Skip to contentSkip to conversation

Fix Adaptive Sync SDP for PR with Link ON + Auxless ALPM

PATCH In review Ankit Nautiyal 2 participants 38 messages Apr 2 – Apr 7

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

cover Fix Adaptive Sync SDP for PR with Link ON + Auxless ALPM

Enable Adaptive Sync SDP for Panel replay + auxless ALPM. First few patches are fixes in existing AS SDP enablement. Later patches add the support to send AS SDP for Panel replay with Link ON and with auxless ALPM (Link-Off).

This series is in continuation from discussions in [1] [2] and more recent [3]:

Apart from few patches from [3] new changes include:
- Some clean up in drm/dp Macros
- Add new DPCD bit to get FAVT PAYLOAD FIELDS PARSING SUPPORT.
- Add AS SDP version and PR and VRR specific bits in drm_dp.h.
- Modify AS SDP compute config to accomodate PR with Link On and Link Off.
- Program Downspread Ctrl DPCD bits.
- Make way for T1 and T2 AS SDP transmission timing/position.

[1] https://lore.kernel.org/all/[email protected]/
[2] https://lore.kernel.org/all/[email protected]/
[3] https://lore.kernel.org/intel-gfx/[email protected]/T/#m6e8beab2cc3b6ff9d61f740f107d83a2f4e08114

Rev2:
- Drop the redundant version member for AS SDP, and use the member revision
- Drop member to store AS SDP transmission time, use simple helper instead.
- Use Burst mode to write Panel Replay config DPCDs.
- Split AS SDP configuration in compute_config() and compute_config_late() phase.
- Always enable AS SDP whenever supported by source + sink.

Rev3:
- Use ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR bit also to determine AS SDP v2 support.
- Drop patch to make provision to send AS SDP v1.
- Add patch to include all relevant AS SDP fields in comparison.
- Other refactor improvements suggested by Ville.

Rev4:
- Add new member coasting vtotal in struct drm_dp_as_sdp and program the coasting vtotal in AS SDP.
- Add new helper to get async timing support during PR active.
- Correct the patches to write Panel Replay Config DPCDs.
- Ensure either PR/VRR have potential to use AS SDP before enabling AS SDP, even though we may not be enabling them at modeset.
- Other correction and changes suggested by Ville.

Ankit Nautiyal (23):
drm/dp: Rename and relocate AS SDP payload field masks drm/dp: Clean up DPRX feature enumeration macros
drm/dp: Add bits for AS SDP FAVT Payload Fields Parsing support drm/dp: Add DPCD for configuring AS SDP for PR + VRR drm/dp: Store coasting vtotal in struct drm_dp_as_sdp drm/i915/dp: Fix readback for target_rr in Adaptive Sync SDP drm/i915/vrr: Avoid vrr for PCON with HDMI2.1 sink drm/i915/dp: Account for AS_SDP guardband only when enabled drm/i915/dp: Add a helper to decide if AS SDP can be used drm/i915/dp: Skip AS SDP for DP branch devices drm/i915/dp: Use revision field of AS SDP data structure drm/i915/dp: Include all relevant AS SDP fields in comparison drm/i915/psr: Add helper to get Async Video timing support in PR
active
drm/i915/dp: Add member to intel_dp to store AS SDP v2 support drm/i915/dp: Allow AS SDP only if v2 is supported drm/i915/psr: Write the PR config DPCDs in burst mode drm/i915/display: Add helper for AS SDP transmission time selection drm/i915/psr: Program Panel Replay CONFIG3 using AS SDP transmission
time
drm/i915/dp: Set relevant Downspread Ctrl DPCD bits for PR + Auxless
ALPM
drm/i915/dp: Program AS SDP DB[1:0] for PR with Link off drm/i915/dp: Split AS SDP computation between compute_config and
compute_config_late
drm/i915/dp: Compute and include coasting vtotal for AS SDP drm/i915/dp: Always enable AS SDP if supported by source + sink

 drivers/gpu/drm/display/drm_dp_helper.c       |   1 +
 drivers/gpu/drm/i915/display/intel_alpm.c     |  20 ++-
 drivers/gpu/drm/i915/display/intel_display.c  |   9 +-
 .../drm/i915/display/intel_display_types.h    |   1 +
 drivers/gpu/drm/i915/display/intel_dp.c       | 165 ++++++++++++++----
 drivers/gpu/drm/i915/display/intel_dp.h       |   2 +
 .../drm/i915/display/intel_dp_link_training.c |  12 +-
 .../drm/i915/display/intel_dp_link_training.h |   3 +-
 drivers/gpu/drm/i915/display/intel_dp_mst.c   |   2 +-
 drivers/gpu/drm/i915/display/intel_psr.c      |  37 ++--
 drivers/gpu/drm/i915/display/intel_psr.h      |   1 +
 drivers/gpu/drm/i915/display/intel_vrr.c      |  10 ++
 include/drm/display/drm_dp.h                  |  20 ++-
 include/drm/display/drm_dp_helper.h           |   1 +
 14 files changed, 226 insertions(+), 58 deletions(-)
Signature · 1 line
2.45.2
#1 lore raw

PATCH 1/23 drm/dp: Rename and relocate AS SDP payload field masks

The AS SDP payload field masks were misnamed and placed under the DPRX feature enumeration list. These are not DPRX capability bits, but are payload field masks for the Adaptive Sync SDP.

Relocate both masks next to the AS SDP definitions. Update users to the corrected names. No functional change.

---
 drivers/gpu/drm/i915/display/intel_dp.c | 4 ++--
 include/drm/display/drm_dp.h            | 5 +++--
 2 files changed, 5 insertions(+), 4 deletions(-)
drivers/gpu/drm/i915/display/intel_dp.c +2 -2
index 4955bd8b11d7..602efaf500bd 100644
@@ -5295,8 +5295,8 @@ int intel_dp_as_sdp_unpack(struct drm_dp_as_sdp *as_sdp,
 	if ((sdp->sdp_header.HB3 & 0x3F) != 9)
 		return -EINVAL;
 
-	as_sdp->length = sdp->sdp_header.HB3 & DP_ADAPTIVE_SYNC_SDP_LENGTH;
-	as_sdp->mode = sdp->db[0] & DP_ADAPTIVE_SYNC_SDP_OPERATION_MODE;
+	as_sdp->length = sdp->sdp_header.HB3 & DP_AS_SDP_LENGTH_MASK;
+	as_sdp->mode = sdp->db[0] & DP_AS_SDP_OPERATION_MODE_MASK;
 	as_sdp->vtotal = (sdp->db[2] << 8) | sdp->db[1];
 	as_sdp->target_rr = (u64)sdp->db[3] | ((u64)sdp->db[4] & 0x3);
 	as_sdp->target_rr_divider = sdp->db[4] & 0x20 ? true : false;
include/drm/display/drm_dp.h +3 -2
index 8b15d3eeb716..4ea3b5b08a12 100644
@@ -1204,8 +1204,6 @@
 
 #define DP_DPRX_FEATURE_ENUMERATION_LIST_CONT_1         0x2214 /* 2.0 E11 */
 # define DP_ADAPTIVE_SYNC_SDP_SUPPORTED    (1 << 0)
-# define DP_ADAPTIVE_SYNC_SDP_OPERATION_MODE		GENMASK(1, 0)
-# define DP_ADAPTIVE_SYNC_SDP_LENGTH				GENMASK(5, 0)
 # define DP_AS_SDP_FIRST_HALF_LINE_OR_3840_PIXEL_CYCLE_WINDOW_NOT_SUPPORTED (1 << 1)
 # define DP_VSC_EXT_SDP_FRAMEWORK_VERSION_1_SUPPORTED  (1 << 4)
 
@@ -1870,4 +1868,7 @@ enum operation_mode {
 	DP_AS_SDP_FAVT_TRR_REACHED = 0x03
 };
 
+#define DP_AS_SDP_OPERATION_MODE_MASK	GENMASK(1, 0)
+#define DP_AS_SDP_LENGTH_MASK		GENMASK(5, 0)
+
 #endif /* _DRM_DP_H_ */
Signature · 1 line
2.45.2
#2 lore raw

PATCH 2/23 drm/dp: Clean up DPRX feature enumeration macros

In reply to Ankit Nautiyal

Align the DP_DPRX feature enumeration macros for better readability and consistency, and use the BIT() macro instead of open-coded shifts.

---
 include/drm/display/drm_dp.h | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
include/drm/display/drm_dp.h +4 -4
index 4ea3b5b08a12..49f0154eb93c 100644
@@ -1202,10 +1202,10 @@
 # define DP_DPRX_SLEEP_WAKE_TIMEOUT_PERIOD_80_MS	0x04
 # define DP_DPRX_SLEEP_WAKE_TIMEOUT_PERIOD_100_MS	0x05
 
-#define DP_DPRX_FEATURE_ENUMERATION_LIST_CONT_1         0x2214 /* 2.0 E11 */
-# define DP_ADAPTIVE_SYNC_SDP_SUPPORTED    (1 << 0)
-# define DP_AS_SDP_FIRST_HALF_LINE_OR_3840_PIXEL_CYCLE_WINDOW_NOT_SUPPORTED (1 << 1)
-# define DP_VSC_EXT_SDP_FRAMEWORK_VERSION_1_SUPPORTED  (1 << 4)
+#define DP_DPRX_FEATURE_ENUMERATION_LIST_CONT_1					0x2214 /* 2.0 E11 */
+# define DP_ADAPTIVE_SYNC_SDP_SUPPORTED						BIT(0)
+# define DP_AS_SDP_FIRST_HALF_LINE_OR_3840_PIXEL_CYCLE_WINDOW_NOT_SUPPORTED	BIT(1)
+# define DP_VSC_EXT_SDP_FRAMEWORK_VERSION_1_SUPPORTED				BIT(4)
 
 #define DP_128B132B_SUPPORTED_LINK_RATES       0x2215 /* 2.0 */
 # define DP_UHBR10                             (1 << 0)
Signature · 1 line
2.45.2
#3 lore raw

PATCH 3/23 drm/dp: Add bits for AS SDP FAVT Payload Fields Parsing support

In reply to Ankit Nautiyal

DP v2.1 introduced support for sending AS SDP payload bytes for FAVT. Add the relavant bits for the same.

---
 include/drm/display/drm_dp.h | 1 +
 1 file changed, 1 insertion(+)
include/drm/display/drm_dp.h +1 -0
index 49f0154eb93c..8d172863eba3 100644
@@ -1205,6 +1205,7 @@
 #define DP_DPRX_FEATURE_ENUMERATION_LIST_CONT_1					0x2214 /* 2.0 E11 */
 # define DP_ADAPTIVE_SYNC_SDP_SUPPORTED						BIT(0)
 # define DP_AS_SDP_FIRST_HALF_LINE_OR_3840_PIXEL_CYCLE_WINDOW_NOT_SUPPORTED	BIT(1)
+# define DP_AS_SDP_FAVT_PAYLOAD_FIELDS_PARSING_SUPPORTED			BIT(2) /* 2.1 */
 # define DP_VSC_EXT_SDP_FRAMEWORK_VERSION_1_SUPPORTED				BIT(4)
 
 #define DP_128B132B_SUPPORTED_LINK_RATES       0x2215 /* 2.0 */
Signature · 1 line
2.45.2
#4 lore raw

PATCH 4/23 drm/dp: Add DPCD for configuring AS SDP for PR + VRR

In reply to Ankit Nautiyal

Add additional DPCDs required to be configured to support VRR with Panel Replay. These DPCDs are specifically required for configuring Adaptive Sync SDP and are introduced in DP v2.1.

v2:
- Correct the shift for the bits. (Ville)
- Add DP_PR_ prefix for the PR-related fields.
v3:
- Use macro values in their shifted form to match the convention. (Ville)
v4:
- Add macro for the mask. (Ville)

---
 include/drm/display/drm_dp.h | 6 ++++++
 1 file changed, 6 insertions(+)
include/drm/display/drm_dp.h +6 -0
index 8d172863eba3..829e4d98d61c 100644
@@ -718,6 +718,12 @@
 #define DP_EXTENDED_DPRX_SLEEP_WAKE_TIMEOUT_GRANT	    0x119   /* 1.4a */
 # define DP_DPRX_SLEEP_WAKE_TIMEOUT_PERIOD_GRANTED	    (1 << 0)
 
+#define PANEL_REPLAY_CONFIG3			0x11a /* DP 2.1 */
+# define DP_PR_AS_SDP_SETUP_TIME_MASK		(3 << 6)
+# define DP_PR_AS_SDP_SETUP_TIME_T1		(0 << 6)
+# define DP_PR_AS_SDP_SETUP_TIME_DYNAMIC	(1 << 6) /* DP 2.1 Table 2-227 */
+# define DP_PR_AS_SDP_SETUP_TIME_T2		(2 << 6)
+
 #define DP_FEC_CONFIGURATION		    0x120    /* 1.4 */
 # define DP_FEC_READY			    (1 << 0)
 # define DP_FEC_ERR_COUNT_SEL_MASK	    (7 << 1)
Signature · 1 line
2.45.2
#5 lore raw

PATCH 5/23 drm/dp: Store coasting vtotal in struct drm_dp_as_sdp

In reply to Ankit Nautiyal

Add new field in struct drm_dp_as_sdp to store coasting vtotal. This is used by the sinks that support Panel Replay and Asynchronous timing during PR Active to derive refresh rate, when AS SDP transmission is stopped by the source.

---
 drivers/gpu/drm/display/drm_dp_helper.c | 1 +
 include/drm/display/drm_dp_helper.h     | 1 +
 2 files changed, 2 insertions(+)
drivers/gpu/drm/display/drm_dp_helper.c +1 -0
index a697cc227e28..e29958f8b0b6 100644
@@ -3512,6 +3512,7 @@ void drm_dp_as_sdp_log(struct drm_printer *p, const struct drm_dp_as_sdp *as_sdp
 	drm_printf(p, "    duration_incr_ms: %d\n", as_sdp->duration_incr_ms);
 	drm_printf(p, "    duration_decr_ms: %d\n", as_sdp->duration_decr_ms);
 	drm_printf(p, "    operation_mode: %d\n", as_sdp->mode);
+	drm_printf(p, "    coasting vtotal: %d\n", as_sdp->coasting_vtotal);
 }
 EXPORT_SYMBOL(drm_dp_as_sdp_log);
 
include/drm/display/drm_dp_helper.h +1 -0
index 1d0acd58f486..8c2d77a032f0 100644
@@ -126,6 +126,7 @@ struct drm_dp_as_sdp {
 	int duration_decr_ms;
 	bool target_rr_divider;
 	enum operation_mode mode;
+	int coasting_vtotal;
 };
 
 void drm_dp_as_sdp_log(struct drm_printer *p,
Signature · 1 line
2.45.2
#6 lore raw

PATCH 6/23 drm/i915/dp: Fix readback for target_rr in Adaptive Sync SDP

In reply to Ankit Nautiyal

Correct the bit-shift logic to properly readback the 10 bit target_rr from DB3 and DB4.

v2: Align the style with readback for vtotal. (Ville)

Fixes 12ea89291603 ("drm/i915/dp: Add Read/Write support for Adaptive Sync SDP")
Cc Mitul Golani <[email protected]>
Cc Ankit Nautiyal <[email protected]>
---
 drivers/gpu/drm/i915/display/intel_dp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/i915/display/intel_dp.c +1 -1
index 602efaf500bd..49eeb35973a1 100644
@@ -5298,7 +5298,7 @@ int intel_dp_as_sdp_unpack(struct drm_dp_as_sdp *as_sdp,
 	as_sdp->length = sdp->sdp_header.HB3 & DP_AS_SDP_LENGTH_MASK;
 	as_sdp->mode = sdp->db[0] & DP_AS_SDP_OPERATION_MODE_MASK;
 	as_sdp->vtotal = (sdp->db[2] << 8) | sdp->db[1];
-	as_sdp->target_rr = (u64)sdp->db[3] | ((u64)sdp->db[4] & 0x3);
+	as_sdp->target_rr = ((sdp->db[4] & 0x3) << 8) | sdp->db[3];
 	as_sdp->target_rr_divider = sdp->db[4] & 0x20 ? true : false;
 
 	return 0;
Signature · 1 line
2.45.2
#7 lore raw

PATCH 7/23 drm/i915/vrr: Avoid vrr for PCON with HDMI2.1 sink

In reply to Ankit Nautiyal

Currently we do not support VRR with HDMI so skip vrr compute config step for all DP branch devices.

v2: Restrict VRR on all DP branch devices instead for checking only for
HDMI. (Ville)

---
 drivers/gpu/drm/i915/display/intel_vrr.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
drivers/gpu/drm/i915/display/intel_vrr.c +10 -0
index 8a957804cb97..160e1c24aa4b 100644
@@ -55,6 +55,16 @@ bool intel_vrr_is_capable(struct intel_connector *connector)
 		if (connector->mst.dp)
 			return false;
 		intel_dp = intel_attached_dp(connector);
+		/*
+		 * Among non-MST DP branch devices, only an HDMI 2.1 sink connected
+		 * via a PCON could support VRR. However, supporting VRR through a
+		 * PCON requires non-trivial changes that are not implemented yet.
+		 * Until that support exists, avoid VRR on all DP branch devices.
+		 *
+		 * TODO: Add support for VRR for DP->HDMI 2.1 PCON.
+		 */
+		if (drm_dp_is_branch(intel_dp->dpcd))
+			return false;
 
 		if (!drm_dp_sink_can_do_video_without_timing_msa(intel_dp->dpcd))
 			return false;
Signature · 1 line
2.45.2
#8 lore raw

PATCH 8/23 drm/i915/dp: Account for AS_SDP guardband only when enabled

In reply to Ankit Nautiyal

Currently the intel_dp_sdp_min_guardband() accounts for AS_SDP for all platforms that support adaptive sync SDP even for configurations where it cannot be enabled. Instead account for adaptive sync SDP guardband only when it is enabled.

