Re: [PATCH v10 2/5] sched: CGroup tagging interface for core scheduling

[Peter Zijlstra]

On Fri, Jan 22, 2021 at 08:17:01PM -0500, Joel Fernandes (Google) wrote:
> +int sched_core_share_tasks(struct task_struct *t1, struct task_struct *t2)
> +{
> +	static DEFINE_MUTEX(sched_core_tasks_mutex);
> +	unsigned long cookie;
> +	int ret = -ENOMEM;
> +
> +	mutex_lock(&sched_core_tasks_mutex);
> +
> +	if (!t2) {
> +		if (t1->core_task_cookie) {
> +			sched_core_put_task_cookie(t1->core_task_cookie);
> +			sched_core_update_task_cookie(t1, 0);
> +			sched_core_put();
> +		}

So this seems to be the bit that implements _CLEAR. ISTR there were
security implications / considerations here.

When the machine is vulnerable to L1TF/MDS and the like, clearing the
cookie would gain privilege and should thus be subject to some checks,
but I can'd find anything.

At the very least that deserves a comment I'm thinking.