debian: package building for ffmpeg-mt

The rules from ffmpeg-0.5.1 debian package have been adapted for building ffmpeg-mt.

debian/README.Source

+
+In the past, Debian used to ship "stripped" ffmpeg packages, i.e. the
+source code of the ffmpeg-debian package has been modified to disable
+specific codecs (mostly encoders like H263, H264, MPEG2 video, MPEG4 and
+MS-MPEG4) whose usage may represent patent infringement in certain
+jurisdictions. To serve the purpose to remain free of patented encoding
+technologies, the modifications to the ffmpeg source code have been done
+in a "non-reversible" way by removing several lines from the source code
+before packaging the release tarball. To reflect this divergence from
+upstream ffmpeg, the Debian package has been renamed to ffmpeg-debian.
+However, please note that only the code calling the affected codecs was
+removed in previous versions to make them unavailable to the resulting
+ffmpeg libraries. No encoder code was actually ever removed from the
+Debian packages!
+
+Several complications have come along with the aforementioned measures:
+* Further packaging hacks (e.g. debian/fixup-config.sh) have become
+  necessary in order to build the source code without the stripped
+  encoders.
+* It was impossible to rebuild unstripped packages for private usage
+  from the ffmpeg-debian source code
+* Ffmpeg upstream was not very happy about Debian redistributing a
+  stripped fork of their code, to say the least.
+
+Nowadays, while the situation remains pretty unchanged with regard to
+the patent threat, the ffmpeg build system has seen a lot of
+improvements. It now provides the possibility to explicitly disable
+specific codecs from the libraries at configure time, allowing disabling
+some patent encumbered codecs without the need to strip the source code.
+The effect on the resulting binary packages would remain the same as
+before.
+
+Regarding the source code, however, this would mean some significant
+improvements:
+* There is no more need to strip the source code in order to disable the
+  codecs and to apply further Debian-specific hacks to still make it
+  compile cleanly.
+* In order to rebuild unstripped ffmpeg libraries for private usage, it
+  would be sufficient to comment out a few configure flags.
+* Debian would not need to fork the ffmpeg source code anymore and call
+  their packages by a different name.
+
+Therefore, the pkg-multimedia-maintainers (the maintainers of the ffmpeg
+packages in Debian) decided to not strip the ffmpeg source code in
+further releases anymore but disable the patent encumbered codecs during
+the configure phase of the packages as intended upstream. We are sure
+this is the right thing to do with regard to the aforementioned
+advantages for both our users and ourselves as maintainers.

debian/README.source

+This package uses quilt to manage all modifications to the upstream
+source.  Changes are stored in the source package as diffs in
+debian/patches and applied during the build.
+
+For more information about quilt, see /usr/share/doc/quilt/README.source

debian/README.upstream-upgrade

+
+Checklist and howto for ffmpeg upstream upgrades:
+
+Needed packages:
+
+  apt-get install devscripts git-buildpackage pristine-tar
+
+Stage in a tmp directory:
+
+  mkdir /tmp/f ; cd /tmp/f
+  git clone git://git.debian.org/git/pkg-multimedia/ffmpeg-debian.git
+  cd ffmpeg-debian
+  branches="master.unstripped pristine-tar upstream upstream.unstripped"
+  for b in $branches; do git checkout -b $b origin/$b ; done
+
+Prepare the environment:
+
+  SVNDATE=`date +%Y%m%d`
+  git checkout master
+
+Fetch and commit the new upstream version:
+
+  debian/rules get-orig-source SVN_VERSION=${SVNDATE}
+  git-import-orig \
+	    --no-dch \
+            --upstream-tag=upstream/0.5+svn${SVNDATE} \
+            --upstream-version=0.5+svn${SVNDATE} \
+            ../ffmpeg_0.5+svn${SVNDATE}.orig.tar.gz
+
+Check and note the svn revision numbers from
+ffmpeg/{libswscale,}.svnrevision in debian/changelog
+
+  git diff upstream^ Changelog libavcodec/allcodecs.c libavformat/allformats.c
+
+Document new formats additions in debian/changelog
+
+Build, test, and compare against the the version already in the archive:
+  - headers in the -dev packages with 
+  - soname in the libraries
+  - formats.txt in the libavcodecs package
+
+Finialize debian/changelog, package should be upload ready now