---
 drivers/gpu/drm/i915/display/intel_dp.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
drivers/gpu/drm/i915/display/intel_dp.c +2 -3
index 49eeb35973a1..499850c2abcf 100644
@@ -7374,7 +7374,6 @@ int intel_dp_get_lines_for_sdp(const struct intel_crtc_state *crtc_state, u32 ty
 int intel_dp_sdp_min_guardband(const struct intel_crtc_state *crtc_state,
 			       bool assume_all_enabled)
 {
-	struct intel_display *display = to_intel_display(crtc_state);
 	int sdp_guardband = 0;
 
 	if (assume_all_enabled ||
@@ -7389,8 +7388,8 @@ int intel_dp_sdp_min_guardband(const struct intel_crtc_state *crtc_state,
 		sdp_guardband = max(sdp_guardband,
 				    intel_dp_get_lines_for_sdp(crtc_state, DP_SDP_PPS));
 
-	if ((assume_all_enabled && HAS_AS_SDP(display)) ||
-	    crtc_state->infoframes.enable & intel_hdmi_infoframe_enable(DP_SDP_ADAPTIVE_SYNC))
+	if (crtc_state->infoframes.enable &
+	    intel_hdmi_infoframe_enable(DP_SDP_ADAPTIVE_SYNC))
 		sdp_guardband = max(sdp_guardband,
 				    intel_dp_get_lines_for_sdp(crtc_state, DP_SDP_ADAPTIVE_SYNC));
 
Signature · 1 line
2.45.2
#9 lore raw

PATCH 9/23 drm/i915/dp: Add a helper to decide if AS SDP can be used

In reply to Ankit Nautiyal

Add a helper that determines whether AS SDP can be used for the current DP configuration. For now this is true only when the sink supports AS SDP and VRR is enabled, but more conditions may be added later.

v2:
- Rename to intel_dp_needs_as_sdp(). (Ville)
- Add a #FIXME documenting non-atomic of DP SDP updates. (Ville)

---
 drivers/gpu/drm/i915/display/intel_dp.c | 16 +++++++++++++++-
 1 file changed, 15 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/display/intel_dp.c +15 -1
index 499850c2abcf..3d0ea074e610 100644
@@ -3124,6 +3124,15 @@ static void intel_dp_compute_vsc_colorimetry(const struct intel_crtc_state *crtc
 	vsc->content_type = DP_CONTENT_TYPE_NOT_DEFINED;
 }
 
+static bool intel_dp_needs_as_sdp(struct intel_dp *intel_dp,
+				  struct intel_crtc_state *crtc_state)
+{
+	if (!intel_dp->as_sdp_supported)
+		return false;
+
+	return crtc_state->vrr.enable;
+}
+
 static void intel_dp_compute_as_sdp(struct intel_dp *intel_dp,
 				    struct intel_crtc_state *crtc_state)
 {
@@ -3131,7 +3140,12 @@ static void intel_dp_compute_as_sdp(struct intel_dp *intel_dp,
 	const struct drm_display_mode *adjusted_mode =
 		&crtc_state->hw.adjusted_mode;
 
-	if (!crtc_state->vrr.enable || !intel_dp->as_sdp_supported)
+	/*
+	 * #FIXME: SDP/infoframe updates aren’t truly atomic, and with the new
+	 * cdclk->tc clock crossing we may transiently send a corrupted packet
+	 * if the update lands mid‑transmission.
+	 */
+	if (!intel_dp_needs_as_sdp(intel_dp, crtc_state))
 		return;
 
 	crtc_state->infoframes.enable |= intel_hdmi_infoframe_enable(DP_SDP_ADAPTIVE_SYNC);
Signature · 1 line
2.45.2
#10 lore raw

PATCH 10/23 drm/i915/dp: Skip AS SDP for DP branch devices

In reply to Ankit Nautiyal

Currently, VRR is not implmented for DP branch devices. So skip sending AS SDP for them.

---
 drivers/gpu/drm/i915/display/intel_dp.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/gpu/drm/i915/display/intel_dp.c +6 -0
index 3d0ea074e610..095bde8e97ef 100644
@@ -3130,6 +3130,12 @@ static bool intel_dp_needs_as_sdp(struct intel_dp *intel_dp,
 	if (!intel_dp->as_sdp_supported)
 		return false;
 
+	/*
+	 * #TODO Implement AS SDP for DP branch device.
+	 */
+	if (drm_dp_is_branch(intel_dp->dpcd))
+		return false;
+
 	return crtc_state->vrr.enable;
 }
 
Signature · 1 line
2.45.2
#11 lore raw

PATCH 11/23 drm/i915/dp: Use revision field of AS SDP data structure

In reply to Ankit Nautiyal

Use the revision field of struct drm_dp_as_sdp instead of current hardcoding for the AS SDP revisions.

---
 drivers/gpu/drm/i915/display/intel_display.c | 3 ++-
 drivers/gpu/drm/i915/display/intel_dp.c      | 7 +++----
 2 files changed, 5 insertions(+), 5 deletions(-)
drivers/gpu/drm/i915/display/intel_display.c +2 -1
index 10b6c6fcb03f..57870baebfa5 100644
@@ -4906,7 +4906,8 @@ static bool
 intel_compare_dp_as_sdp(const struct drm_dp_as_sdp *a,
 			const struct drm_dp_as_sdp *b)
 {
-	return a->vtotal == b->vtotal &&
+	return a->revision == b->revision &&
+		a->vtotal == b->vtotal &&
 		a->target_rr == b->target_rr &&
 		a->duration_incr_ms == b->duration_incr_ms &&
 		a->duration_decr_ms == b->duration_decr_ms &&
drivers/gpu/drm/i915/display/intel_dp.c +3 -4
index 095bde8e97ef..cec0f3d03c2f 100644
@@ -3159,6 +3159,7 @@ static void intel_dp_compute_as_sdp(struct intel_dp *intel_dp,
 	as_sdp->sdp_type = DP_SDP_ADAPTIVE_SYNC;
 	as_sdp->length = 0x9;
 	as_sdp->duration_incr_ms = 0;
+	as_sdp->revision = 0x2;
 	as_sdp->vtotal = intel_vrr_vmin_vtotal(crtc_state);
 
 	if (crtc_state->cmrr.enable) {
@@ -5122,7 +5123,7 @@ static ssize_t intel_dp_as_sdp_pack(const struct drm_dp_as_sdp *as_sdp,
 	/* Prepare AS (Adaptive Sync) SDP Header */
 	sdp->sdp_header.HB0 = 0;
 	sdp->sdp_header.HB1 = as_sdp->sdp_type;
-	sdp->sdp_header.HB2 = 0x02;
+	sdp->sdp_header.HB2 = as_sdp->revision;
 	sdp->sdp_header.HB3 = as_sdp->length;
 
 	/* Fill AS (Adaptive Sync) SDP Payload */
@@ -5309,13 +5310,11 @@ int intel_dp_as_sdp_unpack(struct drm_dp_as_sdp *as_sdp,
 	if (sdp->sdp_header.HB1 != DP_SDP_ADAPTIVE_SYNC)
 		return -EINVAL;
 
-	if (sdp->sdp_header.HB2 != 0x02)
-		return -EINVAL;
-
 	if ((sdp->sdp_header.HB3 & 0x3F) != 9)
 		return -EINVAL;
 
 	as_sdp->length = sdp->sdp_header.HB3 & DP_AS_SDP_LENGTH_MASK;
+	as_sdp->revision = sdp->sdp_header.HB2;
 	as_sdp->mode = sdp->db[0] & DP_AS_SDP_OPERATION_MODE_MASK;
 	as_sdp->vtotal = (sdp->db[2] << 8) | sdp->db[1];
 	as_sdp->target_rr = ((sdp->db[4] & 0x3) << 8) | sdp->db[3];
Signature · 1 line
2.45.2
#12 lore raw

PATCH 12/23 drm/i915/dp: Include all relevant AS SDP fields in comparison

In reply to Ankit Nautiyal

Add missing drm_dp_as_sdp header fields to intel_compare_dp_as_sdp() comparison.

---
 drivers/gpu/drm/i915/display/intel_display.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/display/intel_display.c +4 -1
index 57870baebfa5..a0e7ef2574b2 100644
@@ -4906,11 +4906,14 @@ static bool
 intel_compare_dp_as_sdp(const struct drm_dp_as_sdp *a,
 			const struct drm_dp_as_sdp *b)
 {
-	return a->revision == b->revision &&
+	return a->sdp_type == b->sdp_type &&
+		a->revision == b->revision &&
+		a->length == b->length &&
 		a->vtotal == b->vtotal &&
 		a->target_rr == b->target_rr &&
 		a->duration_incr_ms == b->duration_incr_ms &&
 		a->duration_decr_ms == b->duration_decr_ms &&
+		a->target_rr_divider == b->target_rr_divider &&
 		a->mode == b->mode;
 }
 
Signature · 1 line
2.45.2
#13 lore raw

PATCH 13/23 drm/i915/psr: Add helper to get Async Video timing support in PR active

In reply to Ankit Nautiyal

Introduce a helper to check if Panel Replay has Async Video Timing support during PR Active state.

---
 drivers/gpu/drm/i915/display/intel_psr.c | 8 ++++++++
 drivers/gpu/drm/i915/display/intel_psr.h | 1 +
 2 files changed, 9 insertions(+)
drivers/gpu/drm/i915/display/intel_psr.c +8 -0
index 2f1b48cd8efd..782e839b7396 100644
@@ -4623,3 +4623,11 @@ bool intel_psr_use_trans_push(const struct intel_crtc_state *crtc_state)
 
 	return HAS_PSR_TRANS_PUSH_FRAME_CHANGE(display) && crtc_state->has_psr;
 }
+
+bool intel_psr_pr_async_video_timing_supported(struct intel_dp *intel_dp)
+{
+	struct intel_connector *connector = intel_dp->attached_connector;
+
+	return !(connector->dp.panel_replay_caps.dpcd[INTEL_PR_DPCD_INDEX(DP_PANEL_REPLAY_CAP_CAPABILITY)] &
+		 DP_PANEL_REPLAY_ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR);
+}
drivers/gpu/drm/i915/display/intel_psr.h +1 -0
index 394b641840b3..29723e63888f 100644
@@ -86,5 +86,6 @@ void intel_psr_compute_config_late(struct intel_dp *intel_dp,
 				   struct intel_crtc_state *crtc_state);
 int intel_psr_min_guardband(struct intel_crtc_state *crtc_state);
 bool intel_psr_use_trans_push(const struct intel_crtc_state *crtc_state);
+bool intel_psr_pr_async_video_timing_supported(struct intel_dp *intel_dp);
 
 #endif /* __INTEL_PSR_H__ */
Signature · 1 line
2.45.2
#14 lore raw

Re: [PATCH 13/23] drm/i915/psr: Add helper to get Async Video timing support in PR active

Quoted 14 lines · Ankit Nautiyal

Introduce a helper to check if Panel Replay has Async Video Timing support during PR Active state.

---
 drivers/gpu/drm/i915/display/intel_psr.c | 8 ++++++++
 drivers/gpu/drm/i915/display/intel_psr.h | 1 +
 2 files changed, 9 insertions(+)
drivers/gpu/drm/i915/display/intel_psr.c +0 -0
index 2f1b48cd8efd..782e839b7396 100644

drivers/gpu/drm/i915/display/intel_psr.c bool intel_psr_use_trans_push(const struct intel_crtc_state *crtc_state) · line 4630

Quoted 8 lines
drivers/gpu/drm/i915/display/intel_psr.c +5 -0
@@ -4623,3 +4623,11 @@ bool intel_psr_use_trans_push(const struct intel_crtc_state *crtc_state)
 
 	return HAS_PSR_TRANS_PUSH_FRAME_CHANGE(display) && crtc_state->has_psr;
 }
+
+bool intel_psr_pr_async_video_timing_supported(struct intel_dp *intel_dp)
+{
+	struct intel_connector *connector = intel_dp->attached_connector;
+

I think we also need to confirm here that panel replay itself is supported by the sink. Otherwise this bit will for sure be zero and then we'll think that we need to transmit AS SDP even for non-PR sinks.

Quoted 16 lines
drivers/gpu/drm/i915/display/intel_psr.c +3 -0
+	return !(connector->dp.panel_replay_caps.dpcd[INTEL_PR_DPCD_INDEX(DP_PANEL_REPLAY_CAP_CAPABILITY)] &
+		 DP_PANEL_REPLAY_ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR);
+}
drivers/gpu/drm/i915/display/intel_psr.h +1 -0
index 394b641840b3..29723e63888f 100644
@@ -86,5 +86,6 @@ void intel_psr_compute_config_late(struct intel_dp *intel_dp,
 				   struct intel_crtc_state *crtc_state);
 int intel_psr_min_guardband(struct intel_crtc_state *crtc_state);
 bool intel_psr_use_trans_push(const struct intel_crtc_state *crtc_state);
+bool intel_psr_pr_async_video_timing_supported(struct intel_dp *intel_dp);
 
 #endif /* __INTEL_PSR_H__ */
Signature · 1 line
2.45.2
Signature · 2 lines
Ville Syrjälä
Intel
#15 lore raw

PATCH 14/23 drm/i915/dp: Add member to intel_dp to store AS SDP v2 support

In reply to Ankit Nautiyal

DP v2.1 SCR advertises support for FAVT payload fields parsing in DPCD 0x2214 Bit 2. This indicates the support for Adaptive-Sync SDP version 2 (AS SDP v2), which allows the source to set the version in HB2[4:0] and the payload length in HB3[5:0] of the AS SDP header.

DP v2.1 SCR also introduces ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR in the Panel Replay Capability DPCD 0x00b1 (Bit 3). When this bit is set, the sink does not support asynchronous video timing while in a Panel Replay Active state and the source is required to keep transmitting Adaptive-Sync SDPs. The spec mandates that such sinks shall support AS SDP v2.

Infer AS SDP v2 support from these capabilities and store it in struct intel_dp for use by subsequent feature enablement changes.

v2: Include parsing ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR bit to
determine AS SDP v2 support. (Ville)
v3: Use helper to determine asynch video timing support.

