linux-kernel - Re: [PATCH v3] sched_ext: Fix lock imbalance in dispatch_to_local

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <Z5SiY0WNvEHXkrv5@gpd3>
Date: Sat, 25 Jan 2025 09:35:47 +0100
From: Andrea Righi <arighi@...dia.com>
To: Tejun Heo <tj@...nel.org>, David Vernet <void@...ifault.com>,
	Changwoo Min <changwoo@...lia.com>
Cc: linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3] sched_ext: Fix lock imbalance in
 dispatch_to_local_dsq()

On Sat, Jan 25, 2025 at 07:56:08AM +0100, Andrea Righi wrote:
...
> @@ -2557,6 +2567,7 @@ static void dispatch_to_local_dsq(struct rq *rq, struct scx_dispatch_q *dst_dsq,
>  {
>  	struct rq *src_rq = task_rq(p);
>  	struct rq *dst_rq = container_of(dst_dsq, struct rq, scx.local_dsq);
> +	struct rq *locked_rq = rq;

I just noticed that we have an unused variable here with !CONFIG_SMP, so
ignore this. I'll send a new version soon.

Sorry for the noise.
-Andrea

>  
>  	/*
>  	 * We're synchronized against dequeue through DISPATCHING. As @p can't
> @@ -2593,12 +2604,16 @@ static void dispatch_to_local_dsq(struct rq *rq, struct scx_dispatch_q *dst_dsq,
>  	atomic_long_set_release(&p->scx.ops_state, SCX_OPSS_NONE);
>  
>  	/* switch to @src_rq lock */
> -	if (rq != src_rq) {
> -		raw_spin_rq_unlock(rq);
> +	if (locked_rq != src_rq) {
> +		raw_spin_rq_unlock(locked_rq);
> +		locked_rq = src_rq;
>  		raw_spin_rq_lock(src_rq);
>  	}
>  
> -	/* task_rq couldn't have changed if we're still the holding cpu */
> +	/*
> +	 * If p->scx.holding_cpu still matches the current CPU, task_rq(p)
> +	 * has not changed and we can safely move the task to @dst_rq.
> +	 */
>  	if (likely(p->scx.holding_cpu == raw_smp_processor_id()) &&
>  	    !WARN_ON_ONCE(src_rq != task_rq(p))) {
>  		/*