debian/clean

+config-extra-includes.h
+EXTRA
+formats.txt

debian/compat

+7

debian/confflags

+# -*- mode: makefile -*-
+# vim:syntax=make
+
+# build a static version on every architecture in the 'debian' ffmpeg package
+FLAVORS := static
+
+# shared is generic, i.e. without arch specific opcodes
+FLAVORS += shared
+
+export DEB_HOST_GNU_TYPE  ?= $(shell dpkg-architecture -qDEB_HOST_GNU_TYPE)
+export DEB_BUILD_GNU_TYPE ?= $(shell dpkg-architecture -qDEB_BUILD_GNU_TYPE)
+export DEB_HOST_ARCH      ?= $(shell dpkg-architecture -qDEB_HOST_ARCH)
+
+SVNREVISION=$(shell cat .svnrevision 2>/dev/null || echo "UNKNOWN")
+
+# this is only used for the check_asm macro
+ifeq ($(DEB_BUILD_GNU_TYPE),$(DEB_HOST_GNU_TYPE))
+    CC := gcc
+else
+    CC := $(DEB_HOST_GNU_TYPE)-gcc
+endif
+
+# this outputs 0 or 1 depending on whether a piece of assembly can be compiled
+# with the *default* gcc flags; this is used to test the toolchain *default*
+# configuration
+check_asm = $(shell echo 'void foo(void) { __asm__ volatile("$(1)"); }' | $(CC) -x c -c - -o /dev/null 2>/dev/null && echo 1 || echo 0)
+
+# the other flavors always build dynamic versions
+# Also, disable architecture-specific optimizations for default shared build
+ifeq ($(DEB_HOST_ARCH),armel)
+    # whether the toolchain *default* configuration includes vfp and neon
+    vfp_asm := fadds s0, s0, s0
+    has_vfp := $(call check_asm, $(vfp_asm))
+    neon_asm := vadd.i16 q0, q0, q0
+    has_neon := $(call check_asm, $(neon_asm))
+
+    # only build 
+    ifneq ($(has_vfp),1)
+    FLAVORS += vfp
+    endif
+    ifneq ($(has_neon),1)
+    FLAVORS += neon
+    endif
+else ifeq ($(DEB_HOST_ARCH),i386)
+    FLAVORS += cmov
+else ifeq ($(DEB_HOST_ARCH),powerpc)
+    FLAVORS += altivec
+    nooptflags += --disable-altivec
+else ifeq ($(DEB_HOST_ARCH),sparc)
+    FLAVORS += vis
+    nooptflags += --disable-vis
+endif
+
+$(info Building FLAVORS=$(FLAVORS))
+
+define cond_enable
+	$(shell test -r $(1) && echo --enable-$(2) )
+endef
+
+# variant that also require --enable-version3
+define cond_enable_v3
+	$(shell test -r $(1) && echo --enable-$(2) --enable-version3 )
+endef
+
+# variant that also require --enable-nonfree
+define cond_enable_nf
+	$(shell test -r $(1) && echo --enable-$(2) --enable-nonfree )
+endef
+
+# Configuration flags causing the libs to be GPL tainted
+gpl_confflags += --enable-gpl
+gpl_confflags += --enable-postproc
+gpl_confflags += --enable-swscale
+gpl_confflags += --enable-x11grab
+
+# there is no libfaad in ubuntu/main, on in ubuntu/multiverse
+gpl_confflags += $(call cond_enable,/usr/include/faad.h,libfaad)
+
+# Common configuration flags
+confflags += --extra-version='$(DEB_VERSION)'
+confflags += --prefix=/usr
+confflags += --enable-avfilter
+confflags += --enable-avfilter-lavf
+confflags += --enable-vdpau
+confflags += --enable-bzlib
+confflags += --enable-libgsm
+confflags += --enable-libschroedinger
+confflags += --enable-libspeex
+confflags += --enable-libtheora
+confflags += --enable-libvorbis
+confflags += --enable-pthreads
+confflags += --enable-zlib
+confflags += --disable-stripping
+confflags += --enable-runtime-cpudetect
+ifeq      ($(DEB_HOST_ARCH),armel)
+# this is required on Ubuntu lucid as it defaults to thumb2 and ffmpeg has
+# plenty of incompatible assembly; not sure how to detect that properly
+confflags += --extra-cflags="-marm -fPIC -DPIC"
+endif
+confflags += $(extra_common_confflags)
+
+# this part below is intended for the 'ffmpeg' package in ubuntu/multiverse
+gpl_confflags += $(call cond_enable,/usr/include/xvid.h,libxvid)
+confflags += $(call cond_enable,/usr/include/lame/lame.h,libmp3lame)
+gpl_confflags += $(call cond_enable,/usr/include/x264.h,libx264)
+
+confflags += $(call cond_enable,/usr/include/lame/lame.h,libmp3lame)
+confflags += $(call cond_enable_nf,/usr/include/amrnb/sp_dec.h,libamr-nb)
+confflags += $(call cond_enable_nf,/usr/include/amrwb/dec.h,libamr-wb)
+
+# Opencore-amr requires GPL v3
+v3_confflags += $(call cond_enable_v3,/usr/include/opencore-amrnb/interf_dec.h,libopencore-amrnb)
+v3_confflags += $(call cond_enable_v3,/usr/include/opencore-amrwb/dec_if.h,libopencore-amrwb)
+
+# AAC is considered non-free upstream
+confflags += $(call cond_enable_nf,/usr/include/faac.h,libfaac)
+
+# comment out following line for LGPL versions of the libraries
+confflags += $(gpl_confflags)
+confflags += $(v3_confflags)
+
+# Enable IEEE 1394 (FireWire) support on Linux only
+ifneq (,$(findstring linux,$(DEB_HOST_GNU_TYPE)))
+  confflags += --enable-libdc1394
+  lib1394-dev += libraw1394-dev, libdc1394-22-dev
+endif
+
+# XXX this probably needs fixing
+CFLAGS :=
+
+ifneq (,$(findstring noopt,$(DEB_BUILD_OPTIONS)))
+# Various parts of ffmpeg (and swscale) FTBFS when compiling with -fPIC
+# and with mmx code enabled.
+  confflags += --disable-optimizations
+  confflags += --disable-mmx
+endif
+
+# Configuration flags for the static libraries
+static_build_confflags += $(confflags)
+
+# Configuration flags for the non-optimized shared libraries
+shared_build_confflags += $(confflags)
+# amd64 has no problems with optimized shared libs. i386 and arm do.
+ifneq      ($(DEB_HOST_ARCH),amd64)
+shared_build_confflags += $(nooptflags)
+endif
+shared_build_confflags += --enable-shared
+shared_build_confflags += --disable-static
+
+## armel architecture specific
+# Configuration flags for the optimised shared libraries
+vfp_build_confflags += $(confflags)
+vfp_build_confflags += --shlibdir=/usr/lib/vfp
+vfp_build_confflags += --enable-shared
+vfp_build_confflags += --disable-static
+vfp_build_confflags += --extra-cflags="-mfpu=vfp -mfloat-abi=softfp"
+vfp_build_confflags += --disable-ffmpeg
+vfp_build_confflags += --disable-ffplay
+# NB: NEON always implies v7+ and ffmpeg's NEON implementation requires VFP
+neon_build_confflags += $(confflags)
+neon_build_confflags += --shlibdir=/usr/lib/neon/vfp
+neon_build_confflags += --extra-cflags="-mfpu=neon -mfloat-abi=softfp -fPIC -DPIC"
+neon_build_confflags += --enable-shared
+neon_build_confflags += --disable-static
+neon_build_confflags += --disable-ffmpeg
+neon_build_confflags += --disable-ffplay
+
+## i386 architecture specific
+# Configuration flags for the optimized shared libraries
+cmov_build_confflags += $(confflags)
+cmov_build_confflags += $(nooptflags)
+cmov_build_confflags += --shlibdir=/usr/lib/i686/cmov
+cmov_build_confflags += --cpu='i686'
+cmov_build_confflags += --enable-shared
+cmov_build_confflags += --disable-static
+cmov_build_confflags += --disable-ffmpeg
+cmov_build_confflags += --disable-ffplay
+
+## powerpc architecture specific
+# Configuration flags for the optimized shared libraries
+altivec_build_confflags += $(confflags)
+altivec_build_confflags += --shlibdir=/usr/lib/altivec
+altivec_build_confflags += --cpu='g4'
+altivec_build_confflags += --enable-shared
+altivec_build_confflags += --disable-static
+altivec_build_confflags += --enable-altivec
+altivec_build_confflags += --disable-ffmpeg
+altivec_build_confflags += --disable-ffplay
+
+## sparc architecture specific
+# Configuration flags for the optimized shared libraries
+vis_build_confflags += $(confflags)
+vis_build_confflags += --shlibdir=/usr/lib/v9
+vis_build_confflags += --cpu='sparc64'
+vis_build_confflags += --enable-shared
+vis_build_confflags += --disable-static
+vis_build_confflags += --extra-cflags="-fPIC -DPIC"
+vis_build_confflags += --disable-ffmpeg
+vis_build_confflags += --disable-ffplay
+
+# Additional documentation for PowerPC
+ifneq (,$(findstring powerpc,$(DEB_BUILD_GNU_TYPE)))
+  extradoc := doc/ffmpeg_powerpc_performance_evaluation_howto.txt
+endif
+

