From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jesse Barnes Subject: Re: linux-next: pci tree build failure Date: Tue, 1 Jul 2008 19:11:39 -0700 Message-ID: <200807011911.39511.jbarnes@virtuousgeek.org> References: Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Return-path: Received: from outbound-mail-07.bluehost.com ([69.89.17.207]:34861 "HELO outbound-mail-07.bluehost.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1757416AbYGBCLv (ORCPT ); Tue, 1 Jul 2008 22:11:51 -0400 In-Reply-To: Content-Disposition: inline Sender: linux-next-owner@vger.kernel.org List-ID: To: Alan Stern Cc: Stephen Rothwell , linux-next@vger.kernel.org, Greg KH , "Rafael J. Wysocki" On Monday, June 30, 2008 7:11 am Alan Stern wrote: > On Mon, 30 Jun 2008, Stephen Rothwell wrote: > > Hi Jesse, > > > > Today's linux-next build (powerpc ppc64_defconfig) failed like this: > > > > drivers/usb/core/driver.c: In function 'do_unbind_rebind': > > drivers/usb/core/driver.c:865: error: 'struct dev_pm_info' has no member > > named 'sleeping' > > > > Due to commit 1eede070a59e1cc73da51e1aaa00d9ab86572cfc ("Introduce new > > top level suspend and hibernation callbacks") from the pci tree (which > > removed the sleeping field of the dev_pm_info structure) and commit > > fafa0857b87640d4c297777d432174b8778ac4fd ("USB: Force unbinding of > > drivers lacking reset_resume or other methods") from the usb tree which > > adds a reference. > > I knew this was going to happen. There was no way around it. > > > For now I have reverted commit fafa0857b87640d4c297777d432174b8778ac4fd > > ("USB: Force unbinding of drivers lacking reset_resume or other methods") > > from the usb tree since that gives me a building tree without the pain of > > trying to revert 1eede070a59e1cc73da51e1aaa00d9ab86572cfc (which has > > several dependent commits). > > > > A better solution would be nice (hint, hint). > > Instead of reverting that USB commit, take a look at it. There's a big > FIXME comment at the point where the obsolete field member is used. > For now you can insert a patch to remove that line of code. > > That's not a correct fix, but it will allow things to build and run. > Unfortunately it's not possible at this time to submit a correct fix > because the clash occurs between patches in two different development > trees. In addition, the correct fix involves a non-trivial rewrite of > the USB-PM interface; it can't simply be dashed off. We had some ACPI cross-dependencies like that, so I ended up pulling directly from Len's ACPI tree into my PCI tree, but that only worked because we both use git. Doing it with Greg's stuff would probably be harder, especially if that area is in flux. But as long as some sort of correct fix is in progress we can put it off until the merge window... Jesse