From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail.emea.novell.com ([130.57.118.101]:49482 "EHLO mail.emea.novell.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751240AbbCXHi2 convert rfc822-to-8bit (ORCPT ); Tue, 24 Mar 2015 03:38:28 -0400 Message-Id: <5511227F020000780006CD25@mail.emea.novell.com> Date: Tue, 24 Mar 2015 07:38:23 +0000 From: "Jan Beulich" Subject: Re: [PATCH] Kconfig: drop bogus default values References: <5500584D02000078000688F5@mail.emea.novell.com> <1426162307.5304.41.camel@x220> <178407860.0zoJnDfCo1@tacticalops> In-Reply-To: <178407860.0zoJnDfCo1@tacticalops> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 8BIT Content-Disposition: inline Sender: linux-kbuild-owner@vger.kernel.org List-ID: To: Paul Bolle , Martin Walch Cc: akpm@linux-foundation.org, Sam Ravnborg , Michal Marek , linux-kbuild@vger.kernel.org, linux-kernel@vger.kernel.org >>> On 23.03.15 at 22:08, wrote: > On Thursday 12 March 2015 13:11:47 Paul Bolle wrote: >> On Wed, 2015-03-11 at 13:59 +0000, Jan Beulich wrote: >> > Default "no" is pretty pointless for options without (visible) prompts: >> >> Related: is there ever a situation where using "default n" or "def_bool >> n" makes sense (whether or not the entry has a prompt)? I think I once >> thought of one but I can't remember it at all, so I guess my memory is >> fooling me. > > Your memory is right. It is rarely used, but there is an application for > using a plain "default n": to overwrite an existing other default value. > Particularly in one special case this is desired: Let us say there is a > symbol that may lack a visible prompt, but has the default value y set in > a Kconfig file that is used across all architectures. If there is a single > architecture that must have the default value n then it is possible to > override the default y in the global file with a default n in the > architecture specific file. > > A real world case is PCI_QUIRKS in the mainline kernel: > > init/Kconfig:1554: default y > arch/s390/Kconfig:59: def_bool n > > When setting PCI!=n && EXPERT=n then on each architecture PCI_QUIRKS=y > except on s390 where PCI_QUIRKS=n. But iirc such redundant defaults yield warnings (or at least at some point in the past they did). Jan