Changes in Makefile

d781d1b78acf547bafeb1592e8ba4020dce515c6 - selftests/sched_ext: Check the cmask cid-form ops.enable() receives

Sat, 19 Sep 2026 16:08:20 +0200

selftests/sched_ext: Check the cmask cid-form ops.enable() receivescid-form ops.enable() now hands the task's cmask to the scheduler andset_cmask() repeats it right after. Add a cid-form selftest that checks bothagainst p->cpus_ptr, that they match each other, that the initialset_cmask() lands before set_weight() and before the task first becomesrunnable, and that set_cmask() never precedes enable(), across class-switchenables, fork-path enables and live affinity changes.v2: Mismatch details returned through a caller-local struct instead ofglobals, alloc_words validated in the header check, loop bounded by nr_cidsdirectly (Andrea Righi).Signed-off-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile

9ec7ba20c97d92fa59b351832aeeb934b3b16177 - selftests/sched_ext: Test that ops.dequeue() can iterate the consumed DSQ

Thu, 17 Sep 2026 09:52:42 +0200

selftests/sched_ext: Test that ops.dequeue() can iterate the consumed DSQAdd a scheduler whose ops.dequeue() iterates the source user DSQ withbpf_iter_scx_dsq. The iteration takes the DSQ's raw spinlock; on akernel that runs ops.dequeue() while the consume path still holds thatlock, the first task consumed self-deadlocks the CPU with IRQsdisabled. The watchdog cannot recover from that state, so on anunfixed kernel this test wedges the system instead of failing cleanly.On a fixed kernel the scheduler runs clean and the test passes.Signed-off-by: fangqiurong Signed-off-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile

cfe950d79f524e72bc263f2b153a6d905a75a794 - selftests/sched_ext: Verify nohz_full tick behavior

Wed, 08 Jul 2026 09:46:49 +0200

selftests/sched_ext: Verify nohz_full tick behaviorFinite-slice EXT tasks need the periodic scheduler tick to expire theirslices even when nohz_full is enabled.Add a regression test that selects a nohz_full CPU and exercises bothinfinite-to-finite and finite-to-finite slice transitions across an idleinterval. For each finite task, verify that its ops.tick() callback isinvoked.Skip the test when an allowed nohz_full CPU and a separate housekeepingCPU are not available.Signed-off-by: Andrea Righi Signed-off-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile

5897ca15d2c444af95eaae5f0a384401765afa00 - selftests/sched_ext: Add non_scx_kfunc_deny test

Mon, 20 Apr 2026 11:28:48 +0200

selftests/sched_ext: Add non_scx_kfunc_deny testVerify that the BPF verifier rejects a non-SCX struct_ops program(tcp_congestion_ops) that attempts to call an SCX kfunc (scx_bpf_kick_cpu).The test expects the load to fail with -EACCES from scx_kfunc_context_filter.Signed-off-by: Cheng-Yang Chou Signed-off-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile

94555ca6d0bb41a7cb3c12ad2a7cacb6fdbdca0b - Merge branch 'for-7.0-fixes' into for-7.1

Mon, 30 Mar 2026 20:52:33 +0200

Merge branch 'for-7.0-fixes' into for-7.1Conflict in kernel/sched/ext.c init_sched_ext_class() between: 415cb193bb97 ("sched_ext: Fix SCX_KICK_WAIT deadlock by deferring wait to balance callback")which adds cpus_to_sync cpumask allocation, and: 84b1a0ea0b7c ("sched_ext: Implement scx_bpf_dsq_reenq() for user DSQs") 8c1b9453fde6 ("sched_ext: Convert deferred_reenq_locals from llist to regular list")which add deferred_reenq init code at the same location. Both areindependent additions. Include both.Signed-off-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile

090d34f0f0285124452373225bcc520a31e305e4 - selftests/sched_ext: Add cyclic SCX_KICK_WAIT stress test