debian/control

+Source: ffmpeg
+Section: libs
+Priority: optional
+Maintainer: iphion BV <techlist@iphion.nl>
+XSBC-Original-Maintainer: Debian multimedia packages maintainers <pkg-multimedia-maintainers@lists.alioth.debian.org>
+Uploaders: Sam Hocevar (Debian packages) <sam+deb@zoy.org>,
+           Loic Minier <lool@dooz.org>,
+           Reinhard Tartler <siretart@tauware.de>,
+           Fabian Greffrath <fabian+debian@greffrath.com>,
+           Andres Mejia <mcitadel@gmail.com>
+DM-Upload-Allowed: yes
+Standards-Version: 3.8.3
+Vcs-Git: git://git.m2x.eu/ffmpeg-mt.git
+Vcs-Browser: http://git.m2x.eu/git/cgi-bin/gitweb.cgi?p=ffmpeg-mt.git;a=summary
+Homepage: http://gitorious.org/ffmpeg/ffmpeg-mt
+Build-Depends-Indep: doxygen
+Build-Depends: debhelper (>= 7),
+               libasound2-dev [!kfreebsd-i386 !kfreebsd-amd64 !hurd-i386],
+               libbz2-dev,
+               libdc1394-22-dev [!kfreebsd-i386 !kfreebsd-amd64 !hurd-i386],
+               libfreetype6-dev,
+               libgsm1-dev,
+               libimlib2-dev,
+               libraw1394-dev [!kfreebsd-i386 !kfreebsd-amd64 !hurd-i386],
+               libschroedinger-dev,
+               libsdl1.2-dev,
+               libspeex-dev,
+               libtheora-dev (>> 0.0.0.alpha4),
+               libvorbis-dev,
+               libx11-dev,
+               libxext-dev,
+               libvdpau-dev,
+               libxvmc-dev,
+               quilt,
+               texi2html,
+               yasm [i386 amd64],
+               zlib1g-dev
+
+Package: ffmpeg
+Section: video
+Architecture: any
+Depends: ${shlibs:Depends},
+         ${misc:Depends}
+Description: multimedia player, server and encoder
+ This package contains the ffplay multimedia player, the ffserver streaming
+ server and the ffmpeg audio and video encoder. They support most existing
+ file formats (AVI, MPEG, OGG, Matroska, ASF...) and encoding formats (MPEG,
+ DivX, MPEG4, AC3, DV...).
+
+Package: ffmpeg-dbg
+Section: debug
+Priority: extra
+Architecture: any
+Depends: libavutil49 (= ${binary:Version}),
+         libavcodec52 (= ${binary:Version}),
+         libavdevice52 (= ${binary:Version}),
+         libpostproc51 (= ${binary:Version}),
+         libavformat52 (= ${binary:Version}),
+         libswscale0 (= ${binary:Version}),
+         ffmpeg (= ${binary:Version}),
+         ${misc:Depends}
+Description: Debug symbols for ffmpeg related packages
+ This package contains debug data of the ffmpeg related shared libraries.
+ .
+ Most people will not need this package. Please install it to produce useful
+ stacktraces to help debugging the ffmpeg library.
+
+Package: ffmpeg-doc
+Section: doc
+Architecture: all
+Depends: ${misc:Depends}
+Description: documentation of the ffmpeg API
+ This package contains the html doxygen documentation of the ffmpeg API.
+
+Package: libavutil49
+Architecture: any
+Depends: ${shlibs:Depends},
+         ${misc:Depends}
+Description: ffmpeg utility library
+ This is the common utility library from the ffmpeg project. It is required
+ by all other ffmpeg libraries.
+ .
+ This package contains a Debian-specific version of the libavutil shared
+ object that should only be used by Debian packages.
+
+Package: libavcodec52
+Architecture: any
+Depends: ${shlibs:Depends},
+         ${misc:Depends}
+Replaces: ffmpeg (<< 4:0.5.1-1)
+Description: ffmpeg codec library
+ This is the codec library from the ffmpeg project which suports multi-threaded
+ decoding. It supports most existing encoding formats (MPEG, DivX, MPEG4, AC3, DV...)
+ .
+ This package contains a Debian-specific version of the libavcodec shared
+ object that should only be used by Debian packages.
+
+Package: libavdevice52
+Architecture: any
+Depends: ${shlibs:Depends},
+         ${misc:Depends}
+Description: ffmpeg device handling library
+ This is the device handling library from the ffmpeg project.
+ .
+ This package contains a Debian-specific version of the libavdevice shared
+ object that should only be used by Debian packages.
+
+Package: libavformat52
+Architecture: any
+Depends: ${shlibs:Depends},
+         ${misc:Depends}
+Breaks: libavcodec51 (<< 3:0.svn20090303-1)
+Description: ffmpeg file format library
+ This is the demuxer library from the ffmpeg project. It supports most
+ existing file formats (AVI, MPEG, OGG, Matroska, ASF...).
+ .
+ This package contains a Debian-specific version of the libavformat shared
+ object that should only be used by Debian packages.
+
+Package: libavfilter0
+Architecture: any
+Depends: ${shlibs:Depends},
+         ${misc:Depends}
+Description: ffmpeg video filtering library
+ This is the video filtering library from the ffmpeg project.
+ .
+ This package contains a Debian-specific version of the libavfilter shared
+ object that should only be used by Debian packages.
+
+Package: libpostproc51
+Architecture: any
+Depends: ${shlibs:Depends},
+         ${misc:Depends}
+Description: ffmpeg video postprocessing library
+ This is the video postprocessing library from the ffmpeg project.
+ .
+ This package contains a Debian-specific version of the libpostproc shared
+ object that should only be used by Debian packages.
+
+Package: libswscale0
+Architecture: any
+Depends: ${shlibs:Depends},
+         ${misc:Depends}
+Description: ffmpeg video scaling library
+ This is the video scaling library from the ffmpeg project.
+ .
+ This package contains a Debian-specific version of the libswscale shared
+ object that should only be used by Debian packages.
+
+Package: libavutil-dev
+Section: libdevel
+Architecture: any
+Depends: libavutil49 (>= ${binary:Version}) | libavutil-extra-49 (>= ${source:Upstream-Version}),
+         libavutil49 (<= ${source:Upstream-Version}-99) | libavutil-extra-49 (<= ${source:Upstream-Version}-99),
+         ${misc:Depends}
+Description: development files for libavutil
+ This is the common utility library from the ffmpeg project. It is required
+ by all other ffmpeg libraries.
+ .
+ This package contains the header files and static libraries needed to
+ compile applications or shared objects that use libavutil.
+
+Package: libavcodec-dev
+Section: libdevel
+Architecture: any
+Depends: libavcodec52 (>= ${binary:Version}) | libavcodec-extra-52 (>= ${source:Upstream-Version}),
+         libavcodec52 (<= ${source:Upstream-Version}-99) | libavcodec-extra-52 (<= ${source:Upstream-Version}-99),
+         libavutil-dev (= ${binary:Version}),
+         ${misc:Depends}
+Suggests: libfaad-dev,
+          libgsm1-dev,
+          libogg-dev,
+          libschroedinger-dev,
+          libspeex-dev,
+          libtheora-dev (>> 0.0.0.alpha4),
+          libvorbis-dev,
+          libx11-dev,
+          libxext-dev,
+          zlib1g-dev,
+          ${lib1394-dev}
+Description: development files for libavcodec
+ This is the codec library from the ffmpeg project. It supports most existing
+ encoding formats (MPEG, DivX, MPEG4, AC3, DV...).
+ .
+ This package contains the header files and static libraries needed to
+ compile applications or shared objects that use libavcodec.
+
+Package: libavdevice-dev
+Section: libdevel
+Architecture: any
+Depends: libavdevice52 (>= ${binary:Version}) | libavdevice-extra-52 (>= ${source:Upstream-Version}),
+         libavdevice52 (<= ${source:Upstream-Version}-99) | libavdevice-extra-52 (<= ${source:Upstream-Version}-99),
+         libavformat-dev (= ${binary:Version}),
+         ${misc:Depends}
+Description: development files for libavdevice
+ This is the device handling library from the ffmpeg project.
+ .
+ This package contains the header files and static libraries needed to
+ compile applications or shared objects that use libavdevice.
+
+Package: libavformat-dev
+Section: libdevel
+Architecture: any
+Depends: libavformat52 (>= ${binary:Version}) | libavformat-extra-52 (>= ${source:Upstream-Version}),
+         libavformat52 (<= ${source:Upstream-Version}-99) | libavformat-extra-52 (<= ${source:Upstream-Version}-99),
+         libavcodec-dev (= ${binary:Version}),
+         ${misc:Depends}
+Description: development files for libavformat
+ This is the demuxer library from the ffmpeg project. It supports most
+ existing file formats (AVI, MPEG, OGG, Matroska, ASF...).
+ .
+ This package contains the header files and static libraries needed to
+ compile applications or shared objects that use libavformat.
+
+Package: libavfilter-dev
+Section: libdevel
+Architecture: any
+Depends: libavfilter0 (>= ${binary:Version}) | libavfilter-extra-0 (>= ${source:Upstream-Version}),
+         libavfilter0 (<= ${source:Upstream-Version}-99) | libavfilter-extra-0 (<= ${source:Upstream-Version}-99),
+         libavcodec-dev (= ${binary:Version}),
+         ${misc:Depends}
+Description: development files for libavfilter
+ This is the video filtering library from the ffmpeg project.
+ .
+ This package contains the header files and static libraries needed to
+ compile applications or shared objects that use libavfilter.
+
+Package: libpostproc-dev
+Section: libdevel
+Architecture: any
+Depends: libpostproc51 (>= ${binary:Version}) | libpostproc-extra-51 (>= ${source:Upstream-Version}),
+         libpostproc51 (<= ${source:Upstream-Version}-99) | libpostproc-extra-51 (<= ${source:Upstream-Version}-99),
+         libavutil-dev (= ${binary:Version}),
+         ${misc:Depends}
+Description: development files for libpostproc
+ This is the video postprocessing library from the ffmpeg project.
+ .
+ This package contains the header files and static libraries needed to
+ compile applications or shared objects that use libpostproc.
+
+Package: libswscale-dev
+Section: libdevel
+Architecture: any
+Depends: libswscale0 (>= ${binary:Version}) | libswscale-extra-0 (>= ${source:Upstream-Version}),
+         libswscale0 (<= ${source:Upstream-Version}-99) | libswscale-extra-0 (<= ${source:Upstream-Version}-99),
+         libavutil-dev (= ${binary:Version}),
+         ${misc:Depends}
+Description: development files for libswscale
+ This is the video scaling library from the ffmpeg project.
+ .
+ This package contains the header files and static libraries needed to
+ compile applications or shared objects that use libswscale.

