From mboxrd@z Thu Jan 1 00:00:00 1970 From: Greg KH Subject: Re: [BUILD_FAILURE] linux-next: Tree for March 31 build fails with CONFIG_SLICOSS=y Date: Mon, 31 Mar 2008 08:55:54 -0700 Message-ID: <20080331155554.GE3696@kroah.com> References: <20080331181118.7964d4a9.sfr@canb.auug.org.au> <47F0E456.5080201@linux.vnet.ibm.com> <20080401003110.d4834b07.sfr@canb.auug.org.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-next@vger.kernel.org, LKML , netdev@vger.kernel.org, Andy Whitcroft , Christopher Harrer , Kamalesh Babulal To: Stephen Rothwell Return-path: Received: from pentafluge.infradead.org ([213.146.154.40]:56253 "EHLO pentafluge.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752642AbYCaPud (ORCPT ); Mon, 31 Mar 2008 11:50:33 -0400 Content-Disposition: inline In-Reply-To: <20080401003110.d4834b07.sfr@canb.auug.org.au> Sender: netdev-owner@vger.kernel.org List-ID: On Tue, Apr 01, 2008 at 12:31:10AM +1100, Stephen Rothwell wrote: > Hi Greg, > > Strike two! Again, these drivers need to build (or not try to build) for > all architectures ... Its out until you tell me the above is true. > > #if defined(CONFIG_X86_64) > hcmd->cmdsize = (ulong32) ((((ulong64) & ihcmd->u.slic_buffers.bufs[1] - > (ulong64) hcmd) + 31) >> 5); > #elif defined(CONFIG_X86) > hcmd->cmdsize = ((((ulong32) & ihcmd->u.slic_buffers.bufs[1] - > (ulong32) hcmd) + 31) >> 5); > #else > Stop Compilation; Ah, missed that one, that's why it built on my machines (x86-64 and i386), but would die on anything else. That test is just wrong, I'll fix it up to get it to build on all arches (doesn't mean it will run though, it assumes that port io is memory mapped flat on all machines and still needs a lot of work to be fixed. thanks, greg k-h