---
 .../drm/i915/display/intel_display_types.h    |  1 +
 drivers/gpu/drm/i915/display/intel_dp.c       | 33 +++++++++++++++++++
 2 files changed, 34 insertions(+)
drivers/gpu/drm/i915/display/intel_display_types.h +1 -0
index e2496db1642a..efc609eef4f5 100644
@@ -1870,6 +1870,7 @@ struct intel_dp {
 	/* connector directly attached - won't be use for modeset in mst world */
 	struct intel_connector *attached_connector;
 	bool as_sdp_supported;
+	bool as_sdp_v2_supported;
 
 	struct drm_dp_tunnel *tunnel;
 	bool tunnel_suspended:1;
drivers/gpu/drm/i915/display/intel_dp.c +33 -0
index cec0f3d03c2f..9fd768447f28 100644
@@ -6306,6 +6306,36 @@ intel_dp_unset_edid(struct intel_dp *intel_dp)
 					       false);
 }
 
+static bool
+intel_dp_sink_supports_as_sdp_v2(struct intel_dp *intel_dp)
+{
+	u8 rx_features;
+
+	/*
+	 * The DP spec does not explicitly provide the AS SDP v2 capability.
+	 * So based on the DP v2.1 SCR, we infer it from the following bits:
+	 *
+	 * DP_AS_SDP_FAVT_PAYLOAD_FIELDS_PARSING_SUPPORTED indicates support for
+	 * FAVT, which is explicitly defined to use AS SDP v2.
+	 *
+	 * DP_ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR indicates that the sink
+	 * does not support asynchronous video timing while in PR Active,
+	 * requiring the source to keep transmitting Adaptive-Sync SDPs. The
+	 * spec mandates that such sinks shall support AS SDP v2.
+	 */
+	if (drm_dp_dpcd_read_byte(&intel_dp->aux,
+				  DP_DPRX_FEATURE_ENUMERATION_LIST_CONT_1,
+				  &rx_features) == 1) {
+		if (rx_features & DP_AS_SDP_FAVT_PAYLOAD_FIELDS_PARSING_SUPPORTED)
+			return true;
+	}
+
+	if (!intel_psr_pr_async_video_timing_supported(intel_dp))
+		return true;
+
+	return false;
+}
+
 static void
 intel_dp_detect_sdp_caps(struct intel_dp *intel_dp)
 {
@@ -6313,6 +6343,9 @@ intel_dp_detect_sdp_caps(struct intel_dp *intel_dp)
 
 	intel_dp->as_sdp_supported = HAS_AS_SDP(display) &&
 		drm_dp_as_sdp_supported(&intel_dp->aux, intel_dp->dpcd);
+
+	intel_dp->as_sdp_v2_supported = intel_dp->as_sdp_supported &&
+					intel_dp_sink_supports_as_sdp_v2(intel_dp);
 }
 
 static bool intel_dp_needs_dpcd_probe(struct intel_dp *intel_dp, bool force_on_external)
Signature · 1 line
2.45.2
#16 lore raw

Re: [PATCH 14/23] drm/i915/dp: Add member to intel_dp to store AS SDP v2 support

Quoted 5 lines · Ankit Nautiyal

DP v2.1 SCR advertises support for FAVT payload fields parsing in DPCD 0x2214 Bit 2. This indicates the support for Adaptive-Sync SDP version 2 (AS SDP v2), which allows the source to set the version in HB2[4:0] and the payload length in HB3[5:0] of the AS SDP header.

Quoted 8 lines

DP v2.1 SCR also introduces ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR in the Panel Replay Capability DPCD 0x00b1 (Bit 3). When this bit is set, the sink does not support asynchronous video timing while in a Panel Replay Active state and the source is required to keep transmitting Adaptive-Sync SDPs. The spec mandates that such sinks shall support AS SDP v2.

Infer AS SDP v2 support from these capabilities and store it in struct intel_dp for use by subsequent feature enablement changes.

Hmm. After some more reading I think we might actually want to consult the DisplayID for this. The presence of the new adaptive sync block there might be a good indication for AS SDP v2 support. While the DisplayID 2.1 spec itself doesn't say that AS SDP v2 must be supported when the block is present, the DP 2.1 spec does at least say the converse. Ie. adaptive sync capable DP 2.1 sink must have the DisplayID block. And DP 2.1 seems to be all about the AS SDP v2 and not AS SDP v1 (that's a DP 2.0 thing really).

I think there was a patch posted very recently on dri-devel, adding the adaptive sync DisplayID block parsing. Or at least some DisplayID block related to this.

The intel_psr_pr_async_video_timing_supported() check I think we want to keep regardless, because you could have a sink with PR but no adaptive sync. But the FAVT check could perhaps be dropped if we add the DisplayID check.

Quoted 76 lines

v2: Include parsing ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR bit to
determine AS SDP v2 support. (Ville)
v3: Use helper to determine asynch video timing support.

---
 .../drm/i915/display/intel_display_types.h    |  1 +
 drivers/gpu/drm/i915/display/intel_dp.c       | 33 +++++++++++++++++++
 2 files changed, 34 insertions(+)
drivers/gpu/drm/i915/display/intel_display_types.h +1 -0
index e2496db1642a..efc609eef4f5 100644
@@ -1870,6 +1870,7 @@ struct intel_dp {
 	/* connector directly attached - won't be use for modeset in mst world */
 	struct intel_connector *attached_connector;
 	bool as_sdp_supported;
+	bool as_sdp_v2_supported;
 
 	struct drm_dp_tunnel *tunnel;
 	bool tunnel_suspended:1;
drivers/gpu/drm/i915/display/intel_dp.c +33 -0
index cec0f3d03c2f..9fd768447f28 100644
@@ -6306,6 +6306,36 @@ intel_dp_unset_edid(struct intel_dp *intel_dp)
 					       false);
 }
 
+static bool
+intel_dp_sink_supports_as_sdp_v2(struct intel_dp *intel_dp)
+{
+	u8 rx_features;
+
+	/*
+	 * The DP spec does not explicitly provide the AS SDP v2 capability.
+	 * So based on the DP v2.1 SCR, we infer it from the following bits:
+	 *
+	 * DP_AS_SDP_FAVT_PAYLOAD_FIELDS_PARSING_SUPPORTED indicates support for
+	 * FAVT, which is explicitly defined to use AS SDP v2.
+	 *
+	 * DP_ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR indicates that the sink
+	 * does not support asynchronous video timing while in PR Active,
+	 * requiring the source to keep transmitting Adaptive-Sync SDPs. The
+	 * spec mandates that such sinks shall support AS SDP v2.
+	 */
+	if (drm_dp_dpcd_read_byte(&intel_dp->aux,
+				  DP_DPRX_FEATURE_ENUMERATION_LIST_CONT_1,
+				  &rx_features) == 1) {
+		if (rx_features & DP_AS_SDP_FAVT_PAYLOAD_FIELDS_PARSING_SUPPORTED)
+			return true;
+	}
+
+	if (!intel_psr_pr_async_video_timing_supported(intel_dp))
+		return true;
+
+	return false;
+}
+
 static void
 intel_dp_detect_sdp_caps(struct intel_dp *intel_dp)
 {
@@ -6313,6 +6343,9 @@ intel_dp_detect_sdp_caps(struct intel_dp *intel_dp)
 
 	intel_dp->as_sdp_supported = HAS_AS_SDP(display) &&
 		drm_dp_as_sdp_supported(&intel_dp->aux, intel_dp->dpcd);
+
+	intel_dp->as_sdp_v2_supported = intel_dp->as_sdp_supported &&
+					intel_dp_sink_supports_as_sdp_v2(intel_dp);
 }
 
 static bool intel_dp_needs_dpcd_probe(struct intel_dp *intel_dp, bool force_on_external)
Signature · 1 line
2.45.2
Signature · 2 lines
Ville Syrjälä
Intel
#17 lore raw

Re: [PATCH 14/23] drm/i915/dp: Add member to intel_dp to store AS SDP v2 support

Quoted 15 lines · Ville Syrjälä
Quoted 5 lines · Ankit Nautiyal

DP v2.1 SCR advertises support for FAVT payload fields parsing in DPCD 0x2214 Bit 2. This indicates the support for Adaptive-Sync SDP version 2 (AS SDP v2), which allows the source to set the version in HB2[4:0] and the payload length in HB3[5:0] of the AS SDP header.

Quoted 8 lines

DP v2.1 SCR also introduces ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR in the Panel Replay Capability DPCD 0x00b1 (Bit 3). When this bit is set, the sink does not support asynchronous video timing while in a Panel Replay Active state and the source is required to keep transmitting Adaptive-Sync SDPs. The spec mandates that such sinks shall support AS SDP v2.

Infer AS SDP v2 support from these capabilities and store it in struct intel_dp for use by subsequent feature enablement changes.

Quoted 8 lines

Hmm. After some more reading I think we might actually want to consult the DisplayID for this. The presence of the new adaptive sync block there might be a good indication for AS SDP v2 support. While the DisplayID 2.1 spec itself doesn't say that AS SDP v2 must be supported when the block is present, the DP 2.1 spec does at least say the converse. Ie. adaptive sync capable DP 2.1 sink must have the DisplayID block. And DP 2.1 seems to be all about the AS SDP v2 and not AS SDP v1 (that's a DP 2.0 thing really).

Hmm. But the EDID comes from the display, and we could have a PCON in between that doesn't understand AS SDP v2. Sigh. What a mess.

Quoted 90 lines

I think there was a patch posted very recently on dri-devel, adding the adaptive sync DisplayID block parsing. Or at least some DisplayID block related to this.

The intel_psr_pr_async_video_timing_supported() check I think we want to keep regardless, because you could have a sink with PR but no adaptive sync. But the FAVT check could perhaps be dropped if we add the DisplayID check.

Quoted 76 lines

v2: Include parsing ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR bit to
determine AS SDP v2 support. (Ville)
v3: Use helper to determine asynch video timing support.

---
 .../drm/i915/display/intel_display_types.h    |  1 +
 drivers/gpu/drm/i915/display/intel_dp.c       | 33 +++++++++++++++++++
 2 files changed, 34 insertions(+)
drivers/gpu/drm/i915/display/intel_display_types.h +1 -0
index e2496db1642a..efc609eef4f5 100644
@@ -1870,6 +1870,7 @@ struct intel_dp {
 	/* connector directly attached - won't be use for modeset in mst world */
 	struct intel_connector *attached_connector;
 	bool as_sdp_supported;
+	bool as_sdp_v2_supported;
 
 	struct drm_dp_tunnel *tunnel;
 	bool tunnel_suspended:1;
drivers/gpu/drm/i915/display/intel_dp.c +33 -0
index cec0f3d03c2f..9fd768447f28 100644
@@ -6306,6 +6306,36 @@ intel_dp_unset_edid(struct intel_dp *intel_dp)
 					       false);
 }
 
+static bool
+intel_dp_sink_supports_as_sdp_v2(struct intel_dp *intel_dp)
+{
+	u8 rx_features;
+
+	/*
+	 * The DP spec does not explicitly provide the AS SDP v2 capability.
+	 * So based on the DP v2.1 SCR, we infer it from the following bits:
+	 *
+	 * DP_AS_SDP_FAVT_PAYLOAD_FIELDS_PARSING_SUPPORTED indicates support for
+	 * FAVT, which is explicitly defined to use AS SDP v2.
+	 *
+	 * DP_ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR indicates that the sink
+	 * does not support asynchronous video timing while in PR Active,
+	 * requiring the source to keep transmitting Adaptive-Sync SDPs. The
+	 * spec mandates that such sinks shall support AS SDP v2.
+	 */
+	if (drm_dp_dpcd_read_byte(&intel_dp->aux,
+				  DP_DPRX_FEATURE_ENUMERATION_LIST_CONT_1,
+				  &rx_features) == 1) {
+		if (rx_features & DP_AS_SDP_FAVT_PAYLOAD_FIELDS_PARSING_SUPPORTED)
+			return true;
+	}
+
+	if (!intel_psr_pr_async_video_timing_supported(intel_dp))
+		return true;
+
+	return false;
+}
+
 static void
 intel_dp_detect_sdp_caps(struct intel_dp *intel_dp)
 {
@@ -6313,6 +6343,9 @@ intel_dp_detect_sdp_caps(struct intel_dp *intel_dp)
 
 	intel_dp->as_sdp_supported = HAS_AS_SDP(display) &&
 		drm_dp_as_sdp_supported(&intel_dp->aux, intel_dp->dpcd);
+
+	intel_dp->as_sdp_v2_supported = intel_dp->as_sdp_supported &&
+					intel_dp_sink_supports_as_sdp_v2(intel_dp);
 }
 
 static bool intel_dp_needs_dpcd_probe(struct intel_dp *intel_dp, bool force_on_external)
Signature · 1 line
2.45.2
Signature · 2 lines
Ville Syrjälä
Intel
Signature · 2 lines
Ville Syrjälä
Intel
#18 lore raw

Re: [PATCH 14/23] drm/i915/dp: Add member to intel_dp to store AS SDP v2 support

Quoted 23 lines · Ville Syrjälä
Quoted 14 lines · Ville Syrjälä
Quoted 5 lines · Ankit Nautiyal

DP v2.1 SCR advertises support for FAVT payload fields parsing in DPCD 0x2214 Bit 2. This indicates the support for Adaptive-Sync SDP version 2 (AS SDP v2), which allows the source to set the version in HB2[4:0] and the payload length in HB3[5:0] of the AS SDP header.

Quoted 8 lines

DP v2.1 SCR also introduces ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR in the Panel Replay Capability DPCD 0x00b1 (Bit 3). When this bit is set, the sink does not support asynchronous video timing while in a Panel Replay Active state and the source is required to keep transmitting Adaptive-Sync SDPs. The spec mandates that such sinks shall support AS SDP v2.

Infer AS SDP v2 support from these capabilities and store it in struct intel_dp for use by subsequent feature enablement changes.

Quoted 8 lines

Hmm. After some more reading I think we might actually want to consult the DisplayID for this. The presence of the new adaptive sync block there might be a good indication for AS SDP v2 support. While the DisplayID 2.1 spec itself doesn't say that AS SDP v2 must be supported when the block is present, the DP 2.1 spec does at least say the converse. Ie. adaptive sync capable DP 2.1 sink must have the DisplayID block. And DP 2.1 seems to be all about the AS SDP v2 and not AS SDP v1 (that's a DP 2.0 thing really).

Quoted 2 lines

Hmm. But the EDID comes from the display, and we could have a PCON in between that doesn't understand AS SDP v2. Sigh. What a mess.

hmm yeah. Is it alright to have a #TODO something like:

/*
* TODO: Check the Adaptive-Sync DisplayID 2.1 block once DisplayID parsing
* is available. This may help detect AS SDP v2 support for native DP 2.1
* sinks that do not expose FAVT or PR-based capability bits.
*
* In the presence of PCONs, check PCON support from DPCD and sink support from Display ID. */

Regards,

Ankit

Quoted 87 lines
Quoted 87 lines

I think there was a patch posted very recently on dri-devel, adding the adaptive sync DisplayID block parsing. Or at least some DisplayID block related to this.

The intel_psr_pr_async_video_timing_supported() check I think we want to keep regardless, because you could have a sink with PR but no adaptive sync. But the FAVT check could perhaps be dropped if we add the DisplayID check.

Quoted 75 lines

v2: Include parsing ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR bit to
determine AS SDP v2 support. (Ville)
v3: Use helper to determine asynch video timing support.

---
  .../drm/i915/display/intel_display_types.h    |  1 +
  drivers/gpu/drm/i915/display/intel_dp.c       | 33 +++++++++++++++++++
  2 files changed, 34 insertions(+)
drivers/gpu/drm/i915/display/intel_display_types.h +1 -0
index e2496db1642a..efc609eef4f5 100644
@@ -1870,6 +1870,7 @@ struct intel_dp {
  	/* connector directly attached - won't be use for modeset in mst world */
  	struct intel_connector *attached_connector;
  	bool as_sdp_supported;
+	bool as_sdp_v2_supported;
 
  	struct drm_dp_tunnel *tunnel;
  	bool tunnel_suspended:1;
drivers/gpu/drm/i915/display/intel_dp.c +33 -0
index cec0f3d03c2f..9fd768447f28 100644
@@ -6306,6 +6306,36 @@ intel_dp_unset_edid(struct intel_dp *intel_dp)
  					       false);
  }
 
+static bool
+intel_dp_sink_supports_as_sdp_v2(struct intel_dp *intel_dp)
+{
+	u8 rx_features;
+
+	/*
+	 * The DP spec does not explicitly provide the AS SDP v2 capability.
+	 * So based on the DP v2.1 SCR, we infer it from the following bits:
+	 *
+	 * DP_AS_SDP_FAVT_PAYLOAD_FIELDS_PARSING_SUPPORTED indicates support for
+	 * FAVT, which is explicitly defined to use AS SDP v2.
+	 *
+	 * DP_ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR indicates that the sink
+	 * does not support asynchronous video timing while in PR Active,
+	 * requiring the source to keep transmitting Adaptive-Sync SDPs. The
+	 * spec mandates that such sinks shall support AS SDP v2.
+	 */
+	if (drm_dp_dpcd_read_byte(&intel_dp->aux,
+				  DP_DPRX_FEATURE_ENUMERATION_LIST_CONT_1,
+				  &rx_features) == 1) {
+		if (rx_features & DP_AS_SDP_FAVT_PAYLOAD_FIELDS_PARSING_SUPPORTED)
+			return true;
+	}
+
+	if (!intel_psr_pr_async_video_timing_supported(intel_dp))
+		return true;
+
+	return false;
+}
+
  static void
  intel_dp_detect_sdp_caps(struct intel_dp *intel_dp)
  {
@@ -6313,6 +6343,9 @@ intel_dp_detect_sdp_caps(struct intel_dp *intel_dp)
 
  	intel_dp->as_sdp_supported = HAS_AS_SDP(display) &&
  		drm_dp_as_sdp_supported(&intel_dp->aux, intel_dp->dpcd);
+
+	intel_dp->as_sdp_v2_supported = intel_dp->as_sdp_supported &&
+					intel_dp_sink_supports_as_sdp_v2(intel_dp);
  }
 
  static bool intel_dp_needs_dpcd_probe(struct intel_dp *intel_dp, bool force_on_external)
Signature · 1 line
2.45.2
Signature · 2 lines
Ville Syrjälä
Intel
#19 lore raw

Re: [PATCH 14/23] drm/i915/dp: Add member to intel_dp to store AS SDP v2 support

Quoted 31 lines · Ankit K
Quoted 23 lines · Ville Syrjälä
Quoted 14 lines · Ville Syrjälä
Quoted 5 lines · Ankit Nautiyal

DP v2.1 SCR advertises support for FAVT payload fields parsing in DPCD 0x2214 Bit 2. This indicates the support for Adaptive-Sync SDP version 2 (AS SDP v2), which allows the source to set the version in HB2[4:0] and the payload length in HB3[5:0] of the AS SDP header.