debian/copyright

+Git snapshots are downloaded with git from ffmpeg-mt at:
+    http://gitorious.org/ffmpeg/ffmpeg-mt
+    or copy at
+    http://git.m2x.eu/git/cgi-bin/gitweb.cgi?p=ffmpeg-mt.git;a=summary
+
+SVN snapshots are downloaded with subversion from the ffmpeg SVN at:
+    <svn://svn.ffmpeg.org/ffmpeg/>
+
+Upstream Authors: Fabrice Bellard
+                  Alex Beregszaszi
+                  BERO
+                  Mario Brito
+                  Ronald Bultje
+                  Tim Ferguson
+                  Brian Foley
+                  Arpad Gereoffy
+                  Philip Gladstone
+                  Vladimir Gneushev
+                  Wolfgang Hesseler
+                  Falk Hueffner
+                  Zdenek Kabelac
+                  Robin Kay
+                  Todd Kirby
+                  Nick Kurshev
+                  Mike Melanson
+                  Michael Niedermayer
+                  François Revol
+                  Roman Shaposhnik
+                  Dieter Shirley
+                  Juan J. Sierralta
+                  Ewald Snel
+                  Leon van Stuivenberg
+                  Roberto Togni
+                  Lionel Ulmer
+
+Copyright (c) 2000-2004 Fabrice Bellard et al.
+
+The following files are licensed under the GNU GPL, as clarified below:
+
+ * ffmpeg.c
+ * libavcodec:
+   + dtsdec.c
+   + i386/idct_mmx.c
+   + liba52/*.[ch]
+ * libavformat:
+   + x11grab.c
+   + gxfenc.c
+ * libpostproc:
+   + postprocess_internal.h
+   + postprocess_altivec_template.c
+   + postprocess.h
+   + postprocess_template.c
+   + postprocess.c
+   + mangle.h
+ * libswscale:
+   + swscale.c
+   + swscale-example.c
+   + yuv2rgb_template.c
+   + swscale_altivec_template.c
+   + yuv2rgb_altivec.c
+   + swscale_template.c
+   + rgb2rgb_template.c
+   + rgb2rgb.c
+   + cs_test.c
+   + yuv2rgb_mlib.c
+   + yuv2rgb.c
+
+ | This library is free software; you can redistribute it and/or
+ | modify it under the terms of the GNU General Public License as
+ | published by the Free Software Foundation; either version 2 of
+ | the License, or (at your option) any later version.
+ |
+ | This library is distributed in the hope that it will be useful,
+ | but WITHOUT ANY WARRANTY; without even the implied warranty of
+ | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ | Lesser General Public License for more details.
+ |
+ | You should have received a copy of the GNU General Public License
+ | along with this program; if not, write to the Free Software
+ | Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301 USA
+
+The rest of the code is licensed under the GNU LGPL:
+
+ | This library is free software; you can redistribute it and/or
+ | modify it under the terms of the GNU Lesser General Public License as
+ | published by the Free Software Foundation; either version 2.1 of
+ | the License, or (at your option) any later version.
+ |
+ | This library is distributed in the hope that it will be useful,
+ | but WITHOUT ANY WARRANTY; without even the implied warranty of
+ | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ | Lesser General Public License for more details.
+ |
+ | You should have received a copy of the GNU General Public License
+ | along with this program; if not, write to the Free Software
+ | Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301 USA
+
+On Debian GNU/Linux systems, the complete text of the GNU General Public
+License can be found in `/usr/share/common-licenses/GPL' and the text of the
+GNU Lesser General Public License is in `/usr/share/common-licenses/LGPL'.
+

debian/ffmpeg-doc.doc-base

+Document: ffmpeg-doc
+Title: ffmpeg API Documentation
+Author: FFmpeg Developers
+Abstract: This is the main documentation for the ffmpeg API.
+Section: Programming
+
+Format: HTML
+Index: /usr/share/doc/ffmpeg-doc/html/index.html
+Files: /usr/share/doc/ffmpeg-doc/html/*.html

