[WHAT]
Add KUnit tests for dm_force_atomic_commit() covering a successful
restore of the connector, its CRTC and the CRTC primary plane, and the
error path taken when the connector state cannot be duplicated.
Assisted-by: Copilot:Claude-Opus-5
Reviewed-by: Bhawanpreet Lakha <bhawanpreet.lakha@amd.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Chenyu Chen <chen-yu.chen@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[WHAT]
Add KUnit tests for amdgpu_dm_connector_init() covering a HDMI link
that is fully brought up, a DVI link that maps to a DVI-D connector
without a CEC notifier, and the get_modes() helper hook it installs.
Assisted-by: Copilot:Claude-Opus-5
Reviewed-by: Bhawanpreet Lakha <bhawanpreet.lakha@amd.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Chenyu Chen <chen-yu.chen@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[WHAT]
Add KUnit tests for amdgpu_dm_update_freesync_caps() covering the
DisplayPort invalid-MSA refresh range along with its narrow-range and
valid-MSA-required rejections, an eDP panel advertising replay, the
HDMI AMD VSDB range with and without a parser, a whitelisted DP to HDMI
converter, the forced FreeSync minimum quirk, the MCCS capability
withdrawal and the replay teardown.
v2: rebase (Alex)
Assisted-by: Copilot:Claude-Opus-5
Reviewed-by: Bhawanpreet Lakha <bhawanpreet.lakha@amd.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Chenyu Chen <chen-yu.chen@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[WHAT]
Add a KUnit test for amdgpu_dm_initialize_hdmi_connector() verifying
that a CEC notifier is created for the connector.
Assisted-by: Copilot:Claude-Opus-5
Reviewed-by: Bhawanpreet Lakha <bhawanpreet.lakha@amd.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Chenyu Chen <chen-yu.chen@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[WHAT]
Add KUnit tests for amdgpu_dm_connector_init_helper() covering the
common connector defaults and attached properties, the HDMI,
DisplayPort, DVI-D, VGA, eDP and unhandled connector types, an MST
branch connector that skips the root-only properties, the content
protection property and the default HPD debounce delay.
Assisted-by: Copilot:Claude-Opus-5
Reviewed-by: Bhawanpreet Lakha <bhawanpreet.lakha@amd.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Chenyu Chen <chen-yu.chen@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[WHAT]
Add KUnit tests for create_stream_for_sink() covering the preferred
mode timing recalculation, the scaled path that keeps the previous
refresh rate, the automated test timing override, the DisplayPort VSC
SDP colorimetry decision with and without the disable-colorimetry
quirk, and the HDMI vendor specific infoframe path.
Assisted-by: Copilot:Claude-Opus-5
Reviewed-by: Bhawanpreet Lakha <bhawanpreet.lakha@amd.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Chenyu Chen <chen-yu.chen@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[why & how]
DCN IP team recommends SDPIF_PORT_CONTROL=0 as POR value for APUs.
The new POR value ensures that DCHUBBUB will not disconnect the SDP port to DF unless it explicitly
receives the pre_cstate_req signal. SDPIF_PORT_CONTROL=1 will only be used for IPS entry sequences,
where it will be restored to 0 upon IPS exit.
Reviewed-by: Leo Chen <leo.chen@amd.com>
Signed-off-by: Ovidiu Bunea <ovidiu.bunea@amd.com>
Signed-off-by: Chenyu Chen <chen-yu.chen@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[WHY]
The RMCM logic was embedded inside the MPC block, sharing its source files,
functions, register definitions and other logic. This tight coupling
created several problems:
- The RMCM vtable was bolted onto struct mpc_funcs, making it impossible
to instantiate RMCM independently of MPC
- RMCM resource acquisition had no tracking - there was no mechanism to
allocate the 2 available RMCM instances to pipes or release them
[HOW]
RMCM is now a separate module:
- dc/rmcm/ directory with per-ASIC implementations
- new struct rmcm with its own funcs vtable containing shaper,
3DLUT, and fast-load operations
- removed all RMCM code from MPC
- added RMCM resource lifecycle management in dc_resource.c:
- resource_assign_rmcm(): Assigns the instances for a whole context, so a
plane that is only now asking never evicts one already using an instance
- resource_release_rmcm(): Clears the pipe's allocation.
- updated resource pool constructors (dcn42, dcn42b, dcn60) to create
RMCM instances via the new dcnXX_rmcm_create() and destroy them in
pool teardown
Reviewed-by: Sridevi Arvindekar <sridevi.arvindekar@amd.com>
Signed-off-by: Rafal Ostrowski <rafal.ostrowski@amd.com>
Signed-off-by: Chenyu Chen <chen-yu.chen@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
add energy accumulator on pmfw 0x00685000 and above version.
