sched_ext: Fix rq lock state in hotplug ops
PATCH 1/1 sched_ext: Fix rq lock state in hotplug ops
The ops.cpu_online() and ops.cpu_offline() callbacks incorrectly assume that the rq involved in the operation is locked, which is not the case during hotplug, triggering the following warning:
WARNING: CPU: 1 PID: 20 at kernel/sched/sched.h:1504 handle_hotplug+0x280/0x340
Fix by not tracking the target rq as locked in the context of ops.cpu_online() and ops.cpu_offline().
--- kernel/sched/ext.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
kernel/sched/ext.c +2 -2
@@ -3477,9 +3477,9 @@ static void handle_hotplug(struct rq *rq, bool online)scx_idle_update_selcpu_topology (&scx_ops);if (online &&SCX_HAS_OP (cpu_online)) -SCX_CALL_OP (SCX_KF_UNLOCKED , cpu_online, rq, cpu); +SCX_CALL_OP (SCX_KF_UNLOCKED , cpu_online,NULL , cpu);else if (!online &&SCX_HAS_OP (cpu_offline)) -SCX_CALL_OP (SCX_KF_UNLOCKED , cpu_offline, rq, cpu); +SCX_CALL_OP (SCX_KF_UNLOCKED , cpu_offline,NULL , cpu);else scx_ops_exit(SCX_ECODE_ACT_RESTART | SCX_ECODE_RSN_HOTPLUG, "cpu %d going %s, exiting scheduler", cpu,
Signature · 1 line
2.49.0
Hi Andrea,
Thanks for the fix. Now it cleanly passes the hotplug selftest.
Quoted 32 lines · Andrea Righi
The
ops.cpu_online()andops.cpu_offline()callbacks incorrectly assume that the rq involved in the operation is locked, which is not the case during hotplug, triggering the following warning:WARNING: CPU: 1 PID: 20 at kernel/sched/sched.h:1504 handle_hotplug+0x280/0x340Fix by not tracking the target rq as locked in the context of
ops.cpu_online()andops.cpu_offline().Fixes 18853ba782bef ("sched_ext: Track currently locked rq")Reported-by Tejun Heo <[email protected]>Signed-off-by Andrea Righi <[email protected]>--- kernel/sched/ext.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)kernel/sched/ext.c +2
-2@@ -3477,9 +3477,9 @@ static void handle_hotplug(struct rq *rq, bool online)scx_idle_update_selcpu_topology (&scx_ops);if (online &&SCX_HAS_OP (cpu_online)) -SCX_CALL_OP (SCX_KF_UNLOCKED , cpu_online, rq, cpu); +SCX_CALL_OP (SCX_KF_UNLOCKED , cpu_online,NULL , cpu);else if (!online &&SCX_HAS_OP (cpu_offline)) -SCX_CALL_OP (SCX_KF_UNLOCKED , cpu_offline, rq, cpu); +SCX_CALL_OP (SCX_KF_UNLOCKED , cpu_offline,NULL , cpu);else scx_ops_exit(SCX_ECODE_ACT_RESTART | SCX_ECODE_RSN_HOTPLUG, "cpu %d going %s, exiting scheduler", cpu,
Tejun Heo
AppliedQuoted 4 lines · Andrea Righi
The
ops.cpu_online()andops.cpu_offline()callbacks incorrectly assume that the rq involved in the operation is locked, which is not the case during hotplug, triggering the following warning:
Quoted 8 lines
WARNING: CPU: 1 PID: 20 at kernel/sched/sched.h:1504 handle_hotplug+0x280/0x340Fix by not tracking the target rq as locked in the context of
ops.cpu_online()andops.cpu_offline().Fixes 18853ba782bef ("sched_ext: Track currently locked rq")Reported-by Tejun Heo <[email protected]>Signed-off-by Andrea Righi <[email protected]>
Applied to sched_ext/for-6.15-fixes.
Thanks.
Signature · 1 line
tejun