| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
When GPU suspend is aborted, do the same for dGPU as APU to reset
soc15 asic. Otherwise it may cause following errors:
[ 547.229463] amdgpu 0001:81:00.0: [drm:amdgpu_ring_test_helper [amdgpu]] *ERROR* ring kiq_0.2.1.0 test failed (-110)
[ 555.126827] amdgpu 0000:0a:00.0: [drm:amdgpu_ring_test_helper [amdgpu]] *ERROR* ring kiq_0.2.1.0 test failed (-110)
[ 555.126901] [drm:amdgpu_gfx_enable_kcq [amdgpu]] *ERROR* KCQ enable failed
[ 555.126957] [drm:amdgpu_device_ip_resume_phase2 [amdgpu]] *ERROR* resume of IP block <gfx_v9_4_3> failed -110
[ 555.126959] amdgpu 0000:0a:00.0: amdgpu: amdgpu_device_ip_resume failed (-110).
[ 555.126965] PM: dpm_run_callback(): pci_pm_resume+0x0/0xe0 returns -110
[ 555.126966] PM: Device 0000:0a:00.0 failed to resume async: error -110
This fix has been tested on Mi308X.
Signed-off-by: Jiang Liu <gerry@linux.alibaba.com>
Tested-by: Shuo Liu <shuox.liu@linux.alibaba.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Link: https://lore.kernel.org/r/2462b4b12eb9d025e82525178d568cbaa4c223ff.1736739303.git.gerry@linux.alibaba.com
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Using mmio to do queue reset
v2: Alignment the function with gfx9/gfx9.4.3.
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>
|
|
Using mmio to do queue reset.
v2: Alignment this function with gfx9/gfx9.4.3.
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>
|
|
There could be configs where some UMC instances are harvested. This
information is obtained through discovery data and populated in
umc.active_mask. Avoid reassigning this as AID mask, instead use the
mask directly while iterating through umc instances. This is to avoid
accesses to harvested UMC instances.
v2: fix warning (Alex)
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>
|
|
Set GC 9.5.0 noretry default as 1 for better performance. It can be
changed by the administrator using amdgpu.noretry=0 or by the user using
HSA_XNACK=1 environment variable.
Signed-off-by: Amber Lin <Amber.Lin@amd.com>
Reviewed-by: Harish Kasiviswanathan <Harish.Kasiviwanathan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Harvest table is applied for gfx950.
Signed-off-by: Le Ma <le.ma@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Some chips have a larger VBIOS file so raise the size limit to support
the flashing tool.
Signed-off-by: Shiwu Zhang <shiwu.zhang@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Enable dynamic workload profile switching for gfx12.
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Enable dynamic workload profile switching for gfx11.
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Enable dynamic workload profile switching for gfx10.
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Add helpers to switch the workload profile dynamically when
commands are submitted. This allows us to switch to
the FULLSCREEN3D or COMPUTE profile when work is submitted.
Add a delayed work handler to delay switching out of the
selected profile if additional work comes in. This works
the same as the VIDEO profile for VCN. This lets dynamically
enable workload profiles on the fly and then move back
to the default when there is no work.
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Current kfd driver has its own PASID value for a kfd process and uses it to
locate vm at interrupt handler or mapping between kfd process and vm. That
design is not working when a physical gpu device has multiple spatial
partitions, ex: adev in CPX mode. This patch has kfd driver use same pasid
values that graphic driver generated which is per vm per pasid.
These pasid values are passed to fw/hardware. We do not need change interrupt
handler though more pasid values are used. Also, pasid values at log are
replaced by user process pid; pasid values are not exposed to user. Users see
their process pids that have meaning in user space.
Signed-off-by: Xiaogang Chen <xiaogang.chen@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
RRMT could get dynamically enabled/disabled by PSP firmware. Read the
status from register for reading RRMT status. For VFs, this is not
accessible, hence assume that it's always disabled for now.
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Sathishkumar S <sathishkumar.sundararaju@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
RRMT could get dynamically enabled/disabled by PSP firmware. Read the
status from register for reading RRMT status. For VFs, this is not
accessible, hence assume that it's always disabled for now.
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Sathishkumar S <sathishkumar.sundararaju@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
This initializes PSP IP version 14.0.5.
