From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Gleixner Subject: Re: linux-next: build failure after merge of the final tree (powerpc tree related) Date: Mon, 14 Mar 2011 22:17:56 +0100 (CET) Message-ID: References: <20110314203827.4b61818e.sfr@canb.auug.org.au> <1300135074.22236.788.camel@pasglop> <20110314.135116.112600339.davem@davemloft.net> Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Return-path: Received: from www.tglx.de ([62.245.132.106]:55989 "EHLO www.tglx.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755631Ab1CNVTL (ORCPT ); Mon, 14 Mar 2011 17:19:11 -0400 In-Reply-To: <20110314.135116.112600339.davem@davemloft.net> Sender: linux-next-owner@vger.kernel.org List-ID: To: David Miller Cc: Benjamin Herrenschmidt , Stephen Rothwell , Paul Mackerras , linuxppc-dev@lists.ozlabs.org, linux-next@vger.kernel.org, LKML , buytenh@secretlab.ca, grant.likely@secretlab.ca, Ingo Molnar , Andrew Morton , Feng Tang , Alan Cox On Mon, 14 Mar 2011, David Miller wrote: > From: Benjamin Herrenschmidt > Date: Tue, 15 Mar 2011 07:37:54 +1100 > > > On Mon, 2011-03-14 at 20:38 +1100, Stephen Rothwell wrote: > >> Hi all, > >> > >> After merging the final tree, today's linux-next build (powerpc > >> allyesconfig) failed like this: > >> > >> drivers/gpio/langwell_gpio.c: In function 'lnw_irq_handler': > >> drivers/gpio/langwell_gpio.c:210: error: 'struct irq_desc' has no member named 'chip' > >> drivers/gpio/langwell_gpio.c:211: error: 'struct irq_desc' has no member named 'chip' > >> > >> Caused by commit 17b9f9e2653a ("powerpc: Enable > >> GENERIC_HARDIRQS_NO_DEPRECATED") enabling > >> CONFIG_GENERIC_HARDIRQS_NO_DEPRECATED for powerpc without previously > >> fixing up the above driver. > >> > >> I have reverted that commit for today. > > > > Except that the above driver has nothing to do with powerpc, it's some > > Intel Moorestown stuff... > > > > It should be fixed regardless I suppose, CC'ing Thomas. > > I've had to make fixes to this driver in sparc64 allmodconfig builds > too :-) Sigh. Yes, this wants to depend on x86 in the first place. But this is the second instance of blindly and mindlessly changing eoi to irq_eoi w/o looking at the reason for this change. Those patches should have been rejected based on their changelog in the first place: Latest kernel has many changes in IRQ subsystem and its interfaces, like adding "irq_eoi" for struct irq_chip, this patch is a follow up change for that. When noone beats me to fix that mess, I'll do it tomorrow morning. Thanks, tglx