Sun, 29 Mar 2026 01:18:56 +0100

selftests/sched_ext: Add cyclic SCX_KICK_WAIT stress testAdd a test that creates a 3-CPU kick_wait cycle (A->B->C->A). A BPFscheduler kicks the next CPU in the ring with SCX_KICK_WAIT on everyenqueue while userspace workers generate continuous scheduling churn viasched_yield(). Without the preceding fix, this hangs the machine within seconds.Signed-off-by: Tejun Heo Reviewed-by: Christian Loehle Tested-by: Christian Loehle List of files: /linux/tools/testing/selftests/sched_ext/Makefile

c6f99d0ecc900ceea27eeaba6c431ae97e7b5599 - Revert "selftests/sched_ext: Add tests for SCX_ENQ_IMMED and scx_bpf_dsq_reenq()"

Fri, 27 Mar 2026 01:45:37 +0100

Revert "selftests/sched_ext: Add tests for SCX_ENQ_IMMED and scx_bpf_dsq_reenq()"This reverts commit c50dcf533149.The tests are superficial, likely AI-generated slop, and flaky. Theydon't add actual value and just churn the selftests.Signed-off-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile

c50dcf533149e313a61d483769eb48682a1b0cdd - selftests/sched_ext: Add tests for SCX_ENQ_IMMED and scx_bpf_dsq_reenq()

Sun, 22 Mar 2026 08:35:33 +0100

selftests/sched_ext: Add tests for SCX_ENQ_IMMED and scx_bpf_dsq_reenq()Add three selftests covering features introduced in v7.1:- dsq_reenq: Verify scx_bpf_dsq_reenq() on user DSQs triggers ops.enqueue() with SCX_ENQ_REENQ and SCX_TASK_REENQ_KFUNC in p->scx.flags.- enq_immed: Verify SCX_OPS_ALWAYS_ENQ_IMMED slow path where tasks dispatched to a busy CPU's local DSQ are re-enqueued through ops.enqueue() with SCX_TASK_REENQ_IMMED.- consume_immed: Verify SCX_ENQ_IMMED via the consume path using scx_bpf_dsq_move_to_local___v2() with explicit SCX_ENQ_IMMED.All three tests skip gracefully on kernels that predate the requiredfeatures by checking availability via __COMPAT_has_ksym() /__COMPAT_read_enum() before loading.Signed-off-by: zhidao su Signed-off-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile

32e940f2bd3b16551f23ea44be47f6f5d1746d64 - Merge branch 'for-7.0-fixes' into for-7.1

Fri, 06 Mar 2026 18:46:32 +0100

Merge branch 'for-7.0-fixes' into for-7.1To prepare for hierarchical scheduling patchset which will cause multipleconflicts otherwise.Signed-off-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile

01a867c2e090cb440c8f27158e8650c8ddefec8e - selftests/sched_ext: Add -fms-extensions to bpf build flags

Tue, 03 Mar 2026 08:23:16 +0100

selftests/sched_ext: Add -fms-extensions to bpf build flagsSimilar to commit 835a50753579 ("selftests/bpf: Add -fms-extensions tobpf build flags") and commit 639f58a0f480 ("bpftool: Fix build warningsdue to MS extensions")Fix "declaration does not declare anything" warning by using-fms-extensions and -Wno-microsoft-anon-tag flags to build bpf programsthat #include "vmlinux.h"Signed-off-by: Zhao Mengmeng Reviewed-by: Andrea Righi Signed-off-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile

658ad2259b3e95aea21e548f7ca3440f620bf95f - selftests/sched_ext: Add test to validate ops.dequeue() semantics

