Changes in scx_qmap.h

efe86f088f48f18c27b648e5724048947f3b7fb4 - Merge drm/drm-next into drm-xe-next

Tue, 14 Jul 2026 15:40:31 +0200

Merge drm/drm-next into drm-xe-nextSync some i915/display changesSigned-off-by: Rodrigo Vivi List of files: /linux/tools/sched_ext/scx_qmap.h

6d177908bad5992e17669030db41caab88041d5e - Merge drm/drm-next into drm-intel-next

Tue, 30 Jun 2026 14:55:43 +0200

Merge drm/drm-next into drm-intel-nextSync with v7.2-rc1.Signed-off-by: Jani Nikula List of files: /linux/tools/sched_ext/scx_qmap.h

827b9aabd3ea3e96d5e48abed9f44dbd1e550d4e - Merge drm/drm-next into drm-misc-next

Tue, 30 Jun 2026 10:16:00 +0200

Merge drm/drm-next into drm-misc-nextBackmerging to get drm-misc-next to v7.2-rc1.Signed-off-by: Thomas Zimmermann List of files: /linux/tools/sched_ext/scx_qmap.h

7dd19adaaac9a161723a24de7c270861063ec6ac - Merge tag 'v7.2-rc1' into drm-rust-next

Sun, 28 Jun 2026 22:39:43 +0200

Merge tag 'v7.2-rc1' into drm-rust-nextMerge v7.2-rc1 into drm-rust-next to start the next cycle.Devres now requires T: 'static, which conflicts with shmem::Object usingDevres> without that bound. Resolve by adding 'staticas a supertrait to DriverObject and DeviceContext, which does notrestrict any current use cases since DriverObject is always implementedon owned, refcounted types and DeviceContext is a sealed marker traitwith only unit-type implementors.If DriverObject (or DeviceContext) ever becomes lifetime-parameterized(e.g. via a GAT for TTM backed objects), the 'static supertrait can berelaxed at that point.Signed-off-by: Danilo Krummrich List of files: /linux/tools/sched_ext/scx_qmap.h

11260c335ec6071af5543aef73000b28f041c124 - Merge tag 'sched_ext-for-7.3' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext

Thu, 20 Aug 2026 20:01:37 +0200

Merge tag 'sched_ext-for-7.3' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_extPull sched_ext updates from Tejun Heo: "Most of this cycle completes the enqueue-path support for hierarchical sub-scheduling, which makes sub-scheduler support feature complete: a root BPF scheduler can now hand a cgroup subtree over to a nested sub-scheduler together with revocable CPU grants, and the sub-scheduler owns all scheduling decisions for its tasks on those CPUs. Development volume was high and a number of changes plugging holes in the new support landed late in the cycle. Also included are core scheduling fixes that were completed too late for the v7.2 release and are routed through this pull request. Sub-scheduler CPU delegation: - Parent schedulers now grant and revoke per-CPU capabilities (enqueueing, preemption, CPU frequency control) on their children, enforced on every path a scheduler can reach a CPU through. Previously only dispatching could be delegated; this lets sub-schedulers fully schedule their CPUs. - Rescue execution: a task whose scheduler doesn't have access to the CPUs the task needs to run on starved until the watchdog ejected the whole scheduler. The kernel now runs such tasks directly on a small bandwidth budget, turning a scheduler-killing failure into bounded degradation. - Cgroup integration: tasks migrating across a sub-scheduler boundary weren't re-homed to the new owner, causing wrong-scheduler scheduling and a use-after-free. Sub-schedulers now take over their cgroup subtree and receive its cgroup callbacks. - Arena objects now cross the kernel/BPF boundary as typed pointer arguments, translated transparently by the BPF tree's new arena argument support, replacing untyped arguments with manual translation. - scx_qmap now demonstrates full hierarchical sub-scheduling. Other fixes and updates: - Robustness improvements: the abort path is now NMI-safe, fixing deadlocks when errors are raised from NMI context and making hardlockup recovery direct. Reenqueue loops that could monopolize a CPU ahead of the watchdog now eject the offending scheduler, and stalls are blamed on the scheduler actually responsible. - Hardening: BPF-writable arena memory is validated before kernel use, and task slice and vtime writes got explicit synchronization rules, closing corruption vectors open to buggy or malicious schedulers. - Core scheduling: sched_ext dispatching can drop the rq lock inside the core-wide pick, which let interleaving selections corrupt each other's state and hard-hang the machine. The selection now restarts when the lock was released. The task ordering callback was also invoked with its arguments swapped, and the default ordering is updated to work across sub-scheduler boundaries. The fixes are marked for stable. - Other fixes headed for stable: a task init leak on fork failure during enable, tooling compat macros that silently failed to detect newer kernels, and a crash on reenqueueing against a destroyed dispatch queue. - Tooling: scx_pair moves off deprecated callbacks, and the deprecated scx_bpf_cpu_rq() kfunc is removed"* tag 'sched_ext-for-7.3' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext: (144 commits) sched_ext: Drop the dead SCX_DEQ_CORE_SCHED_EXEC test in dequeue_task_scx() sched_ext: Make core-sched task ordering hierarchy-aware sched_ext: Use runnable_at for the default core-sched task ordering sched_ext: Fix inverted ops.core_sched_before() invocation sched_ext: Move the config-off sub-cap kfunc stubs into sub.c sched_ext: Rename balance-era identifiers to dispatch terms sched_ext: Drop the stale keep_prev fixup in dispatch_pick() sched_ext: Keep kick_sync waiting on the rq's own CPU sched_ext: Make SCHED_CLASS_EXT select GENERIC_ALLOCATOR sched_ext/scx_flatcg: Fix cvtime true-up on slice expiry sched_ext: Don't BUG_ON a destroyed DSQ in process_deferred_reenq_users sched_ext: Fix scx_bpf_dsq_move_to_local___v2 compat detection sched_ext: Make scx_bpf_events() read the calling scheduler's counters sched_ext: Drop unlocked scx_rq_clock_invalidate() from scx_root_disable() selftests/sched_ext: Fix flaky ddsp failure tests on busy systems selftests/sched_ext: Make numa idle validation race-free sched_ext: Fix scx_bpf_dsq_reenq___compat kfunc extern prototype sched_ext/scx_flatcg: expire cached hweights on weight changes sched_ext: Fix exit_task leak on fork failure during enable sched_ext: fix stale references in doc comments ... List of files: /linux/tools/sched_ext/scx_qmap.h

fab183d632628381b466a41479489541ac0e29a0 - sched_ext: Merge branch 'for-7.3-arena-args' into for-7.3

Mon, 17 Aug 2026 23:20:34 +0200

sched_ext: Merge branch 'for-7.3-arena-args' into for-7.3Pull to receive the __arena argument conversion: 67f1f4a48c24 ("sched_ext: Pass kernel arena pointers to ops_cid callbacks") a8dc810968af ("sched_ext: Convert sub-cap kfuncs to __arena cmask arguments") a05c5b5cb5cf ("sched_ext: Convert scx_bpf_cid_override() to __arena array arguments")along with the bpf-next branch carrying the __arena argument support theydepend on.Conflict in kernel/sched/ext/ext.c between: c384ab8a0b13 ("sched_ext: Move the config-off sub-cap kfunc stubs into sub.c")and: a8dc810968af ("sched_ext: Convert sub-cap kfuncs to __arena cmask arguments")which updated the stubs in their old ext.c location. Resolved by keepingext.c without the stubs and applying the prototype conversion to therelocated stubs in sub.c.Signed-off-by: Tejun Heo List of files: /linux/tools/sched_ext/scx_qmap.h

a05c5b5cb5cfc2c2b27ce05a690dd0af1bcdf099 - sched_ext: Convert scx_bpf_cid_override() to __arena array arguments

Wed, 12 Aug 2026 21:55:33 +0200

sched_ext: Convert scx_bpf_cid_override() to __arena array argumentsscx_bpf_cid_override() predates the cid-form arena transition and takes itsarrays as verifier-checked mem+size buffers, forcing scx_qmap to keep thecpu_to_cid and shard_start arrays in writable bss while the rest of itsstate lives in the arena. Unify on arena arguments before cid-formschedulers start seeing real use.BPF now translates between BPF and kernel arena addresses for __arenaarguments. Take the arrays as __arena arguments, with the counts passed inentries. The counts now size the snapshot copies and are bounds-checkedbefore them.scx_qmap moves the arrays into struct qmap_arena. As the arena is mmapped atload, the loader populates them between load and attach instead of beforeload.The arena argument address translation is currently implemented only onx86-64. Schedulers calling this kfunc load only there for now.Signed-off-by: Tejun Heo List of files: /linux/tools/sched_ext/scx_qmap.h

e158e309cd90249a90cc84e6d29e46a3d5551b7b - sched_ext: scx_qmap - Add rescue support