Quoted 8 lines

DP v2.1 SCR also introduces ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR in the Panel Replay Capability DPCD 0x00b1 (Bit 3). When this bit is set, the sink does not support asynchronous video timing while in a Panel Replay Active state and the source is required to keep transmitting Adaptive-Sync SDPs. The spec mandates that such sinks shall support AS SDP v2.

Infer AS SDP v2 support from these capabilities and store it in struct intel_dp for use by subsequent feature enablement changes.

Quoted 8 lines

Hmm. After some more reading I think we might actually want to consult the DisplayID for this. The presence of the new adaptive sync block there might be a good indication for AS SDP v2 support. While the DisplayID 2.1 spec itself doesn't say that AS SDP v2 must be supported when the block is present, the DP 2.1 spec does at least say the converse. Ie. adaptive sync capable DP 2.1 sink must have the DisplayID block. And DP 2.1 seems to be all about the AS SDP v2 and not AS SDP v1 (that's a DP 2.0 thing really).

Quoted 2 lines

Hmm. But the EDID comes from the display, and we could have a PCON in between that doesn't understand AS SDP v2. Sigh. What a mess.

hmm yeah. Is it alright to have a #TODO something like:

Quoted 8 lines

/*
* TODO: Check the Adaptive-Sync DisplayID 2.1 block once DisplayID parsing
* is available. This may help detect AS SDP v2 support for native DP 2.1
* sinks that do not expose FAVT or PR-based capability bits.
*
* In the presence of PCONs, check PCON support from DPCD and sink support from Display ID. */

Going throught the eDP spec, it seems for eDP,  if AS SDP is supported it implicitly means AS SDP v2.

From eDP spec 2.0

Section 2.8 Adaptive-Sync

"eDP may optionally support the Adaptive-Sync operation to maintain video timing synchronization between Source and Sink devices (see DP Standard, “Adaptive-Sync Operation for VESA AdaptiveSync”). If supported, the Adaptive-Sync SDP version shall be v2 or higher (HB2[4:0] = 02h or higher, and HB3[5:0] = 09h or higher)."

Also, the FAVT Payload fields parsing support is not used for eDP and will always read 0.

This however doesn't mean that FAVT is not supported, the support is implicit.

Section 8.7.1 Video Timing Synchronization while in a PR Active State with ALPM Enabled

...Source-to-Sink device timing synchronization during a PR Active state that uses the Adaptive-Sync SDP may be enabled in Adaptive-Sync (FAVT or AVT mode) operation (DPCD 00107h[7] = 1) –or– when
DPCD 00107h[7] = 0..

So, I will update the logic for eDP here.

Regards,

Ankit

Quoted 96 lines

Regards,

Ankit

Quoted 91 lines
Quoted 91 lines

I think there was a patch posted very recently on dri-devel, adding the adaptive sync DisplayID block parsing. Or at least some DisplayID block related to this.

The intel_psr_pr_async_video_timing_supported() check I think we want to keep regardless, because you could have a sink with PR but no adaptive sync. But the FAVT check could perhaps be dropped if we add the DisplayID check.

Quoted 79 lines

v2: Include parsing ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR bit to determine AS SDP v2 support. (Ville)
v3: Use helper to determine asynch video timing support.

---
  .../drm/i915/display/intel_display_types.h    |  1 + drivers/gpu/drm/i915/display/intel_dp.c       | 33
+++++++++++++++++++

  2 files changed, 34 insertions(+)
drivers/gpu/drm/i915/display/intel_display_types.h +1 -0
index e2496db1642a..efc609eef4f5 100644
@@ -1870,6 +1870,7 @@ struct intel_dp {
       /* connector directly attached - won't be use for modeset in mst world */
       struct intel_connector *attached_connector;
       bool as_sdp_supported;
+    bool as_sdp_v2_supported;
         struct drm_dp_tunnel *tunnel;
       bool tunnel_suspended:1;
drivers/gpu/drm/i915/display/intel_dp.c +0 -0
index cec0f3d03c2f..9fd768447f28 100644
@@ -6306,6 +6306,36 @@ intel_dp_unset_edid(struct intel_dp *intel_dp)
                              false);
   }

  +static bool

drivers/gpu/drm/i915/display/intel_dp.c +6 -0
+intel_dp_sink_supports_as_sdp_v2(struct intel_dp *intel_dp)
+{
+    u8 rx_features;
+
+    /*
+     * The DP spec does not explicitly provide the AS SDP v2 

capability.
+     * So based on the DP v2.1 SCR, we infer it from the following bits:
+     *
+     * DP_AS_SDP_FAVT_PAYLOAD_FIELDS_PARSING_SUPPORTED indicates support for
+     * FAVT, which is explicitly defined to use AS SDP v2. +     *
+     * DP_ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR indicates that the sink
+     * does not support asynchronous video timing while in PR Active, +     * requiring the source to keep transmitting Adaptive-Sync SDPs. The
+     * spec mandates that such sinks shall support AS SDP v2. +     */
+    if (drm_dp_dpcd_read_byte(&intel_dp->aux,
+                  DP_DPRX_FEATURE_ENUMERATION_LIST_CONT_1, +                  &rx_features) == 1) {
+        if (rx_features & DP_AS_SDP_FAVT_PAYLOAD_FIELDS_PARSING_SUPPORTED)
+            return true;
+    }

drivers/gpu/drm/i915/display/intel_dp.c +7 -0
+
+    if (!intel_psr_pr_async_video_timing_supported(intel_dp))
+        return true;
+
+    return false;
+}
+

  static void
  intel_dp_detect_sdp_caps(struct intel_dp *intel_dp) {

drivers/gpu/drm/i915/display/intel_dp.c +0 -0
@@ -6313,6 +6343,9 @@ intel_dp_detect_sdp_caps(struct intel_dp 
 *intel_dp)

        intel_dp->as_sdp_supported = HAS_AS_SDP(display) && drm_dp_as_sdp_supported(&intel_dp->aux, intel_dp->dpcd);
+
+    intel_dp->as_sdp_v2_supported = intel_dp->as_sdp_supported &&
+ intel_dp_sink_supports_as_sdp_v2(intel_dp);
  }
    static bool intel_dp_needs_dpcd_probe(struct intel_dp *intel_dp, bool force_on_external)

Signature · 1 line
2.45.2
Signature · 2 lines
Ville Syrjälä
Intel
#20 lore raw

Re: [PATCH 14/23] drm/i915/dp: Add member to intel_dp to store AS SDP v2 support

Quoted 52 lines · Ankit K
Quoted 39 lines · Ankit K
Quoted 31 lines · Ville Syrjälä
Quoted 19 lines · Ville Syrjälä
Quoted 16 lines · Ankit Nautiyal

DP v2.1 SCR advertises support for FAVT payload fields parsing in DPCD 0x2214 Bit 2. This indicates the support for Adaptive-Sync SDP version 2
(AS SDP v2), which allows the source to set the version in HB2[4:0] and the
payload length in HB3[5:0] of the AS SDP header.

DP v2.1 SCR also introduces ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR in the
Panel Replay Capability DPCD 0x00b1 (Bit 3). When this bit is set, the sink
does not support asynchronous video timing while in a Panel Replay Active
state and the source is required to keep transmitting Adaptive-Sync SDPs. The spec mandates that such sinks shall support AS SDP v2.

Quoted 2 lines

Infer AS SDP v2 support from these capabilities and store it in struct intel_dp for use by subsequent feature enablement changes.

Quoted 11 lines

Hmm. After some more reading I think we might actually want to consult the DisplayID for this. The presence of the new adaptive sync block there might be a good indication for AS SDP v2 support. While the DisplayID 2.1 spec itself doesn't say that AS SDP v2 must be supported when the block is present, the DP 2.1 spec does at least say the converse.
Ie. adaptive sync capable DP 2.1 sink must have the DisplayID block. And
DP 2.1 seems to be all about the AS SDP v2 and not AS SDP v1 (that's a DP
2.0 thing really).

Quoted 2 lines

Hmm. But the EDID comes from the display, and we could have a PCON in between that doesn't understand AS SDP v2. Sigh. What a mess.

hmm yeah. Is it alright to have a #TODO something like:

Quoted 10 lines

/*
* TODO: Check the Adaptive-Sync DisplayID 2.1 block once DisplayID parsing
* is available. This may help detect AS SDP v2 support for native DP 2.1
* sinks that do not expose FAVT or PR-based capability bits. *
* In the presence of PCONs, check PCON support from DPCD and sink support from Display ID.
*/

Quoted 4 lines

Going throught the eDP spec, it seems for eDP,  if AS SDP is supported it implicitly means AS SDP v2.

From eDP spec 2.0

eDP 2.0 does not seem relevant for us. But IIRC eDP v1.5 also implies AS SDP v2 support. And that we can actually check if necessary.

Though I suppose the only real concern is branch devices. Maybe we should use the FAVT check only for those. For everything else it should be sufficient to have either the DisplayID adaptive sync block or PR ALPM support.

Quoted 134 lines

Section 2.8 Adaptive-Sync

"eDP may optionally support the Adaptive-Sync operation to maintain video timing
synchronization between Source and Sink devices (see DP Standard, “Adaptive-Sync Operation
for VESA AdaptiveSync”). If supported, the Adaptive-Sync SDP version shall be v2 or higher
(HB2[4:0] = 02h or higher, and HB3[5:0] = 09h or higher)."

Also, the FAVT Payload fields parsing support is not used for eDP and will always read 0.

This however doesn't mean that FAVT is not supported, the support is implicit.

Section 8.7.1 Video Timing Synchronization while in a PR Active State with ALPM Enabled

...Source-to-Sink device timing synchronization during a PR Active state that uses the Adaptive-Sync SDP may be enabled in Adaptive-Sync (FAVT or AVT mode) operation (DPCD 00107h[7] = 1) –or– when DPCD 00107h[7] = 0..

So, I will update the logic for eDP here.

Regards,

Ankit

Quoted 100 lines

Regards,

Ankit

Quoted 95 lines
Quoted 95 lines

I think there was a patch posted very recently on dri-devel, adding the adaptive sync DisplayID block parsing. Or at least some DisplayID block related to this.

The intel_psr_pr_async_video_timing_supported() check I think we want to keep regardless, because you could have a sink with PR but no adaptive sync. But the FAVT check could perhaps be dropped if we add the DisplayID check.

Quoted 83 lines

v2: Include parsing ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR bit to determine AS SDP v2 support. (Ville)
v3: Use helper to determine asynch video timing support.

---
  .../drm/i915/display/intel_display_types.h    |  1 + drivers/gpu/drm/i915/display/intel_dp.c       | 33
+++++++++++++++++++

  2 files changed, 34 insertions(+)
drivers/gpu/drm/i915/display/intel_display_types.h +0 -0
index e2496db1642a..efc609eef4f5 100644
@@ -1870,6 +1870,7 @@ struct intel_dp {
       /* connector directly attached - won't be use for modeset in 

mst world */
      struct intel_connector *attached_connector;
      bool as_sdp_supported;
+    bool as_sdp_v2_supported;
        struct drm_dp_tunnel *tunnel;
      bool tunnel_suspended:1;

drivers/gpu/drm/i915/display/intel_dp.c +0 -0
index cec0f3d03c2f..9fd768447f28 100644
@@ -6306,6 +6306,36 @@ intel_dp_unset_edid(struct intel_dp *intel_dp)
                              false);
   }

  +static bool

drivers/gpu/drm/i915/display/intel_dp.c +6 -0
+intel_dp_sink_supports_as_sdp_v2(struct intel_dp *intel_dp)
+{
+    u8 rx_features;
+
+    /*
+     * The DP spec does not explicitly provide the AS SDP v2 

capability.
+     * So based on the DP v2.1 SCR, we infer it from the following bits:
+     *
+     * DP_AS_SDP_FAVT_PAYLOAD_FIELDS_PARSING_SUPPORTED indicates support for
+     * FAVT, which is explicitly defined to use AS SDP v2. +     *
+     * DP_ASYNC_VIDEO_TIMING_NOT_SUPPORTED_IN_PR indicates that the sink
+     * does not support asynchronous video timing while in PR Active, +     * requiring the source to keep transmitting Adaptive-Sync SDPs. The
+     * spec mandates that such sinks shall support AS SDP v2. +     */
+    if (drm_dp_dpcd_read_byte(&intel_dp->aux,
+                  DP_DPRX_FEATURE_ENUMERATION_LIST_CONT_1, +                  &rx_features) == 1) {
+        if (rx_features & DP_AS_SDP_FAVT_PAYLOAD_FIELDS_PARSING_SUPPORTED)
+            return true;
+    }

drivers/gpu/drm/i915/display/intel_dp.c +7 -0
+
+    if (!intel_psr_pr_async_video_timing_supported(intel_dp))
+        return true;
+
+    return false;
+}
+

  static void
  intel_dp_detect_sdp_caps(struct intel_dp *intel_dp) {

drivers/gpu/drm/i915/display/intel_dp.c +0 -0
@@ -6313,6 +6343,9 @@ intel_dp_detect_sdp_caps(struct intel_dp 
 *intel_dp)

        intel_dp->as_sdp_supported = HAS_AS_SDP(display) && drm_dp_as_sdp_supported(&intel_dp->aux, intel_dp->dpcd); +
+    intel_dp->as_sdp_v2_supported = intel_dp->as_sdp_supported &&
+ intel_dp_sink_supports_as_sdp_v2(intel_dp);
  }
    static bool intel_dp_needs_dpcd_probe(struct intel_dp *intel_dp, bool force_on_external)

Signature · 1 line
2.45.2
Signature · 2 lines
Ville Syrjälä
Intel
Signature · 2 lines
Ville Syrjälä
Intel
#21 lore raw

PATCH 15/23 drm/i915/dp: Allow AS SDP only if v2 is supported

In reply to Ankit Nautiyal

We do not support AS SDP version 1, so allow AS SDP only if AS SDP v2 is supported.

---
 drivers/gpu/drm/i915/display/intel_dp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/i915/display/intel_dp.c +1 -1
index 9fd768447f28..e97233805348 100644
@@ -3127,7 +3127,7 @@ static void intel_dp_compute_vsc_colorimetry(const struct intel_crtc_state *crtc
 static bool intel_dp_needs_as_sdp(struct intel_dp *intel_dp,
 				  struct intel_crtc_state *crtc_state)
 {
-	if (!intel_dp->as_sdp_supported)
+	if (!intel_dp->as_sdp_v2_supported)
 		return false;
 
 	/*
Signature · 1 line
2.45.2
#22 lore raw

PATCH 16/23 drm/i915/psr: Write the PR config DPCDs in burst mode

In reply to Ankit Nautiyal

Replace the consecutive single-byte writes to PANEL_REPLAY_CONFIG and CONFIG2 with one drm_dp_dpcd_write() burst starting at PANEL_REPLAY_CONFIG, reducing AUX transactions.

v2: Drop extra conditions, and optimize variables. (Ville)
v3: Drop the error check after write. (Ville)

---
 drivers/gpu/drm/i915/display/intel_psr.c | 26 ++++++++++++------------
 1 file changed, 13 insertions(+), 13 deletions(-)
drivers/gpu/drm/i915/display/intel_psr.c +13 -13
index 782e839b7396..3fe7691c8b77 100644
@@ -782,27 +782,27 @@ static bool psr2_su_region_et_valid(struct intel_connector *connector, bool pane
 static void _panel_replay_enable_sink(struct intel_dp *intel_dp,
 				      const struct intel_crtc_state *crtc_state)
 {
-	u8 val = DP_PANEL_REPLAY_ENABLE |
-		DP_PANEL_REPLAY_VSC_SDP_CRC_EN |
-		DP_PANEL_REPLAY_UNRECOVERABLE_ERROR_EN |
-		DP_PANEL_REPLAY_RFB_STORAGE_ERROR_EN |
-		DP_PANEL_REPLAY_ACTIVE_FRAME_CRC_ERROR_EN;
-	u8 panel_replay_config2 = DP_PANEL_REPLAY_CRC_VERIFICATION;
+	u8 panel_replay_config[2];
+
+	panel_replay_config[0] = DP_PANEL_REPLAY_ENABLE |
+				 DP_PANEL_REPLAY_VSC_SDP_CRC_EN |
+				 DP_PANEL_REPLAY_UNRECOVERABLE_ERROR_EN |
+				 DP_PANEL_REPLAY_RFB_STORAGE_ERROR_EN |
+				 DP_PANEL_REPLAY_ACTIVE_FRAME_CRC_ERROR_EN;
+	panel_replay_config[1] = DP_PANEL_REPLAY_CRC_VERIFICATION;
 
 	if (crtc_state->has_sel_update)
-		val |= DP_PANEL_REPLAY_SU_ENABLE;
+		panel_replay_config[0] |= DP_PANEL_REPLAY_SU_ENABLE;
 
 	if (crtc_state->enable_psr2_su_region_et)
-		val |= DP_PANEL_REPLAY_ENABLE_SU_REGION_ET;
+		panel_replay_config[0] |= DP_PANEL_REPLAY_ENABLE_SU_REGION_ET;
 
 	if (crtc_state->req_psr2_sdp_prior_scanline)
-		panel_replay_config2 |=
+		panel_replay_config[1] |=
 			DP_PANEL_REPLAY_SU_REGION_SCANLINE_CAPTURE;
 
-	drm_dp_dpcd_writeb(&intel_dp->aux, PANEL_REPLAY_CONFIG, val);
-
-	drm_dp_dpcd_writeb(&intel_dp->aux, PANEL_REPLAY_CONFIG2,
-			   panel_replay_config2);
+	drm_dp_dpcd_write(&intel_dp->aux, PANEL_REPLAY_CONFIG,
+			  panel_replay_config, sizeof(panel_replay_config));
 }
 
 static void _psr_enable_sink(struct intel_dp *intel_dp,
Signature · 1 line
2.45.2
#23 lore raw

PATCH 17/23 drm/i915/display: Add helper for AS SDP transmission time selection

In reply to Ankit Nautiyal

AS SDP may be transmitted at T1 or T2 depending on Panel Replay and Adaptive Sync SDP configuration as per DP 2.1. Current we are using T1 only, but future PR/AS SDP modes/features may require T2 or dynamic selection.

Introduce a helper to return the appropriate AS SDP transmission time so that a single value is consistently used for programming PR_ALPM. For now this returns T1.

v2: Avoid adding new member to crtc_state; use a helper. (Ville)
v3: Clarify why AS SDP transmission time is fixed to T1. (Ville)
v4: Return u8 from intel_dp_as_sdp_transmission_time(). (Ville)

Bspec: 68920

---
 drivers/gpu/drm/i915/display/intel_alpm.c | 20 +++++++++++++++++++-
 drivers/gpu/drm/i915/display/intel_dp.c   | 11 +++++++++++
 drivers/gpu/drm/i915/display/intel_dp.h   |  2 ++
 3 files changed, 32 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/display/intel_alpm.c +19 -1
index a7350ce8e716..c6963ea420cc 100644
@@ -11,6 +11,7 @@
 #include "intel_crtc.h"
 #include "intel_de.h"
 #include "intel_display_types.h"
+#include "intel_display_utils.h"
 #include "intel_dp.h"
 #include "intel_dp_aux.h"
 #include "intel_psr.h"
@@ -359,6 +360,23 @@ void intel_alpm_lobf_compute_config(struct intel_dp *intel_dp,
 	crtc_state->has_lobf = true;
 }
 
+static u32 get_pr_alpm_as_sdp_transmission_time(const struct intel_crtc_state *crtc_state)
+{
+	u8 as_sdp_setup_time = intel_dp_as_sdp_transmission_time();
+
+	switch (as_sdp_setup_time) {
+	case DP_PR_AS_SDP_SETUP_TIME_T1:
+		return PR_ALPM_CTL_ADAPTIVE_SYNC_SDP_POSITION_T1;
+	case DP_PR_AS_SDP_SETUP_TIME_DYNAMIC:
+		return PR_ALPM_CTL_ADAPTIVE_SYNC_SDP_POSITION_T1_OR_T2;
+	case DP_PR_AS_SDP_SETUP_TIME_T2:
+		return PR_ALPM_CTL_ADAPTIVE_SYNC_SDP_POSITION_T2;
+	default:
+		MISSING_CASE(as_sdp_setup_time);
+		return PR_ALPM_CTL_ADAPTIVE_SYNC_SDP_POSITION_T1;
+	}
+}
+
 static void lnl_alpm_configure(struct intel_dp *intel_dp,
 			       const struct intel_crtc_state *crtc_state)
 {
@@ -382,7 +400,7 @@ static void lnl_alpm_configure(struct intel_dp *intel_dp,
 			ALPM_CTL_AUX_LESS_WAKE_TIME(crtc_state->alpm_state.aux_less_wake_lines);
 
 		if (intel_dp->as_sdp_supported) {
-			u32 pr_alpm_ctl = PR_ALPM_CTL_ADAPTIVE_SYNC_SDP_POSITION_T1;
+			u32 pr_alpm_ctl = get_pr_alpm_as_sdp_transmission_time(crtc_state);
 
 			if (crtc_state->link_off_after_as_sdp_when_pr_active)
 				pr_alpm_ctl |= PR_ALPM_CTL_ALLOW_LINK_OFF_BETWEEN_AS_SDP_AND_SU;
drivers/gpu/drm/i915/display/intel_dp.c +11 -0
index e97233805348..87388694ce17 100644
@@ -7466,3 +7466,14 @@ bool intel_dp_joiner_candidate_valid(struct intel_connector *connector,
 
 	return true;
 }
+
+u8 intel_dp_as_sdp_transmission_time(void)
+{
+	/*
+	 * DP allows AS SDP position to move during PR active in some cases, but
+	 * software-controlled refresh rate changes with DC6v / ALPM require the
+	 * AS SDP to remain at T1. Use T1 unconditionally for now.
+	 */
+
+	return DP_PR_AS_SDP_SETUP_TIME_T1;
+}
drivers/gpu/drm/i915/display/intel_dp.h +2 -0
index 2849b9ecdc71..66e7a89bcb64 100644
@@ -238,4 +238,6 @@ bool intel_dp_joiner_candidate_valid(struct intel_connector *connector,
 	for ((__num_joined_pipes) = 1; (__num_joined_pipes) <= (I915_MAX_PIPES); (__num_joined_pipes)++) \
 		for_each_if(intel_dp_joiner_candidate_valid(__connector, (__mode)->hdisplay, __num_joined_pipes))
 
+u8 intel_dp_as_sdp_transmission_time(void);
+
 #endif /* __INTEL_DP_H__ */
Signature · 1 line
2.45.2
#24 lore raw

PATCH 18/23 drm/i915/psr: Program Panel Replay CONFIG3 using AS SDP transmission time

In reply to Ankit Nautiyal

Panel Replay requires the AS SDP transmission time to be written into PANEL_REPLAY_CONFIG3. This field was previously not programmed.

Use the AS SDP transmission-time helper to populate CONFIG3.

---
 drivers/gpu/drm/i915/display/intel_psr.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/display/intel_psr.c +4 -1
index 3fe7691c8b77..a9032b7ff814 100644
@@ -783,6 +783,7 @@ static void _panel_replay_enable_sink(struct intel_dp *intel_dp,
 				      const struct intel_crtc_state *crtc_state)
 {
 	u8 panel_replay_config[2];
+	u8 panel_replay_config_3;
 
 	panel_replay_config[0] = DP_PANEL_REPLAY_ENABLE |
 				 DP_PANEL_REPLAY_VSC_SDP_CRC_EN |
@@ -790,7 +791,6 @@ static void _panel_replay_enable_sink(struct intel_dp *intel_dp,
 				 DP_PANEL_REPLAY_RFB_STORAGE_ERROR_EN |
 				 DP_PANEL_REPLAY_ACTIVE_FRAME_CRC_ERROR_EN;
 	panel_replay_config[1] = DP_PANEL_REPLAY_CRC_VERIFICATION;
-
 	if (crtc_state->has_sel_update)
 		panel_replay_config[0] |= DP_PANEL_REPLAY_SU_ENABLE;
 
@@ -803,6 +803,9 @@ static void _panel_replay_enable_sink(struct intel_dp *intel_dp,
 
 	drm_dp_dpcd_write(&intel_dp->aux, PANEL_REPLAY_CONFIG,
 			  panel_replay_config, sizeof(panel_replay_config));
+
+	panel_replay_config_3 = intel_dp_as_sdp_transmission_time();
+	drm_dp_dpcd_writeb(&intel_dp->aux, PANEL_REPLAY_CONFIG3, panel_replay_config_3);
 }
 
 static void _psr_enable_sink(struct intel_dp *intel_dp,
Signature · 1 line
2.45.2
Reviewed-by on v1 18/23
#25 lore raw

Re: [PATCH 18/23] drm/i915/psr: Program Panel Replay CONFIG3 using AS SDP transmission time

Quoted 6 lines · Ankit Nautiyal

Panel Replay requires the AS SDP transmission time to be written into PANEL_REPLAY_CONFIG3. This field was previously not programmed.

Use the AS SDP transmission-time helper to populate CONFIG3.

Quoted 36 lines
---
 drivers/gpu/drm/i915/display/intel_psr.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/display/intel_psr.c +4 -1
index 3fe7691c8b77..a9032b7ff814 100644
@@ -783,6 +783,7 @@ static void _panel_replay_enable_sink(struct intel_dp *intel_dp,
 				      const struct intel_crtc_state *crtc_state)
 {
 	u8 panel_replay_config[2];
+	u8 panel_replay_config_3;
 
 	panel_replay_config[0] = DP_PANEL_REPLAY_ENABLE |
 				 DP_PANEL_REPLAY_VSC_SDP_CRC_EN |
@@ -790,7 +791,6 @@ static void _panel_replay_enable_sink(struct intel_dp *intel_dp,
 				 DP_PANEL_REPLAY_RFB_STORAGE_ERROR_EN |
 				 DP_PANEL_REPLAY_ACTIVE_FRAME_CRC_ERROR_EN;
 	panel_replay_config[1] = DP_PANEL_REPLAY_CRC_VERIFICATION;
-
 	if (crtc_state->has_sel_update)
 		panel_replay_config[0] |= DP_PANEL_REPLAY_SU_ENABLE;
 
@@ -803,6 +803,9 @@ static void _panel_replay_enable_sink(struct intel_dp *intel_dp,
 
 	drm_dp_dpcd_write(&intel_dp->aux, PANEL_REPLAY_CONFIG,
 			  panel_replay_config, sizeof(panel_replay_config));
+
+	panel_replay_config_3 = intel_dp_as_sdp_transmission_time();
+	drm_dp_dpcd_writeb(&intel_dp->aux, PANEL_REPLAY_CONFIG3, panel_replay_config_3);
 }
 
 static void _psr_enable_sink(struct intel_dp *intel_dp,
Signature · 1 line
2.45.2
Signature · 2 lines
Ville Syrjälä
Intel
#26 lore raw

PATCH 19/23 drm/i915/dp: Set relevant Downspread Ctrl DPCD bits for PR + Auxless ALPM

In reply to Ankit Nautiyal

If a Panel Replay capable sink, supports Async Video timing in PR active state, then source does not necessarily need to send AS SDPs during PR active.

However, if asynchronous video timing is not supported, then for PR with Aux-less ALPM, the source must transmit Adaptive-Sync SDPs for video timing synchronization while PR is active.

If the source needs to send AS SDP during PR active, this requires setting DPCD 0x0107[6] (FIXED_VTOTAL_AS_SDP_EN_IN_PR_ACTIVE). This applies whether VRR is enabled (AVT/FAVT) or fixed-timing mode is used.

This bit defines AS SDP timing behavior during PR Active, even if AS SDPs are briefly suspended.

Program the relevant Downspread Ctrl DPCD bits accordingly.

v2: Instead of Panel Replay check simply use AS SDP enable check. (Ville)

---
 .../gpu/drm/i915/display/intel_dp_link_training.c    | 12 ++++++++++--
 .../gpu/drm/i915/display/intel_dp_link_training.h    |  3 ++-
 drivers/gpu/drm/i915/display/intel_dp_mst.c          |  2 +-
 3 files changed, 13 insertions(+), 4 deletions(-)
drivers/gpu/drm/i915/display/intel_dp_link_training.c +10 -2
index a26094223f78..8b21c479ebfc 100644
@@ -27,6 +27,7 @@
 #include <drm/display/drm_dp_helper.h>
 #include <drm/drm_print.h>
 
+#include "intel_alpm.h"
 #include "intel_display_core.h"
 #include "intel_display_jiffies.h"
 #include "intel_display_types.h"
@@ -34,6 +35,7 @@
 #include "intel_dp.h"
 #include "intel_dp_link_training.h"
 #include "intel_encoder.h"
+#include "intel_hdmi.h"
 #include "intel_hotplug.h"
 #include "intel_panel.h"
 
@@ -710,11 +712,14 @@ static bool intel_dp_link_max_vswing_reached(struct intel_dp *intel_dp,
 	return true;
 }
 
-void intel_dp_link_training_set_mode(struct intel_dp *intel_dp, int link_rate, bool is_vrr)
+void intel_dp_link_training_set_mode(struct intel_dp *intel_dp, int link_rate,
+				     bool is_vrr,
+				     bool as_sdp_enable)
 {
 	u8 link_config[2];
 
 	link_config[0] = is_vrr ? DP_MSA_TIMING_PAR_IGNORE_EN : 0;
+	link_config[0] |= as_sdp_enable ? DP_FIXED_VTOTAL_AS_SDP_EN_IN_PR_ACTIVE : 0;
 	link_config[1] = drm_dp_is_uhbr_rate(link_rate) ?
 			 DP_SET_ANSI_128B132B : DP_SET_ANSI_8B10B;
 	drm_dp_dpcd_write(&intel_dp->aux, DP_DOWNSPREAD_CTRL, link_config, 2);
@@ -737,7 +742,10 @@ static void intel_dp_update_downspread_ctrl(struct intel_dp *intel_dp,
 	  * especially on the first real commit when clearing the inherited flag.
 	  */
 	intel_dp_link_training_set_mode(intel_dp,
-					crtc_state->port_clock, crtc_state->vrr.in_range);
+					crtc_state->port_clock,
+					crtc_state->vrr.in_range,
+					crtc_state->infoframes.enable &
+					intel_hdmi_infoframe_enable(DP_SDP_ADAPTIVE_SYNC));
 }
 
 void intel_dp_link_training_set_bw(struct intel_dp *intel_dp,
drivers/gpu/drm/i915/display/intel_dp_link_training.h +2 -1
index 33dcbde6a408..d3ae8ee38a75 100644
@@ -18,7 +18,8 @@ int intel_dp_init_lttpr_and_dprx_caps(struct intel_dp *intel_dp);
 bool intel_dp_lttpr_transparent_mode_enabled(struct intel_dp *intel_dp);
 
 void intel_dp_link_training_set_mode(struct intel_dp *intel_dp,
-				     int link_rate, bool is_vrr);
+				     int link_rate, bool is_vrr,
+				     bool as_sdp_enable);
 void intel_dp_link_training_set_bw(struct intel_dp *intel_dp,
 				   int link_bw, int rate_select, int lane_count,
 				   bool enhanced_framing, bool post_lt_adj_req);
drivers/gpu/drm/i915/display/intel_dp_mst.c +1 -1
index e8de17834dcd..ffd1cf0aad9a 100644
@@ -2142,7 +2142,7 @@ void intel_dp_mst_prepare_probe(struct intel_dp *intel_dp)
 
 	intel_dp_compute_rate(intel_dp, link_rate, &link_bw, &rate_select);
 
-	intel_dp_link_training_set_mode(intel_dp, link_rate, false);
+	intel_dp_link_training_set_mode(intel_dp, link_rate, false, false);
 	intel_dp_link_training_set_bw(intel_dp, link_bw, rate_select, lane_count,
 				      drm_dp_enhanced_frame_cap(intel_dp->dpcd), false);
 
Signature · 1 line
2.45.2
#27 lore raw

Re: [PATCH 19/23] drm/i915/dp: Set relevant Downspread Ctrl DPCD bits for PR + Auxless ALPM

Quoted 63 lines · Ankit Nautiyal

If a Panel Replay capable sink, supports Async Video timing in PR active state, then source does not necessarily need to send AS SDPs during PR active.

However, if asynchronous video timing is not supported, then for PR with Aux-less ALPM, the source must transmit Adaptive-Sync SDPs for video timing synchronization while PR is active.

If the source needs to send AS SDP during PR active, this requires setting DPCD 0x0107[6] (FIXED_VTOTAL_AS_SDP_EN_IN_PR_ACTIVE). This applies whether VRR is enabled (AVT/FAVT) or fixed-timing mode is used.

This bit defines AS SDP timing behavior during PR Active, even if AS SDPs are briefly suspended.

