From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751477AbdFFMQp (ORCPT ); Tue, 6 Jun 2017 08:16:45 -0400 Received: from Galois.linutronix.de ([146.0.238.70]:57096 "EHLO Galois.linutronix.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751305AbdFFMQo (ORCPT ); Tue, 6 Jun 2017 08:16:44 -0400 Date: Tue, 6 Jun 2017 14:16:38 +0200 (CEST) From: Thomas Gleixner To: LKML cc: linux-rt-users , Sebastian Sewior , Steven Rostedt , Peter Zijlstra , Mathias Koehrer , David Hauck Subject: Re: [PATCH RT] sched: Prevent task state corruption by spurious lock wakeup In-Reply-To: Message-ID: References: User-Agent: Alpine 2.20 (DEB 67 2015-01-07) MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Sat, 3 Jun 2017, Thomas Gleixner wrote: > If the to be woken task has state TASK_TRACED then the wakeup state check > should not match. But for some stupid reason wake_up_lock_sleeper() should > not use TASK_ALL. The lock sleepers are in state UNINTERRUPTIBLE, so the > wake state should be UNINTERRUPTIBLE as well. > > The extra check in try_to_wake_up() should stay though as it prevents > spurious wake ups in general. Further experimentation shows that it's not worth it. I'll send a V2 of that patch with the proper fix. Thanks, tglx