From: <IVAN.LIPSKI@amd.com>
To: <amd-gfx@lists.freedesktop.org>
Cc: Harry Wentland <harry.wentland@amd.com>,
Leo Li <sunpeng.li@amd.com>,
Aurabindo Pillai <aurabindo.pillai@amd.com>,
Roman Li <roman.li@amd.com>, Wayne Lin <wayne.lin@amd.com>,
Tom Chung <chiahsuan.chung@amd.com>,
"Fangzhi Zuo" <jerry.zuo@amd.com>,
Dan Wheeler <daniel.wheeler@amd.com>, Ray Wu <Ray.Wu@amd.com>,
Ivan Lipski <ivan.lipski@amd.com>, Alex Hung <alex.hung@amd.com>,
James Lin <PingLei.Lin@amd.com>,
Chenyu Chen <Chen-Yu.Chen@amd.com>,
James Lin <pinglei.lin@amd.com>
Subject: [PATCH 04/82] drm/amd/display: Use fast update path for address-only plane flips
Date: Tue, 18 Aug 2026 16:14:56 -0400 [thread overview]
Message-ID: <20260818202139.4172592-5-IVAN.LIPSKI@amd.com> (raw)
In-Reply-To: <20260818202139.4172592-1-IVAN.LIPSKI@amd.com>
From: James Lin <pinglei.lin@amd.com>
[Why]
amdgpu_dm_commit_planes() unconditionally attaches plane_info and
scaling_info to every dc_surface_update, even for pure page flips that
only change the surface address. In DC, full_update_required() treats a
non-NULL plane_info or scaling_info pointer as a reason to force a full
update, so fast_update_only() always returns false and every commit is
routed through commit_planes_for_stream() instead of the lightweight
commit_planes_for_stream_fast() path. The heavy path acquires the DMU
global lock and waits for its ACK on every commit (spinning ~27ms),
serializing the kernel against the firmware even though the flip is
address-only.
Tracing the two commit paths during a steady 60Hz flip on two outputs
shows the heavy path is taken on every frame, while the update_type that
DC computed for those same commits is already UPDATE_TYPE_FAST - i.e.
nothing actually changed and the full path is chosen solely because the
optional pointers are present:
before: commit_planes_for_stream (FULL) = 550
commit_planes_for_stream_fast (FAST) = 0
(all commits carried update_type == UPDATE_TYPE_FAST)
[How]
Only attach plane_info and scaling_info to the surface update when the
update is not UPDATE_TYPE_FAST. fill_dc_plane_info_and_addr() is still
called so the flip address is populated for the fast path, and the
scaling info is still computed into the bundle; only the pointers that
would otherwise force a full update are withheld for fast updates.
This is safe because every consumer already NULL-checks these optional
pointers (copy_surface_update_to_plane(), get_plane_info_update_type()
via det_surface_update(), full_update_required() and
commit_planes_for_stream()), det_surface_update() clears
surface->update_bits on every check so no stale state is left behind, and
UPDATE_TYPE_FAST guarantees no plane_info/scaling_info field has changed.
It also mirrors the existing flip_immediate logic, which already relies
on acrtc_state->update_type == UPDATE_TYPE_FAST. Non-fast updates
(modeset, rotation, format, DCC, scaling, MPO, cursor, PSR/Replay) keep
carrying plane_info/scaling_info and are unaffected.
With the change the same workload takes the fast path on every flip:
after: commit_planes_for_stream (FULL) = 0
commit_planes_for_stream_fast (FAST) = 449
Reviewed-by: Sun peng (Leo) Li <sunpeng.li@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Signed-off-by: Ivan Lipski <ivan.lipski@amd.com>
---
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 10 ++++++----
1 file changed, 6 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 567bf04979ee5..e1fdd0cb65b34 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -3884,8 +3884,9 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_commit *state,
amdgpu_dm_plane_fill_dc_scaling_info(dm->adev, new_plane_state,
&bundle->scaling_infos[planes_count]);
- bundle->surface_updates[planes_count].scaling_info =
- &bundle->scaling_infos[planes_count];
+ if (acrtc_state->update_type != UPDATE_TYPE_FAST)
+ bundle->surface_updates[planes_count].scaling_info =
+ &bundle->scaling_infos[planes_count];
plane_needs_flip = old_plane_state->fb && new_plane_state->fb;
@@ -3906,8 +3907,9 @@ static void amdgpu_dm_commit_planes(struct drm_atomic_commit *state,
new_plane_state->plane->index,
bundle->plane_infos[planes_count].dcc.enable);
- bundle->surface_updates[planes_count].plane_info =
- &bundle->plane_infos[planes_count];
+ if (acrtc_state->update_type != UPDATE_TYPE_FAST)
+ bundle->surface_updates[planes_count].plane_info =
+ &bundle->plane_infos[planes_count];
if (acrtc_state->stream->link->psr_settings.psr_feature_enabled ||
acrtc_state->stream->link->replay_settings.replay_feature_enabled) {
--
2.43.0
next prev parent reply other threads:[~2026-08-18 20:23 UTC|newest]
Thread overview: 86+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-08-18 20:14 [PATCH 00/82] DC Patches August 17, 2026 IVAN.LIPSKI
2026-08-18 20:14 ` [PATCH 01/82] drm/amd/display: Fall back to overlay cursor on dcn4x when top plane doesn't fill CRTC IVAN.LIPSKI