Signed-off-by: Kevin Wang <kevin.wang@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Retire the legacy design, always enable unified MES on GFX 12.1 and drop
the driver code and MODULE_FIRMWARE entries that only served the legacy
path.
Signed-off-by: Horatio Zhang <hongkun.zhang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Description]
- UPSP should only be enabled when we want linear scaling for 420/422
- In DC for non-FP16 formats that are told to scale in non-linear, program
bypass for PRE_GAM since these formats are already non-linear (i.e.,
YUV420, YUV422, RGB, etc.)
Reviewed-by: Aric Cyr <aric.cyr@amd.com>
Reviewed-by: Taimur Hassan <syed.hassan@amd.com>
Signed-off-by: Alvin Lee <Alvin.Lee2@amd.com>
Signed-off-by: Chenyu Chen <chen-yu.chen@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Hwss executors shouldn't be coupled to complex dc structs such as
pipe_ctx.
[How]
Refactor cursor offload hardware sequencer extension to accept only
the parameters required for programming cursor offload state.
Extract these parameters in the generic HWSS layer for both immediate execution
and block-sequence construction.
Reviewed-by: Ilya Bakoulin <ilya.bakoulin@amd.com>
Signed-off-by: Tomasz Siemek <tomasz.siemek@amd.com>
Signed-off-by: Chenyu Chen <chen-yu.chen@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When the commit b67d2d039f ("drm/gem: Add callback for when handle
count goes to 0") added the handle_free callback, it did not update the
rust abstraction, which will cause a build failure when that is enabled.
Rather than just adding another None initializer, use
..pin_init::zeroed() to make all unlisted callbacks None, and remove the
existing zero-initializations.
Fixes: b67d2d039f ("drm/gem: Add callback for when handle count goes to 0")
Reported-by: Thorsten Leemhuis <linux@leemhuis.info>
Closes: https://lore.kernel.org/all/aaef18cd-ab55-4fb8-8563-ca3a5ea9cbc6@leemhuis.info
Reported-by: Thomas Zimmermann <tzimmermann@suse.de>
Closes: https://lore.kernel.org/all/4da62a4b-c4ef-492a-9a27-b56bd4784acf@suse.de
Suggested-by: Gary Guo <gary@garyguo.net>
Signed-off-by: Klara Modin <klarasmodin@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
add energy accumulator on pmfw 0x004e8600 and above version.
Signed-off-by: Kevin Wang <kevin.wang@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
GPU metrics v1.3 defines energy_accumulator as a 64‑bit field.
The unsupported‑firmware code path assigns UINT_MAX, which is neither the
full‑width invalid value for this field nor its default value.
Fixes: 8de9edb359 ("drm/amd/pm: remove invalid gpu_metrics.energy_accumulator on smu v13.0.x")
Signed-off-by: Kevin Wang <kevin.wang@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
bo_va is assigned from fpriv->csa_va, which is already verified to be
non-NULL by the surrounding if statement. The inner NULL check can
therefore never trigger and is dead code, so remove it.
Signed-off-by: Jiangshan Yi <yijiangshan@kylinos.cn>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When building with -Wincompatible-function-pointer-types-strict, a
warning designed to catch kernel control flow integrity (kCFI) issues at
build time, there is an instance around mp1_v13_0_eeprom_send_msg():
drivers/gpu/drm/amd/amdgpu/../ras/ras_mgr/amdgpu_ras_mp1_v13_0.c:157:25: error: incompatible function pointer types initializing 'int (*)(struct ras_core_context *, u32, uint32_t, uint32_t *)' (aka 'int (*)(struct ras_core_context *, unsigned int, unsigned int, unsigned int *)') with an expression of type 'int (struct ras_core_context *, enum ras_fw_eeprom_cmd, uint32_t, uint32_t *)' (aka 'int (struct ras_core_context *, enum ras_fw_eeprom_cmd, unsigned int, unsigned int *)') [-Werror,-Wincompatible-function-pointer-types-strict]
157 | .mp1_send_eeprom_msg = mp1_v13_0_eeprom_send_msg,
| ^~~~~~~~~~~~~~~~~~~~~~~~~
While 'u32' and 'enum ras_fw_eeprom_cmd' are ABI compatible, hence no
regular warning from -Wincompatible-function-pointer-types, the mismatch
will trigger a kCFI violation when mp1_v13_0_eeprom_send_msg() is called
indirectly.
