| Age | Commit message (Collapse) | Author | Lines |
|---|
|
There is a kernel-doc warning for the scheduler:
Warning: drivers/gpu/drm/scheduler/sched_main.c:367 function parameter 'result' not described in 'drm_sched_job_done'
Fix the warning by describing the undocumented error code.
Fixes: 539f9ee4b52a ("drm/scheduler: properly forward fence errors")
Signed-off-by: Yujie Liu <yujie.liu@intel.com>
[phasta: Flesh out commit message]
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20260227082452.1802922-1-yujie.liu@intel.com
|
|
This converts some of the visually simpler cases that have been split
over multiple lines. I only did the ones that are easy to verify the
resulting diff by having just that final GFP_KERNEL argument on the next
line.
Somebody should probably do a proper coccinelle script for this, but for
me the trivial script actually resulted in an assertion failure in the
middle of the script. I probably had made it a bit _too_ trivial.
So after fighting that far a while I decided to just do some of the
syntactically simpler cases with variations of the previous 'sed'
scripts.
The more syntactically complex multi-line cases would mostly really want
whitespace cleanup anyway.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
This is the result of running the Coccinelle script from
scripts/coccinelle/api/kmalloc_objs.cocci. The script is designed to
avoid scalar types (which need careful case-by-case checking), and
instead replace kmalloc-family calls that allocate struct or union
object instances:
Single allocations: kmalloc(sizeof(TYPE), ...)
are replaced with: kmalloc_obj(TYPE, ...)
Array allocations: kmalloc_array(COUNT, sizeof(TYPE), ...)
are replaced with: kmalloc_objs(TYPE, COUNT, ...)
Flex array allocations: kmalloc(struct_size(PTR, FAM, COUNT), ...)
are replaced with: kmalloc_flex(*PTR, FAM, COUNT, ...)
(where TYPE may also be *VAR)
The resulting allocations no longer return "void *", instead returning
"TYPE *".
Signed-off-by: Kees Cook <kees@kernel.org>
|
|
Let's kickstart the v6.20 (7.0?) release cycle.
Signed-off-by: Maxime Ripard <mripard@kernel.org>
|
|
Stop open coding pending job list in drivers. Add pending job list
iterator which safely walks DRM scheduler list asserting DRM scheduler
is stopped.
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Link: https://patch.msgid.link/20251209200039.1366764-3-matthew.brost@intel.com
|
|
In the past, drivers used to reach into scheduler internals—this must
end because it makes it difficult to change scheduler internals, as
driver-side code must also be updated.
Add helpers to check if the scheduler is stopped and to query a job’s
signaled state to avoid reaching into scheduler internals. These are
expected to be used driver-side in recovery and debug flows.
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Link: https://patch.msgid.link/20251209200039.1366764-2-matthew.brost@intel.com
|
|
Linux 6.18-rc6
Backmerge in order to merge msm next
Signed-off-by: Dave Airlie <airlied@redhat.com>
|
|
In the general workqueue implementation, if a user enqueues a work item
using schedule_delayed_work() the used wq is "system_wq" (per-cpu wq)
while queue_delayed_work() use WORK_CPU_UNBOUND (used when a cpu is not
specified). The same applies to schedule_work() that is using system_wq
and queue_work(), that makes use again of WORK_CPU_UNBOUND.
This lack of consistency cannot be addressed without refactoring the API.
For more details see the Link tag below.
This continues the effort to refactor worqueue APIs, which has begun
with the change introducing new workqueues and a new alloc_workqueue flag:
commit 128ea9f6ccfb ("workqueue: Add system_percpu_wq and system_dfl_wq")
commit 930c2ea566af ("workqueue: Add new WQ_PERCPU flag")
Use the successor of system_wq, system_percpu_wq, for the scheduler's
default timeout_wq. system_wq will be removed in a few release cycles.
Link: https://lore.kernel.org/all/20250221112003.1dSuoGyc@linutronix.de/
Suggested-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Marco Crivellari <marco.crivellari@suse.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20251106150121.256367-1-marco.crivellari@suse.com
|
|
The Mesa issue referenced below pointed out a possible deadlock:
[ 1231.611031] Possible interrupt unsafe locking scenario:
[ 1231.611033] CPU0 CPU1
[ 1231.611034] ---- ----
[ 1231.611035] lock(&xa->xa_lock#17);
[ 1231.611038] local_irq_disable();
[ 1231.611039] lock(&fence->lock);
[ 1231.611041] lock(&xa->xa_lock#17);
[ 1231.611044] <Interrupt>
[ 1231.611045] lock(&fence->lock);
[ 1231.611047]
*** DEADLOCK ***
In this example, CPU0 would be any function accessing job->dependencies
through the xa_* functions that don't disable interrupts (eg:
drm_sched_job_add_dependency(), drm_sched_entity_kill_jobs_cb()).
