linux-kernel - Re: [PATCH v3 3/7] perf: Simplify perf_event_free

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

Message-ID: <20250402091512.GV5880@noisy.programming.kicks-ass.net>
Date: Wed, 2 Apr 2025 11:15:12 +0200
From: Peter Zijlstra <peterz@...radead.org>
To: Ravi Bangoria <ravi.bangoria@....com>
Cc: mingo@...nel.org, lucas.demarchi@...el.com,
	linux-kernel@...r.kernel.org, acme@...nel.org, namhyung@...nel.org,
	mark.rutland@....com, alexander.shishkin@...ux.intel.com,
	jolsa@...nel.org, irogers@...gle.com, adrian.hunter@...el.com,
	kan.liang@...ux.intel.com
Subject: Re: [PATCH v3 3/7] perf: Simplify perf_event_free_task() wait

On Mon, Mar 17, 2025 at 12:19:07PM +0530, Ravi Bangoria wrote:
> Hi Peter,
> 
> > @@ -1223,8 +1223,14 @@ static void put_ctx(struct perf_event_co
> >  	if (refcount_dec_and_test(&ctx->refcount)) {
> >  		if (ctx->parent_ctx)
> >  			put_ctx(ctx->parent_ctx);
> > -		if (ctx->task && ctx->task != TASK_TOMBSTONE)
> > -			put_task_struct(ctx->task);
> > +		if (ctx->task) {
> > +			if (ctx->task == TASK_TOMBSTONE) {
> > +				smp_mb(); /* pairs with wait_var_event() */
> > +				wake_up_var(&ctx->refcount);
> 
> perf_event_free_task() waits on "ctx->refcount == 1". But moving
> wake_up_var() under refcount_dec_and_test() will cause
> perf_event_free_task() to wait indefinitely. Right? So, shouldn't
> wake_up_var() be outside? something like:
> 
> --- a/kernel/events/core.c
> +++ b/kernel/events/core.c
> @@ -1281,15 +1281,14 @@ static void put_ctx(struct perf_event_context *ctx)
>  	if (refcount_dec_and_test(&ctx->refcount)) {
>  		if (ctx->parent_ctx)
>  			put_ctx(ctx->parent_ctx);
> -		if (ctx->task) {
> -			if (ctx->task == TASK_TOMBSTONE) {
> -				smp_mb(); /* pairs with wait_var_event() */
> -				wake_up_var(&ctx->refcount);
> -			} else {
> -				put_task_struct(ctx->task);
> -			}
> -		}
> +		if (ctx->task && ctx->task != TASK_TOMBSTONE)
> +			put_task_struct(ctx->task);
>  		call_rcu(&ctx->rcu_head, free_ctx);
> +	} else {
> +		if (ctx->task == TASK_TOMBSTONE) {
> +			smp_mb(); /* pairs with wait_var_event() */
> +			wake_up_var(&ctx->refcount);
> +		}
>  	}
>  }

Yes, you're quite right indeed. Thanks!