[PATCH] selftests/sched_ext: Fix flaky ddsp failure tests on busy systems

[PATCH] selftests/sched_ext: Fix flaky ddsp failure tests on busy systems

[Michal Blaszczyk]

The ddsp_vtimelocal_fail and ddsp_bogus_dsq_fail tests skip calling
scx_bpf_dsq_insert_vtime() if scx_bpf_pick_idle_cpu() fails to find
an idle CPU (returns -1). On loaded systems, this results in the tests
skipping the very assertions they are meant to verify.

Eliminate this flakiness by falling back to prev_cpu if no idle CPU is
found, ensuring the illegal dispatch operations are unconditionally
attempted and tested.

Fixes: a5db7817af78 ("sched_ext: Add selftests")
Signed-off-by: Michal Blaszczyk <michalblk@google.com>
---
 .../sched_ext/ddsp_bogus_dsq_fail.bpf.c       | 20 +++++++++----------
 .../sched_ext/ddsp_vtimelocal_fail.bpf.c      | 13 ++++++------
 2 files changed, 15 insertions(+), 18 deletions(-)

diff --git a/tools/testing/selftests/sched_ext/ddsp_bogus_dsq_fail.bpf.c b/tools/testing/selftests/sched_ext/ddsp_bogus_dsq_fail.bpf.c
index 6f4c3f5a1c5d..7ef9de7b27eb 100644
--- a/tools/testing/selftests/sched_ext/ddsp_bogus_dsq_fail.bpf.c
+++ b/tools/testing/selftests/sched_ext/ddsp_bogus_dsq_fail.bpf.c
@@ -14,18 +14,16 @@ s32 BPF_STRUCT_OPS(ddsp_bogus_dsq_fail_select_cpu, struct task_struct *p,
 		   s32 prev_cpu, u64 wake_flags)
 {
 	s32 cpu = scx_bpf_pick_idle_cpu(p->cpus_ptr, 0);
+	if (cpu < 0)
+		cpu = prev_cpu;
 
-	if (cpu >= 0) {
-		/*
-		 * If we dispatch to a bogus DSQ that will fall back to the
-		 * builtin global DSQ, we fail gracefully.
-		 */
-		scx_bpf_dsq_insert_vtime(p, 0xcafef00d, SCX_SLICE_DFL,
-				       p->scx.dsq_vtime, 0);
-		return cpu;
-	}
-
-	return prev_cpu;
+	/*
+	 * If we dispatch to a bogus DSQ that will fall back to the
+	 * builtin global DSQ, we fail gracefully.
+	 */
+	scx_bpf_dsq_insert_vtime(p, 0xcafef00d, SCX_SLICE_DFL,
+				 p->scx.dsq_vtime, 0);
+	return cpu;
 }
 
 void BPF_STRUCT_OPS(ddsp_bogus_dsq_fail_exit, struct scx_exit_info *ei)
diff --git a/tools/testing/selftests/sched_ext/ddsp_vtimelocal_fail.bpf.c b/tools/testing/selftests/sched_ext/ddsp_vtimelocal_fail.bpf.c
index e4a55027778f..82dca4cdc0a6 100644
--- a/tools/testing/selftests/sched_ext/ddsp_vtimelocal_fail.bpf.c
+++ b/tools/testing/selftests/sched_ext/ddsp_vtimelocal_fail.bpf.c
@@ -14,15 +14,14 @@ s32 BPF_STRUCT_OPS(ddsp_vtimelocal_fail_select_cpu, struct task_struct *p,
 		   s32 prev_cpu, u64 wake_flags)
 {
 	s32 cpu = scx_bpf_pick_idle_cpu(p->cpus_ptr, 0);
+	if (cpu < 0)
+		cpu = prev_cpu;
 
-	if (cpu >= 0) {
-		/* Shouldn't be allowed to vtime dispatch to a builtin DSQ. */
-		scx_bpf_dsq_insert_vtime(p, SCX_DSQ_LOCAL, SCX_SLICE_DFL,
-					 p->scx.dsq_vtime, 0);
-		return cpu;
-	}
+	/* Shouldn't be allowed to vtime dispatch to a builtin DSQ. */
+	scx_bpf_dsq_insert_vtime(p, SCX_DSQ_LOCAL, SCX_SLICE_DFL,
+				 p->scx.dsq_vtime, 0);
 
-	return prev_cpu;
+	return cpu;
 }
 
 void BPF_STRUCT_OPS(ddsp_vtimelocal_fail_exit, struct scx_exit_info *ei)
-- 
2.55.0.679.g6767b8d81c-goog

Re: [PATCH] selftests/sched_ext: Fix flaky ddsp failure tests on busy systems

[Tejun Heo]

On Tue, Aug 11, 2026 at 02:13:58PM +0000, Michal Blaszczyk wrote:
> The ddsp_vtimelocal_fail and ddsp_bogus_dsq_fail tests skip calling
> scx_bpf_dsq_insert_vtime() if scx_bpf_pick_idle_cpu() fails to find
> an idle CPU (returns -1). On loaded systems, this results in the tests
> skipping the very assertions they are meant to verify.
>
> Eliminate this flakiness by falling back to prev_cpu if no idle CPU is
> found, ensuring the illegal dispatch operations are unconditionally
> attempted and tested.

Applied to sched_ext/for-7.3 with the description updated to say that
scx_bpf_pick_idle_cpu() returns a negative error code on failure rather
than -1.

Thanks.

-- 
tejun