Lin.Cao [Tue, 24 Jun 2025 09:05:34 +0000 (17:05 +0800)]
drm/amdgpu: Fix memory leak in amdgpu_ctx_mgr_entity_fini
patch dd64956685fa ("drm/amdgpu: Remove duplicated "context still
alive" check") removed ctx put, which will cause amdgpu_ctx_fini()
cannot be called and then cause some finished fence that added by
amdgpu_ctx_add_fence() cannot be released and cause memleak.
Fixes: dd64956685fa ("drm/amdgpu: Remove duplicated "context still alive" check") Signed-off-by: Lin.Cao <lincao12@amd.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Avoid dereferencing struct drm_gem_object.import_attach for the
imported dma-buf. The dma_buf field in the GEM object instance refers
to the same buffer. Prepares to make import_attach an implementation
detail of PRIME.
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Philip Yang [Fri, 20 Jun 2025 22:32:32 +0000 (18:32 -0400)]
drm/amdkfd: Don't call mmput from MMU notifier callback
If the process is exiting, the mmput inside mmu notifier callback from
compactd or fork or numa balancing could release the last reference
of mm struct to call exit_mmap and free_pgtable, this triggers deadlock
with below backtrace.
The deadlock will leak kfd process as mmu notifier release is not called
and cause VRAM leaking.
The fix is to take mm reference mmget_non_zero when adding prange to the
deferred list to pair with mmput in deferred list work.
If prange split and add into pchild list, the pchild work_item.mm is not
used, so remove the mm parameter from svm_range_unmap_split and
svm_range_add_child.
Fixes: fa582c6f3684 ("drm/amdkfd: Use mmget_not_zero in MMU notifier") Signed-off-by: Philip Yang <Philip.Yang@amd.com> Reviewed-by: Felix Kuehling <felix.kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Kent Russell [Tue, 24 Jun 2025 15:42:06 +0000 (11:42 -0400)]
drm/amdgpu: Include sdma_4_4_4.bin
This got missed during SDMA 4.4.4 support.
Signed-off-by: Kent Russell <kent.russell@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
André Almeida [Fri, 13 Jun 2025 18:26:51 +0000 (15:26 -0300)]
drm/amd: Include <linux/export.h> when needed
Fix the following compile time warning when building with W=1:
warning: EXPORT_SYMBOL() is used, but #include <linux/export.h> is missing
Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: André Almeida <andrealmeid@igalia.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
André Almeida [Fri, 13 Jun 2025 18:26:50 +0000 (15:26 -0300)]
drm/amd: Do not include <linux/export.h> when unused
Fix the following compile time warning when building with W=1:
warning: EXPORT_SYMBOL() is not used, but #include <linux/export.h> is present
Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: André Almeida <andrealmeid@igalia.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Mon, 16 Jun 2025 18:04:54 +0000 (14:04 -0400)]
drm/amdgpu/sdma5.x: suspend KFD queues in ring reset
SDMA 5.x only supports engine soft reset which resets
all queues on the engine. As such, we need to suspend
KFD queues around resets like we do for SDMA 4.x.
Reviewed-by: Jesse Zhang <Jesse.Zhang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Takashi Iwai [Mon, 16 Jun 2025 16:08:41 +0000 (18:08 +0200)]
drm/amd/display: Add sanity checks for drm_edid_raw()
When EDID is retrieved via drm_edid_raw(), it doesn't guarantee to
return proper EDID bytes the caller wants: it may be either NULL (that
leads to an Oops) or with too long bytes over the fixed size raw_edid
array (that may lead to memory corruption). The latter was reported
actually when connected with a bad adapter.
Add sanity checks for drm_edid_raw() to address the above corner
cases, and return EDID_BAD_INPUT accordingly.