Update the second parameter of mp1_v13_0_eeprom_send_msg()' to be
'u32 msg_id' to match the prototype in 'struct ras_mp1_sys_func' (which
was recently changed to support mp1_v15_0), clearing up the warning and
kCFI violation.
Fixes: 11a948c781 ("drm/amd/ras: Support retrieving bad page info from mp1_v15_0")
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
resource_build_info_frame() derives colorimetry and RGB quantization from
stream->output_color_space. A Broadcast RGB-only atomic commit updates
that field and reprograms the output CSC, but none of the InfoFrame update
predicates include output_color_space. The sink can therefore retain the
previous AVI InfoFrame range while the source starts transmitting a
different pixel range.
Treat an output color space change as an InfoFrame change in update
classification and in both stream programming paths.
Hardware testing on an HDMI 2.1 television confirmed that its automatic
black-level selection follows Full to Limited and Limited to Full
transitions in SDR, HDR, and HDR with VRR active, without a modeset or
visible link blank.
Fixes: 6eb4c13a38 ("drm/amd/display: Support "Broadcast RGB" drm property")
Signed-off-by: Satyajit Roy <sroy14@alum.utk.edu>
Reviewed-by: Alex Hung <alex.hung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
amdgpu_dm_get_output_color_space() applies the Broadcast RGB connector
property to default RGB output, but always selects full-range output for
BT.2020 RGB. Consequently, explicitly selecting Limited has no effect on
the output CSC or AVI InfoFrame when HDR uses BT.2020 RGB.
Select COLOR_SPACE_2020_RGB_LIMITEDRANGE when the output encoding is RGB
and Broadcast RGB is Limited. Keep Automatic and Full at full range, and
leave YCbCr output unchanged.
Add KUnit coverage for limited-range RGB output through both BT.2020
connector colorspace values.
Fixes: 6eb4c13a38 ("drm/amd/display: Support "Broadcast RGB" drm property")
Signed-off-by: Satyajit Roy <sroy14@alum.utk.edu>
Reviewed-by: Alex Hung <alex.hung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
DC uses dc_edid_caps.qs_bit when constructing the HDMI AVI InfoFrame
quantization-range field. Although DRM parses the sink capability into
drm_display_info, DM never copies it into the DC EDID capabilities. The
field therefore remains zero and the AVI quantization range stays at its
default value.
Copy rgb_quant_range_selectable for HDMI sinks and extend the existing
EDID-capability KUnit test to cover it.
Fixes: 6eb4c13a38 ("drm/amd/display: Support "Broadcast RGB" drm property")
Signed-off-by: Satyajit Roy <sroy14@alum.utk.edu>
Reviewed-by: Alex Hung <alex.hung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This reverts commit c119d05a36.
It removes the newline even when there are no fences attached to a
struct dma_resv, leading to multiple BOs being output on the same line,
making the debug file less readable, not more as the commit intended.
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@igalia.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
gfx_v8_0_init_microcode() allocates rlc.register_list_format before
loading the MEC firmware.
If loading one of the subsequent firmware files fails, the function
returns through the error path without freeing register_list_format.
Use gfx_v8_0_free_microcode() in the error path so all previously
allocated firmware resources, including register_list_format, are
released.
Signed-off-by: Willian Oliveira <williandossantosdeoliveira287@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This is now set for all GMC 12.x variants in
early_init() so drop the duplicate setting.
Reviewed-by: Mukul Joshi <mukul.joshi@amd.com>
Reviewed-by: Prike Liang <Prike.Liang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
If sending TLB shootdown to a remote GPU fails, reset the connection and
mark the connection as NOT_READY.
Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Reviewed-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
On the LSDMA path amdgpu_ualink_send_command() ignored the return of
amdgpu_lsdma_copy_mem() and polled the remote completion for the full
timeout even when the copy failed. With a peer down, every NPA-REVOKE
send in amdgpu_ualink_exp_cleanup_worker() wasted ~2s, and the peer
stayed ESTABLISHED so each subsequent handle cleanup timed out again,
piling up past the 120s hung-task threshold.
Return immediately when an LSDMA copy fails, and mark the connection
NOT_READY on a failed revoke send so other cleanups skip the unreachable
peer via amdgpu_ualink_check_conn_ready().
Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Reviewed-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
If the userq doesn't support in a system. then there's no
valid userq_doorbell_xa entry to walk over and then has a
no-op.