Program the relevant Downspread Ctrl DPCD bits accordingly.

v2: Instead of Panel Replay check simply use AS SDP enable check. (Ville)

---
 .../gpu/drm/i915/display/intel_dp_link_training.c    | 12 ++++++++++--
 .../gpu/drm/i915/display/intel_dp_link_training.h    |  3 ++-
 drivers/gpu/drm/i915/display/intel_dp_mst.c          |  2 +-
 3 files changed, 13 insertions(+), 4 deletions(-)
drivers/gpu/drm/i915/display/intel_dp_link_training.c +6 -1
index a26094223f78..8b21c479ebfc 100644
@@ -27,6 +27,7 @@
 #include <drm/display/drm_dp_helper.h>
 #include <drm/drm_print.h>
 
+#include "intel_alpm.h"
 #include "intel_display_core.h"
 #include "intel_display_jiffies.h"
 #include "intel_display_types.h"
@@ -34,6 +35,7 @@
 #include "intel_dp.h"
 #include "intel_dp_link_training.h"
 #include "intel_encoder.h"
+#include "intel_hdmi.h"
 #include "intel_hotplug.h"
 #include "intel_panel.h"
 
@@ -710,11 +712,14 @@ static bool intel_dp_link_max_vswing_reached(struct intel_dp *intel_dp,
 	return true;
 }
 
-void intel_dp_link_training_set_mode(struct intel_dp *intel_dp, int link_rate, bool is_vrr)
+void intel_dp_link_training_set_mode(struct intel_dp *intel_dp, int link_rate,
+				     bool is_vrr,
+				     bool as_sdp_enable)
 {
 	u8 link_config[2];
 
 	link_config[0] = is_vrr ? DP_MSA_TIMING_PAR_IGNORE_EN : 0;
+	link_config[0] |= as_sdp_enable ? DP_FIXED_VTOTAL_AS_SDP_EN_IN_PR_ACTIVE : 0;
 	link_config[1] = drm_dp_is_uhbr_rate(link_rate) ?
 			 DP_SET_ANSI_128B132B : DP_SET_ANSI_8B10B;
 	drm_dp_dpcd_write(&intel_dp->aux, DP_DOWNSPREAD_CTRL, link_config, 2);

drivers/gpu/drm/i915/display/intel_dp_link_training.c static void intel_dp_update_downspread_ctrl(struct intel_dp *intel_dp, · line 748

Quoted 8 lines
drivers/gpu/drm/i915/display/intel_dp_link_training.c +4 -1
@@ -737,7 +742,10 @@ static void intel_dp_update_downspread_ctrl(struct intel_dp *intel_dp,
 	  * especially on the first real commit when clearing the inherited flag.
 	  */
 	intel_dp_link_training_set_mode(intel_dp,
-					crtc_state->port_clock, crtc_state->vrr.in_range);
+					crtc_state->port_clock,
+					crtc_state->vrr.in_range,
+					crtc_state->infoframes.enable &
+					intel_hdmi_infoframe_enable(DP_SDP_ADAPTIVE_SYNC));

The bit is only documented to be valid for panel replay capable sinks. So we should probably not set it otherwise.

But the weird thing is that the AS SDP DB0[1:0] description does not provide for a set of valid values for that case (DPCD 107h[7:6]=10b). Either they assumed that AS SDP v1 is used in that case (in which case the we'd not be able to support FAVT and maybe some other stuff without PR), or it's just an oversight and the same values apply as for 107h[7:6]=10b as they do for 107h[7:6]=11b.

Quoted 32 lines
drivers/gpu/drm/i915/display/intel_dp_link_training.c +0 -0
 }
 
 void intel_dp_link_training_set_bw(struct intel_dp *intel_dp,
drivers/gpu/drm/i915/display/intel_dp_link_training.h +2 -1
index 33dcbde6a408..d3ae8ee38a75 100644
@@ -18,7 +18,8 @@ int intel_dp_init_lttpr_and_dprx_caps(struct intel_dp *intel_dp);
 bool intel_dp_lttpr_transparent_mode_enabled(struct intel_dp *intel_dp);
 
 void intel_dp_link_training_set_mode(struct intel_dp *intel_dp,
-				     int link_rate, bool is_vrr);
+				     int link_rate, bool is_vrr,
+				     bool as_sdp_enable);
 void intel_dp_link_training_set_bw(struct intel_dp *intel_dp,
 				   int link_bw, int rate_select, int lane_count,
 				   bool enhanced_framing, bool post_lt_adj_req);
drivers/gpu/drm/i915/display/intel_dp_mst.c +1 -1
index e8de17834dcd..ffd1cf0aad9a 100644
@@ -2142,7 +2142,7 @@ void intel_dp_mst_prepare_probe(struct intel_dp *intel_dp)
 
 	intel_dp_compute_rate(intel_dp, link_rate, &link_bw, &rate_select);
 
-	intel_dp_link_training_set_mode(intel_dp, link_rate, false);
+	intel_dp_link_training_set_mode(intel_dp, link_rate, false, false);
 	intel_dp_link_training_set_bw(intel_dp, link_bw, rate_select, lane_count,
 				      drm_dp_enhanced_frame_cap(intel_dp->dpcd), false);
 
Signature · 1 line
2.45.2
Signature · 2 lines
Ville Syrjälä
Intel
#28 lore raw

Re: [PATCH 19/23] drm/i915/dp: Set relevant Downspread Ctrl DPCD bits for PR + Auxless ALPM

Quoted 75 lines · Ville Syrjälä
Quoted 63 lines · Ankit Nautiyal

If a Panel Replay capable sink, supports Async Video timing in PR active state, then source does not necessarily need to send AS SDPs during PR active.

However, if asynchronous video timing is not supported, then for PR with Aux-less ALPM, the source must transmit Adaptive-Sync SDPs for video timing synchronization while PR is active.

If the source needs to send AS SDP during PR active, this requires setting DPCD 0x0107[6] (FIXED_VTOTAL_AS_SDP_EN_IN_PR_ACTIVE). This applies whether VRR is enabled (AVT/FAVT) or fixed-timing mode is used.

This bit defines AS SDP timing behavior during PR Active, even if AS SDPs are briefly suspended.

Program the relevant Downspread Ctrl DPCD bits accordingly.

v2: Instead of Panel Replay check simply use AS SDP enable check. (Ville)

---
  .../gpu/drm/i915/display/intel_dp_link_training.c    | 12 ++++++++++--
  .../gpu/drm/i915/display/intel_dp_link_training.h    |  3 ++-
  drivers/gpu/drm/i915/display/intel_dp_mst.c          |  2 +-
  3 files changed, 13 insertions(+), 4 deletions(-)
drivers/gpu/drm/i915/display/intel_dp_link_training.c +6 -1
index a26094223f78..8b21c479ebfc 100644
@@ -27,6 +27,7 @@
  #include <drm/display/drm_dp_helper.h>
  #include <drm/drm_print.h>
 
+#include "intel_alpm.h"
  #include "intel_display_core.h"
  #include "intel_display_jiffies.h"
  #include "intel_display_types.h"
@@ -34,6 +35,7 @@
  #include "intel_dp.h"
  #include "intel_dp_link_training.h"
  #include "intel_encoder.h"
+#include "intel_hdmi.h"
  #include "intel_hotplug.h"
  #include "intel_panel.h"
 
@@ -710,11 +712,14 @@ static bool intel_dp_link_max_vswing_reached(struct intel_dp *intel_dp,
  	return true;
  }
 
-void intel_dp_link_training_set_mode(struct intel_dp *intel_dp, int link_rate, bool is_vrr)
+void intel_dp_link_training_set_mode(struct intel_dp *intel_dp, int link_rate,
+				     bool is_vrr,
+				     bool as_sdp_enable)
  {
  	u8 link_config[2];
 
  	link_config[0] = is_vrr ? DP_MSA_TIMING_PAR_IGNORE_EN : 0;
+	link_config[0] |= as_sdp_enable ? DP_FIXED_VTOTAL_AS_SDP_EN_IN_PR_ACTIVE : 0;
  	link_config[1] = drm_dp_is_uhbr_rate(link_rate) ?
  			 DP_SET_ANSI_128B132B : DP_SET_ANSI_8B10B;
  	drm_dp_dpcd_write(&intel_dp->aux, DP_DOWNSPREAD_CTRL, link_config, 2);

drivers/gpu/drm/i915/display/intel_dp_link_training.c static void intel_dp_update_downspread_ctrl(struct intel_dp *intel_dp, · line 748

Quoted 8 lines
drivers/gpu/drm/i915/display/intel_dp_link_training.c +4 -1
@@ -737,7 +742,10 @@ static void intel_dp_update_downspread_ctrl(struct intel_dp *intel_dp,
  	  * especially on the first real commit when clearing the inherited flag.
  	  */
  	intel_dp_link_training_set_mode(intel_dp,
-					crtc_state->port_clock, crtc_state->vrr.in_range);
+					crtc_state->port_clock,
+					crtc_state->vrr.in_range,
+					crtc_state->infoframes.enable &
+					intel_hdmi_infoframe_enable(DP_SDP_ADAPTIVE_SYNC));

The bit is only documented to be valid for panel replay capable sinks. So we should probably not set it otherwise.

Quoted 6 lines

But the weird thing is that the AS SDP DB0[1:0] description does not provide for a set of valid values for that case (DPCD 107h[7:6]=10b). Either they assumed that AS SDP v1 is used in that case (in which case the we'd not be able to support FAVT and maybe some other stuff without PR), or it's just an oversight and the same values apply as for 107h[7:6]=10b as they do for 107h[7:6]=11b.

Yeah you are right, the spec doesnt say any specific thing for AS SDP DB0[1:0] when DPCD 107h[7:6]=10b.

It says for values 11b and 01b. As you said, I agree we should have a Panel Replay check also.

I'll just use:

intel_alpm_is_alpm_aux_less() &&
crtc_state->infoframes.enable & intel_hdmi_infoframe_enable(DP_SDP_ADAPTIVE_SYNC))

Regards,

Ankit

Quoted 33 lines
Quoted 33 lines
drivers/gpu/drm/i915/display/intel_dp_link_training.c +0 -0
  }
 
  void intel_dp_link_training_set_bw(struct intel_dp *intel_dp,
drivers/gpu/drm/i915/display/intel_dp_link_training.h +2 -1
index 33dcbde6a408..d3ae8ee38a75 100644
@@ -18,7 +18,8 @@ int intel_dp_init_lttpr_and_dprx_caps(struct intel_dp *intel_dp);
  bool intel_dp_lttpr_transparent_mode_enabled(struct intel_dp *intel_dp);
 
  void intel_dp_link_training_set_mode(struct intel_dp *intel_dp,
-				     int link_rate, bool is_vrr);
+				     int link_rate, bool is_vrr,
+				     bool as_sdp_enable);
  void intel_dp_link_training_set_bw(struct intel_dp *intel_dp,
  				   int link_bw, int rate_select, int lane_count,
  				   bool enhanced_framing, bool post_lt_adj_req);
drivers/gpu/drm/i915/display/intel_dp_mst.c +1 -1
index e8de17834dcd..ffd1cf0aad9a 100644
@@ -2142,7 +2142,7 @@ void intel_dp_mst_prepare_probe(struct intel_dp *intel_dp)
 
  	intel_dp_compute_rate(intel_dp, link_rate, &link_bw, &rate_select);
 
-	intel_dp_link_training_set_mode(intel_dp, link_rate, false);
+	intel_dp_link_training_set_mode(intel_dp, link_rate, false, false);
  	intel_dp_link_training_set_bw(intel_dp, link_bw, rate_select, lane_count,
  				      drm_dp_enhanced_frame_cap(intel_dp->dpcd), false);
 
Signature · 1 line
2.45.2
#29 lore raw

Re: [PATCH 19/23] drm/i915/dp: Set relevant Downspread Ctrl DPCD bits for PR + Auxless ALPM

Quoted 85 lines · Ankit K
Quoted 75 lines · Ville Syrjälä
Quoted 62 lines · Ankit Nautiyal

If a Panel Replay capable sink, supports Async Video timing in PR active state, then source does not necessarily need to send AS SDPs during PR active.

However, if asynchronous video timing is not supported, then for PR with Aux-less ALPM, the source must transmit Adaptive-Sync SDPs for video timing synchronization while PR is active.

If the source needs to send AS SDP during PR active, this requires setting DPCD 0x0107[6] (FIXED_VTOTAL_AS_SDP_EN_IN_PR_ACTIVE). This applies whether VRR is enabled (AVT/FAVT) or fixed-timing mode is used.

This bit defines AS SDP timing behavior during PR Active, even if AS SDPs are briefly suspended.

Program the relevant Downspread Ctrl DPCD bits accordingly.

v2: Instead of Panel Replay check simply use AS SDP enable check. (Ville)

---
  .../gpu/drm/i915/display/intel_dp_link_training.c    | 12
++++++++++--
  .../gpu/drm/i915/display/intel_dp_link_training.h    |  3 ++- drivers/gpu/drm/i915/display/intel_dp_mst.c          |  2 +-

  3 files changed, 13 insertions(+), 4 deletions(-)
drivers/gpu/drm/i915/display/intel_dp_link_training.c +0 -0
index a26094223f78..8b21c479ebfc 100644
@@ -27,6 +27,7 @@
   #include <drm/display/drm_dp_helper.h>
   #include <drm/drm_print.h>

  +#include "intel_alpm.h"
  #include "intel_display_core.h"
  #include "intel_display_jiffies.h"
  #include "intel_display_types.h"

drivers/gpu/drm/i915/display/intel_dp_link_training.c +1 -0
@@ -34,6 +35,7 @@
   #include "intel_dp.h"
   #include "intel_dp_link_training.h"
   #include "intel_encoder.h"
+#include "intel_hdmi.h"
   #include "intel_hotplug.h"
   #include "intel_panel.h"
   @@ -710,11 +712,14 @@ static bool intel_dp_link_max_vswing_reached(struct intel_dp *intel_dp,

      return true;
  }
  -void intel_dp_link_training_set_mode(struct intel_dp *intel_dp, int link_rate, bool is_vrr) +void intel_dp_link_training_set_mode(struct intel_dp *intel_dp, int
link_rate,
+                     bool is_vrr,
+                     bool as_sdp_enable)
  {
      u8 link_config[2];
        link_config[0] = is_vrr ? DP_MSA_TIMING_PAR_IGNORE_EN : 0;
+    link_config[0] |= as_sdp_enable ?
DP_FIXED_VTOTAL_AS_SDP_EN_IN_PR_ACTIVE : 0;
      link_config[1] = drm_dp_is_uhbr_rate(link_rate) ?
               DP_SET_ANSI_128B132B : DP_SET_ANSI_8B10B; drm_dp_dpcd_write(&intel_dp->aux, DP_DOWNSPREAD_CTRL, link_config, 2);

drivers/gpu/drm/i915/display/intel_dp_link_training.c static void · line 749

Quoted 9 lines
drivers/gpu/drm/i915/display/intel_dp_link_training.c +4 -1
@@ -737,7 +742,10 @@ static void 
 intel_dp_update_downspread_ctrl(struct intel_dp *intel_dp,
         * especially on the first real commit when clearing the inherited flag.
         */
       intel_dp_link_training_set_mode(intel_dp,
-                    crtc_state->port_clock, crtc_state->vrr.in_range);
+                    crtc_state->port_clock,
+                    crtc_state->vrr.in_range,
+                    crtc_state->infoframes.enable &
+ intel_hdmi_infoframe_enable(DP_SDP_ADAPTIVE_SYNC));

The bit is only documented to be valid for panel replay capable sinks. So we should probably not set it otherwise.

Quoted 6 lines

But the weird thing is that the AS SDP DB0[1:0] description does not provide for a set of valid values for that case (DPCD 107h[7:6]=10b). Either they assumed that AS SDP v1 is used in that case (in which case the we'd not be able to support FAVT and maybe some other stuff without PR), or it's just an oversight and the same values apply as for 107h[7:6]=10b as they do for 107h[7:6]=11b.

Quoted 7 lines

Yeah you are right, the spec doesnt say any specific thing for AS SDP DB0[1:0] when DPCD 107h[7:6]=10b.

It says for values 11b and 01b. As you said, I agree we should have a Panel Replay check also.

I'll just use:

intel_alpm_is_alpm_aux_less() &&

I think I need to use intel_psr_needs_alpm_aux_less()  instead.

Quoted 38 lines

crtc_state->infoframes.enable & intel_hdmi_infoframe_enable(DP_SDP_ADAPTIVE_SYNC))

Regards,

Ankit

Quoted 31 lines
Quoted 31 lines
drivers/gpu/drm/i915/display/intel_dp_link_training.c +0 -0
   }
     void intel_dp_link_training_set_bw(struct intel_dp *intel_dp,
drivers/gpu/drm/i915/display/intel_dp_link_training.h +2 -1
index 33dcbde6a408..d3ae8ee38a75 100644
@@ -18,7 +18,8 @@ int intel_dp_init_lttpr_and_dprx_caps(struct 
 intel_dp *intel_dp);
   bool intel_dp_lttpr_transparent_mode_enabled(struct intel_dp *intel_dp);
     void intel_dp_link_training_set_mode(struct intel_dp *intel_dp,
-                     int link_rate, bool is_vrr);
+                     int link_rate, bool is_vrr,
+                     bool as_sdp_enable);
   void intel_dp_link_training_set_bw(struct intel_dp *intel_dp,
                      int link_bw, int rate_select, int lane_count,
                      bool enhanced_framing, bool post_lt_adj_req);
drivers/gpu/drm/i915/display/intel_dp_mst.c +1 -0
index e8de17834dcd..ffd1cf0aad9a 100644
@@ -2142,7 +2142,7 @@ void intel_dp_mst_prepare_probe(struct 
 intel_dp *intel_dp)
         intel_dp_compute_rate(intel_dp, link_rate, &link_bw, &rate_select);
   -    intel_dp_link_training_set_mode(intel_dp, link_rate, false);
+    intel_dp_link_training_set_mode(intel_dp, link_rate, false, 
 false);
       intel_dp_link_training_set_bw(intel_dp, link_bw, rate_select, lane_count,
 drm_dp_enhanced_frame_cap(intel_dp->dpcd), false);

  --
2.45.2

#30 lore raw

PATCH 20/23 drm/i915/dp: Program AS SDP DB[1:0] for PR with Link off

In reply to Ankit Nautiyal

For Panel Replay with AUX-less ALPM (link-off PR), the source must send Adaptive-Sync SDP v2. Program DB[1:0] per DP spec v2.1:
- VRR AVT: 00b (variable VTotal)
- VRR FAVT: 10b/11b (TRR not reached/reached)
- Fixed timing with PR link-off (VRR off): 01b (AS disabled; VTotal fixed)

Also, drop the redundant target_rr assignment.

v2: Fix the else case. (Ville)

---
 drivers/gpu/drm/i915/display/intel_dp.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/display/intel_dp.c +3 -2
index 87388694ce17..ddaa187371b8 100644
@@ -3166,9 +3166,10 @@ static void intel_dp_compute_as_sdp(struct intel_dp *intel_dp,
 		as_sdp->mode = DP_AS_SDP_FAVT_TRR_REACHED;
 		as_sdp->target_rr = drm_mode_vrefresh(adjusted_mode);
 		as_sdp->target_rr_divider = true;
-	} else {
+	} else if (crtc_state->vrr.enable) {
 		as_sdp->mode = DP_AS_SDP_AVT_DYNAMIC_VTOTAL;
-		as_sdp->target_rr = 0;
+	} else {
+		as_sdp->mode = DP_AS_SDP_AVT_FIXED_VTOTAL;
 	}
 }
 
Signature · 1 line
2.45.2
#31 lore raw

PATCH 21/23 drm/i915/dp: Split AS SDP computation between compute_config and compute_config_late

In reply to Ankit Nautiyal

Currently we enable AS SDP only when VRR is enabled. As we start using AS SDP for other features, this becomes a problem. The AS SDP configuration can change dynamically based on VRR, CMRR, PR, ALPM, etc. Since these features may be enabled or disabled after the initial configuration, the AS SDP parameters need to be computed later in the pipeline.

However, not all of the AS SDP logic can be moved to the late stage:
the VRR guardband optimization depends on knowing early whether AS SDP can be used. Without this, we would end up accounting for AS SDP on all platforms that support it, even for panels that do not support AS SDP. Therefore we set the infoframe enable bit for AS SDP during compute_config(), before the guardband is computed.

To handle these constraints, split the AS SDP programming into two phases:

- intel_dp_compute_as_sdp()
Runs during compute_config().
Sets only the infoframe enable bit so that the guardband logic can account for AS SDP requirements.

- intel_dp_as_sdp_compute_config_late()
Runs during compute_config_late().
Computes all remaining AS SDP fields based on the features that need it.

The late-stage computation is called from intel_dp_sdp_compute_config_late(), before computing the minimum guardband for SDPs.

This is a preparatory change. A subsequent patches will compute PR related AS SDP fields and enable AS SDP not only for VRR but for other features as well.

---
 drivers/gpu/drm/i915/display/intel_dp.c | 66 ++++++++++++++++---------
 1 file changed, 43 insertions(+), 23 deletions(-)
drivers/gpu/drm/i915/display/intel_dp.c +43 -23
index ddaa187371b8..902c09e0780f 100644
@@ -3142,10 +3142,6 @@ static bool intel_dp_needs_as_sdp(struct intel_dp *intel_dp,
 static void intel_dp_compute_as_sdp(struct intel_dp *intel_dp,
 				    struct intel_crtc_state *crtc_state)
 {
-	struct drm_dp_as_sdp *as_sdp = &crtc_state->infoframes.as_sdp;
-	const struct drm_display_mode *adjusted_mode =
-		&crtc_state->hw.adjusted_mode;
-
 	/*
 	 * #FIXME: SDP/infoframe updates aren’t truly atomic, and with the new
 	 * cdclk->tc clock crossing we may transiently send a corrupted packet
@@ -3154,23 +3150,13 @@ static void intel_dp_compute_as_sdp(struct intel_dp *intel_dp,
 	if (!intel_dp_needs_as_sdp(intel_dp, crtc_state))
 		return;
 
+	/*
+	 * Only set the infoframes.enable flag here. The remaining AS SDP fields
+	 * are programmed in the compute_config_late() phase. We need this flag
+	 * early so that the VRR guardband calculation can properly account for
+	 * AS SDP requirements.
+	 */
 	crtc_state->infoframes.enable |= intel_hdmi_infoframe_enable(DP_SDP_ADAPTIVE_SYNC);
-
-	as_sdp->sdp_type = DP_SDP_ADAPTIVE_SYNC;
-	as_sdp->length = 0x9;
-	as_sdp->duration_incr_ms = 0;
-	as_sdp->revision = 0x2;
-	as_sdp->vtotal = intel_vrr_vmin_vtotal(crtc_state);
-
-	if (crtc_state->cmrr.enable) {
-		as_sdp->mode = DP_AS_SDP_FAVT_TRR_REACHED;
-		as_sdp->target_rr = drm_mode_vrefresh(adjusted_mode);
-		as_sdp->target_rr_divider = true;
-	} else if (crtc_state->vrr.enable) {
-		as_sdp->mode = DP_AS_SDP_AVT_DYNAMIC_VTOTAL;
-	} else {
-		as_sdp->mode = DP_AS_SDP_AVT_FIXED_VTOTAL;
-	}
 }
 
 static void intel_dp_compute_vsc_sdp(struct intel_dp *intel_dp,
@@ -7371,11 +7357,45 @@ void intel_dp_mst_resume(struct intel_display *display)
 }
 
 static
-int intel_dp_sdp_compute_config_late(struct intel_crtc_state *crtc_state)
+void intel_dp_as_sdp_compute_config_late(struct intel_dp *intel_dp,
+					 struct intel_crtc_state *crtc_state)
+{
+	struct drm_dp_as_sdp *as_sdp = &crtc_state->infoframes.as_sdp;
+	const struct drm_display_mode *adjusted_mode =
+		&crtc_state->hw.adjusted_mode;
+
+	if ((crtc_state->infoframes.enable &
+	    intel_hdmi_infoframe_enable(DP_SDP_ADAPTIVE_SYNC)) == 0)
+		return;
+
+	as_sdp->sdp_type = DP_SDP_ADAPTIVE_SYNC;
+	as_sdp->length = 0x9;
+	as_sdp->duration_incr_ms = 0;
+	as_sdp->revision = 0x2;
+	as_sdp->vtotal = intel_vrr_vmin_vtotal(crtc_state);
+
+	if (crtc_state->cmrr.enable) {
+		as_sdp->mode = DP_AS_SDP_FAVT_TRR_REACHED;
+		as_sdp->target_rr = drm_mode_vrefresh(adjusted_mode);
+		as_sdp->target_rr_divider = true;
+	} else if (crtc_state->vrr.enable) {
+		as_sdp->mode = DP_AS_SDP_AVT_DYNAMIC_VTOTAL;
+	} else {
+		as_sdp->mode = DP_AS_SDP_AVT_FIXED_VTOTAL;
+	}
+}
+
+static
+int intel_dp_sdp_compute_config_late(struct intel_dp *intel_dp,
+				     struct intel_crtc_state *crtc_state)
 {
 	struct intel_display *display = to_intel_display(crtc_state);
 	int guardband = intel_crtc_vblank_length(crtc_state);
-	int min_sdp_guardband = intel_dp_sdp_min_guardband(crtc_state, false);
+	int min_sdp_guardband;
+
+	intel_dp_as_sdp_compute_config_late(intel_dp, crtc_state);
+
+	min_sdp_guardband = intel_dp_sdp_min_guardband(crtc_state, false);
 
 	if (guardband < min_sdp_guardband) {
 		drm_dbg_kms(display->drm, "guardband %d < min sdp guardband %d\n",
@@ -7395,7 +7415,7 @@ int intel_dp_compute_config_late(struct intel_encoder *encoder,
 
 	intel_psr_compute_config_late(intel_dp, crtc_state);
 
-	ret = intel_dp_sdp_compute_config_late(crtc_state);
+	ret = intel_dp_sdp_compute_config_late(intel_dp, crtc_state);
 	if (ret)
 		return ret;
 
Signature · 1 line
2.45.2
#32 lore raw

PATCH 22/23 drm/i915/dp: Compute and include coasting vtotal for AS SDP

In reply to Ankit Nautiyal

DP v2.1 allows the source to temporarily suspend Adaptive-Sync SDP transmission while Panel Replay is active when the sink supports asynchronous video timing.

In such cases, the sink relies on the last transmitted AS SDP timing information to maintain the refresh rate. To support this behavior, compute and populate the coasting vtotal field in the AS SDP payload.

Include coasting vtotal in AS SDP packing, unpacking, and comparison, and set it during late AS SDP configuration for PR with Aux-less ALPM when asynchronous video timing is supported.

---
 drivers/gpu/drm/i915/display/intel_display.c |  3 ++-
 drivers/gpu/drm/i915/display/intel_dp.c      | 19 +++++++++++++++++++
 2 files changed, 21 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/display/intel_display.c +2 -1
index a0e7ef2574b2..747dd3112d66 100644
@@ -4914,7 +4914,8 @@ intel_compare_dp_as_sdp(const struct drm_dp_as_sdp *a,
 		a->duration_incr_ms == b->duration_incr_ms &&
 		a->duration_decr_ms == b->duration_decr_ms &&
 		a->target_rr_divider == b->target_rr_divider &&
-		a->mode == b->mode;
+		a->mode == b->mode &&
+		a->coasting_vtotal == b->coasting_vtotal;
 }
 
 static bool
drivers/gpu/drm/i915/display/intel_dp.c +19 -0
index 902c09e0780f..de6f88a5400d 100644
@@ -5123,6 +5123,9 @@ static ssize_t intel_dp_as_sdp_pack(const struct drm_dp_as_sdp *as_sdp,
 	if (as_sdp->target_rr_divider)
 		sdp->db[4] |= 0x20;
 
+	sdp->db[7] = as_sdp->coasting_vtotal & 0xFF;
+	sdp->db[8] = (as_sdp->coasting_vtotal >> 8) & 0xFF;
+
 	return length;
 }
 
@@ -5306,6 +5309,7 @@ int intel_dp_as_sdp_unpack(struct drm_dp_as_sdp *as_sdp,
 	as_sdp->vtotal = (sdp->db[2] << 8) | sdp->db[1];
 	as_sdp->target_rr = ((sdp->db[4] & 0x3) << 8) | sdp->db[3];
 	as_sdp->target_rr_divider = sdp->db[4] & 0x20 ? true : false;
+	as_sdp->coasting_vtotal = (sdp->db[8] << 8) | sdp->db[7];
 
 	return 0;
 }
@@ -7383,6 +7387,21 @@ void intel_dp_as_sdp_compute_config_late(struct intel_dp *intel_dp,
 	} else {
 		as_sdp->mode = DP_AS_SDP_AVT_FIXED_VTOTAL;
 	}
+
+	/*
+	 * For Panel Replay with Async Video Timing support, the source can
+	 * disable sending the AS SDP during PR Active state. In that case,
+	 * the sink needs the coasting vtotal value to maintain the refresh
+	 * rate.
+	 *
+	 * #TODO:
+	 * If we ever advertise support for coasting at other refresh targets,
+	 * this logic could be revisited. For now, use the minimum refresh rate
+	 * as the only safe coasting value.
+	 */
+	if (intel_alpm_is_alpm_aux_less(intel_dp, crtc_state) &&
+	    intel_psr_pr_async_video_timing_supported(intel_dp))
+		as_sdp->coasting_vtotal = crtc_state->vrr.vmax;
 }
 
 static
Signature · 1 line
2.45.2
#33 lore raw

Re: [PATCH 22/23] drm/i915/dp: Compute and include coasting vtotal for AS SDP

Quoted 63 lines · Ankit Nautiyal

DP v2.1 allows the source to temporarily suspend Adaptive-Sync SDP transmission while Panel Replay is active when the sink supports asynchronous video timing.

In such cases, the sink relies on the last transmitted AS SDP timing information to maintain the refresh rate. To support this behavior, compute and populate the coasting vtotal field in the AS SDP payload.

Include coasting vtotal in AS SDP packing, unpacking, and comparison, and set it during late AS SDP configuration for PR with Aux-less ALPM when asynchronous video timing is supported.

---
 drivers/gpu/drm/i915/display/intel_display.c |  3 ++-
 drivers/gpu/drm/i915/display/intel_dp.c      | 19 +++++++++++++++++++
 2 files changed, 21 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/display/intel_display.c +2 -1
index a0e7ef2574b2..747dd3112d66 100644
@@ -4914,7 +4914,8 @@ intel_compare_dp_as_sdp(const struct drm_dp_as_sdp *a,
 		a->duration_incr_ms == b->duration_incr_ms &&
 		a->duration_decr_ms == b->duration_decr_ms &&
 		a->target_rr_divider == b->target_rr_divider &&
-		a->mode == b->mode;
+		a->mode == b->mode &&
+		a->coasting_vtotal == b->coasting_vtotal;
 }
 
 static bool
drivers/gpu/drm/i915/display/intel_dp.c +9 -0
index 902c09e0780f..de6f88a5400d 100644
@@ -5123,6 +5123,9 @@ static ssize_t intel_dp_as_sdp_pack(const struct drm_dp_as_sdp *as_sdp,
 	if (as_sdp->target_rr_divider)
 		sdp->db[4] |= 0x20;
 
+	sdp->db[7] = as_sdp->coasting_vtotal & 0xFF;
+	sdp->db[8] = (as_sdp->coasting_vtotal >> 8) & 0xFF;
+
 	return length;
 }
 
@@ -5306,6 +5309,7 @@ int intel_dp_as_sdp_unpack(struct drm_dp_as_sdp *as_sdp,
 	as_sdp->vtotal = (sdp->db[2] << 8) | sdp->db[1];
 	as_sdp->target_rr = ((sdp->db[4] & 0x3) << 8) | sdp->db[3];
 	as_sdp->target_rr_divider = sdp->db[4] & 0x20 ? true : false;
+	as_sdp->coasting_vtotal = (sdp->db[8] << 8) | sdp->db[7];
 
 	return 0;
 }
@@ -7383,6 +7387,21 @@ void intel_dp_as_sdp_compute_config_late(struct intel_dp *intel_dp,
 	} else {
 		as_sdp->mode = DP_AS_SDP_AVT_FIXED_VTOTAL;
 	}
+
+	/*
+	 * For Panel Replay with Async Video Timing support, the source can
+	 * disable sending the AS SDP during PR Active state. In that case,
+	 * the sink needs the coasting vtotal value to maintain the refresh

drivers/gpu/drm/i915/display/intel_dp.c void intel_dp_as_sdp_compute_config_late(struct intel_dp *intel_dp, · line 7404

Quoted 10 lines
drivers/gpu/drm/i915/display/intel_dp.c +10 -0
+	 * rate.
+	 *
+	 * #TODO:
+	 * If we ever advertise support for coasting at other refresh targets,
+	 * this logic could be revisited. For now, use the minimum refresh rate
+	 * as the only safe coasting value.
+	 */
+	if (intel_alpm_is_alpm_aux_less(intel_dp, crtc_state) &&
+	    intel_psr_pr_async_video_timing_supported(intel_dp))
+		as_sdp->coasting_vtotal = crtc_state->vrr.vmax;

Seems reasonable.

Is this always under our control or could the hardware overwrite this with the current vtotal at the time of PR entry? Assuming we can enter PR before the vtotal goes back to vmax on its own anyway.