Fixes: 48edb2a4256e ("drm/amd/display: switch amdgpu_dm_connector to use struct drm_edid") Link: https://bugzilla.suse.com/show_bug.cgi?id=1236415 Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Kenneth Feng [Mon, 23 Jun 2025 09:02:51 +0000 (17:02 +0800)]
drm/amd/pm: revise the pcie dpm parameters
revise the pcie dpm parameters
Signed-off-by: Kenneth Feng <kenneth.feng@amd.com> Reviewed-by: Yang Wang <kevinyang.wang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/display: Add a trace event for brightness programming
[Why]
Brightness programming may involve a conversion of a user requested
brightness against what was in a custom brightness curve. The values
might not match what a user programmed.
[How]
Add a new trace event to show specific converted brightness values.
[Why]
commit 16dc8bc27c2a ("drm/amd/display: Export full brightness range to
userspace") adjusted the brightness range to scale to larger values, but
missed updating AMDGPU_MAX_BL_LEVEL which is needed to make sure that
scaling works properly with custom brightness curves.
[How]
As the change for max brightness of 0xFFFF only applies to devices
supporting DC, use existing DC define MAX_BACKLIGHT_LEVEL.
Fixes: 16dc8bc27c2a ("drm/amd/display: Export full brightness range to userspace") Acked-by: Alex Deucher <alexander.deucher@amd.com> Link: https://lore.kernel.org/r/20250623171114.1156451-1-mario.limonciello@amd.com Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Lijo Lazar [Mon, 2 Jun 2025 07:25:14 +0000 (12:55 +0530)]
drm/amdgpu: Add more checks to PSP mailbox
Instead of checking the response flag, use status mask also to check
against any unexpected failures like a device drop. Also, log error if
waiting on a psp response fails/times out.
Likun Gao [Fri, 13 Jun 2025 14:30:34 +0000 (22:30 +0800)]
drm/amdgpu: make ip map init to common function
IP instance map init function can be an common function
instead of operation on different ASIC.
V2: Create amdgpu_ip.[ch] file for ip related functions.
Patrick Lerda [Tue, 10 Jun 2025 19:12:22 +0000 (21:12 +0200)]
drm/radeon/evergreen_cs: implement cond_exec and cond_write
This change implements the support of PACKET3_COND_EXEC and
PACKET3_COND_WRITE which are required to implement
ARB_indirect_parameters. ARB_indirect_parameters is
part of OpenGL 4.6.
Pratap Nirujogi [Thu, 29 May 2025 19:21:48 +0000 (15:21 -0400)]
drm/amd/amdgpu: Add ISP Generic PM Domain (genpd) support
AMDISP I2C device requires to power on ISP HW to probe the sensor
device. Instead of using the exported symbols from ISP driver to
control the power and clocks remotely,added Generic PM Domain (genpd)
support in amdgpu_isp device for its child devices (amd_isp_capture,
amd_isp_i2c_designware) to set power and clocks using PM methods.
Co-developed-by: Bin Du <bin.du@amd.com> Signed-off-by: Bin Du <bin.du@amd.com> Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Pratap Nirujogi <pratap.nirujogi@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Pratap Nirujogi [Mon, 16 Jun 2025 18:16:43 +0000 (14:16 -0400)]
drm/amd/pm: Add support to set min ISP clocks
Add support to set ISP clocks for SMU v14.0.0. ISP driver
uses amdgpu_dpm_set_soft_freq_range() API to set clocks via
SMU interface than communicating with PMFW directly.
amdgpu_dpm_set_soft_freq_range() is updated to take in any
pp_clock_type than limiting to support only PP_SCLK to allow
ISP and other driver modules to set the min/max clocks. Any
clock specific restrictions are expected to be taken care in
SOC specific SMU implementations instead of generic amdgpu_dpm
and amdgpu_smu interfaces.
Reviewed-by: Xiaojian Du <xiaojian.du@amd.com> Reviewed-by: Lijo Lazar <lijo.lazar@amd.com> Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Pratap Nirujogi <pratap.nirujogi@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Pratap Nirujogi [Mon, 16 Jun 2025 19:07:57 +0000 (15:07 -0400)]
drm/amd/pm: Add support to set ISP Power
Add support to set ISP power for SMU v14.0.0. ISP driver
uses amdgpu_dpm_set_powergating_by_smu() API to
enable / disable power via SMU interface than communicating
with PMFW directly.