debian/ffmpeg-doc.docs

+debian/tmp/usr/share/doc/ffmpeg-doc/html

debian/ffmpeg.docs

+debian/tmp/usr/share/doc/ffmpeg/html

debian/ffmpeg.install

+etc
+usr/bin
+usr/share/man

debian/files

+ffmpeg_0.5.1-1ubuntu1_i386.deb video optional
+ffmpeg-dbg_0.5.1-1ubuntu1_i386.deb debug extra
+ffmpeg-doc_0.5.1-1ubuntu1_all.deb doc optional
+libavutil49_0.5.1-1ubuntu1_i386.deb libs optional
+libavcodec52_0.5.1-1ubuntu1_i386.deb libs optional
+libavdevice52_0.5.1-1ubuntu1_i386.deb libs optional
+libavformat52_0.5.1-1ubuntu1_i386.deb libs optional
+libavfilter0_0.5.1-1ubuntu1_i386.deb libs optional
+libpostproc51_0.5.1-1ubuntu1_i386.deb libs optional
+libswscale0_0.5.1-1ubuntu1_i386.deb libs optional
+libavutil-dev_0.5.1-1ubuntu1_i386.deb libdevel optional
+libavcodec-dev_0.5.1-1ubuntu1_i386.deb libdevel optional
+libavdevice-dev_0.5.1-1ubuntu1_i386.deb libdevel optional
+libavformat-dev_0.5.1-1ubuntu1_i386.deb libdevel optional
+libavfilter-dev_0.5.1-1ubuntu1_i386.deb libdevel optional
+libpostproc-dev_0.5.1-1ubuntu1_i386.deb libdevel optional
+libswscale-dev_0.5.1-1ubuntu1_i386.deb libdevel optional

debian/gbp.conf

+[DEFAULT]
+upstream-branch = upstream-0.5
+debian-branch = ubuntu
+upstream-tag = upstream/%(version)s
+debian-tag = debian/%(version)s
+pristine-tar = True

debian/get-orig-source.sh

+#!/bin/sh
+#
+#  Script to create a 'pristine' tarball for the debian ffmpeg source package
+#  Copyright (C) 2008, 2009, 2010 Reinhard Tartler
+#
+#  This program is free software; you can redistribute it and/or modify
+#  it under the terms of the GNU General Public License as published by
+#  the Free Software Foundation; either version 2 of the License, or
+#  (at your option) any later version.
+#
+#  This program is distributed in the hope that it will be useful,
+#  but WITHOUT ANY WARRANTY; without even the implied warranty of
+#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+#  GNU General Public License for more details.
+#
+#  You should have received a copy of the GNU General Public License along
+#  with this program; if not, write to the Free Software Foundation, Inc.,
+#  51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+set -eu
+
+usage() {
+	cat >&2 <<EOF
+usage: $0 [-dh]
+  -h : display help
+  -d : date of svn snapshot
+  -o : output tarball name
+  -c : path to cleanup script
+EOF
+}
+
+debug () {
+	$DEBUG && echo "DEBUG: $*" >&2
+}
+
+error () {
+	echo "$1" >&2
+	exit 1;
+}
+
+set +e
+PARAMS=`getopt hd: "$@"`
+if test $? -ne 0; then usage; exit 1; fi;
+set -e
+
+eval set -- "$PARAMS"
+
+DEBUG=false
+SVNDATE=
+
+while test $# -gt 0
+do
+	case $1 in
+		-h) usage; exit 1 ;;
+		-d) SVNDATE=$2; shift ;;
+		--) shift ; break ;;
+		*)  echo "Internal error!" ; exit 1 ;;
+	esac
+	shift
+done
+
+# sanity checks now
+dh_testdir
+
+if [ -z $SVNDATE ]; then
+	error "you need to specify an svn date. e.g. 20081230 for Dec 29. 2008"
+fi
+
+TARBALL=../ffmpeg_0.5+svn${SVNDATE}.orig.tar.gz
+PACKAGENAME=ffmpeg
+
+TMPDIR=`mktemp -d`
+trap 'rm -rf ${TMPDIR}'  EXIT
+
+baseurl="svn://svn.ffmpeg.org/ffmpeg/branches/0.5"
+
+svn export -r{${SVNDATE}} \
+	--ignore-externals \
+	${baseurl}  \
+	${TMPDIR}/${PACKAGENAME}
+
+svn info -r{${SVNDATE}} \
+	${baseurl} \
+	| awk '/^Revision/ {print $2}' \
+	> ${TMPDIR}/${PACKAGENAME}/.svnrevision
+
+tar czf ${TARBALL} -C ${TMPDIR} ${PACKAGENAME}
+

debian/libavcodec-dev.install

+usr/include/libavcodec
+usr/lib/libavcodec.a
+usr/lib/{,*/,*/*/}libavcodec.so
+usr/lib/pkgconfig/libavcodec.pc

debian/libavcodec52.install

+usr/lib/{,*/,*/*/}libavcodec.so.*
+usr/share/ffmpeg/*.ffpreset

debian/libavcodec52.lintian-overrides

+# Overriding these fpic lintian errors. Please see bug #528080.
+libavcodec52: shlib-with-non-pic-code usr/lib/i686/cmov/libavcodec.so.52.20.0
+libavcodec52: shlib-with-non-pic-code usr/lib/libavcodec.so.52.20.0

debian/libavdevice-dev.install

+usr/include/libavdevice
+usr/lib/libavdevice.a
+usr/lib/{,*/,*/*/}libavdevice.so
+usr/lib/pkgconfig/libavdevice.pc

debian/libavdevice52.install

+usr/lib/{,*/,*/*/}libavdevice.so.*

debian/libavdevice52.lintian-overrides

+# Overriding these fpic lintian errors. Please see bug #528080.
+libavdevice52: shlib-with-non-pic-code usr/lib/i686/cmov/libavdevice.so.52.1.0
+libavdevice52: shlib-with-non-pic-code usr/lib/libavdevice.so.52.1.0

debian/libavfilter-dev.install

+usr/include/libavfilter
+usr/lib/libavfilter.a
+usr/lib/{,*/,*/*/}libavfilter.so
+usr/lib/pkgconfig/libavfilter.pc

debian/libavfilter0.install

+usr/lib/{,*/,*/*/}libavfilter.so.*

debian/libavfilter0.lintian-overrides

+# Overriding these fpic lintian errors. Please see bug #528080.
+libavfilter0: shlib-with-non-pic-code usr/lib/i686/cmov/libavfilter.so.0.4.0
+libavfilter0: shlib-with-non-pic-code usr/lib/libavfilter.so.0.4.0

debian/libavformat-dev.install

+usr/include/libavformat
+usr/lib/libavformat.a
+usr/lib/{,*/,*/*/}libavformat.so
+usr/lib/pkgconfig/libavformat.pc

debian/libavformat52.install

+usr/lib/{,*/,*/*/}libavformat.so.*

debian/libavformat52.lintian-overrides

+# Overriding these fpic lintian errors. Please see bug #528080.
+libavformat52: shlib-with-non-pic-code usr/lib/i686/cmov/libavformat.so.52.31.0
+libavformat52: shlib-with-non-pic-code usr/lib/libavformat.so.52.31.0

