Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752289AbaBSCLQ (ORCPT ); Tue, 18 Feb 2014 21:11:16 -0500 Received: from mail.linuxfoundation.org ([140.211.169.12]:52255 "EHLO mail.linuxfoundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750904AbaBSCLP (ORCPT ); Tue, 18 Feb 2014 21:11:15 -0500 Date: Tue, 18 Feb 2014 18:13:44 -0800 From: Greg KH To: Arve =?iso-8859-1?B?SGr4bm5lduVn?= Cc: John Stultz , LKML , Serban Constantinescu , Colin Cross , Android Kernel Team Subject: Re: [PATCH 12/14] staging: binder: Fix ABI for 64bit Android Message-ID: <20140219021344.GA21896@kroah.com> References: <1392674322-9036-1-git-send-email-john.stultz@linaro.org> <1392674322-9036-13-git-send-email-john.stultz@linaro.org> <20140218190828.GA10376@kroah.com> <20140218194936.GA8584@kroah.com> <20140218203239.GA18852@kroah.com> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: User-Agent: Mutt/1.5.22 (2013-10-16) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Feb 18, 2014 at 04:08:20PM -0800, Arve Hj?nnev?g wrote: > On Tue, Feb 18, 2014 at 12:32 PM, Greg KH wrote: > > On Tue, Feb 18, 2014 at 12:02:07PM -0800, John Stultz wrote: > >> On Tue, Feb 18, 2014 at 11:49 AM, Greg KH wrote: > >> > On Tue, Feb 18, 2014 at 11:30:26AM -0800, John Stultz wrote: > >> >> On Tue, Feb 18, 2014 at 11:08 AM, Greg KH wrote: > >> >> > On Mon, Feb 17, 2014 at 01:58:40PM -0800, John Stultz wrote: > >> >> >> From: Serban Constantinescu > >> >> >> > >> >> >> This patch fixes the ABI for 64bit Android userspace. > >> >> >> BC_REQUEST_DEATH_NOTIFICATION and BC_CLEAR_DEATH_NOTIFICATION claim > >> >> >> to be using struct binder_ptr_cookie, but they are using a 32bit handle > >> >> >> and a pointer. > >> >> >> > >> >> >> On 32bit systems the payload size is the same as the size of struct > >> >> >> binder_ptr_cookie, however for 64bit systems this will differ. This > >> >> >> patch adds struct binder_handle_cookie that fixes this issue for 64bit > >> >> >> Android. > >> >> >> > >> >> >> Since there are no 64bit users of this interface that we know of this > >> >> >> change should not affect any existing systems. > >> >> > > >> >> > But you are changing the ioctl structures here, what is that going to > >> >> > cause with old programs? > >> >> > >> >> So I'd be glad for Serban or Arve to clarify, but my understanding > >> >> (and as is described in the commit message) is that the assumption is > >> >> there are no 64bit binder users at this point, and the ioctl structure > >> >> changes are made such that existing 32bit applications are unaffected. > >> > > >> > How does changing the structure size, and contents, not affect any > >> > applications or the kernel code? What am I missing here? > >> > >> On 32bit pointers and ints are the same size? (Years ago I sat through > >> your presentation on this, so I'm worried I'm missing something here > >> :) > >> > >> struct binder_ptr_cookie { > >> void *ptr; > >> void *cookie; > >> }; > >> > >> struct binder_handle_cookie { > >> __u32 handle; > >> void *cookie; > >> } __attribute__((packed)); > >> > >> > >> On 32bit systems these are the same size. Now on 64bit systems, this > >> changes things, and would break users, but the assumption here is > >> there are no pre-existing 64bit binder users. > > > > But you added a field to the existing structure, right? I don't really > > remember the patch, it was a few hundred back in my review of stuff > > today, sorry... > > > > greg k-h > > The existing structure is not changed. These two commands were defined > with wrong structure that did not match the code. Since a binder > pointer and handle are the same size on 32 bit systems, this change > does not affect them. On 64 bit systems, the ioctl number does change, > but these systems need the next patch to run 32 bit processes anyway, > so I don't expect anyone to ship a system without this change. The > main purpose of this patch is to add the binder_handle_cookie struct > so the service manager does not have to define its own version > (libbinder writes one field at a time so it does not use the struct). Ah, ok, that makes more sense, can someone put it in the changelog information so that I don't have to reject the patch for the same reason again? thanks, greg k-h -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/