From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ingo Molnar Subject: Re: Next April 24: [S390] allmodconfig build failure (trace/events) Date: Wed, 29 Apr 2009 14:04:54 +0200 Message-ID: <20090429120454.GF11586@elte.hu> References: <20090424150456.ff35e4ea.sfr@canb.auug.org.au> <49F16654.8040208@in.ibm.com> <20090424072533.GE24912@elte.hu> <20090424104108.0f2c0f05@osiris.boeblingen.de.ibm.com> <49F8232B.4090007@in.ibm.com> <20090429135139.5fac79b8@osiris.boeblingen.de.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20090429135139.5fac79b8@osiris.boeblingen.de.ibm.com> Sender: linux-next-owner@vger.kernel.org List-Archive: List-Post: To: Heiko Carstens Cc: Sachin Sant , Steven Rostedt , =?iso-8859-1?Q?Fr=E9d=E9ric?= Weisbecker , linux-s390@vger.kernel.org, linux-next@vger.kernel.org, LKML List-ID: * Heiko Carstens wrote: > On Wed, 29 Apr 2009 15:21:39 +0530 > Sachin Sant wrote: > > Heiko Carstens wrote: > > > On Fri, 24 Apr 2009 09:25:33 +0200 > > > Ingo Molnar wrote: > > >> * Sachin Sant wrote: > > >>> Today's next tree build(s390 allmodconfig) failed with > > >>> > > >>> kernel/built-in.o: In function `trace_softirq_entry' > > >>> include/trace/events/irq.h:42: undefined reference to > > >>> `' > > >>> include/trace/events/irq.h:42: undefined reference to > > >>> `__tracepoint_softirq_entry' > > >>> kernel/built-in.o: In function `trace_softirq_exit': > > >>> include/trace/events/irq.h:48: undefined reference to > > >>> `__tracepoint_softirq_exit' > > >>> include/trace/events/irq.h:48: undefined reference to > > >>> `__tracepoint_softirq_exit' > > >>> > > >> Hm, that's weird - s390 does not build kernel/softirq.o? Hm, it does > > >> - softirq.o is an obj-y object. > > > > > > s390 does build kernel/softirq.o. However it's anything but obvious to > > > me how the tracepoint infrastructure works. Too many #ifdefs, #define's > > > and #undefine's... > > > > > > I would expect that struct __tracepoint_softirq_entry somehow gets > > > defined via one of the TRACE_FORMAT macros, no? > > Today's next tree also has this failure. Any solution for this problem ? > > Ingo, could you pick up the patch below please? applied, thanks Heiko! Ingo