From mboxrd@z Thu Jan 1 00:00:00 1970 From: Randy Dunlap Subject: Re: linux-next: Tree for November 2 (usb/whci) Date: Tue, 17 Nov 2009 07:51:01 -0800 Message-ID: <20091117075101.0bf72415.randy.dunlap@oracle.com> References: <20091102173845.210d1c57.sfr@canb.auug.org.au> <20091102091339.479a3ddb.randy.dunlap@oracle.com> <20091111090403.5fb0229e.randy.dunlap@oracle.com> <20091111173015.GA9641@kroah.com> <4B028BBB.1080803@csr.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Return-path: Received: from rcsinet11.oracle.com ([148.87.113.123]:32147 "EHLO rgminet11.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750905AbZKQPvj (ORCPT ); Tue, 17 Nov 2009 10:51:39 -0500 In-Reply-To: <4B028BBB.1080803@csr.com> Sender: linux-next-owner@vger.kernel.org List-ID: To: David Vrabel Cc: Greg KH , lud , Stephen Rothwell , linux-next@vger.kernel.org, LKML On Tue, 17 Nov 2009 11:40:43 +0000 David Vrabel wrote: > Greg KH wrote: > > On Wed, Nov 11, 2009 at 09:04:03AM -0800, Randy Dunlap wrote: > >> On Mon, 2 Nov 2009 09:13:39 -0800 Randy Dunlap wrote: > >> > >>> On Mon, 2 Nov 2009 17:38:45 +1100 Stephen Rothwell wrote: > >>> > >>>> Hi all, > >>>> > >>>> Changes since 20091030: <<<<<<<<<<<<<<<<<<<<<<<<<< > >>> > >>> whci build fails on i386 due to 64-bit multiply & divide: > >>> > >>> ERROR: "__udivdi3" [drivers/usb/host/whci/whci-hcd.ko] undefined! > >>> ERROR: "__umoddi3" [drivers/usb/host/whci/whci-hcd.ko] undefined! > >> > >> Still present in linux-next-20091111. > > > > David, can you please look into this and resolve it? > > Yes, I've been on holiday and haven't had a change to look into it yet. > > I can't immediately think of any 64 bit maths and I don't understand why > I haven't seen it. I always build for i386. When the driver is built-in (not a loadable module), the messages are: drivers/built-in.o: In function `qset_add_urb': (.text+0x4595f8): undefined reference to `__umoddi3' drivers/built-in.o: In function `qset_add_urb': (.text+0x45969a): undefined reference to `__udivdi3' Those could still refer to other (inline) functions... --- ~Randy