Signed-off-by: Tim Huang <tim.huang@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
This initializes SMU IP version 14.0.5.
Signed-off-by: Tim Huang <tim.huang@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Enable VCN/JPEG CGPG for ASIC with GFX version 11.5.3.
Signed-off-by: Saleemkhan Jamadar <saleemkhan.jamadar@amd.com>
Signed-off-by: Tim Huang <tim.huang@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
This initializes MMHUB IP version 3.3.2.
Signed-off-by: Tim Huang <tim.huang@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
This initializes NBIO IP version 7.11.2.
Signed-off-by: Tim Huang <tim.huang@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
This initializes SDMA IP version 6.1.3.
Signed-off-by: Tim Huang <tim.huang@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
This initializes GC IP version 11.5.3.
Signed-off-by: Tim Huang <tim.huang@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
It uses the VGADCC bus. DC doesn't use this bus, so it
is safe to add it here.
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
No functional change. Rework the code to allow for
adding some additional i2c buses in conjunction with DC
in the future.
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
This was leftover from when amdgpu was forked from radeon.
The function is empty so drop it.
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Allow the creation of an OEM i2c bus and use the proper
DC helpers for that case.
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
increase max rings to 132 to support all JPEG5_0_1 cores, else
ring_init fails due to ring count exceeding maximum limit.
Signed-off-by: Sathishkumar S <sathishkumar.sundararaju@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
In function psp_init_cap_microcode(), it should bail out when failed to
load firmware, otherwise it may cause invalid memory access.
Fixes: 07dbfc6b102e ("drm/amd: Use `amdgpu_ucode_*` helpers for PSP")
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Jiang Liu <gerry@linux.alibaba.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Bump the driver version for RV/PCO compute stability fix
so mesa can use this check to enable compute queues on
RV/PCO.
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org # 6.12.x
|
|
When mesa started using compute queues more often
we started seeing additional hangs with compute queues.
Disabling gfxoff seems to mitigate that. Manually
control gfxoff and gfx pg with command submissions to avoid
any issues related to gfxoff. KFD already does the same
thing for these chips.
v2: limit to compute
v3: limit to APUs
v4: limit to Raven/PCO
v5: only update the compute ring_funcs
v6: Disable GFX PG
v7: adjust order
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Suggested-by: Błażej Szczygieł <mumei6102@gmail.com>
Suggested-by: Sergey Kovalenko <seryoga.engineering@gmail.com>
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/3861
Link: https://lists.freedesktop.org/archives/amd-gfx/2025-January/119116.html
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org # 6.12.x
|
|
drm_sched_init() has a great many parameters and upcoming new
functionality for the scheduler might add even more. Generally, the
great number of parameters reduces readability and has already caused
one missnaming, addressed in:
commit 6f1cacf4eba7 ("drm/nouveau: Improve variable name in
nouveau_sched_init()").
Introduce a new struct for the scheduler init parameters and port all
users.
Reviewed-by: Liviu Dudau <liviu.dudau@arm.com>
Acked-by: Matthew Brost <matthew.brost@intel.com> # for Xe
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com> # for Panfrost and Panthor
Reviewed-by: Christian Gmeiner <cgmeiner@igalia.com> # for Etnaviv
Reviewed-by: Frank Binns <frank.binns@imgtec.com> # for Imagination
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com> # for Sched
Reviewed-by: Maíra Canal <mcanal@igalia.com> # for v3d
Reviewed-by: Danilo Krummrich <dakr@kernel.org>
Reviewed-by: Lizhi Hou <lizhi.hou@amd.com> # for amdxdna
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20250211111422.21235-2-phasta@kernel.org
|
|
Bring rc1 to start the new release dev.
Signed-off-by: Maxime Ripard <mripard@kernel.org>
|
|
Vulkan can't support DCC and Z/S compression on GFX12 without
WRITE_COMPRESS_DISABLE in this commit or a completely different DCC
interface.
AMDGPU_TILING_GFX12_SCANOUT is added because it's already used by userspace.