CPU1 is executing drm_sched_entity_kill_jobs_cb() as a fence signalling
callback so in an interrupt context. It will deadlock when trying to
grab the xa_lock which is already held by CPU0.
Replacing all xa_* usage by their xa_*_irq counterparts would fix
this issue, but Christian pointed out another issue: dma_fence_signal
takes fence.lock and so does dma_fence_add_callback.
dma_fence_signal() // locks f1.lock
-> drm_sched_entity_kill_jobs_cb()
-> foreach dependencies
-> dma_fence_add_callback() // locks f2.lock
This will deadlock if f1 and f2 share the same spinlock.
To fix both issues, the code iterating on dependencies and re-arming them
is moved out to drm_sched_entity_kill_jobs_work().
Cc: stable@vger.kernel.org # v6.2+
Fixes: 2fdb8a8f07c2 ("drm/scheduler: rework entity flush, kill and fini")
Link: https://gitlab.freedesktop.org/mesa/mesa/-/issues/13908
Reported-by: Mikhail Gavrilov <mikhail.v.gavrilov@gmail.com>
Suggested-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
[phasta: commit message nits]
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20251104095358.15092-1-pierre-eric.pelloux-prayer@amd.com
|
|
The assembled developers agreed at the X.Org Developers Conference 2025
that the hack added for amdgpu in drm_sched_fini() shall be removed. It
shouldn't be needed by amdgpu anymore.
As it's unclear whether all drivers really follow the life time rule of
entities having to be torn down before their scheduler, it is reasonable
to warn for a while before removing the hack.
Add a warning in drm_sched_fini() that fires if an entity is still
active.
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20251023123429.139848-2-phasta@kernel.org
|
|
drm_sched_run_job_work() contains a comment which explains that an
entity being NULL means that there is no more work to do. It can,
however, also mean that there is work, but the scheduler doesn't have
enough credits to process the jobs right now.
Provide this detail in the comment.
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20251028134602.94125-3-phasta@kernel.org
|
|
The DRM scheduler tracks who last uses an entity and when that process
is killed blocks all further submissions to that entity.
The problem is that we didn't track who initially created an entity, so
when a process accidently leaked its file descriptor to a child and
that child got killed, we killed the parent's entities.
Avoid that and instead initialize the entities last user on entity
creation. This also allows to drop the extra NULL check.
Signed-off-by: David Rosca <david.rosca@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4568
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
CC: stable@vger.kernel.org
Acked-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20251015140128.1470-1-christian.koenig@amd.com
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20251015140128.1470-1-christian.koenig@amd.com
|
|
In a past bug fix it was forgotten that entity access must be protected
by the entity lock. That's a data race and potentially UB.
Move the spin_unlock() to the appropriate position.
Cc: stable@vger.kernel.org # v5.13+
Fixes: ac4eb83ab255 ("drm/sched: select new rq even if there is only one v3")
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://patch.msgid.link/20251022063402.87318-2-phasta@kernel.org
|
|
https://gitlab.freedesktop.org/drm/misc/kernel into drm-next
drm-misc-next for v6.19:
UAPI Changes:
Cross-subsystem Changes:
- fbcon cleanups.
- Make drivers depend on FB_TILEBLITTING instead of selecting it,
and hide FB_MODE_HELPERS.
Core Changes:
- More preparations for rust.
- Throttle dirty worker with vblank
- Use drm_for_each_bridge_in_chain_scoped in drm's bridge code and
assorted fixes.
- Ensure drm_client_modeset tests are enabled in UML.
- Rename ttm_bo_put to ttm_bo_fini, as a further step in removing the
TTM bo refcount.
- Add POST_LT_ADJ_REQ training sequence.
- Show list of removed but still allocated bridges.
- Add a simulated vblank interrupt for hardware without it,
and add some helpers to use them in vkms and hypervdrm.
Driver Changes:
- Assorted small fixes, cleanups and updates to host1x, tegra,
panthor, amdxdna, gud, vc4, ssd130x, ivpu, panfrost, panthor,
sysfb, bridge/sn65dsi86, solomon, ast, tidss.
- Convert drivers from using .round_rate() to .determine_rate()
- Add support for KD116N3730A07/A12, chromebook mt8189, JT101TM023,
LQ079L1SX01, raspberrypi 5" panels.
- Improve reclocking on tegra186+ with nouveau.
- Improve runtime pm in amdxdna.
- Add support for HTX_PAI in imx.
- Use a helper to calculate dumb buffer sizes in most drivers.
