[why & how]
Pipe power gating is causing regressions.
Revert to unblock testing and promotion
This reverts commit 2eb0681ea7.
Reviewed-by: Ovidiu Bunea <ovidiu.bunea@amd.com>
Signed-off-by: Leo Chen <leo.chen@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[why & how]
This change causes regressions in ACPI and display off/on testing.
Revert the change to unblock testing.
This reverts commit 5f6937c1af.
Reviewed-by: Aric Cyr <aric.cyr@amd.com>
Signed-off-by: Ovidiu Bunea <ovidiu.bunea@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why & How]
this is to force ODM 2:1 on single eDP to lower dispclk/dppclk.
Reviewed-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Charlene Liu <Charlene.Liu@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
warnings were triggered by enum forward declarations that are not
valid in C++ without an explicit underlying type.
[How]
- Replace problematic enum forward declarations with C++-safe forms where
applicable.
- Use plain integer types for interface-only declarations that do not
require strong enum typing.
- Update dependent winterface signatures and related type usage
consistently.
- Add required include and type-visibility fixes to avoid follow-on parse
and type-resolution issues.
Reviewed-by: Aric Cyr <aric.cyr@amd.com>
Signed-off-by: Gaghik Khachatrian <gaghik.khachatrian@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why & How]
Fully enable IPS to achieve higher power savings.
Reviewed-by: Sunpeng Li <sunpeng.li@amd.com>
Signed-off-by: Ivan Lipski <ivan.lipski@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Multiple paths issue DMUB commands without managing IPS state, causing
dc_wake_and_execute_gpint/dmub_cmd to internally wake from IPS and
reallow idle. This flips idle_allowed back to true while
idle_optimizations_allowed remains false during in-flight commits,
desynchronizing the two flags.
Affected paths:
- amdgpu_dm_psr_set_event() and amdgpu_dm_replay_set_event() calls from
amdgpu_dm_handle_vrr_transition(), amdgpu_dm_commit_planes() and
amdgpu_dm_mod_power_update_streams(), that are invoked on atomic commits.
- debugfs psr_get(), psr_read_residency(), replay_get_state(),
replay_set_residency() access hardware without holding dc_lock or
disabling IPS.
[How]
- Explicitly exit IPS before PSR/Replay set_event w/ hw_programming,
called within atomic commit.
- Wrap debugfs PSR/Replay state getters and setters with IPS exit/entry +
dc_lock.
Reviewed-by: Sunpeng Li <sunpeng.li@amd.com>
Signed-off-by: Ivan Lipski <ivan.lipski@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Improves code organization by separating Adaptive Backlight
Modulation functionality from general power management.
This modular approach enhances maintainability and makes the
codebase easier to navigate.
[How]
Create new power_abm.c file containing all ABM-related functions moved from power.c.
Remove static qualifier from shared functions to enable cross-file access:
- initialize_backlight_caps: Initialize backlight capabilities
- validate_ext_backlight_caps: Validate external backlight capabilities
- backlight_millipercent_to_pwm: Convert brightness percent to PWM
- backlight_millipercent_to_millinit: Convert brightness percent to nits
- fill_backlight_level_params: Populate backlight level parametersAdd function
declarations to mod_power.h header. Update CMakeLists.txt to include power_abm.c in build.
Maintain forward declaration of struct core_power for type compatibility.
Rename struct core_power field from 'public' to 'mod_public'.
Move internal structures (backlight_state, backlight_properties,
dmcu_varibright_cached_properties, core_power) to power_helpers.h to
ensure consistent memory layouts across compilation units.
Reviewed-by: Martin Leung <martin.leung@amd.com>
Signed-off-by: Lohita Mudimela <lohita.mudimela@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
A recent type change made the lower-bound part of the OTG instance
assert redundant, which can trigger static-analysis noise and distract
from actionable diagnostics.
[How]
Kept the meaningful upper-bound range validation required for safe
narrowing to uint8_t. Removed the redundant non-negative portion of the
assert so the check matches current type semantics. Revalidated with the
latest debug build log: no warnings and no build-failure markers.
Reviewed-by: Aric Cyr <aric.cyr@amd.com>
Signed-off-by: Gaghik Khachatrian <gaghik.khachatrian@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Introduce dc_plane_cm struct.
TRIM_CM2 flag is used to synchronize between old and new implementation.
[How]
Introduce dc_plane_cm struct.
Reviewed-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Rafal Ostrowski <rafal.ostrowski@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why & How]
The logic for choosing between the dce_aux_transfer function variants is
moved into dce_aux.c rather than link_ddc.c.
The "dce_aux_transfer_with_retries" function now uses
dce_aux_transfer_raw in its implementation as the logic is equivalent.