Cc: stable@vger.kernel.org # 6.12.x
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
change the config of cgcg on gfx12
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>
Cc: stable@vger.kernel.org # 6.12.x
|
|
MES internal will check CP_MES_MSCRATCH_LO/HI register to set scratch
data location during ucode start, driver side need to start the MES
one by one with different setting for each pipe
Signed-off-by: Shaoyun Liu <shaoyun.liu@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Effectively amdgpu.gttsize gets set to ~1/2 of RAM, but that's controlled
by what the TTM page limit is set to. Clarify the kdoc.
Reviewed-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>
|
|
If the parent is NULL, adev->pdev is used to retrieve the PCIe speed and
width, ensuring that the function can still determine these
capabilities from the device itself.
Fixes the below:
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:6193 amdgpu_device_gpu_bandwidth()
error: we previously assumed 'parent' could be null (see line 6180)
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
6170 static void amdgpu_device_gpu_bandwidth(struct amdgpu_device *adev,
6171 enum pci_bus_speed *speed,
6172 enum pcie_link_width *width)
6173 {
6174 struct pci_dev *parent = adev->pdev;
6175
6176 if (!speed || !width)
6177 return;
6178
6179 parent = pci_upstream_bridge(parent);
6180 if (parent && parent->vendor == PCI_VENDOR_ID_ATI) {
^^^^^^
If parent is NULL
6181 /* use the upstream/downstream switches internal to dGPU */
6182 *speed = pcie_get_speed_cap(parent);
6183 *width = pcie_get_width_cap(parent);
6184 while ((parent = pci_upstream_bridge(parent))) {
6185 if (parent->vendor == PCI_VENDOR_ID_ATI) {
6186 /* use the upstream/downstream switches internal to dGPU */
6187 *speed = pcie_get_speed_cap(parent);
6188 *width = pcie_get_width_cap(parent);
6189 }
6190 }
6191 } else {
6192 /* use the device itself */
--> 6193 *speed = pcie_get_speed_cap(parent);
^^^^^^ Then we are toasted here.
6194 *width = pcie_get_width_cap(parent);
6195 }
6196 }
Fixes: 757e8b951ce2 ("drm/amdgpu: cache gpu pcie link width")
Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Suggested-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
commit 26c95e838e63 ("drm/amdgpu: set the VM pointer to NULL in
amdgpu_job_prepare") set job->vm as NULL if there is no fence. It will
cause emit switch buffer be skippen if job->vm set as NULL.
Check job rather than vm could solve this problem.
Fixes: 26c95e838e63 ("drm/amdgpu: set the VM pointer to NULL in amdgpu_job_prepare")
Signed-off-by: Lin.Cao <lincao12@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Combine the platform and GPU caps like we do for PCIe Gen.
This aligns properly with expectations and documentation
for the interface.
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/3820
Reviewed-by: Yang Wang <kevinyang.wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Get the PCIe link with of the device itself (or it's
integrated upstream bridge) and cache that.
v2: fix typo
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/3820
Reviewed-by: Yang Wang <kevinyang.wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
'add ip block' causes a confusion if the blocks are disabled later with
ip_block_mask. Instead change to 'detected' and also add device context.
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Asad Kamal <asad.kamal@amd.com>
Suggested-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Context empty interrupt is enabled for SDMA 4.4.2. Add a handler for
context empty interrupt so that it is disposed of fast, and not
propagated to KFD layer.
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Suggested-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
This commit enables the cleaner shader feature for GFX12.0 and GFX12.0.1
GPUs. The cleaner shader is important for clearing GPU resources such as
Local Data Share (LDS), Vector General Purpose Registers (VGPRs), and
Scalar General Purpose Registers (SGPRs) between workloads.
- This feature ensures that GPU resources are reset between workloads,
preventing data leaks and ensuring accurate computation.
By enabling the cleaner shader, this update enhances the security and
reliability of GPU operations on GFX12.0 hardware.
Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Mark options only meant to be used for debugging as unsafe so that the
kernel is tainted when they are used.
Signed-off-by: Kent Russell <kent.russell@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
FW attestation was disabled on MP0_14_0_{2/3}.