Sat, 21 Feb 2026 21:02:20 +0100

selftests/sched_ext: Add test to validate ops.dequeue() semanticsAdd a new kselftest to validate that the new ops.dequeue() semanticswork correctly for all task lifecycle scenarios, including thedistinction between terminal DSQs (where BPF scheduler is done with thetask), user DSQs (where BPF scheduler manages the task lifecycle) andBPF data structures, regardless of which event performs the dispatch.The test validates the following scenarios: - From ops.select_cpu(): - scenario 0 (local DSQ): tasks dispatched to the local DSQ bypass the BPF scheduler entirely; they never enter BPF custody, so ops.dequeue() is not called, - scenario 1 (global DSQ): tasks dispatched to SCX_DSQ_GLOBAL also bypass the BPF scheduler, like the local DSQ; ops.dequeue() is not called, - scenario 2 (user DSQ): tasks dispatched to user DSQs from ops.select_cpu(): tasks enter BPF scheduler's custody with full enqueue/dequeue lifecycle tracking and state machine validation, expects 1:1 enqueue/dequeue pairing, - From ops.enqueue(): - scenario 3 (local DSQ): same behavior as scenario 0, - scenario 4 (global DSQ): same behavior as scenario 1, - scenario 5 (user DSQ): same behavior as scenario 2, - scenario 6 (BPF internal queue): tasks are stored in a BPF queue from ops.enqueue() and consumed from ops.dispatch(); similarly to scenario 5, tasks enter BPF scheduler's custody with full lifecycle tracking and 1:1 enqueue/dequeue validation.This verifies that: - terminal DSQ dispatch (local, global) don't trigger ops.dequeue(), - tasks dispatched to user DSQs, either from ops.select_cpu() or ops.enqueue(), enter BPF scheduler's custody and have exact 1:1 enqueue/dequeue pairing, - tasks stored to internal BPF data structures from ops.enqueue() enter BPF scheduler's custody and have exact 1:1 enqueue/dequeue pairing, - dispatch dequeues have no flags (normal workflow), - property change dequeues have the %SCX_DEQ_SCHED_CHANGE flag set, - no duplicate enqueues or invalid state transitions are happening.Cc: Tejun Heo Cc: Emil Tsalapatis Cc: Kuba Piecuch Reviewed-by: Daniel Jordan Signed-off-by: Andrea Righi Signed-off-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile

dd6a37e8faa723c680cb8615efa5b042691b927f - selftests/sched_ext: Add test for DL server total_bw consistency

Mon, 26 Jan 2026 10:59:05 +0100

selftests/sched_ext: Add test for DL server total_bw consistencyAdd a new kselftest to verify that the total_bw value in/sys/kernel/debug/sched/debug remains consistent across all CPUsunder different sched_ext BPF program states:1. Before a BPF scheduler is loaded2. While a BPF scheduler is loaded and active3. After a BPF scheduler is unloadedThe test runs CPU stress threads to ensure DL server bandwidthvalues stabilize before checking consistency. This helps catchpotential issues with DL server bandwidth accounting duringsched_ext transitions.Co-developed-by: Andrea Righi Signed-off-by: Andrea Righi Signed-off-by: Joel Fernandes Signed-off-by: Peter Zijlstra (Intel) Tested-by: Christian Loehle Link: https://patch.msgid.link/20260126100050.3854740-8-arighi@nvidia.com List of files: /linux/tools/testing/selftests/sched_ext/Makefile

be621a76341caa911ff98175114ff072618d7d4a - selftests/sched_ext: Add test for sched_ext dl_server

Mon, 26 Jan 2026 10:59:04 +0100

selftests/sched_ext: Add test for sched_ext dl_serverAdd a selftest to validate the correct behavior of the deadline serverfor the ext_sched_class.Co-developed-by: Joel Fernandes Signed-off-by: Joel Fernandes Signed-off-by: Andrea Righi Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Emil Tsalapatis Tested-by: Christian Loehle Link: https://patch.msgid.link/20260126100050.3854740-7-arighi@nvidia.com List of files: /linux/tools/testing/selftests/sched_ext/Makefile

5aff3b319997eb8d0e25b8d2e3b5c88ed07395ac - sched_ext: Add a selftest for scx_bpf_dsq_peek

Wed, 15 Oct 2025 17:50:36 +0200

sched_ext: Add a selftest for scx_bpf_dsq_peekThis commit adds two tests. The first is the most basic unit test:make sure an empty queue peeks as empty, and when we put one elementin the queue, make sure peek returns that element.However, even this simple test is a little complicated by the differentbehavior of scx_bpf_dsq_insert in different calling contexts: - insert is for direct dispatch in enqueue - insert is delayed when called from select_cpuIn this case we split the insert and the peek that verifies theresult between enqueue/dispatch.Note: An alternative would be to call `scx_bpf_dsq_move_to_local` on anempty queue, which in turn calls `flush_dispatch_buf`, in order to flushthe buffered insert. Unfortunately, this is not viable within theenqueue path, as it attempts a voluntary context switch within an RCUread-side critical section.The second test is a stress test that performs many peeks on all DSQsand records the observed tasks.Signed-off-by: Ryan Newton Reviewed-by: Christian Loehle Signed-off-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile

36adf6fe6cbe59a22f64c004b201685cb6f38c42 - selftests/sched_ext: Update test enq_select_cpu_fails

Wed, 21 May 2025 09:34:07 +0200

selftests/sched_ext: Update test enq_select_cpu_failsWith commit 08699d20467b6 ("sched_ext: idle: Consolidate default idleCPU selection kfuncs") allowing scx_bpf_select_cpu_dfl() to be invokedfrom multiple contexts, update the test to validate that the kfuncbehaves correctly when used from ops.enqueue() and via BPF test_run.Additionally, rename the test to enq_select_cpu, dropping "fails" fromthe name, as the logic has now been inverted.Signed-off-by: Andrea Righi Signed-off-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile

01d541baedd74530831c8c98b946622cb7487f30 - selftests/sched_ext: Add test for scx_bpf_select_cpu_and()

Sat, 05 Apr 2025 15:39:25 +0200

selftests/sched_ext: Add test for scx_bpf_select_cpu_and()Add a selftest to validate the behavior of the built-in idle CPUselection policy applied to a subset of allowed CPUs, usingscx_bpf_select_cpu_and().Signed-off-by: Andrea Righi Signed-off-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile

5ae5161820e5c21e85a905a1f47e8f28e5042bf5 - selftests/sched_ext: Add NUMA-aware scheduler test

Wed, 26 Feb 2025 07:50:57 +0100

selftests/sched_ext: Add NUMA-aware scheduler testAdd a selftest to validate the behavior of the NUMA-aware schedulerfunctionalities, including idle CPU selection within nodes, per-nodeDSQs and CPU to node mapping.Signed-off-by: Andrea Righi Signed-off-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile

9b3c11a867a82ebee4e096008014417918b82801 - selftests/sched_ext: add order-only dependency of runner.o on BPFOBJ

Tue, 22 Oct 2024 01:16:52 +0200

selftests/sched_ext: add order-only dependency of runner.o on BPFOBJThe runner.o may start building before libbpf headers are installed,and as a result build fails. This happened a couple of times onlibbpf/ci test jobs: * https://github.com/libbpf/ci/actions/runs/11447667257/job/31849533100 * https://github.com/theihor/libbpf-ci/actions/runs/11445162764/job/31841649552Headers are installed in a recipe for $(BPFOBJ) target, and adding anorder-only dependency should ensure this doesn't happen.Signed-off-by: Ihor Solodrai Signed-off-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile

7941b83bceb374c7e81061e0ebb45cd45f3e3517 - selftests: sched_ext: Add sched_ext as proper selftest target

Tue, 08 Oct 2024 17:35:18 +0200

selftests: sched_ext: Add sched_ext as proper selftest targetThe sched_ext selftests is missing proper cross-compilation support, aproper target entry, and out-of-tree build support.When building the kselftest suite, e.g.: make ARCH=riscv CROSS_COMPILE=riscv64-linux-gnu- \ TARGETS=sched_ext SKIP_TARGETS="" O=/output/foo \ -C tools/testing/selftests installor: make ARCH=arm64 LLVM=1 TARGETS=sched_ext SKIP_TARGETS="" \ O=/output/foo -C tools/testing/selftests installThe expectation is that the sched_ext is included, cross-built, thecorrect toolchain is picked up, and placed into /output/foo.In contrast to the BPF selftests, the sched_ext suite does not usebpftool at test run-time, so it is sufficient to build bpftool for thebuild host only.Add ARCH, CROSS_COMPILE, OUTPUT, and TARGETS support to the sched_extselftest. Also, remove some variables that were unused by theMakefile.Signed-off-by: Björn Töpel Reviewed-by: Shuah Khan Acked-by: David Vernet Tested-by: Mark Brown Reviewed-by: Mark Brown Signed-off-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile

a5db7817af780db6a7f290c79677eff4fd13c5fa - sched_ext: Add selftests

Tue, 18 Jun 2024 22:09:21 +0200

sched_ext: Add selftestsAdd basic selftests.Signed-off-by: David Vernet Acked-by: Tejun Heo List of files: /linux/tools/testing/selftests/sched_ext/Makefile