sched_ext: Fix incorrect autogroup migration detection
PATCH 1/1 sched_ext: Fix incorrect autogroup migration detection
scx_move_task() is called from sched_move_task() and tells the BPF scheduler that cgroup migration is being committed. sched_move_task() is used by both cgroup and autogroup migrations and scx_move_task() tried to filter out autogroup migrations by testing the destination cgroup and PF_EXITING but this is not enough. In fact, without explicitly tagging the thread which is doing the cgroup migration, there is no good way to tell apart scx_move_task() invocations for racing migration to the root cgroup and an autogroup migration.
This led to scx_move_task() incorrectly ignoring a migration from non-root cgroup to an autogroup of the root cgroup triggering the following warning:
WARNING: CPU: 7 PID: 1 at kernel/sched/ext.c:3725 scx_cgroup_can_attach+0x196/0x340
...
Call Trace:
<TASK>
cgroup_migrate_execute+0x5b1/0x700
cgroup_attach_task+0x296/0x400
__cgroup_procs_write+0x128/0x140
cgroup_procs_write+0x17/0x30
kernfs_fop_write_iter+0x141/0x1f0
vfs_write+0x31d/0x4a0
__x64_sys_write+0x72/0xf0
do_syscall_64+0x82/0x160
entry_SYSCALL_64_after_hwframe+0x76/0x7eFix it by adding an argument to sched_move_task() that indicates whether the moving is for a cgroup or autogroup migration. After the change, scx_move_task() is called only for cgroup migrations and renamed to scx_cgroup_move_task().
---
Peter, I can also flag the task that's doing the cgroup migration but that seems unnecessarily convoluted. If you're okay with the change, I'll route this through the sched_ext tree.
Thanks.
kernel/sched/autogroup.c | 4 ++-- kernel/sched/core.c | 7 ++++--- kernel/sched/ext.c | 15 +-------------- kernel/sched/ext.h | 4 ++-- kernel/sched/sched.h | 2 +- 5 files changed, 10 insertions(+), 22 deletions(-)
kernel/sched/autogroup.c +2 -2
@@ -150,7 +150,7 @@ void sched_autogroup_exit_task(struct task_struct *p)* see this thread after that: we can no longer use signal->autogroup. * See the PF_EXITING check in task_wants_autogroup(). */ -sched_move_task (p); +sched_move_task (p, true); } static void
@@ -182,7 +182,7 @@ autogroup_move_group(struct task_struct *p, struct autogroup *ag)* sched_autogroup_exit_task(). */ for_each_thread (p, t) -sched_move_task (t); +sched_move_task (t, true);unlock_task_sighand (p, &flags);autogroup_kref_put (prev);
kernel/sched/core.c +4 -3
@@ -9042,7 +9042,7 @@ static void sched_change_group(struct task_struct *tsk, struct task_group *group* now. This function just updates tsk->se.cfs_rq and tsk->se.parent to reflect * its new group. */ -void sched_move_task (struct task_struct *tsk) +void sched_move_task (struct task_struct *tsk,bool for_autogroup) {int queued, running, queue_flags =DEQUEUE_SAVE |DEQUEUE_MOVE |DEQUEUE_NOCLOCK ;
@@ -9071,7 +9071,8 @@ void sched_move_task(struct task_struct *tsk)put_prev_task (rq, tsk);sched_change_group (tsk, group); -scx_move_task (tsk); +if (!for_autogroup) +scx_cgroup_move_task (tsk);if (queued)enqueue_task (rq, tsk, queue_flags);
@@ -9172,7 +9173,7 @@ static void cpu_cgroup_attach(struct cgroup_taskset *tset)struct cgroup_subsys_state *css;cgroup_taskset_for_each (task, css, tset) -sched_move_task (task); +sched_move_task (task, false);scx_cgroup_finish_attach (); }
kernel/sched/ext.c +1 -14
@@ -4323,24 +4323,11 @@ int scx_cgroup_can_attach(struct cgroup_taskset *tset)return ops_sanitize_err ("cgroup_prep_move" , ret); } -void scx_move_task (struct task_struct *p) +void scx_cgroup_move_task (struct task_struct *p) {if (!scx_cgroup_enabled)return ; -/* -* We're called from sched_move_task() which handles both cgroup and -* autogroup moves. Ignore the latter. -* -* Also ignore exiting tasks, because in the exit path tasks transition -* from the autogroup to the root group, so task_group_is_autogroup() -* alone isn't able to catch exiting autogroup tasks. This is safe for -* cgroup_move(), because cgroup migrations never happen for PF_EXITING -* tasks. -*/ -if (task_group_is_autogroup (task_group (p)) || (p->flags &PF_EXITING )) -return ; -/* * @p must have ops.cgroup_prep_move() called on it and thus * cgrp_moving_from set.