V2:
Move check into is_fw_attestation_support func. (Frank)
Remove DRM_WARN log info. (Alex)
Fix format. (Christian)
Signed-off-by: Gui Chengming <Jack.Gui@amd.com>
Reviewed-by: Frank.Min <Frank.Min@amd.com>
Reviewed-by: Christian König <Christian.Koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
That is needed to enforce isolation between contexts.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
We sometimes have people trying to use debugging options in production
environments.
Mark options only meant to be used for debugging as unsafe so that the
kernel is tainted when they are used.
Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Felix Kuehling <felix.kuehling@amd.com>
Acked-by: Simona Vetter <simona.vetter@ffwll.ch>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Lets use the existing helper instead of peeking into the structure
directly.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Danilo Krummrich <dakr@redhat.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Philipp Stanner <pstanner@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
|
|
Disable gfxoff with the compute workload on gfx12. This is a
workaround for the opencl test failure.
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>
|
|
This commit addresses a circular locking dependency issue within the GFX
isolation mechanism. The problem was identified by a warning indicating
a potential deadlock due to inconsistent lock acquisition order.
- The `amdgpu_gfx_enforce_isolation_ring_begin_use` and
`amdgpu_gfx_enforce_isolation_ring_end_use` functions previously
acquired `enforce_isolation_mutex` and called `amdgpu_gfx_kfd_sch_ctrl`,
leading to potential deadlocks. ie., If `amdgpu_gfx_kfd_sch_ctrl` is
called while `enforce_isolation_mutex` is held, and
`amdgpu_gfx_enforce_isolation_handler` is called while `kfd_sch_mutex` is
held, it can create a circular dependency.
By ensuring consistent lock usage, this fix resolves the issue:
[ 606.297333] ======================================================
[ 606.297343] WARNING: possible circular locking dependency detected
[ 606.297353] 6.10.0-amd-mlkd-610-311224-lof #19 Tainted: G OE
[ 606.297365] ------------------------------------------------------
[ 606.297375] kworker/u96:3/3825 is trying to acquire lock:
[ 606.297385] ffff9aa64e431cb8 ((work_completion)(&(&adev->gfx.enforce_isolation[i].work)->work)){+.+.}-{0:0}, at: __flush_work+0x232/0x610
[ 606.297413]
but task is already holding lock:
[ 606.297423] ffff9aa64e432338 (&adev->gfx.kfd_sch_mutex){+.+.}-{3:3}, at: amdgpu_gfx_kfd_sch_ctrl+0x51/0x4d0 [amdgpu]
[ 606.297725]
which lock already depends on the new lock.
[ 606.297738]
the existing dependency chain (in reverse order) is:
[ 606.297749]
-> #2 (&adev->gfx.kfd_sch_mutex){+.+.}-{3:3}:
[ 606.297765] __mutex_lock+0x85/0x930
[ 606.297776] mutex_lock_nested+0x1b/0x30
[ 606.297786] amdgpu_gfx_kfd_sch_ctrl+0x51/0x4d0 [amdgpu]
[ 606.298007] amdgpu_gfx_enforce_isolation_ring_begin_use+0x2a4/0x5d0 [amdgpu]
[ 606.298225] amdgpu_ring_alloc+0x48/0x70 [amdgpu]
[ 606.298412] amdgpu_ib_schedule+0x176/0x8a0 [amdgpu]
[ 606.298603] amdgpu_job_run+0xac/0x1e0 [amdgpu]