Reviewed-by: Nevenko Stupar <nevenko.stupar@amd.com>
Reviewed-by: Gabe Teeger <gabe.teeger@amd.com>
Signed-off-by: Matthew Stewart <Matthew.Stewart2@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why & How]
for fault safe case: only call pmfw if smu present flags true
and default to 2 channle for bios intergration info table error.
Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Signed-off-by: Charlene Liu <Charlene.Liu@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why & How]
Fix various warnings related to unsigned/signed mismatches
- Consistently use the same signedness for a given value
- Explcitly cast between types when needed
Reviewed-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Clay King <clayking@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Unreachable Code;
Copy Constructor Deleted;
Local Declaration Hides Parameter;
Local Declaration Hides Outer Scope;
Uninitialized or Suspicious Memory Use.
[How]
- Removed or refactored unreachable code paths
- Ensured proper copy constructors in C++ classes
- Renamed local variables that shadowed function parameters
- Renamed inner loop/block variables to avoid shadowing outer scope
Fixed in 8 files across several FPU layers
Also fixed in color_gamma and cs_funcs modules
- Reordered guard conditions to validate pipe type before accessing stream
- Ensures safe memory access patterns in DC DMUB service layer
All changes maintain backward compatibility and preserve functional behavior.
Reviewed-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Gaghik Khachatrian <gaghik.khachatrian@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why & How]
In dml2_translation_helper.c, rename the inner loop index inside
dml2_init_soc_states() for several project cases
to avoid shadowing the outer function-scope index variable.
In display_mode_core.c, replace shift-based power-of-two expressions
used to compute dpte_row_height and dpte_row_height_linear with an
equivalent floating-point power function, consistent with existing
usage elsewhere in the file.
Behavior for valid inputs is preserved in both cases.
Reviewed-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Gaghik Khachatrian <gaghik.khachatrian@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Address signed/unsigned comparison warnings in DC paths
to keep builds warning-clean and improve type safety at comparison boundaries.
Most warnings came from signed loop/index temporaries compared against unsigned
counters (for example pipe_count, num_states, and resource-cap counters), plus a
small number of mixed signed/unsigned checks in writeback and clock-related assertions.
[How]
Aligned iterator and temporary variable types with the semantic type of the compared
bounds. Used unsigned indices for loops bounded by unsigned counters, and retained signed
types where values are semantically signed (for example arithmetic with sentinel or signed
intermediate values). Where mixed signed/unsigned comparisons are intentional, applied
explicit boundary casts or split assertions (for example non-negative signed-cap
checks before unsigned comparisons) instead of broad type changes.
No functional behavior changes are intended; this is a warning-resolution and
type-alignment cleanup.
Reviewed-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Gaghik Khachatrian <gaghik.khachatrian@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Opening the CRC data file during active rendering can fail with -EINVAL.
The wait for commit->hw_done returns remaining jiffies on success, but
the CRC path was treating that as an error.
[How]
Handle wait_for_completion_interruptible_timeout() correctly:
positive return as success, 0 as timeout, and negative as error.
Reviewed-by: Ray Wu <ray.wu@amd.com>
Signed-off-by: Tom Chung <chiahsuan.chung@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
During mode change, replay_event_general_ui may remain set on the old
stream while replay_event_hw_programming is set. This can re-enable
Replay too early before hardware programming is complete.
[How]
Clear replay_event_general_ui in the mode-change path when setting
replay_event_hw_programming to keep Replay blocked until programming
finishes, avoiding white screen on OLED panels after boot.
Reviewed-by: Sunpeng Li <sunpeng.li@amd.com>
Signed-off-by: Ray Wu <ray.wu@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why & How]
fix refresh rate round up case
Reviewed-by: Robin Chen <robin.chen@amd.com>
Signed-off-by: ChunTao Tso <ChunTao.Tso@amd.com>
Signed-off-by: James Lin <pinglei.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The function mes_v12_1_map_test_bo incorrectly returned 0 unconditionallyon error path,
which would hide the real error code and mislead upperlayers about the failure status.
Fix it by returning the correct error code 'r' instead of 0.
Fixes: 44e5195fa3 ("drm/amdgpu/mes_v12_1: add mes self test");
Signed-off-by: Yang Wang <kevinyang.wang@amd.com>
Reviewed-by: Asad Kamal <asad.kamal@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Convert the SMU15.0.8 enabled-feature query to
smu_cmn_send_smc_msg_with_params() so it uses the common SMU
multi-msgs helper.
No functional change intended.
Signed-off-by: Yang Wang <kevinyang.wang@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Convert the SMU15.0.0 table transfer path and enabled-feature query to
smu_cmn_send_smc_msg_with_params() so both paths use the common SMU
multi-msgs helper.
No functional change intended.