debian/libavutil-dev.install

+usr/include/libavutil
+usr/lib/libavutil.a
+usr/lib/{,*/,*/*/}libavutil.so
+usr/lib/pkgconfig/libavutil.pc

debian/libavutil49.install

+usr/lib/{,*/,*/*/}libavutil.so.*

debian/libavutil49.lintian-overrides

+# Overriding these fpic lintian errors. Please see bug #528080.
+libavutil49: shlib-with-non-pic-code usr/lib/i686/cmov/libavutil.so.49.15.0
+libavutil49: shlib-with-non-pic-code usr/lib/libavutil.so.49.15.0

debian/libpostproc-dev.install

+usr/include/libpostproc
+usr/lib/libpostproc.a
+usr/lib/{,*/,*/*/}libpostproc.so
+usr/lib/pkgconfig/libpostproc.pc

debian/libpostproc51.install

+usr/lib/{,*/,*/*/}libpostproc.so.*

debian/libpostproc51.lintian-overrides

+# Overriding these fpic lintian errors. Please see bug #528080.
+libpostproc51: shlib-with-non-pic-code usr/lib/i686/cmov/libpostproc.so.51.2.0
+libpostproc51: shlib-with-non-pic-code usr/lib/libpostproc.so.51.2.0

debian/libswscale-dev.install

+usr/include/libswscale
+usr/lib/libswscale.a
+usr/lib/{,*/,*/*/}libswscale.so
+usr/lib/pkgconfig/libswscale.pc

debian/libswscale0.install

+usr/lib/{,*/,*/*/}libswscale.so.*

debian/libswscale0.lintian-overrides

+# Overriding these fpic lintian errors. Please see bug #528080.
+libswscale0: shlib-with-non-pic-code usr/lib/i686/cmov/libswscale.so.0.7.1
+libswscale0: shlib-with-non-pic-code usr/lib/libswscale.so.0.7.1

debian/patches/900_doxyfile

+Author: Reinhard Tartler <siretart@tauware.de>
+
+Exclude some directories we use for packaging.
+
+==========================================================================
+--- a/Doxyfile
++++ b/Doxyfile
+@@ -359,7 +359,7 @@
+ # excluded from the INPUT source files. This way you can easily exclude a
+ # subdirectory from a directory tree whose root is specified with the INPUT tag.
+ 
+-EXCLUDE                =
++EXCLUDE                = debian debian-shared debian-static debian-cmov .pc .git
+ 
+ # The EXCLUDE_SYMLINKS tag can be used select whether or not files or directories
+ # that are symbolic links (a Unix filesystem feature) are excluded from the input.

debian/patches/901-fix-misc-typos.patch

+From: stefano <stefano@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Wed, 27 Jan 2010 00:31:52 +0000 (+0000)
+Subject: Fix misc typos, patch by
+X-Git-Url: http://git.ffmpeg.org/?p=ffmpeg;a=commitdiff_plain;h=ba4f92903641af2f11d47d1ee845a62abfd5dd38
+
+Fix misc typos, patch by
+Fabian Greffrath base64_decode(PGZhYmlhbkBncmVmZnJhdGguY29tPg==).
+
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@21473 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+
+--- ffmpeg.orig/libavcodec/mjpegdec.c
++++ ffmpeg/libavcodec/mjpegdec.c
+@@ -967,7 +967,7 @@
+     }
+ 
+     if(s->avctx->debug & FF_DEBUG_PICT_INFO)
+-        av_log(s->avctx, AV_LOG_DEBUG, "%s %s p:%d >>:%d ilv:%d bits:%d %s\n", s->lossless ? "lossless" : "sequencial DCT", s->rgb ? "RGB" : "",
++        av_log(s->avctx, AV_LOG_DEBUG, "%s %s p:%d >>:%d ilv:%d bits:%d %s\n", s->lossless ? "lossless" : "sequential DCT", s->rgb ? "RGB" : "",
+                predictor, point_transform, ilv, s->bits,
+                s->pegasus_rct ? "PRCT" : (s->rct ? "RCT" : ""));
+ 
+--- ffmpeg.orig/libavcodec/opt.c
++++ ffmpeg/libavcodec/opt.c
+@@ -424,7 +424,7 @@
+             break;
+             case FF_OPT_TYPE_INT64:
+                 if((double)(opt->default_val+0.6) == opt->default_val)
+-                    av_log(s, AV_LOG_DEBUG, "loss of precission in default of %s\n", opt->name);
++                    av_log(s, AV_LOG_DEBUG, "loss of precision in default of %s\n", opt->name);
+                 av_set_int(s, opt->name, opt->default_val);
+             break;
+             case FF_OPT_TYPE_FLOAT: {
+--- ffmpeg.orig/libavcodec/options.c
++++ ffmpeg/libavcodec/options.c
+@@ -156,7 +156,7 @@
+ {"very", "strictly conform to a older more strict version of the spec or reference software", 0, FF_OPT_TYPE_CONST, FF_COMPLIANCE_VERY_STRICT, INT_MIN, INT_MAX, V|D|E, "strict"},
+ {"strict", "strictly conform to all the things in the spec no matter what consequences", 0, FF_OPT_TYPE_CONST, FF_COMPLIANCE_STRICT, INT_MIN, INT_MAX, V|D|E, "strict"},
+ {"normal", NULL, 0, FF_OPT_TYPE_CONST, FF_COMPLIANCE_NORMAL, INT_MIN, INT_MAX, V|D|E, "strict"},
+-{"inofficial", "allow inofficial extensions", 0, FF_OPT_TYPE_CONST, FF_COMPLIANCE_INOFFICIAL, INT_MIN, INT_MAX, V|D|E, "strict"},
++{"inofficial", "allow unofficial extensions", 0, FF_OPT_TYPE_CONST, FF_COMPLIANCE_INOFFICIAL, INT_MIN, INT_MAX, V|D|E, "strict"},
+ {"experimental", "allow non standardized experimental things", 0, FF_OPT_TYPE_CONST, FF_COMPLIANCE_EXPERIMENTAL, INT_MIN, INT_MAX, V|D|E, "strict"},
+ {"b_qoffset", "qp offset between P and B frames", OFFSET(b_quant_offset), FF_OPT_TYPE_FLOAT, 1.25, -FLT_MAX, FLT_MAX, V|E},
+ {"er", "set error detection aggressivity", OFFSET(error_recognition), FF_OPT_TYPE_INT, FF_ER_CAREFUL, INT_MIN, INT_MAX, A|V|D, "er"},