Signed-off-by: Simona Vetter <simona.vetter@ffwll.ch>
From: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://lore.kernel.org/r/b412fb91-8545-466a-8102-d89c0f2758a7@linux.intel.com
|
|
When adding dependencies with drm_sched_job_add_dependency(), that
function consumes the fence reference both on success and failure, so in
the latter case the dma_fence_put() on the error path (xarray failed to
expand) is a double free.
Interestingly this bug appears to have been present ever since
commit ebd5f74255b9 ("drm/sched: Add dependency tracking"), since the code
back then looked like this:
drm_sched_job_add_implicit_dependencies():
...
for (i = 0; i < fence_count; i++) {
ret = drm_sched_job_add_dependency(job, fences[i]);
if (ret)
break;
}
for (; i < fence_count; i++)
dma_fence_put(fences[i]);
Which means for the failing 'i' the dma_fence_put was already a double
free. Possibly there were no users at that time, or the test cases were
insufficient to hit it.
The bug was then only noticed and fixed after
commit 9c2ba265352a ("drm/scheduler: use new iterator in drm_sched_job_add_implicit_dependencies v2")
landed, with its fixup of
commit 4eaf02d6076c ("drm/scheduler: fix drm_sched_job_add_implicit_dependencies").
At that point it was a slightly different flavour of a double free, which
commit 963d0b356935 ("drm/scheduler: fix drm_sched_job_add_implicit_dependencies harder")
noticed and attempted to fix.
But it only moved the double free from happening inside the
drm_sched_job_add_dependency(), when releasing the reference not yet
obtained, to the caller, when releasing the reference already released by
the former in the failure case.
As such it is not easy to identify the right target for the fixes tag so
lets keep it simple and just continue the chain.
While fixing we also improve the comment and explain the reason for taking
the reference and not dropping it.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Fixes: 963d0b356935 ("drm/scheduler: fix drm_sched_job_add_implicit_dependencies harder")
Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Closes: https://lore.kernel.org/dri-devel/aNFbXq8OeYl3QSdm@stanley.mountain/
Cc: Christian König <christian.koenig@amd.com>
Cc: Rob Clark <robdclark@chromium.org>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Danilo Krummrich <dakr@kernel.org>
Cc: Philipp Stanner <phasta@kernel.org>
Cc: Christian König <ckoenig.leichtzumerken@gmail.com>
Cc: dri-devel@lists.freedesktop.org
Cc: stable@vger.kernel.org # v5.16+
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20251015084015.6273-1-tvrtko.ursulin@igalia.com
|
|
A rework of the scheduler unit tests removed the done_list. That list is
still mentioned in the mock test header.
Remove that relict.
Fixes: 4576de9b7977 ("drm/sched/tests: Implement cancel_job() callback")
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250919064450.147176-2-phasta@kernel.org
|
|
The drm_sched_job_unschedulable trace point can access
entity->dependency after it was cleared by the callback
installed in drm_sched_entity_add_dependency_cb, causing:
BUG: kernel NULL pointer dereference, address: 0000000000000020
[...]
Workqueue: comp_1.1.0 drm_sched_run_job_work [gpu_sched]
RIP: 0010:trace_event_raw_event_drm_sched_job_unschedulable+0x70/0xd0 [gpu_sched]
To fix this we either need to keep a reference to the fence before
setting up the callbacks, or move the trace_drm_sched_job_unschedulable
calls into drm_sched_entity_add_dependency_cb where they can be
done earlier.
Fixes: 76d97c870f29 ("drm/sched: Trace dependencies for GPU jobs")
Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250901124032.1955-1-pierre-eric.pelloux-prayer@amd.com
|
|
In drm_sched_fini() all entities are marked as stopped - without taking
the appropriate lock, because that would deadlock. That means that
drm_sched_fini() and drm_sched_entity_push_job() can race against each
other.
This should most likely be fixed by establishing the rule that all
entities associated with a scheduler must be torn down first. Then,
however, the locking should be removed from drm_sched_fini() alltogether
with an appropriate comment.
Reported-by: James Flowers <bold.zone2373@fastmail.com>
Link: https://lore.kernel.org/dri-devel/20250720235748.2798-1-bold.zone2373@fastmail.com/
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250813085654.102504-2-phasta@kernel.org
|
|
The header file <linux/atomic.h> is already included on line 8. Remove the
redundant include.