Signed-off-by: Yang Wang <kevinyang.wang@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Fix lock inversions pointed out by Prike and Sunil. The hang detection
timeout *CAN'T* grab locks under which we wait for fences, especially
not the userq_mutex lock.
Then instead of this completely broken handling with the
hang_detect_fence just cancel the work when fences are processed and
re-start if necessary.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Sunil Khatri <sunil.khatri@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Well the reset handling seems broken on multiple levels.
As first step of fixing this remove most calls to the hang detection.
That function should only be called after we run into a timeout! And *NOT*
as random check spread over the code in multiple places.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Sunil Khatri <sunil.khatri@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This one was fortunately not looking so bad as the wait ioctl path, but
there were still a few things which could be fixed/improved:
1. Allocating with GFP_ATOMIC was quite unnecessary, we can do that
before taking the userq_lock.
2. Use a new mutex as protection for the fence_drv_xa so that we can do
memory allocations while holding it.
3. Starting the reset timer is unnecessary when the fence is already
signaled when we create it.
4. Cleanup error handling, avoid trying to free the queue when we don't
even got one.
v2: fix incorrect usage of xa_find, destroy the new mutex on error
v3: cleanup ref ordering
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Sunil Khatri <sunil.khatri@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Convert the SMU15 table address messages to
smu_cmn_send_smc_msg_with_params() so they use the common SMU
multi-msgs helper instead of open-coding struct smu_msg_args.
No functional change intended.
Signed-off-by: Yang Wang <kevinyang.wang@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
SMU15 driver messages can carry multiple input parameters and return
values, but callers still have to build struct smu_msg_args directly.
Add common SMU multi-msgs helpers in smu_cmn and reuse them in the
single-parameter wrapper and the shared table transfer path.
Keep smu_cmn_send_smc_msg() semantics unchanged for older callers.
No functional change intended.
Signed-off-by: Yang Wang <kevinyang.wang@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The purpose of a GPU reset is to make sure that fence can be signaled
again and the signal and resume workers can make progress again.
So waiting for the resume worker or any fence in the GPU reset path is
just utterly nonsense.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Prike Liang <Prike.Liang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Fix this by skipping the sysfs disable mapping when the GPU is
currently undergoing a reset or suspend flow.
Additionally, add debug logging in psp_ptl_invoke() to better
trace PTL state and format queries/updates cmd.
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
During RAS UE-triggered GPU reset, gfx_v9_4_3_hw_fini() attempts to
send a PTL disable command to PSP. Since PSP is unresponsive at that
point, this produces spurious error logs on all hive nodes:
PTL command 0xa0000001 failed, PSP response status: 0xFFFFFFFF
PTL initialization failed (-5)
Skip the PTL disable command when GPU reset is in progress, as PTL
will be properly re-initialized during post-reset recovery via
gfx_v9_4_3_late_init().
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Only set the bit when PTL is actually being disabled (state=0)
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Downgrade unhalt_cpsch warning to dev_dbg when sched is already stopped
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Move amdgpu_amdkfd_stop/start_sched calls from kfd_ptl_control()
into amdgpu_ptl_perf_monitor_ctrl() so all PTL callers (KFD ioctl,
sysfs, GFX init) get consistent scheduling management.
Add amdgpu_amdkfd_stop/start_sched_all() wrappers to stop and
restart KFD scheduling on all nodes without assuming node ID ordering.
v3:
* call start/stop for PTL Set Only
v2:
* move the stop/start sched function to
amdgpu_ptl_perf_monitor_ctrl(Lijo)
* add wrapper amdgpu_amdkfd_stop_sched_all and
amdgpu_amdkfd_start_sched_all (Lijo)
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
GC 9.4.4 uses SPI busy status for idle detection instead of GRBM GUI_ACTIVE.
Add version check to use SPI_BUSY for 9.4.4 while keeping GRBM_STATUS
GUI_ACTIVE check for other GC versions.
v2: move this check into amdgpu_ptl_perf_monitor_ctrl(Lijo)
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Add an explicit check on cmd->resp.status after psp_cmd_submit_buf()
returns to ensure PTL state is only updated on actual success.
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ensure GFX engine is idle before switching PTL state to prevent
register access violations and CP hang. This addresses the race
condition where in-flight GPU commands could conflict with PTL
state changes.
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Add F8 and VECTOR to amdgpu_ptl_fmt and PSP format mapping.
Update PTL format strings and GFX format enum to keep PSP/KFD in sync.
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Use a bitmap to track PTL disable requests from sysfs and profiler.
PTL is only re-enabled once all sources have released their disable
requests, avoiding premature enablement.