kernel/sched/ext.h +2 -2
@@ -73,7 +73,7 @@ static inline void scx_update_idle(struct rq *rq, bool idle, bool do_notify) {}int scx_tg_online (struct task_group *tg);void scx_tg_offline (struct task_group *tg);int scx_cgroup_can_attach (struct cgroup_taskset *tset); -void scx_move_task (struct task_struct *p); +void scx_cgroup_move_task (struct task_struct *p);void scx_cgroup_finish_attach (void );void scx_cgroup_cancel_attach (struct cgroup_taskset *tset);void scx_group_set_weight (struct task_group *tg,unsigned long cgrp_weight);
@@ -82,7 +82,7 @@ void scx_group_set_idle(struct task_group *tg, bool idle);static inline int scx_tg_online (struct task_group *tg) {return 0 ; }static inline void scx_tg_offline (struct task_group *tg) {}static inline int scx_cgroup_can_attach (struct cgroup_taskset *tset) {return 0 ; } -static inline void scx_move_task (struct task_struct *p) {} +static inline void scx_cgroup_move_task (struct task_struct *p) {}static inline void scx_cgroup_finish_attach (void ) {}static inline void scx_cgroup_cancel_attach (struct cgroup_taskset *tset) {}static inline void scx_group_set_weight (struct task_group *tg,unsigned long cgrp_weight) {}
kernel/sched/sched.h +1 -1
@@ -572,7 +572,7 @@ extern void sched_online_group(struct task_group *tg,extern void sched_destroy_group (struct task_group *tg);extern void sched_release_group (struct task_group *tg); -extern void sched_move_task (struct task_struct *tsk); +extern void sched_move_task (struct task_struct *tsk,bool for_autogroup);#ifdef CONFIG_FAIR_GROUP_SCHED extern int sched_group_set_shares (struct task_group *tg,unsigned long shares);
Peter Zijlstra
maintainerQuoted 31 lines · Tejun Heo
scx_move_task()is called fromsched_move_task()and tells the BPF scheduler that cgroup migration is being committed.sched_move_task()is used by both cgroup and autogroup migrations andscx_move_task()tried to filter out autogroup migrations by testing the destination cgroup andPF_EXITINGbut this is not enough. In fact, without explicitly tagging the thread which is doing the cgroup migration, there is no good way to tell apartscx_move_task()invocations for racing migration to the root cgroup and an autogroup migration.This led to
scx_move_task()incorrectly ignoring a migration from non-root cgroup to an autogroup of the root cgroup triggering the following warning:WARNING: CPU: 7 PID: 1 at kernel/sched/ext.c:3725 scx_cgroup_can_attach+0x196/0x340 ... Call Trace: <TASK> cgroup_migrate_execute+0x5b1/0x700 cgroup_attach_task+0x296/0x400 __cgroup_procs_write+0x128/0x140 cgroup_procs_write+0x17/0x30 kernfs_fop_write_iter+0x141/0x1f0 vfs_write+0x31d/0x4a0 __x64_sys_write+0x72/0xf0 do_syscall_64+0x82/0x160 entry_SYSCALL_64_after_hwframe+0x76/0x7eFix it by adding an argument to
sched_move_task()that indicates whether the moving is for a cgroup or autogroup migration. After the change,scx_move_task()is called only for cgroup migrations and renamed toscx_cgroup_move_task().
Quoted 8 lines
Fixes 819513666966 ("sched_ext: Add cgroup support")Cc [email protected] # v6.12+Signed-off-by Tejun Heo <[email protected]>---
Peter, I can also flag the task that's doing the cgroup migration but that seems unnecessarily convoluted. If you're okay with the change, I'll route this through thesched_exttree.
ACK, Thanks!
Tejun Heo
AppliedQuoted 31 lines · Tejun Heo
scx_move_task()is called fromsched_move_task()and tells the BPF scheduler that cgroup migration is being committed.sched_move_task()is used by both cgroup and autogroup migrations andscx_move_task()tried to filter out autogroup migrations by testing the destination cgroup andPF_EXITINGbut this is not enough. In fact, without explicitly tagging the thread which is doing the cgroup migration, there is no good way to tell apartscx_move_task()invocations for racing migration to the root cgroup and an autogroup migration.This led to
scx_move_task()incorrectly ignoring a migration from non-root cgroup to an autogroup of the root cgroup triggering the following warning:WARNING: CPU: 7 PID: 1 at kernel/sched/ext.c:3725 scx_cgroup_can_attach+0x196/0x340 ... Call Trace: <TASK> cgroup_migrate_execute+0x5b1/0x700 cgroup_attach_task+0x296/0x400 __cgroup_procs_write+0x128/0x140 cgroup_procs_write+0x17/0x30 kernfs_fop_write_iter+0x141/0x1f0 vfs_write+0x31d/0x4a0 __x64_sys_write+0x72/0xf0 do_syscall_64+0x82/0x160 entry_SYSCALL_64_after_hwframe+0x76/0x7eFix it by adding an argument to
sched_move_task()that indicates whether the moving is for a cgroup or autogroup migration. After the change,scx_move_task()is called only for cgroup migrations and renamed toscx_cgroup_move_task().
Quoted 4 lines
Fixes 819513666966 ("sched_ext: Add cgroup support")Cc [email protected] # v6.12+Signed-off-by Tejun Heo <[email protected]>
Applied to sched_ext/for-6.14-fixes.
Thanks.
Signature · 1 line
tejun