Fixes: 5a99350794fec ("drm/sched: Add scheduler unit testing infrastructure and some basic tests")
Signed-off-by: Liao Yuanhong <liaoyuanhong@vivo.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250819142630.368796-1-liaoyuanhong@vivo.com
|
|
Indirect buffers are an AMD term describing essentialy a job submitted to
the scheduler, just a lower level one. Since scheduler was promoted to be
generic long ago, lets replace those references with jobs.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Danilo Krummrich <dakr@kernel.org>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Philipp Stanner <phasta@kernel.org>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250814133627.2550-1-tvrtko.ursulin@igalia.com
|
|
Updating drm-misc-next to the state of v6.17-rc1. Begins a new release
cycle.
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
|
|
Pull drm updates from Dave Airlie:
"Highlights:
- Intel xe enable Panthor Lake, started adding WildCat Lake
- amdgpu has a bunch of reset improvments along with the usual IP
updates
- msm got VM_BIND support which is important for vulkan sparse memory
- more drm_panic users
- gpusvm common code to handle a bunch of core SVM work outside
drivers.
Detail summary:
Changes outside drm subdirectory:
- 'shrink_shmem_memory()' for better shmem/hibernate interaction
- Rust support infrastructure:
- make ETIMEDOUT available
- add size constants up to SZ_2G
- add DMA coherent allocation bindings
- mtd driver for Intel GPU non-volatile storage
- i2c designware quirk for Intel xe
core:
- atomic helpers: tune enable/disable sequences
- add task info to wedge API
- refactor EDID quirks
- connector: move HDR sink to drm_display_info
- fourcc: half-float and 32-bit float formats
- mode_config: pass format info to simplify
dma-buf:
- heaps: Give CMA heap a stable name
ci:
- add device tree validation and kunit
displayport:
- change AUX DPCD access probe address
- add quirk for DPCD probe
- add panel replay definitions
- backlight control helpers
fbdev:
- make CONFIG_FIRMWARE_EDID available on all arches
fence:
- fix UAF issues
format-helper:
- improve tests
gpusvm:
- introduce devmem only flag for allocation
- add timeslicing support to GPU SVM
ttm:
- improve eviction
sched:
- tracing improvements
- kunit improvements
- memory leak fixes
- reset handling improvements
color mgmt:
- add hardware gamma LUT handling helpers
bridge:
- add destroy hook
- switch to reference counted drm_bridge allocations
- tc358767: convert to devm_drm_bridge_alloc
- improve CEC handling
panel:
- switch to reference counter drm_panel allocations
- fwnode panel lookup
- Huiling hl055fhv028c support
- Raspberry Pi 7" 720x1280 support
- edp: KDC KD116N3730A05, N160JCE-ELL CMN, N116BCJ-EAK
- simple: AUO P238HAN01
- st7701: Winstar wf40eswaa6mnn0
- visionox: rm69299-shift
- Renesas R61307, Renesas R69328 support
- DJN HX83112B
hdmi:
- add CEC handling
- YUV420 output support
xe:
- WildCat Lake support
- Enable PanthorLake by default
- mark BMG as SRIOV capable
- update firmware recommendations
- Expose media OA units
- aux-bux support for non-volatile memory
- MTD intel-dg driver for non-volatile memory
- Expose fan control and voltage regulator in sysfs
- restructure migration for multi-device
- Restore GuC submit UAF fix
- make GEM shrinker drm managed
- SRIOV VF Post-migration recovery of GGTT nodes
- W/A additions/reworks
- Prefetch support for svm ranges
- Don't allocate managed BO for each policy change
- HWMON fixes for BMG
- Create LRC BO without VM
- PCI ID updates
- make SLPC debugfs files optional
- rework eviction rejection of bound external BOs
- consolidate PAT programming logic for pre/post Xe2
- init changes for flicker-free boot
- Enable GuC Dynamic Inhibit Context switch
i915:
- drm_panic support for i915/xe
- initial flip queue off by default for LNL/PNL
- Wildcat Lake Display support
- Support for DSC fractional link bpp
- Support for simultaneous Panel Replay and Adaptive sync
- Support for PTL+ double buffer LUT
- initial PIPEDMC event handling
- drm_panel_follower support
- DPLL interface renames
- allocate struct intel_display dynamically
- flip queue preperation
- abstract DRAM detection better
- avoid GuC scheduling stalls
- remove DG1 force probe requirement
- fix MEI interrupt handler on RT kernels
- use backlight control helpers for eDP
- more shared display code refactoring
amdgpu:
- add userq slot to INFO ioctl
- SR-IOV hibernation support
- Suspend improvements
- Backlight improvements
- Use scaling for non-native eDP modes
- cleaner shader updates for GC 9.x
- Remove fence slab
- SDMA fw checks for userq support
- RAS updates
- DMCUB updates
- DP tunneling fixes
- Display idle D3 support