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com> Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Pratap Nirujogi <pratap.nirujogi@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Vitaly Prosyak [Wed, 18 Jun 2025 23:35:21 +0000 (19:35 -0400)]
drm/amdgpu: fix slab-use-after-free in amdgpu_userq_mgr_fini+0x70c
The issue was reproduced on NV10 using IGT pci_unplug test.
It is expected that `amdgpu_driver_postclose_kms()` is called prior to `amdgpu_drm_release()`.
However, the bug is that `amdgpu_fpriv` was freed in `amdgpu_driver_postclose_kms()`, and then
later accessed in `amdgpu_drm_release()` via a call to `amdgpu_userq_mgr_fini()`.
As a result, KASAN detected a use-after-free condition, as shown in the log below.
The proposed fix is to move the calls to `amdgpu_eviction_fence_destroy()` and
`amdgpu_userq_mgr_fini()` into `amdgpu_driver_postclose_kms()`, so they are invoked before
`amdgpu_fpriv` is freed.
This also ensures symmetry with the initialization path in `amdgpu_driver_open_kms()`,
where the following components are initialized:
- `amdgpu_userq_mgr_init()`
- `amdgpu_eviction_fence_init()`
- `amdgpu_ctx_mgr_init()`
Correspondingly, in `amdgpu_driver_postclose_kms()` we should clean up using:
- `amdgpu_userq_mgr_fini()`
- `amdgpu_eviction_fence_destroy()`
- `amdgpu_ctx_mgr_fini()`
This change eliminates the use-after-free and improves consistency in resource management between open and close paths.
[ +0.000013] The buggy address belongs to the object at ffff88811c068000
which belongs to the cache kmalloc-rnd-01-4k of size 4096
[ +0.000016] The buggy address is located 3168 bytes inside of
freed 4096-byte region [ffff88811c068000, ffff88811c069000)
Cc: Alex Deucher <alexander.deucher@amd.com> Cc: Christian König <christian.koenig@amd.com> Cc: Lijo Lazar <lijo.lazar@amd.com> Cc: Jesse Zhang <Jesse.Zhang@amd.com> Cc: Arvind Yadav <arvind.yadav@amd.com>
v2: drop amdgpu_drm_release() and assign drm_release()
as the callback directly.(Alex)
Fixes: adba0929736a ("drm/amdgpu: Fix Illegal opcode in command stream Error") Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Vitaly Prosyak <vitaly.prosyak@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd: Add missing kdoc for amd_ip_funcs `complete` callback
The `complete` callback should be described in kernel doc.
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au> Closes: https://lore.kernel.org/linux-next/20250619205931.41cf9332@canb.auug.org.au/ Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Link: https://lore.kernel.org/r/20250620041420.3585005-1-superm1@kernel.org Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd: Adjust output for discovery error handling
commit 017fbb6690c2 ("drm/amdgpu/discovery: check ip_discovery fw file
available") added support for reading an amdgpu IP discovery bin file
for some specific products. If it's not found then it will fallback to
hardcoded values. However if it's not found there is also a lot of noise
about missing files and errors.
Adjust the error handling to decrease most messages to DEBUG and to show
users less about missing files.
