Userspace wants to run jobs on a specific sdma ring for verification purposes.
This debugfs entry helps to disable or enable submitting jobs to a specific ring.
This entry is populated only if there are at least two or more cores in the sdma ip.
Signed-off-by: Jesse Zhang <jesse.zhang@amd.com>
Suggested-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Tim Huang <tim.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
compute/gfx may have multiple rings on some hardware.
In some cases, userspace wants to run jobs on a specific ring for validation purposes.
This debugfs entry helps to disable or enable submitting jobs to a specific ring.
This entry is populated only if there are at least two or more cores in the gfx/compute ip.
Signed-off-by: Jesse Zhang <jesse.zhang@amd.com>
Suggested-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Tim Huang <tim.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Under sriov, host driver will save and restore vf pci cfg space during
reset. And during device init, under sriov, pci_restore_state happens after
fullaccess released, and it can have race condition with mmio protection
enable from host side leading to missing interrupts.
So skip amdgpu_device_cache_pci_state for sriov.
Signed-off-by: Victor Zhao <Victor.Zhao@amd.com>
Acked-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
acpi_evaluate_object() may return AE_NOT_FOUND (failure), which
would result in dereferencing buffer.pointer (obj) while being NULL.
Although this case may be unrealistic for the current code, it is
still better to protect against possible bugs.
Bail out also when status is AE_NOT_FOUND.
This fixes 1 FORWARD_NULL issue reported by Coverity
Report: CID 1600951: Null pointer dereferences (FORWARD_NULL)
Signed-off-by: Antonio Quartulli <antonio@mandelbit.com>
Fixes: c9b7c809b8 ("drm/amd: Guard against bad data for ATIF ACPI method")
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Link: https://lore.kernel.org/r/20241031152848.4716-1-antonio@mandelbit.com
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Since 040fdcde28 ("drm/amdgpu: respect the abmlevel module parameter value
if it is set"), the default value for amdgpu.abmlevel was set to -1, or auto.
However, the comment explaining the default value was not updated to reflect
the change (-1, or auto; not -1, or disabled).
Clarify that the default value (-1) means auto.
Fixes: 040fdcde28 ("drm/amdgpu: respect the abmlevel module parameter value if it is set")
Reported-by: Ruikai Liu <rickliu2000@outlook.com>
Signed-off-by: Mingcong Bai <jeffbai@aosc.io>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Since on modern systems all of vram can be made visible anyways, to
simplify the new implementation, drops tracking how much memory is
visible for now. If this is really needed we can add it back on top of
the new implementation, or just report all the BOs as visible.
Signed-off-by: Yunxiang Li <Yunxiang.Li@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The old behavior reports the resident memory usage for this key and the
documentation say so as well. However this was accidentally changed to
include buffers that was evicted.
Fixes: 04bdba4654 ("drm/amdgpu: Use drm_print_memory_stats helper from fdinfo")
Signed-off-by: Yunxiang Li <Yunxiang.Li@amd.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Fix two issues with memory allocation in amdgpu_discovery_get_nps_info()
for mem_ranges:
- Add a check for allocation failure to avoid dereferencing a null
pointer.
- As suggested by Christophe, use kvcalloc() for memory allocation,
which checks for multiplication overflow.
Additionally, assign the output parameters nps_type and range_cnt after
the kvcalloc() call to prevent modifying the output parameters in case
of an error return.
Fixes: b194d21b9b ("drm/amdgpu: Use NPS ranges from discovery table")
Suggested-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Li Huafei <lihuafei1@huawei.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Add messages to make it clear when a per ring reset
happens. This is helpful for debugging and aligns with
other reset methods.
v2: add ring name in success/fail messages (Lijo)
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Kent Russell <kent.russell@amd.com> (v1)
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Make sure KFD gets a turn when serializing access to
the GC IP. Currently non-KFD jobs can starve KFD if they
submit often enough. This patch prevents that by stalling
non-KFD if its time period has elapsed.
v2: fix units
v3: check enablement properly
Acked-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
fdget() is the first thing done in scope, all matching fdput() are
immediately followed by leaving the scope.
Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
drm-misc-next for v6.13:
All of the previous pull request, with MORE!
Core Changes:
- Update documentation for scheduler start/stop and job init.
- Add dedede and sm8350-hdk hardware to ci runs.
Driver Changes:
- Small fixes and cleanups to panfrost, omap, nouveau, ivpu, zynqmp, v3d,
panthor docs, and leadtek-ltk050h3146w.
- Crashdump support for qaic.
- Support DP compliance in zynqmp.
- Add Samsung S6E88A0-AMS427AP24 panel.
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/deeef745-f3fb-4e85-a9d0-e8d38d43c1cf@linux.intel.com
Separated xgmi ta is required for specific APU, and driver needs
parse the ta binary properly with aux xgmi ta packed.
v2: make the check function more generic (Lijo)
Signed-off-by: Le Ma <le.ma@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
To check the status of S3 suspend completion,
use the PM core pm_suspend_global_flags bit(1)
to detect S3 abort events. Therefore, clean up
the AMDGPU driver's private flag suspend_complete.
Signed-off-by: Prike Liang <Prike.Liang@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
In the normal S3 entry, the TOS cycle counter is not
reset during BIOS execution the _S3 method, so it doesn't
determine whether the _S3 method is executed exactly.
Howerver, the PM core performs the S3 suspend will set the
PM_SUSPEND_FLAG_FW_RESUME bit if all the devices suspend
successfully. Therefore, drivers can check the
pm_suspend_global_flags bit(1) to detect the S3 suspend
abort event.
Fixes: 6704dbf719 ("drm/amdgpu: update suspend status for aborting from deeper suspend")
Signed-off-by: Prike Liang <Prike.Liang@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Volatile only prevents the compiler from re-ordering reads and writes.
Since we always only modify the ring buffer from one CPU thread and have
an explicit barrier before signaling the HW this should have no effect at
all and just prevents compiler optimisations.
While at it drop the local variables as well.
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>
If a 2nd fault comes in before the 1st is handled, the 1st fault will
clear out the FAULT STATUS registers before the 2nd fault is handled.
Thus we get a lot of zeroes. If status=0, just skip the L2 fault status
information, to avoid confusion of why some VM fault status prints in
dmesg are all zeroes.
Signed-off-by: Kent Russell <kent.russell@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Host drivers can create partial hives per guest by disabling xgmi sharing
between certain peers in the main hive.
Typically, these partial hives are fully connected per guest session.
In the event that the host makes a mistake by adding a non-shared node
to a guest session, have the KFD reflect sharing disabled by severing
the IO link.
Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Tested-by: James Yao <yiqing.yao@amd.com>
Reviewed-by: Harish Kasiviswanathan <harish.kasiviswanathan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
On a hive, NPS request is placed by the first one for all devices in the
hive. If the request fails, mark the mode as UNKNOWN so that subsequent
devices on unload don't request it. Also, fix the mutex double lock
issue in error condition, should have been mutex_unlock.
Fixes: ee52489d12 ("drm/amdgpu: Place NPS mode request on unload")
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Rajneesh Bhardwaj <rajneesh.bhardwaj@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
On nbio v7.4, ras controller interrupt and athub
interrupt are generated after injecting UE to PCIE,
but gpu reset only needs to be triggered once.
Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
If a BIOS provides bad data in response to an ATIF method call
this causes a NULL pointer dereference in the caller.
```
? show_regs (arch/x86/kernel/dumpstack.c:478 (discriminator 1))
? __die (arch/x86/kernel/dumpstack.c:423 arch/x86/kernel/dumpstack.c:434)
? page_fault_oops (arch/x86/mm/fault.c:544 (discriminator 2) arch/x86/mm/fault.c:705 (discriminator 2))
? do_user_addr_fault (arch/x86/mm/fault.c:440 (discriminator 1) arch/x86/mm/fault.c:1232 (discriminator 1))
? acpi_ut_update_object_reference (drivers/acpi/acpica/utdelete.c:642)
? exc_page_fault (arch/x86/mm/fault.c:1542)
? asm_exc_page_fault (./arch/x86/include/asm/idtentry.h:623)
? amdgpu_atif_query_backlight_caps.constprop.0 (drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c:387 (discriminator 2)) amdgpu
? amdgpu_atif_query_backlight_caps.constprop.0 (drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c:386 (discriminator 1)) amdgpu
```
It has been encountered on at least one system, so guard for it.