Mon, 03 Aug 2026 23:02:03 +0200

sched_ext: scx_qmap - Add rescue supportA sched holds only the cids its parent granted and nothing guarantees thatthey cover its tasks' affinities. A task that can run on none of them hasnowhere to go and qmap stalls out: it force-inserts the task onto its firstallowed cid, but the kernel bounces the insert back and the task parks inSHARED_DSQ, which is drained only on self cids it can't run on.Set SCX_ENQ_RESCUE on these inserts so the kernel diverts such tasks to itsrescue path instead of bouncing them. The force-insert covers scheds withand without children and fires on re-enqueues, and the SHARED_DSQ scan onevery dispatch rescues tasks stranded there - the enqueue-time check missesa task whose cids were lost while it was already queued. The wrong-cid faultinjection carries the flag too and doubles as a deterministic rescue-trafficgenerator.-B and -q set the root-only rescue bandwidth and quantum ops knobs. -B 0maps to SCX_RESCUE_DISABLE and turns rescue off kernel-side. Rescue insertsare counted and reported in the hier stats line.Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi List of files: /linux/tools/sched_ext/scx_qmap.h

b20dfde5ec543597f7f178f7b96cf6ebcc50e5ee - sched_ext: Rename the cid-form cgroup ops to cpuctl_*

Sat, 18 Jul 2026 12:04:25 +0200

sched_ext: Rename the cid-form cgroup ops to cpuctl_*Two unrelated things go by "cgroup" in the cid form. Sub-schedulers attachto cgroups, and the cgroup_*() ops deliver cpu controller events. While theops names suggest cgroup2 hierarchy, they actually operate on the cpucontroller.Rename them to cpuctl_* in struct sched_ext_ops_cid, which has no usersoutside scx_qmap yet. The cpu form is deployed ABI and keeps the old names.The layout is unchanged and the kernel keeps calling through the cpu-formunion view.Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi List of files: /linux/tools/sched_ext/scx_qmap.h

01cad830436468bb9bdc900a64ac2c38d152c917 - tools/sched_ext: scx_qmap - Add init fault injection modes

Sat, 18 Jul 2026 10:12:20 +0200

tools/sched_ext: scx_qmap - Add init fault injection modesAdd -J init-fail which makes ops.init_task() fail with -ENOMEM for taskswhose comm starts with "qmfail", and -J cgrp-init-fail which does the samein ops.cgroup_init() for cgroups named "qmfail*".The former exercises the migration veto path: the cgroup.procs write mustfail with the injected errno while the destination sched stays up and thetask stays put. The latter exercises the ownership-return failure path: aparent failing to re-init a returned cgroup leaves it unowned, and moves andset_* ops against it must be skipped instead of dereferencing the missingowner. Matching on "qmfail" names keeps the injecting scheduler's own enableunaffected.Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi List of files: /linux/tools/sched_ext/scx_qmap.h

29ac3ae9abd1af403d839283a392f0639bf7a735 - tools/sched_ext: scx_qmap - Consume cgroup weights through set_weight

Sat, 18 Jul 2026 10:12:20 +0200

tools/sched_ext: scx_qmap - Consume cgroup weights through set_weightWith the set_* ops delivered to the parent's sched, a parent qmap instancenow receives cgroup_set_weight for its child subs' attach points. Updatethe matching sub_sched_ctx weight and redistribute() in-kernel, and dropthe userspace feed_weights() polling. This exercises the knob routing endto end.The self weight is fixed at 100: a cgroup's weight is its parent's knob andnot the scheduler's own business. This drops the self-weight polling and therepartition PROG_RUN poke with it.sub_attach seeds the slot with the cgroup's current weight, read throughbpf_cgroup_from_id(), so a weight set before the sub attaches is picked up.A write racing the attach can still be lost until the next value-changingcpu.weight write. Acceptable for a demo.While at it, add a traced ops.cgroup_move() so tests can observe movedelivery.Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi List of files: /linux/tools/sched_ext/scx_qmap.h

eb00f4a3620ae8f5e15bc1d6a39277e7d76d2feb - tools/sched_ext: scx_qmap - Add sub-sched cap fault injection