Signed-off-by: Prike Liang <Prike.Liang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
A hung SDMA user queue wedges MES, so the MES packet reset times out and
falls back to a full MODE1 reset - once per in-flight job, a reset storm.
The queue is still on its HW slot at the first hang-detect, so use
detect_hung_queue to recover its slot from the doorbell and reset it over
MMIO, which does not need MES. Mark it HUNG (not UNMAPPED) so the restore
worker does not re-map and re-run the guilty job, and short-circuit the
per-fence hang-detect re-entry once the queue is already reset.
Signed-off-by: Jesse Zhang <Jesse.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Match each SDMA queue's DOORBELL_OFFSET register against the given
doorbell to recover its (instance, queue_id) HW slot. The per-queue
register stride is derived from the named QUEUE0/QUEUE1 registers.
Signed-off-by: Jesse Zhang <Jesse.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Match each SDMA queue's DOORBELL_OFFSET register against the given
doorbell to recover its (instance, queue_id) HW slot. The per-queue
register stride is derived from the named QUEUE0/QUEUE1 registers.
Signed-off-by: Jesse Zhang <Jesse.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
GFX 12.1.0 queues are managed by MES, not by the HWS runlist, so the
127 queue limit does not apply.
The topology advertises 32 CP queues plus 16 SDMA engines with 6 queues
each, so creating every queue needs 128 and the last one fails with
-ENOSPC.
Signed-off-by: Horatio Zhang <hongkun.zhang@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Add an optional per-IP callback that maps a user queue's doorbell index
to its HW slot (instance, queue_id). It is needed to reset a hung SDMA
user queue over MMIO, since struct amdgpu_usermode_queue does not store
its HW slot.
Signed-off-by: Jesse Zhang <Jesse.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This version brings along the following updates:
- Ensure requested LSDMA bandwidth stays within the SOP bound.
- Refactor DC_SEND_CURSOR_INFO_TO_DMU to drop the pipe_ctx param.
- Add an override for LSDMA bandwidth in the QoS table.
- Expand amdgpu_dm KUnit coverage across vblank IRQ, self-refresh, GPU
reset, plane update, MST resume, cached suspend state, writeback
transitions, atomic validation, connector-state restore, EDID CEA
parsing and DM IRQ source handling.
- Keep streamclk gating off when enabling streamclk.
- Remove unnecessary includes.
- Set mpc_tree_params->opp_id during OPP resource construct.
- Add override capability for UTM table params.
- Add immediate restore to FAMS2 for DRR.
- Fix missing APG registers for DCN60 HDMI.
- Exit IPS before connector detection on resume.
- Fix HF-VSDB DSC bpc detection to be cumulative.
- Fix unused params in flip schedule.
- Enable the FW locality check in DCN6.
- Fix DC Hub reference frequency assert range.
- Dump clock registers in HW init instead of SW init.
- Validate the IRQ source in the DM IRQ handler.
- Revert "Fix CalculateFlipSchedule Calculation".
- Assorted DML updates.
- Update urgent burst calculation to account for unbounded requests.
- Add a stressed peak-bandwidth probe with DMA contention.
- Revert "Unify CalculateFlipSchedule Logic".
Acked-by: Tom Chung <chiahsuan.chung@amd.com>
Signed-off-by: Taimur Hassan <Syed.Hassan@amd.com>
Signed-off-by: Ray Wu <ray.wu@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
In cm3_helper_translate_curve_to_degamma_hw_format(), seg_distr[] and
the loop counter k are uint32_t, while -1, region_start/region_end and
i are signed. Comparing them relies on implicit integer conversions,
which some compilers diagnose and which obscures the intent of the
comparisons.
[How]
Add explicit casts so the signedness of each comparison is stated in
the code. No functional change.
Assisted-by: Copilot:Claude-Opus-5
Reviewed-by: Tom Chung <chiahsuan.chung@amd.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Ray Wu <ray.wu@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why & How]
amdgpu_dm_update_freesync_state_on_stream() keyed the VTEM guard off
new_stream->sink->sink_signal, but new_stream->sink can be NULL, leading
to a NULL pointer dereference.
Use new_stream->signal instead, which is a direct field of the already
validated new_stream and carries the same HDMI signal type. This matches
the ALLM block in the same function, which already uses new_stream->signal.
Fixes: 640fd039dc ("drm/amd/display: Emit VTEM for HF-VSDB VRR on TMDS links")
Reviewed-by: Tom Chung <chiahsuan.chung@amd.com>
Signed-off-by: Fangzhi Zuo <Jerry.Zuo@amd.com>
Signed-off-by: Ray Wu <ray.wu@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why & How]
Update fields used for mode support.
