From mboxrd@z Thu Jan 1 00:00:00 1970 From: Heiko Carstens Subject: Re: [-next Nov 16] s390 build failure (arch/s390/kernel/time) Date: Tue, 17 Nov 2009 08:59:17 +0100 Message-ID: <20091117075917.GA5124@osiris.boeblingen.de.ibm.com> References: <20091116181453.b42247cf.sfr@canb.auug.org.au> <4B02283D.8040103@in.ibm.com> <20091117162546.16bb00b0.sfr@canb.auug.org.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from mtagate7.uk.ibm.com ([195.212.29.140]:45233 "EHLO mtagate7.uk.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752488AbZKQH7j (ORCPT ); Tue, 17 Nov 2009 02:59:39 -0500 Content-Disposition: inline In-Reply-To: <20091117162546.16bb00b0.sfr@canb.auug.org.au> Sender: linux-next-owner@vger.kernel.org List-ID: To: Stephen Rothwell Cc: Martin Schwidefsky , linux-s390@vger.kernel.org, linux-next@vger.kernel.org, Tejun Heo , Rusty Russell , Christoph Lameter , Ingo Molnar , Sachin Sant On Tue, Nov 17, 2009 at 04:25:46PM +1100, Stephen Rothwell wrote: > On Tue, 17 Nov 2009 10:06:13 +0530 Sachin Sant wrote: > > > > Next 20091116 build failed on a s390 box with > > > > arch/s390/kernel/time.c: In function 'get_sync_clock': > > arch/s390/kernel/time.c:337: error: 'clock_sync_sync' undeclared (first use in this function) > > arch/s390/kernel/time.c:337: error: (Each undeclared identifier is reported only once > > arch/s390/kernel/time.c:337: error: for each function it appears in.) > > arch/s390/kernel/time.c: In function 'check_sync_clock': > > arch/s390/kernel/time.c:387: error: 'clock_sync_sync' undeclared (first use in this function) > > Caused by commits 8283cb43ab3e92a039d3486a0f6253df95a5fa55 ("[S390] clock > sync mode flags") (which entered Linus' tree during 2.6.30-rc1) and > commit d2fec595511b5718bdb65645b3d5d99800d97943 ("[S390] stp support") > (which entered Linus' tree during 2.6.27-rc1) but only exposed by commit > e0fdb0e050eae331046385643618f12452aa7e73 ("percpu: add __percpu for > sparse") from the percpu tree. > > Needs to be fixed in the s390 tree. (put_per_cpu() now references its > argument, so that had better be a real variable :-)) Weird... at least it wasn't a bug. Patch below will fix compile breakage. --- arch/s390/kernel/time.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) Index: linux-next/arch/s390/kernel/time.c =================================================================== --- linux-next.orig/arch/s390/kernel/time.c +++ linux-next/arch/s390/kernel/time.c @@ -334,7 +334,7 @@ int get_sync_clock(unsigned long long *c sw0 = atomic_read(sw_ptr); *clock = get_clock(); sw1 = atomic_read(sw_ptr); - put_cpu_var(clock_sync_sync); + put_cpu_var(clock_sync_word); if (sw0 == sw1 && (sw0 & 0x80000000U)) /* Success: time is in sync. */ return 0; @@ -384,7 +384,7 @@ static inline int check_sync_clock(void) sw_ptr = &get_cpu_var(clock_sync_word); rc = (atomic_read(sw_ptr) & 0x80000000U) != 0; - put_cpu_var(clock_sync_sync); + put_cpu_var(clock_sync_word); return rc; }