Tue, 14 Jul 2026 10:18:44 +0200

tools/sched_ext: scx_qmap - Add sub-sched cap fault injectionAdd a fault-injection mode to the scx_qmap sub-scheduler that deliberatelydispatches one of its own tasks to a cid it does not hold. The kernel capcheck must reject it and re-enqueue with SCX_TASK_REENQ_CAP, so thenr_inject_attempts counter tracks nr_reenq_cap one to one, exercising thedelivery-time cap enforcement.Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi List of files: /linux/tools/sched_ext/scx_qmap.h

e9151ed5c94480c73a15305bde5e69a2c7bf45e4 - tools/sched_ext: scx_qmap - Expand hierarchical sub-scheduling

Tue, 14 Jul 2026 10:18:44 +0200

tools/sched_ext: scx_qmap - Expand hierarchical sub-schedulingsched_ext sub-scheduling began as dispatch delegation only: a parent couldcall into a child cgroup sub-scheduler's ops.dispatch() from its owndispatch path, but could not delegate cpus to the child for enqueue and theother paths. sched_ext has since gained cap-based cid delegation, where aparent grants and revokes a child's per-cid caps. Expand scx_qmap todemonstrate it.scx_qmap can now delegate the cids it holds exclusively, split among itselfand its children by cpu.weight. Each gets the floor of its share asdedicated cids. The leftover from rounding forms a shared pool,round-robined among them as an ENQ_IMMED time-share.This shape is deliberate. Exclusive cids exercise the basic grant and revokeof ownership, and the shared pool exercises time-sharing one cid acrossseveral schedulers. The implemented policy is impractical, but it coversmost of what a practical sub-scheduler would need without overcomplicatingqmap.Delegation nests. A cid a node receives from its parent only as around-robin share stays self-local and is never re-delegated. A node leftwith no exclusive cid, e.g. after its cpus went offline, evicts itschildren.v5: Highpri dispatch masked with self_cids, single-read dispatch cgroup_id, feed_weights race comment. (sashiko AI)v4: Track all idle cids and mask with self_cids at the dispatch pick, dropping the reseed. (sashiko AI)v3: Dispatch IMMED flags, repartition accounting order, partition-input snapshot. (sashiko AI)v2: Use __sync_fetch_and_add() for the shared nr_dsps counter. (sashiko AI)Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi List of files: /linux/tools/sched_ext/scx_qmap.h

30067643bce0d02cbeb512416372e93272e674f3 - sched_ext: Add shard boundaries to scx_bpf_cid_override()

Tue, 14 Jul 2026 10:18:42 +0200

sched_ext: Add shard boundaries to scx_bpf_cid_override()An overridden cid mapping invalidates the auto-generated shard layout, sothe override call has to provide both. Extend scx_bpf_cid_override() with ashard_start[] array that lists the first cid of each shard (starting at 0,strictly increasing, last shard implicitly extends to num_possible_cpus()).A scheduler that wants only custom shards with the auto-generated cidmapping can read the current mapping and pass it back unchanged.Overridden shards can span NUMA nodes, so scx_shard_node[] is rebuilt bymajority count: each shard is assigned to the node that owns the most cpusin it.v2: Snapshot the caller's cpu_to_cid/shard_start arrays before validating. (sashiko AI)Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi List of files: /linux/tools/sched_ext/scx_qmap.h

874fdc0e86e4d32539026c8e9b3999c0523c96d4 - sched_ext: Add ops.init_cids() to finalize the cid layout before init

