The xcp select_sched function does not need to
remain as a soc specific callback. It can be reused
for future products
v2: bypass the function if xcp_mgr is not available (Likun)
v3: Let caller check the availability of xcp mgr (Lijo)
Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Transfer to use function amdgpu_ip_map_init to map ip
instance for aqua_vanjaram instead of operation on
different ASIC.
Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
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.
Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
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>
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>
We need the driver-core fixes that are in 6.16-rc3 into here as well
to build on top of.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
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: 3f4c175d62 ("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>
(cherry picked from commit bf1cd14f9e)
Cc: stable@vger.kernel.org
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>
(cherry picked from commit 375bf56465)
Cc: stable@vger.kernel.org
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>
(cherry picked from commit 3bab282dfe)
Cc: stable@vger.kernel.org
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>
(cherry picked from commit 5efa6217c2)
Cc: stable@vger.kernel.org
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: 3f4c175d62 ("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>
xgmi doesn't have discrete max speeds defined. Speed numbers can be
arbitrary based on SOC. Deprecate the enum.
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>
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>
when amdgpu_bad_page_threshold == -1 or -2, driver will issue a warning
message when threshold is reached and continue runtime services.
Signed-off-by: ganglxie <ganglxie@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
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>
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: b22659d5d3 ("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>
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>
Implement soft reset engine callback for SDMA 4.4.x IPs. This avoids IP
version check in generic implementation.
V2: Correct physical instance ID calculation in soft_reset_engine (Jesse)
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>
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>
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>
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>
Add debug mask to disable kernel logs of RAS correctable errors,
including both ACA and CE error counter kernel messages.
Signed-off-by: Xiang Liu <xiang.liu@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Once RAS errors are cleared with appropriate recovery mechanism, clear
reset flags also from RAS context. Otherwise, stale flag values could
affect the subsequent RAS reset handling on the device.
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
It doesn't work reliably and we have soft recover and
full adapter reset so drop this.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
It doesn't work reliably and we have soft recover and
full adapter reset so drop this.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
It doesn't work reliably and we have soft recover and
full adapter reset so drop this.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The KFD currently bars a compute partition mode switch while a KFD
process exists.
Since cgroup excluded devices remain excluded for the lifetime of a KFD
process and user space is able to mode switch single devices, allow
users to mode switch a device with any running process that has been
cgroup excluded from this device.
Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Harish Kasiviswanathan <harish.kasiviswanathan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
clear pa and mca record counter when resetting eeprom, so that
ras_num_bad_pages can be calculated correctly
Signed-off-by: ganglxie <ganglxie@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
IH is not working after switching a new gpu index for the first time.
During VM resume, QEMU programming of VF MSIX table (register GFXMSIX_VECT0_ADDR_LO)
may not work.The access could be blocked by nBIF protection as VF isn't in
exclusive access mode. Exclusive access is enabled now, disable/enable MSIX
so that QEMU reprograms MSIX table.
call amdgpu_restore_msix on resume to restore msix table.
Signed-off-by: Samuel Zhang <guoqing.zhang@amd.com>
Acked-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When switching to new GPU index after hibernation and then resume,
VRAM offset of each VRAM BO will be changed, and the cached gpu
addresses needed to updated.
This is to enable pdb0 and switch to use pdb0-based virtual gpu
address by default in amdgpu_bo_create_reserved(). since the virtual
addresses do not change, this can avoid the need to update all
cached gpu addresses all over the codebase.
Signed-off-by: Emily Deng <Emily.Deng@amd.com>
Signed-off-by: Samuel Zhang <guoqing.zhang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