[ 606.298866] drm_sched_run_job_work+0x24f/0x430 [gpu_sched]
[ 606.298880] process_one_work+0x21e/0x680
[ 606.298890] worker_thread+0x190/0x350
[ 606.298899] kthread+0xe7/0x120
[ 606.298908] ret_from_fork+0x3c/0x60
[ 606.298919] ret_from_fork_asm+0x1a/0x30
[ 606.298929]
-> #1 (&adev->enforce_isolation_mutex){+.+.}-{3:3}:
[ 606.298947] __mutex_lock+0x85/0x930
[ 606.298956] mutex_lock_nested+0x1b/0x30
[ 606.298966] amdgpu_gfx_enforce_isolation_handler+0x87/0x370 [amdgpu]
[ 606.299190] process_one_work+0x21e/0x680
[ 606.299199] worker_thread+0x190/0x350
[ 606.299208] kthread+0xe7/0x120
[ 606.299217] ret_from_fork+0x3c/0x60
[ 606.299227] ret_from_fork_asm+0x1a/0x30
[ 606.299236]
-> #0 ((work_completion)(&(&adev->gfx.enforce_isolation[i].work)->work)){+.+.}-{0:0}:
[ 606.299257] __lock_acquire+0x16f9/0x2810
[ 606.299267] lock_acquire+0xd1/0x300
[ 606.299276] __flush_work+0x250/0x610
[ 606.299286] cancel_delayed_work_sync+0x71/0x80
[ 606.299296] amdgpu_gfx_kfd_sch_ctrl+0x287/0x4d0 [amdgpu]
[ 606.299509] amdgpu_gfx_enforce_isolation_ring_begin_use+0x2a4/0x5d0 [amdgpu]
[ 606.299723] amdgpu_ring_alloc+0x48/0x70 [amdgpu]
[ 606.299909] amdgpu_ib_schedule+0x176/0x8a0 [amdgpu]
[ 606.300101] amdgpu_job_run+0xac/0x1e0 [amdgpu]
[ 606.300355] drm_sched_run_job_work+0x24f/0x430 [gpu_sched]
[ 606.300369] process_one_work+0x21e/0x680
[ 606.300378] worker_thread+0x190/0x350
[ 606.300387] kthread+0xe7/0x120
[ 606.300396] ret_from_fork+0x3c/0x60
[ 606.300406] ret_from_fork_asm+0x1a/0x30
[ 606.300416]
other info that might help us debug this:
[ 606.300428] Chain exists of:
(work_completion)(&(&adev->gfx.enforce_isolation[i].work)->work) --> &adev->enforce_isolation_mutex --> &adev->gfx.kfd_sch_mutex
[ 606.300458] Possible unsafe locking scenario:
[ 606.300468] CPU0 CPU1
[ 606.300476] ---- ----
[ 606.300484] lock(&adev->gfx.kfd_sch_mutex);
[ 606.300494] lock(&adev->enforce_isolation_mutex);
[ 606.300508] lock(&adev->gfx.kfd_sch_mutex);
[ 606.300521] lock((work_completion)(&(&adev->gfx.enforce_isolation[i].work)->work));
[ 606.300536]
*** DEADLOCK ***
[ 606.300546] 5 locks held by kworker/u96:3/3825:
[ 606.300555] #0: ffff9aa5aa1f5d58 ((wq_completion)comp_1.1.0){+.+.}-{0:0}, at: process_one_work+0x3f5/0x680
[ 606.300577] #1: ffffaa53c3c97e40 ((work_completion)(&sched->work_run_job)){+.+.}-{0:0}, at: process_one_work+0x1d6/0x680
[ 606.300600] #2: ffff9aa64e463c98 (&adev->enforce_isolation_mutex){+.+.}-{3:3}, at: amdgpu_gfx_enforce_isolation_ring_begin_use+0x1c3/0x5d0 [amdgpu]
[ 606.300837] #3: ffff9aa64e432338 (&adev->gfx.kfd_sch_mutex){+.+.}-{3:3}, at: amdgpu_gfx_kfd_sch_ctrl+0x51/0x4d0 [amdgpu]
[ 606.301062] #4: ffffffff8c1a5660 (rcu_read_lock){....}-{1:2}, at: __flush_work+0x70/0x610
[ 606.301083]
stack backtrace:
[ 606.301092] CPU: 14 PID: 3825 Comm: kworker/u96:3 Tainted: G OE 6.10.0-amd-mlkd-610-311224-lof #19
[ 606.301109] Hardware name: Gigabyte Technology Co., Ltd. X570S GAMING X/X570S GAMING X, BIOS F7 03/22/2024
[ 606.301124] Workqueue: comp_1.1.0 drm_sched_run_job_work [gpu_sched]
[ 606.301140] Call Trace:
[ 606.301146] <TASK>
[ 606.301154] dump_stack_lvl+0x9b/0xf0
[ 606.301166] dump_stack+0x10/0x20