Also expose some functions so they can be re-used for future products
Reviewed-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Austin Zheng <Austin.Zheng@amd.com>
Signed-off-by: Ray Wu <ray.wu@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
dml2_core_dcn6_funcs_mode_programming.c called DCN5 and DCN6
calcs functions directly by name, coupling the mode programming
flow to specific generations. Routing through the calcs table
lets a future core generation reuse this flow without editing it.
[How]
Add a static get_calcs(ctx) helper returning ctx->calcs->dcn6.
Replace every direct dcn5_*/dcn6_* calcs call with
get_calcs(ctx)->slot(...). dcn6_populate_mode_programming() did
not previously receive the mp context, so thread
calc_mp_ctx through it to reach get_calcs() for its three calcs
calls. Assign ctx->calcs from core->calcs in
dcn6_mp_build_calculate_mp_context(). Drop the now-unused
dml2_core_dcn5_calcs_dchub.h, dml2_core_dcn5_calcs_display_pipe.h,
and dml2_core_dcn6_calcs_dchub.h includes in favor of
dml2_core_dcn6_calcs.h.
Reviewed-by: Austin Zheng <austin.zheng@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Ray Wu <ray.wu@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
dml2_core_dcn6_funcs_mode_support.c called DCN5 and DCN6 calcs
functions directly by name, coupling the mode support flow to
specific generations. Routing through the calcs table lets a
future core generation reuse this flow without editing it.
[How]
Add a static get_calcs(ctx) helper returning ctx->calcs->dcn6.
Replace every direct dcn5_*/dcn6_* calcs call with
get_calcs(ctx)->slot(...). Assign calc_ms_ctx->calcs from
core->calcs when the context is built. Drop the now-unused
dml2_core_dcn5_calcs_dchub.h, dml2_core_dcn5_calcs_display_pipe.h,
and dml2_core_dcn6_calcs_dchub.h includes in favor of
dml2_core_dcn6_calcs.h.
Reviewed-by: Austin Zheng <austin.zheng@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Ray Wu <ray.wu@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
DCN6 core mode support/programming call DCN5 and DCN6 calcs
functions directly by name. The calcs layer needs one seam the
funcs layer can go through instead of hardcoding a generation.
[How]
Add union dml2_core_calcs holding one function pointer table
per generation, plus a calcs field on dml2_core_instance and on
the mode support/programming contexts.
Add dml2_core_dcn6_calcs.c/.h defining struct
dml2_core_dcn6_calcs, a flat table covering every DCN6 calcs
function and every DCN5 calcs function DCN6 reuses, and
dml2_core_dcn6_calcs_init() to populate it.
Call dml2_core_dcn6_calcs_init() from
dml2_core_dcn6_funcs_initialize().
Reviewed-by: Austin Zheng <austin.zheng@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Ray Wu <ray.wu@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why & How]
dcn5_calculate_urgent_burst_factor function is aware of unbounded
request mode making the buffer larger
Reviewed-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Austin Zheng <Austin.Zheng@amd.com>
Signed-off-by: Ray Wu <ray.wu@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Revert commit 7bee763002 ("drm/amd/display: Fix CalculateFlipSchedule Calculation")
Because it causes some regression
Reviewed-by: Austin Zheng <austin.zheng@amd.com>
Signed-off-by: Samson Tam <samson.tam@amd.com>
Signed-off-by: Ray Wu <ray.wu@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why & How]
Add dm_test_irq_handler_invalid_source to cover the early return in
amdgpu_dm_irq_handler() when dc_interrupt_to_irq_source() cannot map the
hardware entry.
Reuse the existing unmapped irq_service stub and assert that neither the
high- nor the low-context handler is dispatched.
Reviewed-by: Roman Li <roman.li@amd.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Ray Wu <ray.wu@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
dc_interrupt_to_irq_source() can return DC_IRQ_SOURCE_INVALID for an
unmapped src_id/src_data pair. The value was passed straight to
dc_interrupt_ack() and used to index the DM high/low handler tables.
[How]
Reject sources outside the valid range before using them.
Reviewed-by: Roman Li <roman.li@amd.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Ray Wu <ray.wu@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why & How]
- Dumping clock registers in SW init can get incorrect clock readback due
to other init code (e.g., in DMU) potentially not executed yet
- Move clk_ip register access into HW init to ensure the clock read back
gets the correct values
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alvin Lee <Alvin.Lee2@amd.com>
Signed-off-by: Ray Wu <ray.wu@amd.com>
Tested-by: Dan Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