2026-08-19 7:27 ` Michel Dänzer
2026-08-20 10:04 ` Timur Kristóf
2026-08-18 20:14 ` [PATCH 02/82] drm/amd/display: Fixes for HPO test regressions IVAN.LIPSKI
2026-08-18 20:14 ` [PATCH 03/82] drm/amd/display: Refactor DPP_SET_INPUT_TRANSFER_FUNC to drop pipe_ctx IVAN.LIPSKI
2026-08-18 20:14 ` IVAN.LIPSKI [this message]
2026-08-18 20:14 ` [PATCH 05/82] drm/amd/display: Split OPTC_PIPE_CONTROL_LOCK into smaller HWSS blocks IVAN.LIPSKI
2026-08-18 20:14 ` [PATCH 06/82] drm/amd/display: Fix DPREFCLK override when SMU isn't present or ready for DCN315 IVAN.LIPSKI
2026-08-18 20:14 ` [PATCH 07/82] drm/amd/display: Test writeback connector IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 08/82] drm/amd/display: Test GPU memory allocation IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 09/82] drm/amd/display: Cover MST path in encoder atomic_check IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 10/82] drm/amd/display: Cover amdgpu_dm_encoder_init IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 11/82] drm/amd/display: Cover MST-start failure in detect_mst IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 12/82] drm/amd/display: Cover amdgpu_dm_update_connector_after_detect IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 13/82] drm/amd/display: Cover HDMI infoframe/freesync timing paths IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 14/82] drm/amd/display: Clear HUBPREQ_DEBUG_DB on DCN6 IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 15/82] drm/amd/display: Disable alt-ch until dependencies are ready IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 16/82] drm/amd/display: Fix CalculateFlipSchedule Calculation IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 17/82] drm/amd/display: Test EDID quirks and ACPI EDID read IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 18/82] drm/amd/display: Test execute_synaptics_rc_command failures IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 19/82] drm/amd/display: Test MST stream feature read failure IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 20/82] drm/amd/display: Test dm_helpers_submit_i2c_over_aux IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 21/82] drm/amd/display: Test GPU memory allocate and free helpers IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 22/82] drm/amd/display: Test dm_helpers_dmub_set_config_sync IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 23/82] drm/amd/display: Test dm_helpers_is_dp_sink_present IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 24/82] drm/amd/display: Test dm_helpers_read_local_edid IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 25/82] drm/amd/display: Test dp_handle_test_pattern_request patterns IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 26/82] drm/amd/display: Test DMUB reg callbacks IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 27/82] drm/amd/display: Test VBIOS bounding box IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 28/82] drm/amd/display: Test dm_init_microcode IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 29/82] drm/amd/display: Test dm_dmub_sw_init IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 30/82] drm/amd/display: Add hook to disable alt-ch in PMO IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 31/82] drm/amd/display: Update alt-ch size calculations IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 32/82] drm/amd/display: Enable min dispclk ODM on DCN42 IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 33/82] drm/amd/display: Add amdgpu_dm_connector_poll KUnit tests IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 34/82] drm/amd/display: Cover hide_secondary_tile_from_userspace IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 35/82] drm/amd/display: Cover dm_validate_stream_and_context IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 36/82] drm/amd/display: Cover amdgpu_dm_create_validate_stream_for_sink IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 37/82] drm/amd/display: Cover amdgpu_dm_connector_mode_valid IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 38/82] drm/amd/display: Test MST sideband message ack path IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 39/82] drm/amd/display: Test dm_dp_mst_get_modes without a remote EDID IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 40/82] drm/amd/display: Test dm_dp_mst_get_modes with " IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 41/82] drm/amd/display: Test dm_dp_mst_detect DPCD probe and unplug IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 42/82] drm/amd/display: Test MST connector register and unregister IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 43/82] drm/amd/display: Test dm_dp_mst_connector_destroy IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 44/82] drm/amd/display: Test plane