.performance_trace = false,
 };
 
 static const struct dc_debug debug_defaults_diags = {
drivers/gpu/drm/amd/display/dc/dm_services.h +2 -0
index 61482dc9c39e..2d2db0b0dc1c 100644
@@ -417,6 +417,8 @@ bool dm_dmcu_set_pipe(struct dc_context *ctx, unsigned int controller_id);
 #define dm_log_to_buffer(buffer, size, fmt, args)\
 	vsnprintf(buffer, size, fmt, args)
 
+unsigned long long dm_get_timestamp(struct dc_context *ctx);
+
 /*
  * Debug and verification hooks
  */
drivers/gpu/drm/amd/display/include/logger_interface.h +28 -0
index 5aaf2dacfe38..8e1fe70097be 100644
@@ -44,6 +44,8 @@ struct dal_logger *dal_logger_create(struct dc_context *ctx, uint32_t log_mask);
 
 uint32_t dal_logger_destroy(struct dal_logger **logger);
 
+void dm_logger_flush_buffer(struct dal_logger *logger, bool should_warn);
+
 void dm_logger_write(
 		struct dal_logger *logger,
 		enum dc_log_type log_type,
@@ -157,4 +159,30 @@ void context_clock_trace(
 #define DTN_INFO_END() \
 	dm_dtn_log_end(dc_ctx)
 
+#define PERFORMANCE_TRACE_START() \
+	unsigned long long perf_trc_start_stmp = dm_get_timestamp(dc->ctx); \
+	unsigned long long perf_trc_start_log_msk = dc->ctx->logger->mask; \
+	unsigned int perf_trc_start_log_flags = dc->ctx->logger->flags.value; \
+	if (dc->debug.performance_trace) {\
+		dm_logger_flush_buffer(dc->ctx->logger, false);\
+		dc->ctx->logger->mask = 1<<LOG_PERF_TRACE;\
+		dc->ctx->logger->flags.bits.ENABLE_CONSOLE = 0;\
+		dc->ctx->logger->flags.bits.ENABLE_BUFFER = 1;\
+	}
+
+#define PERFORMANCE_TRACE_END() do {\
+	unsigned long long perf_trc_end_stmp = dm_get_timestamp(dc->ctx);\
+	if (dc->debug.performance_trace) {\
+		dm_logger_write(dc->ctx->logger, \
+				LOG_PERF_TRACE, \
+				"%s duration: %d ticks\n", __func__,\
+				perf_trc_end_stmp - perf_trc_start_stmp); \
+		if (perf_trc_start_log_msk != 1<<LOG_PERF_TRACE) {\
+			dc->ctx->logger->mask = perf_trc_start_log_msk;\
+			dc->ctx->logger->flags.value = perf_trc_start_log_flags;\
+			dm_logger_flush_buffer(dc->ctx->logger, false);\
+		} \
+	} \
+} while (0)
+
 #endif /* __DAL_LOGGER_INTERFACE_H__ */
drivers/gpu/drm/amd/display/include/logger_types.h +34 -2
index 1f22e84cedb9..e2ff8cd423d6 100644
@@ -64,8 +64,7 @@ enum dc_log_type {
 	LOG_EVENT_LINK_LOSS,
 	LOG_EVENT_UNDERFLOW,
 	LOG_IF_TRACE,
-	LOG_HW_MARKS,
-	LOG_PPLIB,
+	LOG_PERF_TRACE,
 
 	LOG_SECTION_TOTAL_COUNT
 };
@@ -131,4 +130,37 @@ struct dc_log_type_info {
 	char name[MAX_NAME_LEN];
 };
 
+/* Structure for keeping track of offsets, buffer, etc */
+
+#define DAL_LOGGER_BUFFER_MAX_SIZE 2048
+
+/*Connectivity log needs to output EDID, which needs at lease 256x3 bytes,
+ * change log line size to 896 to meet the request.
+ */
+#define LOG_MAX_LINE_SIZE 896
+
+struct dal_logger {
+
+	/* How far into the circular buffer has been read by dsat
+	 * Read offset should never cross write offset. Write \0's to
+	 * read data just to be sure?
+	 */
+	uint32_t buffer_read_offset;
+
+	/* How far into the circular buffer we have written
+	 * Write offset should never cross read offset
+	 */
+	uint32_t buffer_write_offset;
+
+	uint32_t open_count;
+
+	char *log_buffer;	/* Pointer to malloc'ed buffer */
+	uint32_t log_buffer_size; /* Size of circular buffer */
+
+	uint32_t mask; /*array of masks for major elements*/
+
+	union logger_flags flags;
+	struct dc_context *ctx;
+};
+
 #endif /* __DAL_LOGGER_TYPES_H__ */
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#106 lore raw

PATCH 101/103 drm/amd/display: Set addressable region as active + border

In reply to a message not in the archive

From: Dmytro Laktyushkin <[email protected]>

This ensures that we do not draw the blank region onscreen, and that we do underscan instead.

---
 drivers/gpu/drm/amd/display/dc/core/dc_resource.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)
drivers/gpu/drm/amd/display/dc/core/dc_resource.c +14 -0
index 303d7ec684d9..454a521a268d 100644
@@ -850,6 +850,20 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx)
 	 */
 	pipe_ctx->plane_res.scl_data.lb_params.depth = LB_PIXEL_DEPTH_30BPP;
 
+	/**
+	 * KMD sends us h and v_addressable without the borders, which causes us sometimes to draw
+	 * the blank region on-screen. Correct for this by adding the borders back to their
+	 * respective addressable values, and by shifting recout.
+	 */
+	timing->h_addressable += timing->h_border_left + timing->h_border_right;
+	timing->v_addressable += timing->v_border_top + timing->v_border_bottom;
+	pipe_ctx->plane_res.scl_data.recout.y += timing->v_border_top;
+	pipe_ctx->plane_res.scl_data.recout.x += timing->h_border_left;
+	timing->v_border_top = 0;
+	timing->v_border_bottom = 0;
+	timing->h_border_left = 0;
+	timing->h_border_right = 0;
+
 	pipe_ctx->plane_res.scl_data.h_active = timing->h_addressable;
 	pipe_ctx->plane_res.scl_data.v_active = timing->v_addressable;
 
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#107 lore raw

PATCH 102/103 drm/amd/display: Report stream as different on scaling change

In reply to a message not in the archive

From: "Leo (Sunpeng) Li" <[email protected]>

When scaling is enabled, our preference is to scale up to the prefered (native) mode. This means that hardware timings will be the same across a modeset.

Therefore, also report mode as changed if source or destination rectangle is different.

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 10 +++++-----
 drivers/gpu/drm/amd/display/dc/core/dc_resource.c | 22 ++++++++++++++++++++++
 drivers/gpu/drm/amd/display/dc/dc.h               |  2 ++
 3 files changed, 29 insertions(+), 5 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +5 -5
index 351e398cb445..6d97486e074b 100644
@@ -4593,13 +4593,13 @@ static int dm_update_crtcs_state(
 			}
 		}
 
-		if (dc_is_stream_unchanged(new_stream,
-				old_acrtc_state->stream)) {
+		if (dc_is_stream_unchanged(new_stream, old_acrtc_state->stream) &&
+				dc_is_stream_scaling_unchanged(new_stream, old_acrtc_state->stream)) {
 
-				crtc_state->mode_changed = false;
+			crtc_state->mode_changed = false;
 
-				DRM_DEBUG_DRIVER("Mode change not required, setting mode_changed to %d",
-					      crtc_state->mode_changed);
+			DRM_DEBUG_DRIVER("Mode change not required, setting mode_changed to %d",
+				         crtc_state->mode_changed);
 		}
 
 
drivers/gpu/drm/amd/display/dc/core/dc_resource.c +22 -0
index 454a521a268d..2a5c91772473 100644
@@ -1335,6 +1335,28 @@ bool dc_is_stream_unchanged(
 	return true;
 }
 
+bool dc_is_stream_scaling_unchanged(
+	struct dc_stream_state *old_stream, struct dc_stream_state *stream)
+{
+	if (old_stream == stream)
+		return true;
+
+	if (old_stream == NULL || stream == NULL)
+		return false;
+
+	if (memcmp(&old_stream->src,
+			&stream->src,
+			sizeof(struct rect)) != 0)
+		return false;
+
+	if (memcmp(&old_stream->dst,
+			&stream->dst,
+			sizeof(struct rect)) != 0)
+		return false;
+
+	return true;
+}
+
 /* Maximum TMDS single link pixel clock 165MHz */
 #define TMDS_MAX_PIXEL_CLOCK_IN_KHZ 165000
 
drivers/gpu/drm/amd/display/dc/dc.h +2 -0
index ace17d13683a..f60191d9c7e5 100644
@@ -627,6 +627,8 @@ struct dc_stream_update {
 
 bool dc_is_stream_unchanged(
 	struct dc_stream_state *old_stream, struct dc_stream_state *stream);
+bool dc_is_stream_scaling_unchanged(
+	struct dc_stream_state *old_stream, struct dc_stream_state *stream);
 
 /*
  * Set up surface attributes and associate to a stream
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx
#108 lore raw

PATCH 103/103 drm/amd/display: ASSERT on dc_create_state failing.

In reply to a message not in the archive

From: "Leo (Sunpeng) Li" <[email protected]>

dc_create_state() returns NULL on allocation failure. Raise warning when that happens.

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 1 +
 1 file changed, 1 insertion(+)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +1 -0
index 6d97486e074b..454422390331 100644
@@ -671,6 +671,7 @@ int amdgpu_dm_display_resume(struct amdgpu_device *adev)
 	dc_release_state(cached_state->context);
 
 	cached_state->context = dc_create_state();
+	ASSERT(cached_state->context);
 	dc_resource_state_copy_construct_current(adev->dm.dc, cached_state->context);
 
 	for_each_new_crtc_in_state(adev->dm.cached_state, crtc, crtc_state, i) {
Signature · 6 lines
2.14.1

_______________________________________________
amd-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/amd-gfx