[ 606.301175] print_circular_bug+0x26c/0x340
[ 606.301187] check_noncircular+0x157/0x170
[ 606.301197] ? register_lock_class+0x48/0x490
[ 606.301213] __lock_acquire+0x16f9/0x2810
[ 606.301230] lock_acquire+0xd1/0x300
[ 606.301239] ? __flush_work+0x232/0x610
[ 606.301250] ? srso_alias_return_thunk+0x5/0xfbef5
[ 606.301261] ? mark_held_locks+0x54/0x90
[ 606.301274] ? __flush_work+0x232/0x610
[ 606.301284] __flush_work+0x250/0x610
[ 606.301293] ? __flush_work+0x232/0x610
[ 606.301305] ? __pfx_wq_barrier_func+0x10/0x10
[ 606.301318] ? mark_held_locks+0x54/0x90
[ 606.301331] ? srso_alias_return_thunk+0x5/0xfbef5
[ 606.301345] cancel_delayed_work_sync+0x71/0x80
[ 606.301356] amdgpu_gfx_kfd_sch_ctrl+0x287/0x4d0 [amdgpu]
[ 606.301661] amdgpu_gfx_enforce_isolation_ring_begin_use+0x2a4/0x5d0 [amdgpu]
[ 606.302050] ? srso_alias_return_thunk+0x5/0xfbef5
[ 606.302069] amdgpu_ring_alloc+0x48/0x70 [amdgpu]
[ 606.302452] amdgpu_ib_schedule+0x176/0x8a0 [amdgpu]
[ 606.302862] ? drm_sched_entity_error+0x82/0x190 [gpu_sched]
[ 606.302890] amdgpu_job_run+0xac/0x1e0 [amdgpu]
[ 606.303366] drm_sched_run_job_work+0x24f/0x430 [gpu_sched]
[ 606.303388] process_one_work+0x21e/0x680
[ 606.303409] worker_thread+0x190/0x350
[ 606.303424] ? __pfx_worker_thread+0x10/0x10
[ 606.303437] kthread+0xe7/0x120
[ 606.303449] ? __pfx_kthread+0x10/0x10
[ 606.303463] ret_from_fork+0x3c/0x60
[ 606.303476] ? __pfx_kthread+0x10/0x10
[ 606.303489] ret_from_fork_asm+0x1a/0x30
[ 606.303512] </TASK>
v2: Refactor lock handling to resolve circular dependency (Alex)
- Introduced a `sched_work` flag to defer the call to
`amdgpu_gfx_kfd_sch_ctrl` until after releasing
`enforce_isolation_mutex`.
- This change ensures that `amdgpu_gfx_kfd_sch_ctrl` is called outside
the critical section, preventing the circular dependency and deadlock.
- The `sched_work` flag is set within the mutex-protected section if
conditions are met, and the actual function call is made afterward.
- This approach ensures consistent lock acquisition order.
Fixes: afefd6f24502 ("drm/amdgpu: Implement Enforce Isolation Handler for KGD/KFD serialization")
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>
|
|
amd-drm-next-6.14-2025-01-10:
amdgpu:
- Fix max surface handling in DC
- clang fixes
- DCN 3.5 fixes
- DCN 4.0.1 fixes
- DC CRC fixes
- DML updates
- DSC fixes
- PSR fixes
- DC add some divide by 0 checks
- SMU13 updates
- SR-IOV fixes
- RAS fixes
- Cleaner shader support for gfx10.3 dGPUs
- fix drm buddy trim handling
- SDMA engine reset updates
_ Fix RB bitmap setup
- Fix doorbell ttm cleanup
- Add CEC notifier support
- DPIA updates
- MST fixes
amdkfd:
- Shader debugger fixes
- Trap handler cleanup
- Cleanup includes
- Eviction fence wq fix
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250110172731.2960668-1-alexander.deucher@amd.com
|
Jiang Liu
Jesse.zhang@amd.com
Lijo Lazar
Amber Lin
Le Ma
Shiwu Zhang
Alex Deucher
Xiaogang Chen
Tim Huang
Sathishkumar S
Philipp Stanner
Maxime Ripard
Marek Olšák
Kenneth Feng
Shaoyun Liu
Mario Limonciello
Srinivasan Shanmugam
Lin.Cao
Kent Russell
Gui Chengming
Christian König
Tvrtko Ursulin
Dave Airlie