Re: [PATCH 1/2] sched_ext: Fix cgroup double-put on sub-sched abort path

[Andrea Righi <arighi@nvidia.com>]

On Sun, Mar 15, 2026 at 07:43:27PM -1000, Tejun Heo wrote:
> The abort path in scx_sub_enable_workfn() fell through to out_put_cgrp,
> double-putting the cgroup ref already owned by sch->cgrp. It also skipped
> kthread_flush_work() needed to flush the disable path.
> 
> Relocate the abort block above err_unlock_and_disable so it falls through to
> err_disable.
> 
> Fixes: ebeca1f930ea ("sched_ext: Introduce cgroup sub-sched support")

Maybe we should point to commit 337ec00b1d9c ("sched_ext: Implement cgroup
sub-sched enabling and disabling")

That's where we added the real sub-sched enable/disable implementation and
the abort block.

Apart from that looks good.

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

Thanks,
-Andrea

> Signed-off-by: Tejun Heo <tj@kernel.org>
> ---
>  kernel/sched/ext.c | 16 +++++++---------
>  1 file changed, 7 insertions(+), 9 deletions(-)
> 
> diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
> index 9202c6d7a771..2f70effcc4a6 100644
> --- a/kernel/sched/ext.c
> +++ b/kernel/sched/ext.c
> @@ -7050,6 +7050,13 @@ static void scx_sub_enable_workfn(struct kthread_work *work)
>  	ret = 0;
>  	goto out_unlock;
>  
> +out_put_cgrp:
> +	cgroup_put(cgrp);
> +out_unlock:
> +	mutex_unlock(&scx_enable_mutex);
> +	cmd->ret = ret;
> +	return;
> +
>  abort:
>  	put_task_struct(p);
>  	scx_task_iter_stop(&sti);
> @@ -7063,15 +7070,6 @@ static void scx_sub_enable_workfn(struct kthread_work *work)
>  		}
>  	}
>  	scx_task_iter_stop(&sti);
> -	scx_cgroup_unlock();
> -	percpu_up_write(&scx_fork_rwsem);
> -out_put_cgrp:
> -	cgroup_put(cgrp);
> -out_unlock:
> -	mutex_unlock(&scx_enable_mutex);
> -	cmd->ret = ret;
> -	return;
> -
>  err_unlock_and_disable:
>  	/* we'll soon enter disable path, keep bypass on */
>  	scx_cgroup_unlock();
> -- 
> 2.53.0
>