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com> Reported-by: Marcus Seyfarth <m.seyfarth@gmail.com> Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4312 Tested-by: Marcus Seyfarth <m.seyfarth@gmail.com> Fixes: 017fbb6690c2 ("drm/amdgpu/discovery: check ip_discovery fw file available") Acked-by: Alex Deucher <alexander.deucher@amd.com> Link: https://lore.kernel.org/r/20250617183052.1692059-1-superm1@kernel.org Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Taimur Hassan [Mon, 16 Jun 2025 02:38:55 +0000 (21:38 -0500)]
drm/amd/display: Promote DAL to 3.2.339
Summary:
* Improve USB4 bandwidth validation
* dml clock calcuation with EQU Prefetch included
* Tweaking udelay time to fix "failed to blank crtc!" error
* Add LSDMA support to DMUB
* Fix Coverity issue
Reviewed-by: Aurabindo Pillai <aurabindo.pillai@amd.com> Signed-off-by: Taimur Hassan <Syed.Hassan@amd.com> Signed-off-by: Fangzhi Zuo <jerry.zuo@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Taimur Hassan [Sun, 15 Jun 2025 15:36:32 +0000 (11:36 -0400)]
drm/amd/display: [FW Promotion] Release 0.1.16.0
Summary for changes in firmware:
* Add DMCUB IPS commands and command parser support
* use OTG count to disable interrupts
* Fix dmub_cmd header data boundary issue
* remove the HW register override
Reviewed-by: Aurabindo Pillai <aurabindo.pillai@amd.com> Signed-off-by: Taimur Hassan <Syed.Hassan@amd.com> Signed-off-by: Fangzhi Zuo <jerry.zuo@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Weiguang Li [Thu, 12 Jun 2025 03:10:12 +0000 (11:10 +0800)]
drm/amd/display: Remove redundant macro of refresh rate
[Why&How]
Found that we add redundant macro on refresh rate when calculating vtotal,
so we remove it.
Reviewed-by: Robin Chen <robin.chen@amd.com> Signed-off-by: Weiguang Li <wei-guang.li@amd.com> Signed-off-by: Fangzhi Zuo <jerry.zuo@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Wen Chen [Mon, 2 Jun 2025 20:37:08 +0000 (16:37 -0400)]
drm/amd/display: Fix 'failed to blank crtc!'
[why]
DCN35 is having “DC: failed to blank crtc!” when running HPO
test cases. It's caused by not having sufficient udelay time.
[how]
Replace the old wait_for_blank_complete function with fsleep function to
sleep just until the next frame should come up. This way it doesn't poll
in case the pixel clock or other clock was bugged or until vactive and
the vblank are hit again.
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Signed-off-by: Wen Chen <Wen.Chen3@amd.com> Signed-off-by: Fangzhi Zuo <jerry.zuo@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Austin Zheng [Thu, 5 Jun 2025 20:37:36 +0000 (16:37 -0400)]
drm/amd/display: Include EQU Prefetch Bandwidth For Bandwidth Calculations
[Why]
Pixel data bandwidth required in mode programming (MP) ends up being
higher than what was calculated in mode support (MS) even though
the prefetch bandwidths calculated in MP are lower than the MS ones.
MP used a different equ prefetch schedule than MS which lead a
slight difference in parameters. This resulted in the pixel data
bandwidth in MP to be higher than MS.
[How]
Rename the RequiredPrefetchBWOTO term so it can be applied generically.
Update the value with the EQU bandwidth if the EQU schedule is used.
Get the max prefetch bandwidth that MS calculated and use it
as part of the calculations for required bandwidth.
Reviewed-by: Dillon Varone <dillon.varone@amd.com> Signed-off-by: Austin Zheng <Austin.Zheng@amd.com> Signed-off-by: Fangzhi Zuo <jerry.zuo@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amdgpu/gfx9: Add Cleaner Shader Support for GFX9.x GPUs
Enable the cleaner shader for other GFX9.x series of GPUs to provide
data isolation between GPU workloads. The cleaner shader is responsible
for clearing the Local Data Store (LDS), Vector General Purpose
Registers (VGPRs), and Scalar General Purpose Registers (SGPRs), which
helps prevent data leakage and ensures accurate computation results.
This update extends cleaner shader support to GFX9.x GPUs, previously
available for GFX9.4.2. It enhances security by clearing GPU memory
between processes and maintains a consistent GPU state across KGD and
KFD workloads.
Cc: Manu Rastogi <manu.rastogi@amd.com> Cc: Christian König <christian.koenig@amd.com> Cc: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Mon, 16 Jun 2025 19:47:13 +0000 (15:47 -0400)]
drm/amdgpu/sdma5.2: init engine reset mutex
Missing the mutex init.