debian/patches/fix-constant-quant-encoding.patch

+From: conrad <conrad@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Wed, 30 Sep 2009 22:25:49 +0000 (+0000)
+Subject: Support constant-quant encoding for libtheora
+X-Git-Url: http://git.ffmpeg.org/?p=ffmpeg;a=commitdiff_plain;h=a5ac197b1e60ca44cbc1ea0f64e25af977f7dae5
+
+Support constant-quant encoding for libtheora
+Based on a patch by j f (gonzaz-z # operamail (d) com)
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@20109 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+
+diff --git a/libavcodec/libtheoraenc.c b/libavcodec/libtheoraenc.c
+index 9da0d19..f028595 100644
+--- a/libavcodec/libtheoraenc.c
++++ b/libavcodec/libtheoraenc.c
+@@ -106,11 +106,9 @@ static av_cold int encode_init(AVCodecContext* avc_context)
+     }
+     t_info.colorspace = OC_CS_UNSPECIFIED;
+     t_info.pixelformat = OC_PF_420;
+-    t_info.target_bitrate = avc_context->bit_rate;
+     t_info.keyframe_frequency = avc_context->gop_size;
+     t_info.keyframe_frequency_force = avc_context->gop_size;
+     t_info.keyframe_mindistance = avc_context->keyint_min;
+-    t_info.quality = 0;
+ 
+     t_info.quick_p = 1;
+     t_info.dropframes_p = 0;
+@@ -120,6 +118,19 @@ static av_cold int encode_init(AVCodecContext* avc_context)
+     t_info.noise_sensitivity = 1;
+     t_info.sharpness = 0;
+ 
++    if (avc_context->flags & CODEC_FLAG_QSCALE) {
++        /* to be constant with the libvorbis implementation, clip global_quality to 0 - 10
++           Theora accepts a quality parameter p, which is:
++                * 0 <= p <=63
++                * an int value
++         */
++        t_info.quality = av_clip(avc_context->global_quality / (float)FF_QP2LAMBDA, 0, 10) * 6.3;
++        t_info.target_bitrate = 0;
++    } else {
++        t_info.target_bitrate = avc_context->bit_rate;
++        t_info.quality = 0;
++    }
++
+     /* Now initialise libtheora */
+     if (theora_encode_init( &(h->t_state), &t_info ) != 0) {
+         av_log(avc_context, AV_LOG_ERROR, "theora_encode_init failed\n");

debian/patches/fix-dv-seeking.patch

+Author: Reinhard Tartler <siretart@tauware.de>
+
+fix seeking in DIF (DV) movies, closes Bug: #540424
+
+Thanks to Dan Dennedy <dan@dennedy.org> for identifying the patch!
+
+------------------------------------------------------------------------
+r19192 | bcoudurier | 2009-06-15 00:34:28 +0200 (Mo, 15. Jun 2009) | 1 line
+
+check if frame size matches old sys and assumes corrupted input, fixes #1192
+------------------------------------------------------------------------
+
+--- a/libavcodec/dv.c
++++ b/libavcodec/dv.c
+@@ -1115,7 +1115,7 @@ static int dvvideo_decode_frame(AVCodecC
+ {
+     DVVideoContext *s = avctx->priv_data;
+ 
+-    s->sys = dv_frame_profile(buf);
++    s->sys = dv_frame_profile(s->sys, buf, buf_size);
+     if (!s->sys || buf_size < s->sys->frame_size || dv_init_dynamic_tables(s->sys))
+         return -1; /* NOTE: we only accept several full frames */
+ 
+--- a/libavcodec/dvdata.h
++++ b/libavcodec/dvdata.h
+@@ -698,7 +698,9 @@ enum dv_pack_type {
+  */
+ #define DV_MAX_BPM 8
+ 
+-static inline const DVprofile* dv_frame_profile(const uint8_t* frame)
++static inline
++const DVprofile* dv_frame_profile(const DVprofile *sys,
++                                  const uint8_t* frame, unsigned buf_size)
+ {
+    int i;
+ 
+@@ -715,6 +717,10 @@ static inline const DVprofile* dv_frame_
+        if (dsf == dv_profiles[i].dsf && stype == dv_profiles[i].video_stype)
+            return &dv_profiles[i];
+ 
++   /* check if old sys matches and assumes corrupted input */
++   if (sys && buf_size == sys->frame_size)
++       return sys;
++
+    return NULL;
+ }
+ 
+--- a/libavformat/dv.c
++++ b/libavformat/dv.c
+@@ -322,7 +322,7 @@ int dv_produce_packet(DVDemuxContext *c,
+     uint8_t *ppcm[4] = {0};
+ 
+     if (buf_size < DV_PROFILE_BYTES ||
+-        !(c->sys = dv_frame_profile(buf)) ||
++        !(c->sys = dv_frame_profile(c->sys, buf, buf_size)) ||
+         buf_size < c->sys->frame_size) {
+           return -1;   /* Broken frame, or not enough data */
+     }
+@@ -421,7 +421,7 @@ static int dv_read_header(AVFormatContex
+         url_fseek(s->pb, -DV_PROFILE_BYTES, SEEK_CUR) < 0)
+         return AVERROR(EIO);
+ 
+-    c->dv_demux->sys = dv_frame_profile(c->buf);
++    c->dv_demux->sys = dv_frame_profile(c->dv_demux->sys, c->buf, DV_PROFILE_BYTES);
+     if (!c->dv_demux->sys) {
+         av_log(s, AV_LOG_ERROR, "Can't determine profile of DV input stream.\n");
+         return -1;

debian/patches/fix-qtrle-encoding-when-previous-frame-linesize.patch

+From: Peter Fritzsche <peter.fritzsche@gmx.de>
+Date: Sat, 23 May 2009 05:55:29 +0000
+Subject: [PATCH] fix qtrle encoding when previous frame linesize differs
+
+Applied upstream as revision 18908:
+
+------------------------------------------------------------------------
+r18908 | bcoudurier | 2009-05-23 07:55:29 +0200 (Sa, 23. Mai 2009) | 1 line
+
+fix qtrle encoding when previous frame linesize differs, fix #998
+
+
+diff --git a/libavcodec/qtrleenc.c b/libavcodec/qtrleenc.c
+index 171ac25..46bb573 100644
+--- a/libavcodec/qtrleenc.c
++++ b/libavcodec/qtrleenc.c
+@@ -121,18 +121,20 @@ static void qtrle_encode_line(QtrleEncContext *s, AVFrame *p, int line, uint8_t
+     /* The cost of the three different possibilities */
+     int total_bulk_cost;
+     int total_skip_cost;
+     int total_repeat_cost;
+ 
+     int temp_cost;
+     int j;
+ 
+-    uint8_t *this_line = p->               data[0] + line*p->linesize[0] + (width - 1)*s->pixel_size;
+-    uint8_t *prev_line = s->previous_frame.data[0] + line*p->linesize[0] + (width - 1)*s->pixel_size;
++    uint8_t *this_line = p->               data[0] + line*p->               linesize[0] +
++        (width - 1)*s->pixel_size;
++    uint8_t *prev_line = s->previous_frame.data[0] + line*s->previous_frame.linesize[0] +
++        (width - 1)*s->pixel_size;
+ 
+     s->length_table[width] = 0;
+     skipcount = 0;
+ 
+     for (i = width - 1; i >= 0; i--) {
+ 
+         if (!s->frame.key_frame && !memcmp(this_line, prev_line, s->pixel_size))
+             skipcount = FFMIN(skipcount + 1, MAX_RLE_SKIP);
+@@ -235,26 +237,27 @@ static void qtrle_encode_line(QtrleEncContext *s, AVFrame *p, int line, uint8_t
+ static int encode_frame(QtrleEncContext *s, AVFrame *p, uint8_t *buf)
+ {
+     int i;
+     int start_line = 0;
+     int end_line = s->avctx->height;
+     uint8_t *orig_buf = buf;
+ 
+     if (!s->frame.key_frame) {
++        unsigned line_size = s->avctx->width * s->pixel_size;
+         for (start_line = 0; start_line < s->avctx->height; start_line++)
+             if (memcmp(p->data[0] + start_line*p->linesize[0],
+-                       s->previous_frame.data[0] + start_line*p->linesize[0],
+-                       p->linesize[0]))
++                       s->previous_frame.data[0] + start_line*s->previous_frame.linesize[0],
++                       line_size))
+                 break;
+ 
+         for (end_line=s->avctx->height; end_line > start_line; end_line--)
+             if (memcmp(p->data[0] + (end_line - 1)*p->linesize[0],
+-                       s->previous_frame.data[0] + (end_line - 1)*p->linesize[0],
+-                       p->linesize[0]))
++                       s->previous_frame.data[0] + (end_line - 1)*s->previous_frame.linesize[0],
++                       line_size))
+                 break;
+     }
+ 
+     bytestream_put_be32(&buf, 0);                         // CHUNK SIZE, patched later
+ 
+     if ((start_line == 0 && end_line == s->avctx->height) || start_line == s->avctx->height)
+         bytestream_put_be16(&buf, 0);                     // header
+     else {
+