state duplicate and destroy IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 45/82] drm/amd/display: Test modifier list de-duplication IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 46/82] drm/amd/display: Test GFX6-8 tiling info from modifiers IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 47/82] drm/amd/display: Test GFX6-8 tile mode and tile split lookups IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 48/82] drm/amd/display: Test GFX6-8 modifier calculation IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 49/82] drm/amd/display: Test GFX6-8 modifier list generation IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 50/82] drm/amd/display: Test framebuffer prepare and cleanup IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 51/82] drm/amd/display: Test cursor update and async plane update IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 52/82] drm/amd/display: Remove RMCM tetrahedral cube from dc_plane_state IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 53/82] drm/amd/display: Cover mode_valid EDID mgmt path IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 54/82] drm/amd/display: Cover amdgpu_dm_fill_hdr_info_packet IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 55/82] drm/amd/display: Cover amdgpu_dm_connector_atomic_check IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 56/82] drm/amd/display: Cover atomic_check modeset triggers IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 57/82] drm/amd/display: Cover funcs_force valid EDID path IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 58/82] drm/amd/display: Cover amdgpu_dm_connector_get_modes IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 59/82] drm/amd/display: Cover create_eml_sink valid EDID path IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 60/82] drm/amd/display: Cover amdgpu_set_panel_orientation IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 61/82] drm/amd/display: Refactor amdgpu_dm_irq_test IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 62/82] drm/amd/display: Test pageflip completion in the high IRQ handlers IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 63/82] drm/amd/display: Test writeback handling in dm_crtc_high_irq IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 64/82] drm/amd/display: Test schedule_dc_vmin_vmax IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 65/82] drm/amd/display: Test handle_hpd_irq_helper detect and debounce exits IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 66/82] drm/amd/display: Test HPD RX, HPD init and DMUB callback branches IVAN.LIPSKI
2026-08-18 20:15 ` [PATCH 67/82] drm/amd/display: Test dm_dmub_outbox1_low_irq drain and work guards IVAN.LIPSKI
2026-08-18 20:16 ` [PATCH 68/82] drm/amd/display: Test amdgpu_dm_dce110_register_irq_handlers IVAN.LIPSKI
2026-08-18 20:16 ` [PATCH 69/82] drm/amd/display: Test amdgpu_dm_dcn10_register_irq_handlers IVAN.LIPSKI
2026-08-18 20:16 ` [PATCH 70/82] drm/amd/display: Fix mismatch number of OPP/DPP accounting IVAN.LIPSKI
2026-08-18 20:16 ` [PATCH 71/82] drm/amd/display: Cover amdgpu_dm_prune_primary_tile_modes IVAN.LIPSKI
2026-08-18 20:16 ` [PATCH 72/82] drm/amd/display: Cover add_fs_modes mode generation IVAN.LIPSKI
2026-08-18 20:16 ` [PATCH 73/82] drm/amd/display: Cover add_fs_modes illegal timing skip IVAN.LIPSKI
2026-08-18 20:16 ` [PATCH 74/82] drm/amd/display: Refactor hdmi_frl_status_polling_work for Kunit testing IVAN.LIPSKI
2026-08-18 20:16 ` [PATCH 75/82] drm/amd/display: Cover hdmi_frl_status_polling_work IVAN.LIPSKI
2026-08-18 20:16 ` [PATCH 76/82] drm/amd/display: Cover amdgpu_dm_i2c_xfer IVAN.LIPSKI
2026-08-18 20:16 ` [PATCH 77/82] drm/amd/display: Cover amdgpu_dm_create_i2c IVAN.LIPSKI
2026-08-18 20:16 ` [PATCH 78/82] drm/amd/display: Add passthrough visual confirm IVAN.LIPSKI
2026-08-18 20:16 ` [PATCH 79/82] drm/amd/display: Populate vblank_nom according to bounding box IVAN.LIPSKI
2026-08-18 20:16 ` [PATCH 80/82] drm/amd/display: Adjust vblank_nom policy for HW SDP tranmission reqs IVAN.LIPSKI
2026-08-18 20:16 ` [PATCH 81/82] drm/amd/display: Guard amdgpu_dm_irq_schedule_work against NULL irq_wq IVAN.LIPSKI
2026-08-18 20:16 ` [PATCH 82/82] drm/amd/display: Promote DC to 3.2.395 IVAN.LIPSKI
2026-08-24 13:46 ` [PATCH 00/82] DC Patches August 17, 2026 Wheeler, Daniel
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260818202139.4172592-5-IVAN.LIPSKI@amd.com \
--to=ivan.lipski@amd.com \
--cc=Chen-Yu.Chen@amd.com \
--cc=PingLei.Lin@amd.com \
--cc=Ray.Wu@amd.com \
--cc=alex.hung@amd.com \
--cc=amd-gfx@lists.freedesktop.org \
--cc=aurabindo.pillai@amd.com \
--cc=chiahsuan.chung@amd.com \
--cc=daniel.wheeler@amd.com \
--cc=harry.wentland@amd.com \
--cc=jerry.zuo@amd.com \
--cc=roman.li@amd.com \
--cc=sunpeng.li@amd.com \
--cc=wayne.lin@amd.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox