From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Gleixner Subject: Re: [PATCH tip v4 2/5] [s]wait: Add compile time type check assertion Date: Fri, 27 Nov 2015 21:10:00 +0100 (CET) Message-ID: References: <1448370187-10075-1-git-send-email-daniel.wagner@bmw-carit.de> <1448370187-10075-3-git-send-email-daniel.wagner@bmw-carit.de> <5656F98D.70701@bmw-carit.de> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Cc: linux-kernel@vger.kernel.org, linux-rt-users@vger.kernel.org, Boqun Feng , Marcelo Tosatti , Paolo Bonzini , "Paul E. McKenney" , Paul Gortmaker , "Peter Zijlstra (Intel)" To: Daniel Wagner Return-path: Received: from www.linutronix.de ([62.245.132.108]:45304 "EHLO Galois.linutronix.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754234AbbK0UKs (ORCPT ); Fri, 27 Nov 2015 15:10:48 -0500 In-Reply-To: <5656F98D.70701@bmw-carit.de> Sender: linux-rt-users-owner@vger.kernel.org List-ID: On Thu, 26 Nov 2015, Daniel Wagner wrote: > On 11/24/2015 02:03 PM, Daniel Wagner wrote: > > The API provided by wait.h and swait.h is very similiar. Most of the > > time your are only one character away from either of it: > > > > wake_up() vs swake_up() > > > > This is on purpose so that we do not have two nearly identical bits of > > infrastructre code with dissimilar names. > > > > A compile time type check assertion ensures that obvious wrong usage > > is caught at early stage. > > Obviously, this didn't really work as one can see with patch #4. That > one just compiled. So I wrapped almost all functions to get a better > check coverage. woken_wake_function(), autoremove_wake_function() and > wake_bit_function() can't be wrapped easily because DEFINE_WAIT and > friends. I just left them out. > > The result looks pretty bad in my opinion. Probably it would be > better do add -Werror=incompatible-pointer-types to the CFLAGS. That's really bad. If we can pull off the -Werror=incompatible-pointer-types trick, that would solve it nicely. Thanks, tglx