Fixes: 47454f2dc0bf ("drm/amdgpu: Register the new sdma function pointers for sdma_v5_2") Reviewed-by: Jesse Zhang <Jesse.Zhang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Jay Cornwall [Wed, 11 Jun 2025 14:52:14 +0000 (09:52 -0500)]
drm/amdkfd: Fix race in GWS queue scheduling
q->gws is not updated atomically with qpd->mapped_gws_queue. If a
runlist is created between pqm_set_gws and update_queue it will
contain a queue which uses GWS in a process with no GWS allocated.
This will result in a scheduler hang.
Use q->properties.is_gws which is changed while holding the DQM lock.
Signed-off-by: Jay Cornwall <jay.cornwall@amd.com> Reviewed-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Mon, 16 Jun 2025 19:43:55 +0000 (15:43 -0400)]
drm/amdgpu/sdma5: init engine reset mutex
Missing the mutex init.
Fixes: e56d4bf57fab ("drm/amdgpu/: drm/amdgpu: Register the new sdma function pointers for sdma_v5_0") Reviewed-by: Jesse Zhang <Jesse.Zhang@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Mon, 2 Jun 2025 15:31:52 +0000 (11:31 -0400)]
drm/amdgpu: switch job hw_fence to amdgpu_fence
Use the amdgpu fence container so we can store additional
data in the fence. This also fixes the start_time handling
for MCBP since we were casting the fence to an amdgpu_fence
and it wasn't.
Fixes: 3f4c175d62d8 ("drm/amdgpu: MCBP based on DRM scheduler (v9)") Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Frank Min [Wed, 21 May 2025 08:46:27 +0000 (16:46 +0800)]
drm/amdgpu: reclaim psp fw reservation memory region
PSP v14 fw update introduced changes on memory reservation region, according
to the change driver reclaim some non-reserved region.
1. introduce 2 new psp commands to query fw reservation regions
2. add a new reservation region for psp
3. reclaim psp non-used region
Signed-off-by: Frank Min <Frank.Min@amd.com> Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Jesse Zhang [Mon, 16 Jun 2025 11:21:41 +0000 (19:21 +0800)]
drm/amdgpu: Fix SDMA UTC_L1 handling during start/stop sequences
This commit makes two key fixes to SDMA v4.4.2 handling:
1. disable UTC_L1 in sdma_cntl register when stopping SDMA engines
by reading the current value before modifying UTC_L1_ENABLE bit.
2. Ensure UTC_L1_ENABLE is consistently managed by:
- Adding the missing register write when enabling UTC_L1 during start
- Keeping UTC_L1 enabled by default as per hardware requirements
v2: Correct SDMA_CNTL setting (Philip)
Suggested-by: Jonathan Kim <jonathan.kim@amd.com> Signed-off-by: Jesse Zhang <Jesse.Zhang@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Kenneth Feng [Fri, 13 Jun 2025 12:55:03 +0000 (20:55 +0800)]
drm/amd/pm: move the dpm table setting back after featureenablement
move the dpm table setting back after featureenablemend due to dependancy.
For SMUv13.0.6, there is no pptable. Those frequency tables are available
through FW metrics and it needs DPM to be enabled.
Signed-off-by: Kenneth Feng <kenneth.feng@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Mon, 9 Jun 2025 16:13:04 +0000 (12:13 -0400)]
drm/amdgpu/sdma: handle paging queues in amdgpu_sdma_reset_engine()
Need to properly start and stop paging queues if they are present.
This is not an issue today since we don't support a paging queue
on any chips with queue reset.
Fixes: b22659d5d352 ("drm/amdgpu: switch amdgpu_sdma_reset_engine to use the new sdma function pointers") Reviewed-by: Jesse Zhang <Jesse.Zhang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Common modes are added to eDP for compatibility in clone mode, but
not all panels support them. Non-native modes were disabled in the past
but this caused problems because compositors didn't use scaling for non
native modes. Now non-native modes on eDP will enable the scaler by
default.
[How]
Check the connector type. If the connector is eDP avoid adding common
modes.