Tue, 14 Jul 2026 10:18:42 +0200

sched_ext: Add ops.init_cids() to finalize the cid layout before initA cid-form scheduler that calls scx_bpf_cid_override() to install a customcid layout can only do so from ops.init(). Enable-path setup that depends onthe cid layout thus has to run after ops.init(), and ops.init() itself can'tuse anything derived from the final layout, which turned out to be toorestrictive.Add an ops.init_cids() callback dedicated to finalizing the cid layout. Itruns before the rest of the enable-path setup, so the final layout is ineffect for everything that follows including ops.init(), which now runsafter the arena pool and cmask scratch allocations.scx_bpf_cid_override() is restricted to ops.init_cids() at load time. Itsits in a kfunc set gated by SCX_KF_ALLOW_INIT_CIDS, a flag set only on theinit_cids op, so the verifier rejects a call from any other context. Theruntime root-only check is dropped as ops.init_cids() only runs during rootenable.The qmap demo moves its override into a dedicated qmap_init_cids() and,while at it, introduces an enum for the cid override modes instead ofhard-coded integers.Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi List of files: /linux/tools/sched_ext/scx_qmap.h

c89b7a09b7a8c5cb5b055a9509a58370abe97b32 - tools/sched_ext: scx_qmap - Use bare u64/u32/s32 integer types

Fri, 10 Jul 2026 00:06:11 +0200

tools/sched_ext: scx_qmap - Use bare u64/u32/s32 integer typesscx_qmap.c and the shared scx_qmap.h mixed __u64/__u32/__s32 with the baretypedefs that scx/common.h provides. Convert the remaining __-prefixedinteger types to the bare forms for consistency. The struct fields becomebare u64 (uint64_t), so the stats printfs that fed them to %llu now cast tounsigned long long. No functional change.Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi List of files: /linux/tools/sched_ext/scx_qmap.h

0eaed89c18aeedf0898baf2dbf5ff027c6795152 - Merge tag 'timers-v7.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/daniel.lezcano/linux into timers/clocksource

Mon, 17 Aug 2026 10:29:52 +0200

Merge tag 'timers-v7.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/daniel.lezcano/linux into timers/clocksource - Use designated initializers for sh_mtu2, sh_cmt, and sh_tmu, and drop the unused initializer in the platform_device_id table for sh_mtu2 (Uwe Kleine-König) - Remove redundant dev_err()/dev_err_probe() messages when devm_request_*_irq() fails, as the helper already logs an error message (Pan Chuang) - Fix a boot hang on Allwinner D1 when a forced minimum delta is used with the sun4i timer (Felix Yan) - Fix an IRQ leak in the cpuhp_setup_state() error path by freeing the IRQ on failure in the NXP PIT driver (WenTao Liang) - Fix incorrect unmapping of shared MMIO between the clocksource and clockevent drivers. If one of them fails to initialize, the error path unmaps the shared MMIO region, leaving the other driver with an invalid mapping on clps711x (Guangshuo Li) - Make the samsung_pwm driver compatible with PREEMPT_RT by replacing regular spinlocks with raw_spinlock_t in atomic contexts (Marek Szyprowski) - Use __raw_readl() and __raw_writel() instead of ioread32() and iowrite32() to support SWAP_IO_SPACE in the rtl-otto driver (Rustam Adilov) - Fix a missing clk_disable_unprepare() call in the timer initialization error path of the Armada driver (Yuho Choi)Link: https://lore.kernel.org/lkml/75feea31-683d-45a1-87f4-ab045e0152ae@oss.qualcomm.com List of files: /linux/tools/sched_ext/scx_qmap.h

b4d85f863e5a6f1fa01b186001af0f68f4f5239e - Merge branch 'next' into for-linus

Sat, 15 Aug 2026 06:26:07 +0200

Merge branch 'next' into for-linusPrepare input updates for 7.3 merge window. List of files: /linux/tools/sched_ext/scx_qmap.h

76904fccf81936c084faebc73ac72c0919a42941 - Merge tag 'v7.2-rc3' into next

Tue, 14 Jul 2026 04:06:00 +0200

Merge tag 'v7.2-rc3' into nextSync up with mainline to pull in stable fixes to avoid merge conflicts. List of files: /linux/tools/sched_ext/scx_qmap.h

00599d4841790d05401820a96cf7edb193888b00 - Merge drm/drm-fixes into drm-misc-fixes

Mon, 29 Jun 2026 17:58:00 +0200

Merge drm/drm-fixes into drm-misc-fixesPull in tag v7.2-rc1 so that drm-misc-fixes becomes useful again,and drm-misc-next-fixes can be closed.Signed-off-by: Maarten Lankhorst List of files: /linux/tools/sched_ext/scx_qmap.h