- Per queue reset improvements
- initial smartmux support
amdkfd:
- enable KFD on loongarch
- mtype fix for ext coherent system memory
radeon:
- CS validation additional GL extensions
- drop console lock during suspend/resume
- bump driver version
msm:
- VM BIND support
- CI: infrastructure updates
- UBWC single source of truth
- decouple GPU and KMS support
- DP: rework I/O accessors
- DPU: SM8750 support
- DSI: SM8750 support
- GPU: X1-45 support and speedbin support for X1-85
- MDSS: SM8750 support
nova:
- register! macro improvements
- DMA object abstraction
- VBIOS parser + fwsec lookup
- sysmem flush page support
- falcon: generic falcon boot code and HAL
- FWSEC-FRTS: fb setup and load/execute
ivpu:
- Add Wildcat Lake support
- Add turbo flag
ast:
- improve hardware generations implementation
imx:
- IMX8qxq Display Controller support
lima:
- Rockchip RK3528 GPU support
nouveau:
- fence handling cleanup
panfrost:
- MT8370 support
- bo labeling
- 64-bit register access
qaic:
- add RAS support
rockchip:
- convert inno_hdmi to a bridge
rz-du:
- add RZ/V2H(P) support
- MIPI-DSI DCS support
sitronix:
- ST7567 support
sun4i:
- add H616 support
tidss:
- add TI AM62L support
- AM65x OLDI bridge support
bochs:
- drm panic support
vkms:
- YUV and R* format support
- use faux device
vmwgfx:
- fence improvements
hyperv:
- move out of simple
- add drm_panic support"
* tag 'drm-next-2025-07-30' of https://gitlab.freedesktop.org/drm/kernel: (1479 commits)
drm/tidss: oldi: convert to devm_drm_bridge_alloc() API
drm/tidss: encoder: convert to devm_drm_bridge_alloc()
drm/amdgpu: move reset support type checks into the caller
drm/amdgpu/sdma7: re-emit unprocessed state on ring reset
drm/amdgpu/sdma6: re-emit unprocessed state on ring reset
drm/amdgpu/sdma5.2: re-emit unprocessed state on ring reset
drm/amdgpu/sdma5: re-emit unprocessed state on ring reset
drm/amdgpu/gfx12: re-emit unprocessed state on ring reset
drm/amdgpu/gfx11: re-emit unprocessed state on ring reset
drm/amdgpu/gfx10: re-emit unprocessed state on ring reset
drm/amdgpu/gfx9.4.3: re-emit unprocessed state on kcq reset
drm/amdgpu/gfx9: re-emit unprocessed state on kcq reset
drm/amdgpu: Add WARN_ON to the resource clear function
drm/amd/pm: Use cached metrics data on SMUv13.0.6
drm/amd/pm: Use cached data for min/max clocks
gpu: nova-core: fix bounds check in PmuLookupTableEntry::new
drm/amdgpu: Replace HQD terminology with slots naming
drm/amdgpu: Add user queue instance count in HW IP info
drm/amd/amdgpu: Add helper functions for isp buffers
drm/amd/amdgpu: Initialize swnode for ISP MFD device
...
|
|
Currently the job free work item will lock sched->job_list_lock first time
to see if there are any jobs, free a single job, and then lock again to
decide whether to re-queue itself if there are more finished jobs.
Since drm_sched_get_finished_job() already looks at the second job in the
queue we can simply add the signaled check and have it return the presence
of more jobs to be freed to the caller. That way the work item does not
have to lock the list again and repeat the signaled check.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Danilo Krummrich <dakr@kernel.org>
Cc: Maíra Canal <mcanal@igalia.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Philipp Stanner <phasta@kernel.org>
Reviewed-by: Maíra Canal <mcanal@igalia.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250716085117.56864-1-tvrtko.ursulin@igalia.com
|
|
When application A submits jobs and application B submits a job with a
dependency on A's fence, the normal flow wakes up the scheduler after
processing each job. However, the optimization in
drm_sched_entity_add_dependency_cb() uses a callback that only clears
dependencies without waking up the scheduler.
When application A is killed before its jobs can run, the callback gets
triggered but only clears the dependency without waking up the scheduler,
causing the scheduler to enter sleep state and application B to hang.
Remove the optimization by deleting drm_sched_entity_clear_dep() and its
usage, ensuring the scheduler is always woken up when dependencies are
cleared.
Fixes: 777dbd458c89 ("drm/amdgpu: drop a dummy wakeup scheduler")
Cc: stable@vger.kernel.org # v4.6+
Signed-off-by: Lin.Cao <lincao12@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250717084453.921097-1-lincao12@amd.com
|
|
The "skip reset" test waits for the timeout handler to run for the
duration of 2 * MOCK_TIMEOUT, and because the mock scheduler opted to
remove the "skip reset" flag once it fires, this gives opportunity for the
timeout handler to run twice. Second time the job will be removed from the
mock scheduler job list and the drm_mock_sched_advance() call in the test
will fail.
