Re: [PATCH 4/4] sched_ext: Fix cgroup exit ordering by moving sched_ext_free() to finish_task_switch()

[Andrea Righi <arighi@nvidia.com>]

Hi Tejun,

On Tue, Oct 28, 2025 at 08:19:18PM -1000, Tejun Heo wrote:
> sched_ext_free() was called from __put_task_struct() when the last reference
> to the task is dropped, which could be long after the task has finished
> running. This causes cgroup-related problems:
> 
> - ops.task_init() can be called on a cgroup which didn't get ops.cgroup_init()'d
>   during scheduler load.

s/task_init/init_task/

Also, it took me a bit to understand this point, maybe we could add
something like this to make it more clear:

 - ops.init_task() can be called on a cgroup which didn't get ops.cgroup_init()'d
   during scheduler load, because the cgroup might be destroyed/unlinked
   while the zombie task is still lingering on the scx_tasks list.

> 
> - ops.cgroup_exit() could be called before ops.exit_task() is called on all
>   member tasks, leading to incorrect exit ordering.
> 
> Fix by moving it to finish_task_switch() to be called right after the final
> context switch away from the dying task, matching when sched_class->task_dead()
> is called. Rename it to sched_ext_dead() to match the new calling context.
> 
> By calling sched_ext_dead() before cgroup_task_dead(), we ensure that:
> 
> - Tasks visible on scx_tasks list have valid cgroups during scheduler load,
>   as cgroup_mutex prevents cgroup destruction while the task is still linked.
> 
> - All member tasks have ops.exit_task() called and are removed from scx_tasks
>   before the cgroup can be destroyed and trigger ops.cgroup_exit().
> 
> This fix is made possible by the cgroup_task_dead() split in the previous patch.
> 
> This also makes more sense resource-wise as there's no point in keeping
> scheduler side resources around for dead tasks.
> 
> Reported-by: Dan Schatzberg <dschatzberg@meta.com>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Signed-off-by: Tejun Heo <tj@kernel.org>

Apart from the minor comment in the description above, everything else
looks good to me.

Reviewed-by: Andrea Righi <arighi@nvidia.com>

Thanks,
-Andrea

> ---
>  include/linux/sched/ext.h | 4 ++--
>  kernel/fork.c             | 1 -
>  kernel/sched/core.c       | 6 ++++++
>  kernel/sched/ext.c        | 2 +-
>  4 files changed, 9 insertions(+), 4 deletions(-)
> 
> diff --git a/include/linux/sched/ext.h b/include/linux/sched/ext.h
> index d82b7a9b0658..d7dd77be571f 100644
> --- a/include/linux/sched/ext.h
> +++ b/include/linux/sched/ext.h
> @@ -207,14 +207,14 @@ struct sched_ext_entity {
>  	struct list_head	tasks_node;
>  };
>  
> -void sched_ext_free(struct task_struct *p);
> +void sched_ext_dead(struct task_struct *p);
>  void print_scx_info(const char *log_lvl, struct task_struct *p);
>  void scx_softlockup(u32 dur_s);
>  bool scx_rcu_cpu_stall(void);
>  
>  #else	/* !CONFIG_SCHED_CLASS_EXT */
>  
> -static inline void sched_ext_free(struct task_struct *p) {}
> +static inline void sched_ext_dead(struct task_struct *p) {}
>  static inline void print_scx_info(const char *log_lvl, struct task_struct *p) {}
>  static inline void scx_softlockup(u32 dur_s) {}
>  static inline bool scx_rcu_cpu_stall(void) { return false; }
> diff --git a/kernel/fork.c b/kernel/fork.c
> index 960c39c9c264..5ae37909a813 100644
> --- a/kernel/fork.c
> +++ b/kernel/fork.c
> @@ -736,7 +736,6 @@ void __put_task_struct(struct task_struct *tsk)
>  	WARN_ON(tsk == current);
>  
>  	unwind_task_free(tsk);
> -	sched_ext_free(tsk);
>  	io_uring_free(tsk);
>  	cgroup_task_free(tsk);
>  	task_numa_free(tsk, true);
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index 40f12e37f60f..d4dbffb27a66 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -5222,6 +5222,12 @@ static struct rq *finish_task_switch(struct task_struct *prev)
>  		if (prev->sched_class->task_dead)
>  			prev->sched_class->task_dead(prev);
>  
> +		/*
> +		 * sched_ext_dead() must come before cgroup_task_dead() to
> +		 * prevent cgroups from being removed while its member tasks are
> +		 * visible to SCX schedulers.
> +		 */
> +		sched_ext_dead(prev);
>  		cgroup_task_dead(prev);
>  
>  		/* Task is done with its stack. */
> diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
> index 2b0e88206d07..840bc76210c8 100644
> --- a/kernel/sched/ext.c
> +++ b/kernel/sched/ext.c
> @@ -2926,7 +2926,7 @@ void scx_cancel_fork(struct task_struct *p)
>  	percpu_up_read(&scx_fork_rwsem);
>  }
>  
> -void sched_ext_free(struct task_struct *p)
> +void sched_ext_dead(struct task_struct *p)
>  {
>  	unsigned long flags;
>  
> -- 
> 2.51.1
>