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Stop the scheduler before releasing the PTL disable request to ensure
the GPU is quiescent during the PTL state transition. This prevents
potential queue preemption failures and GPU resets caused by modifying
PTL state while waves are executing
v1->v2:
only stop/start the scheduler when the PTL state actually needs to transition(Yifan)
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Introduce a new IOCTL option to allow userspace explicit control over
the Peak Tops Limiter (PTL) state for profiling
Link: https://github.com/ROCm/rocm-systems/tree/develop/projects/rocprofiler-sdk
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Add kgd->ptl_ctrl() callback so KFD can query/enable/disable
PTL state through the PSP performance monitor interface.
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Introduce hardware detection, runtime state tracking and a
kgd->ptl_ctrl() callback to enable/disable/query PTL via the
PSP performance-monitor interface (commands 0xA0000000/1).
The driver now exposes PTL capability to KFD and keeps the
software state in sync with the hardware.
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Introduce a Peak Tops Limiter (PTL) driver that dynamically caps
engine frequency to ensure delivered TOPS never exceeds a defined
TOPS_limit. This initial implementation provides core data structures
and kernel-space interfaces (set/get, enable/disable) to manage PTL state.
PTL performs a firmware handshake to initialize its state and update
predefined format types. It supports updating these format types at
runtime while user-space tools automatically switch PTL state, and
also allows explicitly switching PTL state via newly added commands.
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Introduce new psp interfaces and structures for performance
monitoring hardware control.
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
kfd_ioctl_profiler takes a similar approach to that of
kfd_ioctl_dbg_trap (which contains debugger related IOCTL
services) where kfd_ioctl_profiler will contain all profiler
related IOCTL services. The IOCTL is designed to be expanded
as needed to support additional profiler functionality.
The current functionality of the IOCTL is to allow for profilers
which need PMC counters from GPU devices to both signal to other
profilers that may be on the system that the device has active PMC
profiling taking place on it (multiple PMC profilers on the same
device can result in corrupted counter data) and to setup the device
to allow for the collection of SQ PMC data on all queues on the device.
For PMC data for the SQ block (such as SQ_WAVES) to be available
to a profiler, mmPERFCOUNT_ENABLE must be set on the queues. When
profiling a single process, the profiler can inject PM4 packets into
each queue to turn on PERFCOUNT_ENABLE. When profiling system wide,
the profiler does not have this option and must have a way to turn
on profiling for queues in which it cannot inject packets into directly.
Accomplishing this requires a few steps:
1. Checking if the user has the necessary permissions to profile system
wide on the device. This check uses the same check that linux perf
uses to determine if a user has the necessary permissions to profile
at this scope (primarily if the process has CAP_SYS_PERFMON or is root).
2. Locking the device for profiling. This is done by setting a lock bit
on the device struct and storing the process that locked the device.
3. Iterating all queues on the device and issuing an MQD Update to enable
perfcounting on the queues.
4. Actions to cleanup if the process exits or releases the lock.
The IOCTL also contains a link to the existing PC Sampling IOCTL as well.
This is per a suggestion that we should potentially remove the PC Sampling
IOCTL to have it be a part of the profiler IOCTL. This is a future change.
In addition, we do expect to expand the profiler IOCTL to include
additional profiler functionality in the future (which necessitates the
use of a version number).
v2: sqaush in proper IOCTL number
Proposed userpace support:
40abc95a64
Signed-off-by: Benjamin Welton <benjamin.welton@amd.com>
Signed-off-by: Perry Yuan <perry.yuan@amd.com>
Acked-by: Kent Russell <kent.russell@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Currently i915 and xe each implement their own versions of
intel_plane_(un)pin(). Now that we have the fb_pin parent
interface we can consolidate this to a single implementation.
The result is a mixture of the i915 and xe implementations.
The reuse_vma() hack comes from xe (and i915 doesn't implement
that part of the parent interface, and the pin_params are
taken from i915 since the platforms supported by i915 need
more things.
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patch.msgid.link/20260508143426.26504-17-ville.syrjala@linux.intel.com
Use xe_fb_pin_ggtt_pin() instead of intel_fb_pin_to_ggtt() for
the initial FB pin. We want to get rid of intel_fb_pin_to_ggtt()
and just use the new fb_pin parent interface.
This still isn't quite the final solution since we bypass the
actual parent interface and call the implementation directly.
But sorting that out will require more cleanup to the initial
FB code.
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patch.msgid.link/20260508143426.26504-16-ville.syrjala@linux.intel.com
Replace the intel_fb_pin_to_ggtt() and intel_fb_unpin_vma() with the
new abstract parent interface (intel_parent_fb_pin_ggtt_(un)pin()).
xe no longer needs intel_fb_unpin_vma(), and in i915 it now
becomes and internal function to i915_fb_pin.
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patch.msgid.link/20260508143426.26504-15-ville.syrjala@linux.intel.com