Fix it by making the "don't reset" flag persist for the lifetime of the
job and add a new flag to verify that the code path had executed as
expected.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Fixes: 1472e7549f84 ("drm/sched: Add new test for DRM_GPU_SCHED_STAT_NO_HANG")
Cc: Maíra Canal <mcanal@igalia.com>
Cc: Philipp Stanner <phasta@kernel.org>
Reviewed-by: Maíra Canal <mcanal@igalia.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250716084817.56797-1-tvrtko.ursulin@igalia.com
|
|
Add a test to submit a single job against a scheduler with the timeout
configured and verify that if the job is still running, the timeout
handler will skip the reset and allow the job to complete.
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Reviewed-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250714-sched-skip-reset-v6-4-5c5ba4f55039@igalia.com
Signed-off-by: Maíra Canal <mcanal@igalia.com>
|
|
As more KUnit tests are introduced to evaluate the basic capabilities of
the `timedout_job()` hook, the test suite will continue to increase in
duration. To reduce the overall running time of the test suite, decrease
the scheduler's timeout for the timeout tests.
Before this commit:
[15:42:26] Elapsed time: 15.637s total, 0.002s configuring, 10.387s building, 5.229s running
After this commit:
[15:45:26] Elapsed time: 9.263s total, 0.002s configuring, 5.168s building, 4.037s running
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Acked-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250714-sched-skip-reset-v6-3-5c5ba4f55039@igalia.com
Signed-off-by: Maíra Canal <mcanal@igalia.com>
|
|
When the DRM scheduler times out, it's possible that the GPU isn't hung;
instead, a job just took unusually long (longer than the timeout) but is
still running, and there is, thus, no reason to reset the hardware. This
can occur in two scenarios:
1. The job is taking longer than the timeout, but the driver determined
through a GPU-specific mechanism that the hardware is still making
progress. Hence, the driver would like the scheduler to skip the
timeout and treat the job as still pending from then onward. This
happens in v3d, Etnaviv, and Xe.
2. Timeout has fired before the free-job worker. Consequently, the
scheduler calls `sched->ops->timedout_job()` for a job that isn't
timed out.
These two scenarios are problematic because the job was removed from the
`sched->pending_list` before calling `sched->ops->timedout_job()`, which
means that when the job finishes, it won't be freed by the scheduler
though `sched->ops->free_job()` - leading to a memory leak.
To solve these problems, create a new `drm_gpu_sched_stat`, called
DRM_GPU_SCHED_STAT_NO_HANG, which allows a driver to skip the reset. The
new status will indicate that the job must be reinserted into
`sched->pending_list`, and the hardware / driver will still complete that
job.
Reviewed-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250714-sched-skip-reset-v6-2-5c5ba4f55039@igalia.com
Signed-off-by: Maíra Canal <mcanal@igalia.com>
|
|
Among the scheduler's statuses, the only one that indicates an error is
DRM_GPU_SCHED_STAT_ENODEV. Any status other than DRM_GPU_SCHED_STAT_ENODEV
signifies that the operation succeeded and the GPU is in a nominal state.
However, to provide more information about the GPU's status, it is needed
to convey more information than just "OK".
Therefore, rename DRM_GPU_SCHED_STAT_NOMINAL to
DRM_GPU_SCHED_STAT_RESET, which better communicates the meaning of this
status. The status DRM_GPU_SCHED_STAT_RESET indicates that the GPU has
hung, but it has been successfully reset and is now in a nominal state
again.
Reviewed-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250714-sched-skip-reset-v6-1-5c5ba4f55039@igalia.com
Signed-off-by: Maíra Canal <mcanal@igalia.com>
|
|
drm_sched_fini() can leak jobs under certain circumstances.
Warn if that happens.
Acked-by: Danilo Krummrich <dakr@kernel.org>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250710125412.128476-7-phasta@kernel.org
|
|
The scheduler unit tests now provide a new callback, cancel_job(). This
callback gets used by drm_sched_fini() for all still pending jobs to
cancel them.
Implement a new unit test to test this.
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250710125412.128476-6-phasta@kernel.org
|
|
The GPU Scheduler now supports a new callback, cancel_job(), which lets
the scheduler cancel all jobs which might not yet be freed when
drm_sched_fini() runs. Using this callback allows for significantly
simplifying the mock scheduler teardown code.
Implement the cancel_job() callback and adjust the code where necessary.
