Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755988Ab2JKIlk (ORCPT ); Thu, 11 Oct 2012 04:41:40 -0400 Received: from mail-vb0-f46.google.com ([209.85.212.46]:53935 "EHLO mail-vb0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751656Ab2JKIlf (ORCPT ); Thu, 11 Oct 2012 04:41:35 -0400 MIME-Version: 1.0 In-Reply-To: <201210111023.51311.o.endriss@gmx.de> References: <30699.1349789424@warthog.procyon.org.uk> <20121009183908.1e402a43@infradead.org> <201210111023.51311.o.endriss@gmx.de> Date: Thu, 11 Oct 2012 14:11:34 +0530 Message-ID: Subject: Re: Re: [GIT PULL] Disintegrate UAPI for media From: Manu Abraham To: Oliver Endriss Cc: Mauro Carvalho Chehab , David Howells , linux-media@vger.kernel.org, linux-kernel@vger.kernel.org, Hans Verkuil Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 5426 Lines: 128 On Thu, Oct 11, 2012 at 1:53 PM, Oliver Endriss wrote: > Mauro Carvalho Chehab wrote: >> Em Tue, 09 Oct 2012 14:30:24 +0100 >> David Howells escreveu: >> >> > Can you merge the following branch into the media tree please. >> > >> > This is to complete part of the Userspace API (UAPI) disintegration for which >> > the preparatory patches were pulled recently. After these patches, userspace >> > headers will be segregated into: >> > >> > include/uapi/linux/.../foo.h >> > >> > for the userspace interface stuff, and: >> > >> > include/linux/.../foo.h >> > >> > for the strictly kernel internal stuff. >> > >> > --- >> > The following changes since commit 9e2d8656f5e8aa214e66b462680cf86b210b74a8: >> > >> > Merge branch 'akpm' (Andrew's patch-bomb) (2012-10-09 16:23:15 +0900) >> > >> > are available in the git repository at: >> > >> > >> > git://git.infradead.org/users/dhowells/linux-headers.git tags/disintegrate-media-20121009 >> > >> > for you to fetch changes up to 1c436decd49665be131887b08d172a7989cdceee: >> > >> > UAPI: (Scripted) Disintegrate include/linux/dvb (2012-10-09 09:48:42 +0100) >> > >> > ---------------------------------------------------------------- >> > UAPI Disintegration 2012-10-09 >> > >> > ---------------------------------------------------------------- >> > David Howells (1): >> > UAPI: (Scripted) Disintegrate include/linux/dvb >> > >> > include/linux/dvb/Kbuild | 8 - >> > include/linux/dvb/dmx.h | 130 +-------------- >> > include/linux/dvb/video.h | 249 +---------------------------- >> > include/uapi/linux/dvb/Kbuild | 8 + >> > include/{ => uapi}/linux/dvb/audio.h | 0 >> > include/{ => uapi}/linux/dvb/ca.h | 0 >> > include/uapi/linux/dvb/dmx.h | 155 ++++++++++++++++++ >> > include/{ => uapi}/linux/dvb/frontend.h | 0 >> > include/{ => uapi}/linux/dvb/net.h | 0 >> > include/{ => uapi}/linux/dvb/osd.h | 0 >> > include/{ => uapi}/linux/dvb/version.h | 0 >> > include/uapi/linux/dvb/video.h | 274 ++++++++++++++++++++++++++++++++ >> > 12 files changed, 439 insertions(+), 385 deletions(-) >> > rename include/{ => uapi}/linux/dvb/audio.h (100%) >> > rename include/{ => uapi}/linux/dvb/ca.h (100%) >> > create mode 100644 include/uapi/linux/dvb/dmx.h >> > rename include/{ => uapi}/linux/dvb/frontend.h (100%) >> > rename include/{ => uapi}/linux/dvb/net.h (100%) >> > rename include/{ => uapi}/linux/dvb/osd.h (100%) >> > rename include/{ => uapi}/linux/dvb/version.h (100%) >> > create mode 100644 include/uapi/linux/dvb/video.h >> >> Hmm... last year, it was decided that we would be putting the DVB av7110-only >> API files on a separate place, as the API there conflicts with V4L/alsa APIs > > Wrong! Hans Verkuil and you tried to do it, without caring that it would > break userspace, and it was NAKed. > > Btw, if there is an API conflict, you guys created it. > > Anyone, who is interested in the _true_ history, should have a look at > the GIT changelog: > - dvb/{video.h,audio.h,osd.h} was the original decoder API. > - Then Hans extended this API, still using the same files. > - Later the v4l guys decided to create a new API. > - Now they want to (re)move the old one, breaking userspace. > > I explicitly NAK any attempt to break userspace applications and tools! > There is no reason to do it! > >> and are used only by one upstream driver (there were two drivers using them, >> at that time). As you might notice, av7110 hardware is really old, not >> manufactured anymore since maybe 10 years ago, and it is an unmaintained >> driver. > > The driver works fine, and it will continue to do so, unless someone > tampers with it. It does not require maintenance. > The hardware is old, but it is still in use, as it is easy to create a > pc-based settopbox with it. As of writing; the av7110 is the only driver in-kernel, but there is more to it: Such drivers are hard to bring up and takes an awful amount of time. There is already one driver which is nearing completion based on the same interface. > >> Some developers complained, arguing that moving it to a separate file would >> be breaking the compilation on existing tools (they're basically concerned with >> it due to out-of-tree drivers - mostly propietary ones, that use this API). > > It you move the API somewhere else, you will break userspace applications > like VDR. This is not acceptable. > >> Now that we're moving everything, it does make sense to do that, moving >> dvb/(audio|osd|video).h to someplace else (maybe linux/dvb/av7110.h or >> linux/dvb/legacy/*.h). > > As far as I understand the original patchset, it will not break > userspace, as it will simply move all files somewhere else, preserving > file names and the position of the files in the tree. > > Mauro is trying to the move the old decoder API somewhere else, possibly > into a different file, which will definitely break userspace. > NAK for this! I completely agree with Oliver on this and NAK the suggestion put forward by Mauro (and Hans) Regards, Manu -- 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/