From mboxrd@z Thu Jan 1 00:00:00 1970 From: Matthew Wilcox Subject: Re: linux-next: pci tree build failure Date: Tue, 24 Mar 2009 21:30:01 -0700 Message-ID: <20090325043001.GF20264@linux.intel.com> References: <20090323124948.fed84e3a.sfr@canb.auug.org.au> <20090325122121.225a0ccb.sfr@canb.auug.org.au> <20090324183800.6ba50715@hobbes.virtuouswap> <20090325144117.9c54ae5a.sfr@canb.auug.org.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from mga05.intel.com ([192.55.52.89]:8078 "EHLO fmsmga101.fm.intel.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752901AbZCYEaE (ORCPT ); Wed, 25 Mar 2009 00:30:04 -0400 Content-Disposition: inline In-Reply-To: <20090325144117.9c54ae5a.sfr@canb.auug.org.au> Sender: linux-next-owner@vger.kernel.org List-ID: To: Stephen Rothwell Cc: Jesse Barnes , linux-next@vger.kernel.org On Wed, Mar 25, 2009 at 02:41:17PM +1100, Stephen Rothwell wrote: > Hi Jesse, > > On Tue, 24 Mar 2009 18:38:00 -0700 Jesse Barnes wrote: > > > > On Wed, 25 Mar 2009 12:21:21 +1100 > > Stephen Rothwell wrote: > > > > > On Mon, 23 Mar 2009 12:49:48 +1100 Stephen Rothwell > > > wrote: > > > > > > > > Today's linux-next build (powerpc ppc64_defconfig) failed like this: > > > > > > > > arch/powerpc/kernel/msi.c: In function 'arch_msi_check_device': > > > > arch/powerpc/kernel/msi.c:23: error: 'PCI_CAP_ID_MSI' undeclared > > > > (first use in this function) > > > > > > > > Caused by commit 1c8d7b0a562da06d3ebe83f01b1ed553205d1ae4 ("PCI > > > > MSI: Add support for multiple MSI"). PCI_CAP_ID_MSI is defined in > > > > linux/pci_regs.h which is included by linux/pci.h, but neither is > > > > included by arch/powerpc/kernel/msi.c. > > > > > > > > I have used the version of the pci tree from next-20090320 for > > > > today. > > > > > > This is still happening ... I just got back from a long weekend ... > > Does this fix it? Some header file changes must be to blame... > > Yeah, that patch fixes it, but so does just reverting the commit above, > so I guess that powerpc file did not need any of the pci.h infrastructure > before. > > Tested-by: Stephen Rothwell I guess not. I didn't compile-test it (I don't have a powerpc machine or cross-compiler). Seems odd, given that the whole purpose of the file is to implement PCI MSI ... oh well.