Quoted 5 lines
drivers/gpu/drm/i915/display/intel_dp.c +0 -0
 }
 
 static
Signature · 1 line
2.45.2
Signature · 2 lines
Ville Syrjälä
Intel
#34 lore raw

Re: [PATCH 22/23] drm/i915/dp: Compute and include coasting vtotal for AS SDP

Quoted 76 lines · Ville Syrjälä
Quoted 63 lines · Ankit Nautiyal

DP v2.1 allows the source to temporarily suspend Adaptive-Sync SDP transmission while Panel Replay is active when the sink supports asynchronous video timing.

In such cases, the sink relies on the last transmitted AS SDP timing information to maintain the refresh rate. To support this behavior, compute and populate the coasting vtotal field in the AS SDP payload.

Include coasting vtotal in AS SDP packing, unpacking, and comparison, and set it during late AS SDP configuration for PR with Aux-less ALPM when asynchronous video timing is supported.

---
  drivers/gpu/drm/i915/display/intel_display.c |  3 ++-
  drivers/gpu/drm/i915/display/intel_dp.c      | 19 +++++++++++++++++++
  2 files changed, 21 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/display/intel_display.c +2 -1
index a0e7ef2574b2..747dd3112d66 100644
@@ -4914,7 +4914,8 @@ intel_compare_dp_as_sdp(const struct drm_dp_as_sdp *a,
  		a->duration_incr_ms == b->duration_incr_ms &&
  		a->duration_decr_ms == b->duration_decr_ms &&
  		a->target_rr_divider == b->target_rr_divider &&
-		a->mode == b->mode;
+		a->mode == b->mode &&
+		a->coasting_vtotal == b->coasting_vtotal;
  }
 
  static bool
drivers/gpu/drm/i915/display/intel_dp.c +9 -0
index 902c09e0780f..de6f88a5400d 100644
@@ -5123,6 +5123,9 @@ static ssize_t intel_dp_as_sdp_pack(const struct drm_dp_as_sdp *as_sdp,
  	if (as_sdp->target_rr_divider)
  		sdp->db[4] |= 0x20;
 
+	sdp->db[7] = as_sdp->coasting_vtotal & 0xFF;
+	sdp->db[8] = (as_sdp->coasting_vtotal >> 8) & 0xFF;
+
  	return length;
  }
 
@@ -5306,6 +5309,7 @@ int intel_dp_as_sdp_unpack(struct drm_dp_as_sdp *as_sdp,
  	as_sdp->vtotal = (sdp->db[2] << 8) | sdp->db[1];
  	as_sdp->target_rr = ((sdp->db[4] & 0x3) << 8) | sdp->db[3];
  	as_sdp->target_rr_divider = sdp->db[4] & 0x20 ? true : false;
+	as_sdp->coasting_vtotal = (sdp->db[8] << 8) | sdp->db[7];
 
  	return 0;
  }
@@ -7383,6 +7387,21 @@ void intel_dp_as_sdp_compute_config_late(struct intel_dp *intel_dp,
  	} else {
  		as_sdp->mode = DP_AS_SDP_AVT_FIXED_VTOTAL;
  	}
+
+	/*
+	 * For Panel Replay with Async Video Timing support, the source can
+	 * disable sending the AS SDP during PR Active state. In that case,
+	 * the sink needs the coasting vtotal value to maintain the refresh

drivers/gpu/drm/i915/display/intel_dp.c void intel_dp_as_sdp_compute_config_late(struct intel_dp *intel_dp, · line 7404

Quoted 10 lines
drivers/gpu/drm/i915/display/intel_dp.c +10 -0
+	 * rate.
+	 *
+	 * #TODO:
+	 * If we ever advertise support for coasting at other refresh targets,
+	 * this logic could be revisited. For now, use the minimum refresh rate
+	 * as the only safe coasting value.
+	 */
+	if (intel_alpm_is_alpm_aux_less(intel_dp, crtc_state) &&
+	    intel_psr_pr_async_video_timing_supported(intel_dp))
+		as_sdp->coasting_vtotal = crtc_state->vrr.vmax;

Seems reasonable.

Quoted 4 lines

Is this always under our control or could the hardware overwrite this with the current vtotal at the time of PR entry? Assuming we can enter PR before the vtotal goes back to vmax on its own anyway.

What I understand from Bspec in this regard is:

Do not write Adaptive Sync SDP Transmission Disable in PR Active State i.e. DB[2], instead use PR_ALPM_CTL[ AS SDP Transmission in Active Disable ] bit.

HW will sample the PR_ALPM_CTL bit only when PR is active, and it will get reflected in AS SDP payload in an 'appropriate' time.

HW will ignore this bit when PR is Inactive and always send AS SDP.

So I think HW will not touch the coasting vtotal DBs.

Driver should set appropriate coasting Vtotal and set the PR_ALPM_CTL[AS SDP Transmission in Active disable] bit.

During PR active Driver will set the DB[2] bit in payload which will trigger the sink to use Coasting Vtotal.

(Unless we start writing : DPCD 00B1[0] ie. PANEL_REPLAY_CONFIG2[PANEL_REPLAY_SINK_REFRESH_RATE_UNLOCK_GRANTED], in which sink starts using its own logic).

Regards,

Ankit

Quoted 6 lines
Quoted 6 lines
drivers/gpu/drm/i915/display/intel_dp.c +0 -0
  }
 
  static
Signature · 1 line
2.45.2
#35 lore raw

Re: [PATCH 22/23] drm/i915/dp: Compute and include coasting vtotal for AS SDP

Quoted 94 lines · Ankit K
Quoted 80 lines · Ville Syrjälä
Quoted 69 lines · Ankit Nautiyal

DP v2.1 allows the source to temporarily suspend Adaptive-Sync SDP transmission while Panel Replay is active when the sink supports asynchronous video timing.

In such cases, the sink relies on the last transmitted AS SDP timing information to maintain the refresh rate. To support this behavior, compute and populate the coasting vtotal field in the AS SDP payload.

Include coasting vtotal in AS SDP packing, unpacking, and comparison, and set it during late AS SDP configuration for PR with Aux-less ALPM when asynchronous video timing is supported.

---
  drivers/gpu/drm/i915/display/intel_display.c |  3 ++- drivers/gpu/drm/i915/display/intel_dp.c      | 19 +++++++++++++++++++

  2 files changed, 21 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/display/intel_display.c +0 -0
index a0e7ef2574b2..747dd3112d66 100644
@@ -4914,7 +4914,8 @@ intel_compare_dp_as_sdp(const struct 
 drm_dp_as_sdp *a,

          a->duration_incr_ms == b->duration_incr_ms && a->duration_decr_ms == b->duration_decr_ms && a->target_rr_divider == b->target_rr_divider && -        a->mode == b->mode;
+        a->mode == b->mode &&
+        a->coasting_vtotal == b->coasting_vtotal; }
    static bool

drivers/gpu/drm/i915/display/intel_dp.c +2 -0
index 902c09e0780f..de6f88a5400d 100644
@@ -5123,6 +5123,9 @@ static ssize_t intel_dp_as_sdp_pack(const 
 struct drm_dp_as_sdp *as_sdp,
       if (as_sdp->target_rr_divider)
           sdp->db[4] |= 0x20;
   +    sdp->db[7] = as_sdp->coasting_vtotal & 0xFF;
+    sdp->db[8] = (as_sdp->coasting_vtotal >> 8) & 0xFF;
+
       return length;
   }

  @@ -5306,6 +5309,7 @@ int intel_dp_as_sdp_unpack(struct drm_dp_as_sdp *as_sdp, as_sdp->vtotal = (sdp->db[2] << 8) | sdp->db[1]; as_sdp->target_rr = ((sdp->db[4] & 0x3) << 8) | sdp->db[3]; as_sdp->target_rr_divider = sdp->db[4] & 0x20 ? true : false; +    as_sdp->coasting_vtotal = (sdp->db[8] << 8) | sdp->db[7]; return 0;
  }

drivers/gpu/drm/i915/display/intel_dp.c +3 -0
@@ -7383,6 +7387,21 @@ void 
 intel_dp_as_sdp_compute_config_late(struct intel_dp *intel_dp,
       } else {
           as_sdp->mode = DP_AS_SDP_AVT_FIXED_VTOTAL;
       }
+
+    /*
+     * For Panel Replay with Async Video Timing support, the source 

can
+     * disable sending the AS SDP during PR Active state. In that case,
+     * the sink needs the coasting vtotal value to maintain the refresh
+     * rate.
+     *
+     * #TODO:
+     * If we ever advertise support for coasting at other refresh

Quoted 8 lines

targets,
+     * this logic could be revisited. For now, use the minimum refresh rate
+     * as the only safe coasting value.
+     */
+    if (intel_alpm_is_alpm_aux_less(intel_dp, crtc_state) && +        intel_psr_pr_async_video_timing_supported(intel_dp)) +        as_sdp->coasting_vtotal = crtc_state->vrr.vmax;

Seems reasonable.

Quoted 4 lines

Is this always under our control or could the hardware overwrite this with the current vtotal at the time of PR entry? Assuming we can enter PR before the vtotal goes back to vmax on its own anyway.

What I understand from Bspec in this regard is:

Do not write Adaptive Sync SDP Transmission Disable in PR Active State i.e. DB[2], instead use PR_ALPM_CTL[ AS SDP Transmission in Active Disable ] bit.

HW will sample the PR_ALPM_CTL bit only when PR is active, and it will get reflected in AS SDP payload in an 'appropriate' time.

Quoted 7 lines

HW will ignore this bit when PR is Inactive and always send AS SDP.

So I think HW will not touch the coasting vtotal DBs.

Driver should set appropriate coasting Vtotal and set the PR_ALPM_CTL[AS SDP Transmission in Active disable] bit.

During PR active Driver will set the DB[2] bit in payload which will trigger the sink to use Coasting Vtotal.

..HW will set the DB[2] bit ... (facepalm)

Quoted 15 lines

(Unless we start writing : DPCD 00B1[0] ie. PANEL_REPLAY_CONFIG2[PANEL_REPLAY_SINK_REFRESH_RATE_UNLOCK_GRANTED], in which sink starts using its own logic).

Regards,

Ankit

Quoted 5 lines
Quoted 5 lines

  }
    static

Signature · 1 line
2.45.2
Reviewed-by
#36 lore raw

Re: [PATCH 22/23] drm/i915/dp: Compute and include coasting vtotal for AS SDP

Quoted 106 lines · Ankit K
Quoted 104 lines · Ankit K
Quoted 83 lines · Ville Syrjälä
Quoted 72 lines · Ankit Nautiyal

DP v2.1 allows the source to temporarily suspend Adaptive-Sync SDP transmission while Panel Replay is active when the sink supports asynchronous video timing.

In such cases, the sink relies on the last transmitted AS SDP timing information to maintain the refresh rate. To support this behavior, compute and populate the coasting vtotal field in the AS SDP payload.

Include coasting vtotal in AS SDP packing, unpacking, and comparison, and set it during late AS SDP configuration for PR with Aux-less ALPM when asynchronous video timing is supported.

---
  drivers/gpu/drm/i915/display/intel_display.c |  3 ++- drivers/gpu/drm/i915/display/intel_dp.c      | 19 +++++++++++++++++++

  2 files changed, 21 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/display/intel_display.c +0 -0
index a0e7ef2574b2..747dd3112d66 100644
@@ -4914,7 +4914,8 @@ intel_compare_dp_as_sdp(const struct 
 drm_dp_as_sdp *a,

          a->duration_incr_ms == b->duration_incr_ms && a->duration_decr_ms == b->duration_decr_ms && a->target_rr_divider == b->target_rr_divider && -        a->mode == b->mode;
+        a->mode == b->mode &&
+        a->coasting_vtotal == b->coasting_vtotal; }
    static bool

drivers/gpu/drm/i915/display/intel_dp.c +2 -0
index 902c09e0780f..de6f88a5400d 100644
@@ -5123,6 +5123,9 @@ static ssize_t intel_dp_as_sdp_pack(const 
 struct drm_dp_as_sdp *as_sdp,
       if (as_sdp->target_rr_divider)
           sdp->db[4] |= 0x20;
   +    sdp->db[7] = as_sdp->coasting_vtotal & 0xFF;
+    sdp->db[8] = (as_sdp->coasting_vtotal >> 8) & 0xFF;
+
       return length;
   }

  @@ -5306,6 +5309,7 @@ int intel_dp_as_sdp_unpack(struct drm_dp_as_sdp *as_sdp,
      as_sdp->vtotal = (sdp->db[2] << 8) | sdp->db[1]; as_sdp->target_rr = ((sdp->db[4] & 0x3) << 8) | sdp->db[3]; as_sdp->target_rr_divider = sdp->db[4] & 0x20 ? true : false; +    as_sdp->coasting_vtotal = (sdp->db[8] << 8) | sdp->db[7]; return 0;
  }

drivers/gpu/drm/i915/display/intel_dp.c +3 -0
@@ -7383,6 +7387,21 @@ void 
 intel_dp_as_sdp_compute_config_late(struct intel_dp *intel_dp,
       } else {
           as_sdp->mode = DP_AS_SDP_AVT_FIXED_VTOTAL;
       }
+
+    /*
+     * For Panel Replay with Async Video Timing support, the source 

can
+     * disable sending the AS SDP during PR Active state. In that case,
+     * the sink needs the coasting vtotal value to maintain the refresh
+     * rate.
+     *
+     * #TODO:
+     * If we ever advertise support for coasting at other refresh

Quoted 8 lines

targets,
+     * this logic could be revisited. For now, use the minimum refresh rate
+     * as the only safe coasting value.
+     */
+    if (intel_alpm_is_alpm_aux_less(intel_dp, crtc_state) && +        intel_psr_pr_async_video_timing_supported(intel_dp)) +        as_sdp->coasting_vtotal = crtc_state->vrr.vmax;

Seems reasonable.

Quoted 4 lines

Is this always under our control or could the hardware overwrite this with the current vtotal at the time of PR entry? Assuming we can enter PR before the vtotal goes back to vmax on its own anyway.

What I understand from Bspec in this regard is:

Do not write Adaptive Sync SDP Transmission Disable in PR Active State i.e. DB[2], instead use PR_ALPM_CTL[ AS SDP Transmission in Active Disable ] bit.

HW will sample the PR_ALPM_CTL bit only when PR is active, and it will get reflected in AS SDP payload in an 'appropriate' time.

HW will ignore this bit when PR is Inactive and always send AS SDP.

So I think HW will not touch the coasting vtotal DBs.

Quoted 8 lines
Quoted 5 lines

Driver should set appropriate coasting Vtotal and set the PR_ALPM_CTL[AS SDP Transmission in Active disable] bit.

During PR active Driver will set the DB[2] bit in payload which will trigger the sink to use Coasting Vtotal.

..HW will set the DB[2] bit ... (facepalm)

OK. Might be good to note in the commit message that this fully under driver control. I suppose it would be good to make that statement (one way or the other) for any additions we do to the AS SDP payload...

Quoted 19 lines
Quoted 19 lines

(Unless we start writing : DPCD 00B1[0] ie. PANEL_REPLAY_CONFIG2[PANEL_REPLAY_SINK_REFRESH_RATE_UNLOCK_GRANTED], in which sink starts using its own logic).

Regards,

Ankit

Quoted 5 lines
Quoted 5 lines

  }
    static

Signature · 1 line
2.45.2
Signature · 2 lines
Ville Syrjälä
Intel
#37 lore raw

PATCH 23/23 drm/i915/dp: Always enable AS SDP if supported by source + sink

In reply to Ankit Nautiyal

Currently AS SDP is only configured when VRR is enabled. However, other use cases like CMRR, Panel Replay, etc. also send information to the sink via AS SDPs.

With optimized guardband, we also need to account for wakeup time and other relevant details that depend on the AS SDP position whenever AS SDP is enabled. If a feature enabling AS SDP gets turned on later (after modeset), the guardband might not be sufficient and may need to increase, triggering a full modeset.

To avoid this, always send AS SDP whenever:
- the source and sink both support it, AND,
- there is a possibility to use it for VRR and Panel Replay for synchronization.

v2: Check if AS SDP can be used for synchronization for VRR or PR. (Ville)

---
 drivers/gpu/drm/i915/display/intel_dp.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/gpu/drm/i915/display/intel_dp.c +5 -1
index f5d96074c48e..5fc545d95cae 100644
@@ -3136,7 +3136,11 @@ static bool intel_dp_needs_as_sdp(struct intel_dp *intel_dp,
 	if (drm_dp_is_branch(intel_dp->dpcd))
 		return false;
 
-	return crtc_state->vrr.enable;
+	if (intel_alpm_is_alpm_aux_less(intel_dp, crtc_state) &&
+	    !intel_psr_pr_async_video_timing_supported(intel_dp))
+		return true;
+
+	return intel_vrr_possible(crtc_state);
 }
 
 static void intel_dp_compute_as_sdp(struct intel_dp *intel_dp,
Signature · 1 line
2.45.2