linux-kernel - Re: [PATCH 3/3] rtmutex: Reduce top-waiter blocking on a lock

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: <20160314132356.GY6356@twins.programming.kicks-ass.net>
Date:	Mon, 14 Mar 2016 14:23:56 +0100
From:	Peter Zijlstra <peterz@...radead.org>
To:	Davidlohr Bueso <dave@...olabs.net>
Cc:	tglx@...utronix.de, mingo@...nel.org, bigeasy@...utronix.de,
	umgwanakikbuti@...il.com, paulmck@...ux.vnet.ibm.com,
	linux-kernel@...r.kernel.org, Davidlohr Bueso <dbueso@...e.de>
Subject: Re: [PATCH 3/3] rtmutex: Reduce top-waiter blocking on a lock

On Tue, Mar 08, 2016 at 10:20:23AM -0800, Davidlohr Bueso wrote:
>  	lock->owner = (void *) RT_MUTEX_HAS_WAITERS;

>  static inline struct task_struct *rt_mutex_owner(struct rt_mutex *lock)
>  {
>  	return (struct task_struct *)
> -		((unsigned long)lock->owner & ~RT_MUTEX_OWNER_MASKALL);
> +		((unsigned long)READ_ONCE(lock->owner) & ~RT_MUTEX_OWNER_MASKALL);
>  }

If you READ_ONCE(), you should also WRITE_ONCE(), because while the
write is under the proper locks, our friendly compiler might still
choose to emit the store in a random sequence of byte stores, rendering
our READ_ONCE() pointless.