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250710125412.128476-5-phasta@kernel.org
|
|
Since its inception, the GPU scheduler can leak memory if the driver
calls drm_sched_fini() while there are still jobs in flight.
The simplest way to solve this in a backwards compatible manner is by
adding a new callback, drm_sched_backend_ops.cancel_job(), which
instructs the driver to signal the hardware fence associated with the
job. Afterwards, the scheduler can safely use the established free_job()
callback for freeing the job.
Implement the new backend_ops callback cancel_job().
Suggested-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Link: https://lore.kernel.org/dri-devel/20250418113211.69956-1-tvrtko.ursulin@igalia.com/
Reviewed-by: Maíra Canal <mcanal@igalia.com>
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250710125412.128476-4-phasta@kernel.org
|
|
Extract out two copies of the identical code to
drm_sched_rq_select_entity_rr()'s epilogue to make it smaller and more
readable.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Danilo Krummrich <dakr@kernel.org>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Philipp Stanner <phasta@kernel.org>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
[phasta: commit message]
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250708122121.75689-1-tvrtko.ursulin@igalia.com
|
|
Move work queue allocation into a helper for a more streamlined function
body.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Danilo Krummrich <dakr@kernel.org>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Philipp Stanner <phasta@kernel.org>
Reviewed-by: Maíra Canal <mcanal@igalia.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250704130754.89935-1-tvrtko.ursulin@igalia.com
|
|
Since the drm_mock_scheduler does not have real users in userspace, nor
does it have real hardware or firmware rings, it's not necessary to
signal timedout fences nor free jobs - from a functional standpoint.
Still, the dma_fence framework establishes the hard rule that all fences
must always get signaled.
The unit tests, moreover, should as much as possible represent the
intended usage of the scheduler API.
Furthermore, this later enables simplifying the mock scheduler's
teardown code path.
Make sure timed out hardware fences get signaled with the appropriate
error code.
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Acked-by: Danilo Krummrich <dakr@kernel.org>
Link: https://lore.kernel.org/r/20250605134154.191764-2-phasta@kernel.org
|
|
Fix the compile-time warnings
drivers/gpu/drm/scheduler/sched_entity.c: warning: EXPORT_SYMBOL() is used, but #include <linux/export.h> is missing
drivers/gpu/drm/scheduler/sched_fence.c: warning: EXPORT_SYMBOL() is used, but #include <linux/export.h> is missing
drivers/gpu/drm/scheduler/sched_main.c: warning: EXPORT_SYMBOL() is used, but #include <linux/export.h> is missing
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Fixes: a934a57a42f6 ("scripts/misc-check: check missing #include <linux/export.h> when W=1")
Reviewed-by: André Almeida <andrealmeid@igalia.com>
Cc: Masahiro Yamada <masahiroy@kernel.org>
Cc: Nathan Chancellor <nathan@kernel.org>
Link: https://lore.kernel.org/r/20250612121633.229222-9-tzimmermann@suse.de
|
|
Backmerging to forward to v6.16-rc1
Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
|
|
Pull more drm fixes from Simona Vetter:
"Another small batch of drm fixes, this time with a different baseline
and hence separate.
Drivers:
- ivpu:
- dma_resv locking
- warning fixes
- reset failure handling
- improve logging
- update fw file names
- fix cmdqueue unregister
- panel-simple: add Evervision VGG644804
Core Changes:
- sysfb: screen_info type check
- video: screen_info for relocated pci fb
- drm/sched: signal fence of killed job
- dummycon: deferred takeover fix"
* tag 'drm-fixes-2025-06-06' of https://gitlab.freedesktop.org/drm/kernel:
sysfb: Fix screen_info type check for VGA
video: screen_info: Relocate framebuffers behind PCI bridges
accel/ivpu: Fix warning in ivpu_gem_bo_free()
accel/ivpu: Trigger device recovery on engine reset/resume failure
accel/ivpu: Use dma_resv_lock() instead of a custom mutex
drm/panel-simple: fix the warnings for the Evervision VGG644804
accel/ivpu: Reorder Doorbell Unregister and Command Queue Destruction
accel/ivpu: Use firmware names from upstream repo
accel/ivpu: Improve buffer object logging
dummycon: Trigger redraw when switching consoles with deferred takeover
drm/scheduler: signal scheduled fence when kill job
|
|
There is no need for separate locks for single jobs and the entire
scheduler. The dma_fence context can be protected by the scheduler lock,
allowing for removing the jobs' locks. This simplifies things and
reduces the likelyhood of deadlocks etc.
Replace the jobs' locks with the mock scheduler lock.