Fixes: d38ceaf99e ("drm/amdgpu: add core driver (v4)")
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit c9b7c809b8)
Cc: stable@vger.kernel.org
VCN shared memory is in framebuffer and there are some flags initialized
during sw_init. Ideally, such programming should be during hw_init.
Make sure the flags are saved during reset on initialization since that
reset will affect frame buffer region. For clarity, separate it out to
another function.
Fixes: 1e4acf4d93 ("drm/amdgpu: Add reset on init handler for XGMI")
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reported-by: Hao Zhou <hao.zhou@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Port this change to vega20_ih.c:
commit afbf7955ff ("drm/amdgpu: clear RB_OVERFLOW bit when enabling interrupts")
Original commit message:
"Why:
Setting IH_RB_WPTR register to 0 will not clear the RB_OVERFLOW bit
if RB_ENABLE is not set.
How to fix:
Set WPTR_OVERFLOW_CLEAR bit after RB_ENABLE bit is set.
The RB_ENABLE bit is required to be set, together with
WPTR_OVERFLOW_ENABLE bit so that setting WPTR_OVERFLOW_CLEAR bit
would clear the RB_OVERFLOW."
Signed-off-by: Victor Lu <victorchengchi.lu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Before making a function call to resume, validate
the function pointer like we do in sw_init.
Use the helper function amdgpu_ip_block_resume where
same checks and calls are repeated.
Signed-off-by: Sunil Khatri <sunil.khatri@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Before making a function call to suspend, validate
the function pointer like we do in sw_init.
Use the helper function amdgpu_ip_block_suspend where
same checks and calls are repeated.
Signed-off-by: Sunil Khatri <sunil.khatri@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This commit adds the cleaner shader microcode for GFX9.4.2 GPUs. The
cleaner shader is a piece of GPU code that is used to clear or
initialize certain GPU resources, such as Local Data Share (LDS), Vector
General Purpose Registers (VGPRs), and Scalar General Purpose Registers
(SGPRs).
Clearing these resources is important for ensuring data isolation
between different workloads running on the GPU. Without the cleaner
shader, residual data from a previous workload could potentially be
accessed by a subsequent workload, leading to data leaks and incorrect
computation results.
The cleaner shader microcode is represented as an array of 32-bit words
(`gfx_9_4_2_cleaner_shader_hex`). This array is the binary
representation of the cleaner shader code, which is written in a
low-level GPU instruction set.
Also, this patch updates the `gfx_v9_0_sw_init` function to initialize
the cleaner shader if the MEC firmware version is 88 or higher. It sets
the `cleaner_shader_ptr` and `cleaner_shader_size` to the appropriate
values and attempts to initialize the cleaner shader.
When the cleaner shader feature is enabled, the AMDGPU driver loads this
array into a specific location in the GPU memory. The GPU then reads
this memory location to fetch and execute the cleaner shader
instructions.
The cleaner shader is executed automatically by the GPU at the end of
each workload, before the next workload starts. This ensures that all
GPU resources are in a clean state before the start of each workload.
This change ensures that the GPU memory is properly cleared between
different processes, preventing data leakage and enhancing security. It
also aligns with the serialization mechanism between KGD and KFD,
ensuring that the GPU state is consistent across different workloads.
Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Suggested-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
There is random data corruption caused by const fill, this is caused by
write compression mode not correctly configured.
So correct compression mode for const fill.
Signed-off-by: Frank Min <Frank.Min@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
In certain cases - ex: when a reset is required on initialization - XCP
manager won't have a valid partition mode. In such cases, use SPX as the
default selected mode for which partition configuration details are
populated.
Fixes: 4ae86dc878 ("drm/amdgpu: Add sysfs nodes to get xcp details")
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reported-by: Hao Zhou <hao.zhou@amd.com>
Reviewed-by: Asad Kamal <asad.kamal@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