Reviewed-by: Roman Li <roman.li@amd.com> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/display: Use scaling for non-native resolutions on eDP
[Why]
Common resolutions are added to supported modes to enable compatibility
scenarios that compositors may use to do things like clone displays. There
is no guarantee however that the panel will natively support these modes.
[How]
If the compositor hasn't enabled scaling but a non-native resolution has
been picked for an eDP panel turn the scaler on anyway. This will ensure
compatibility.
Reviewed-by: Roman Li <roman.li@amd.com> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
In DML2 (not DML2.1), DCN35 and DCN351 have the
default value for use_native_pstate_optimization
set to true. The code path where this bit is
false is not used.
[How]
Remove the bit and the corresponding code path
when it is set to false.
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Reviewed-by: Aric Cyr <aric.cyr@amd.com> Signed-off-by: Yan Li <yan.li@amd.com> Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Yan Li [Wed, 21 May 2025 20:02:55 +0000 (16:02 -0400)]
drm/amd/display: apply two different methods to validate modes
[Why]
In DML2, the current method to determine a
mode is supported involves checking the
voltage levels sequentially from the lowest,
until one is found that can support the
mode. It causes cursor lag due to low
performance.
[How]
We apply two methods for mode validation.
1) DC_VALIDATE_MODE_ONLY: only the
maximum voltage level is checked to
determine whether the mode is supported, which
improves performance and eliminate cursor lag.
2) DC_VALIDATE_MODE_AND_STATE_INDEX: when
the optimal voltage level is required, check
the voltage level from the lowest until
a suitable one is found found and returns
its index.
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Reviewed-by: Aric Cyr <aric.cyr@amd.com> Signed-off-by: Yan Li <yan.li@amd.com> Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Emily Deng [Tue, 27 May 2025 03:42:11 +0000 (11:42 +0800)]
drm/amdkfd: Move the process suspend and resume out of full access
For the suspend and resume process, exclusive access is not required.
Therefore, it can be moved out of the full access section to reduce the
duration of exclusive access.
v3:
Move suspend processes before hardware fini.
Remove twice call for bare metal.
v4:
Refine code
Signed-off-by: Emily Deng <Emily.Deng@amd.com> Acked-by: Lijo Lazar <lijo.lazar@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Lijo Lazar [Thu, 12 Jun 2025 12:13:58 +0000 (17:43 +0530)]
drm/amd/pm: Use pointer type for typecheck()
typecheck creates local variables based on the type passed. That could
result in stack frame size warnings like below in certain configs:
drivers/gpu/drm/amd/amdgpu/../pm/swsmu/smu13/smu_v13_0_6_ppt.c:2885:1: error: the frame size of 8304 bytes is larger than 2048 bytes [-Werror=frame-larger-than=]
Checking against the pointer type is sufficient for the purpose of
getting a diagnostic message during build time.
drm/amd: Allow printing Renoir OD SCLK levels without setting dpm to manual
Several other ASICs allow printing OD SCLK levels without setting DPM
control to manual. When OD is disabled it will show the range the
hardware supports. When OD is enabled it will show what values have been
programmed. Adjust Renoir to work the same.
Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Reported-by: Vicki Pfau <vi@endrift.com> Link: https://lore.kernel.org/r/20250609031227.479079-2-superm1@kernel.org Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd: Allow printing VanGogh OD SCLK levels without setting dpm to manual
Several other ASICs allow printing OD SCLK levels without setting DPM
control to manual. When OD is disabled it will show the range the
hardware supports. When OD is enabled it will show what values have
been programmed. Adjust VanGogh to work the same.