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250527101029.56491-2-phasta@kernel.org
|
|
This commit adds a document section in drm-uapi.rst about tracepoints,
and mark the events gpu_scheduler_trace.h as stable uAPI.
The goal is to explicitly state that tools can rely on the fields,
formats and semantics of these events.
Acked-by: Lucas Stach <l.stach@pengutronix.de>
Acked-by: Maíra Canal <mcanal@igalia.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250526125505.2360-10-pierre-eric.pelloux-prayer@amd.com
|
|
Its only purpose was for trace events, but jobs can already be
uniquely identified using their fence.
The downside of using the fence is that it's only available
after 'drm_sched_job_arm' was called which is true for all trace
events that used job.id so they can safely switch to using it.
Suggested-by: Tvrtko Ursulin <tursulin@igalia.com>
Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Arvind Yadav <arvind.yadav@amd.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250526125505.2360-9-pierre-eric.pelloux-prayer@amd.com
|
|
All events now start with the same prefix (drm_sched_job_).
drm_sched_job_wait_dep was misleading because it wasn't waiting
at all. It's now replaced by trace_drm_sched_job_unschedulable,
which is only traced if the job cannot be scheduled.
For moot dependencies, nothing is traced.
Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250526125505.2360-8-pierre-eric.pelloux-prayer@amd.com
|
|
For processes with multiple drm_file instances, the drm_client_id is
the only way to map jobs back to their unique owner.
It's even more useful if drm client_name is set, because now a tool
can map jobs to the client name instead of only having access to
the process name.
Reviewed-by: Christian König <christian.koenig@amd.com>
Acked-by: Philipp Stanner <phasta@kernel.org>
Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250526125505.2360-7-pierre-eric.pelloux-prayer@amd.com
|
|
We can't trace dependencies from drm_sched_job_add_dependency
because when it's called the job's fence is not available yet.
So instead each dependency is traced individually when
drm_sched_entity_push_job is used.
Tracing the dependencies allows tools to analyze the dependencies
between the jobs (previously it was only possible for fences
traced by drm_sched_job_wait_dep).
Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250526125505.2360-6-pierre-eric.pelloux-prayer@amd.com
|
|
A fence uniquely identify a job, so this commits updates the places
where a kernel pointer was used as an identifier by:
"fence=%llu:%llu"
Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250526125505.2360-5-pierre-eric.pelloux-prayer@amd.com
|
|
Since switching the scheduler from using kthreads to workqueues in
commit a6149f039369 ("drm/sched: Convert drm scheduler to use a work
queue rather than kthread") userspace applications cannot determine
the device from the PID of the threads sending the trace events
anymore.
Each queue had its own kthread which had a given PID for the whole
time. So, at least for amdgpu, it was possible to associate a PID
to the hardware queues of each GPU in the system. Then, when a
drm_run_job trace event was received by userspace, the source PID
allowed to associate it back to the correct GPU.
With workqueues this is not possible anymore, so the event needs to
contain the dev_name() to identify the device.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250526125505.2360-4-pierre-eric.pelloux-prayer@amd.com
|
|
This will be used in a later commit to trace the drm client_id in
some of the gpu_scheduler trace events.
This requires changing all the users of drm_sched_job_init to
add an extra parameter.
The newly added drm_client_id field in the drm_sched_fence is a bit
of a duplicate of the owner one. One suggestion I received was to
merge those 2 fields - this can't be done right now as amdgpu uses
some special values (AMDGPU_FENCE_OWNER_*) that can't really be
translated into a client id. Christian is working on getting rid of
those; when it's done we should be able to squash owner/drm_client_id
together.
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250526125505.2360-3-pierre-eric.pelloux-prayer@amd.com
|
|
When an entity from application B is killed, drm_sched_entity_kill()
removes all jobs belonging to that entity through
drm_sched_entity_kill_jobs_work(). If application A's job depends on a
scheduled fence from application B's job, and that fence is not properly
signaled during the killing process, application A's dependency cannot be
cleared.
This leads to application A hanging indefinitely while waiting for a
dependency that will never be resolved. Fix this issue by ensuring that
scheduled fences are properly signaled when an entity is killed, allowing
dependent applications to continue execution.
Signed-off-by: Lin.Cao <lincao12@amd.com>
Reviewed-by: Philipp Stanner <phasta@kernel.org>
Signed-off-by: Christian König <christian.koenig@amd.com>
Link: https://lore.kernel.org/r/20250515020713.1110476-1-lincao12@amd.com
|
|
The kthread header doesn't need to be included anymore. It's a relict
from commit a6149f039369 ("drm/sched: Convert drm scheduler to use a
work queue rather than kthread").
Remove the unneeded includes.
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250314101023.111248-3-phasta@kernel.org
|