Cc: Pierre-Loup A. Griffais <pgriffais@valvesoftware.com> Reported-by: Vicki Pfau <vi@endrift.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Link: https://lore.kernel.org/r/20250609031227.479079-1-superm1@kernel.org Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Kenneth Feng [Thu, 12 Jun 2025 07:57:40 +0000 (15:57 +0800)]
drm/amd/pm: update pcie dpm parameters before smu feature enablement
update pcie dpm parameters before smu feature enablement
Signed-off-by: Kenneth Feng <kenneth.feng@amd.com> Reviewed-by: Yang Wang <kevinyang.wang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Kenneth Feng [Thu, 12 Jun 2025 07:56:11 +0000 (15:56 +0800)]
drm/amd/pm: override pcie dpm parameters only if it is necessary
override pcie dpm parameters only if it is necessary
Signed-off-by: Kenneth Feng <kenneth.feng@amd.com> Reviewed-by: Yang Wang <kevinyang.wang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Jesse Zhang [Wed, 11 Jun 2025 07:07:11 +0000 (15:07 +0800)]
drm/amdgpu: Use logical instance ID for SDMA v4_4_2 queue operations
Simplify SDMA v4_4_2 queue reset and stop operations by:
1. Removing GET_INST(SDMA0) conversion for ring->me
2. Using the logical instance ID (ring->me) directly
3. Maintaining consistent behavior with other SDMA queue operations
This change aligns with the existing queue handling logic where
ring->me already represents the correct instance identifier.
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com> 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>
Jesse Zhang [Wed, 11 Jun 2025 07:02:09 +0000 (15:02 +0800)]
drm/amdgpu: Fix SDMA engine reset with logical instance ID
This commit makes the following improvements to SDMA engine reset handling:
1. Clarifies in the function documentation that instance_id refers to a logical ID
2. Adds conversion from logical to physical instance ID before performing reset
using GET_INST(SDMA0, instance_id) macro
3. Improves error messaging to indicate when a logical instance reset fails
4. Adds better code organization with blank lines for readability
The change ensures proper SDMA engine reset by using the correct physical
instance ID while maintaining the logical ID interface for callers.
V2: Remove harvest_config check and convert directly to physical instance (Lijo)
Suggested-by: Jonathan Kim <jonathan.kim@amd.com> 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>
Lijo Lazar [Tue, 3 Jun 2025 06:13:36 +0000 (11:43 +0530)]
drm/amdgpu: Suspend IH during mode-2 reset
On multi-aid SOCs, there could be a continuous stream of interrupts from
GC after poison consumption. Suspend IH to disable them before doing
mode-2 reset. This avoids conflicts in hardware accesses during
interrupt handlers while a reset is ongoing.
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com> Reviewed-by: Asad Kamal <asad.kamal@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/display: Destroy cached state in complete() callback
[Why]
When the suspend sequence has been aborted after prepare() but
before suspend() the resume() callback never gets called. The PM core
will call complete() when this happens. As the state has been cached
in prepare() it needs to be destroyed in complete() if it's still around.
[How]
Create a helper for destroying cached state and call it both in resume()
and complete() callbacks. If resume has been called the state will be
destroyed and it's a no-op for complete(). If resume hasn't been called
(such as an aborted suspend) then destroy the state in complete().
Fixes: 50e0bae34fa6 ("drm/amd/display: Add and use new dm_prepare_suspend() callback") Reviewed-by: Alex Hung <alex.hung@amd.com> Link: https://lore.kernel.org/r/20250602014432.3538345-4-superm1@kernel.org Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/display: Stop storing failures into adev->dm.cached_state
If drm_atomic_helper_suspend() has failed for any reason, it's stored
in adev->dm.cached_state. This isn't expected because the resume
(or complete()) sequence will attempt to use the stored state to
resume.
complete() callbacks are supposed to handle reversing anything
that occurred during prepare() callbacks. They'll be called on every
power state transition, and will also be called if the sequence is
failed (such as an aborted suspend).
Lijo Lazar [Fri, 6 Jun 2025 06:48:57 +0000 (12:18 +0530)]
drm/amd/pm: Show default gfx clock levels
For SMU v13.0.6 SOCs, always show default clock levels for gfx in
pp_dpm_sclk. Any custom min/max levels set by user will be available in
pp_od_clk_voltage
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com> Reviewed-by: Asad Kamal <asad.kamal@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Frank Min <Frank.Min@amd.com> Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
projects / thirdparty / kernel / linux.git / log