debian/patches/increase-VOFW-on-x86.patch

+From: michael <michael@b3059339-0415-0410-9bf9-f77b7e298cf2>
+Date: Tue, 5 May 2009 01:34:16 +0000 (+0000)
+Subject: Change VOFW for x86 to 5120, it allows larger images to be scaled and was
+X-Git-Url: http://git.ffmpeg.org/?p=libswscale;a=commitdiff_plain;h=c9d08ac2f27b3c7a18db9d2ef417673f77be405f
+
+Change VOFW for x86 to 5120, it allows larger images to be scaled and was
+not slower. Other archs are not changed as the larger VOFW was slower on PPC.
+
+
+git-svn-id: file:///var/local/repositories/mplayer/trunk/libswscale@29256 b3059339-0415-0410-9bf9-f77b7e298cf2
+---
+
+diff --git a/swscale_internal.h b/swscale_internal.h
+index 3995887..eb11166 100644
+--- a/libswscale/swscale_internal.h
++++ b/libswscale/swscale_internal.h
+@@ -33,7 +33,12 @@
+ 
+ #define MAX_FILTER_SIZE 256
+ 
+-#define VOFW 2048
++#if ARCH_X86
++#define VOFW 5120
++#else
++#define VOFW 2048 // faster on PPC and not tested on others
++#endif
++
+ #define VOF  (VOFW*2)
+ 
+ #ifdef WORDS_BIGENDIAN

debian/patches/neon/0001-ARM-NEON-optimised-add_pixels_clamped.patch

+From 75eadd829625f3ef75fad613846ff98773e547ca Mon Sep 17 00:00:00 2001
+From: mru <mru@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Sat, 4 Apr 2009 20:18:58 +0000
+Subject: [PATCH 01/27] ARM: NEON optimised add_pixels_clamped
+
+Based on patch by David Conrad.
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@18332 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+ libavcodec/arm/dsputil_neon.c   |    4 +++
+ libavcodec/arm/dsputil_neon_s.S |   45 +++++++++++++++++++++++++++++++++++++++
+ 2 files changed, 49 insertions(+), 0 deletions(-)
+
+diff --git a/libavcodec/arm/dsputil_neon.c b/libavcodec/arm/dsputil_neon.c
+index e18a487..2e56308 100644
+--- a/libavcodec/arm/dsputil_neon.c
++++ b/libavcodec/arm/dsputil_neon.c
+@@ -41,6 +41,8 @@ void ff_put_pixels8_xy2_no_rnd_neon(uint8_t *, const uint8_t *, int, int);
+ 
+ void ff_avg_pixels16_neon(uint8_t *, const uint8_t *, int, int);
+ 
++void ff_add_pixels_clamped_neon(const DCTELEM *, uint8_t *, int);
++
+ void ff_put_h264_qpel16_mc00_neon(uint8_t *, uint8_t *, int);
+ void ff_put_h264_qpel16_mc10_neon(uint8_t *, uint8_t *, int);
+ void ff_put_h264_qpel16_mc20_neon(uint8_t *, uint8_t *, int);
+@@ -176,6 +178,8 @@ void ff_dsputil_init_neon(DSPContext *c, AVCodecContext *avctx)
+ 
+     c->avg_pixels_tab[0][0] = ff_avg_pixels16_neon;
+ 
++    c->add_pixels_clamped = ff_add_pixels_clamped_neon;
++
+     c->put_h264_chroma_pixels_tab[0] = ff_put_h264_chroma_mc8_neon;
+     c->put_h264_chroma_pixels_tab[1] = ff_put_h264_chroma_mc4_neon;
+ 
+diff --git a/libavcodec/arm/dsputil_neon_s.S b/libavcodec/arm/dsputil_neon_s.S
+index 3b39d2e..c305210 100644
+--- a/libavcodec/arm/dsputil_neon_s.S
++++ b/libavcodec/arm/dsputil_neon_s.S
+@@ -273,6 +273,51 @@ function ff_put_h264_qpel8_mc00_neon, export=1
+         pixfunc2 put_ pixels8_y2,   _no_rnd, vhadd.u8
+         pixfunc2 put_ pixels8_xy2,  _no_rnd, vshrn.u16, 1
+ 
++function ff_add_pixels_clamped_neon, export=1
++        mov             r3, r1
++        vld1.64         {d16},   [r1,:64], r2
++        vld1.64         {d0-d1}, [r0,:128]!
++        vaddw.u8        q0, q0, d16
++        vld1.64         {d17},   [r1,:64], r2
++        vld1.64         {d2-d3}, [r0,:128]!
++        vqmovun.s16     d0, q0
++        vld1.64         {d18},   [r1,:64], r2
++        vaddw.u8        q1, q1, d17
++        vld1.64         {d4-d5}, [r0,:128]!
++        vaddw.u8        q2, q2, d18
++        vst1.64         {d0},    [r3,:64], r2
++        vqmovun.s16     d2, q1
++        vld1.64         {d19},   [r1,:64], r2
++        vld1.64         {d6-d7}, [r0,:128]!
++        vaddw.u8        q3, q3, d19
++        vqmovun.s16     d4, q2
++        vst1.64         {d2},    [r3,:64], r2
++        vld1.64         {d16},   [r1,:64], r2
++        vqmovun.s16     d6, q3
++        vld1.64         {d0-d1}, [r0,:128]!
++        vaddw.u8        q0, q0, d16
++        vst1.64         {d4},    [r3,:64], r2
++        vld1.64         {d17},   [r1,:64], r2
++        vld1.64         {d2-d3}, [r0,:128]!
++        vaddw.u8        q1, q1, d17
++        vst1.64         {d6},    [r3,:64], r2
++        vqmovun.s16     d0, q0
++        vld1.64         {d18},   [r1,:64], r2
++        vld1.64         {d4-d5}, [r0,:128]!
++        vaddw.u8        q2, q2, d18
++        vst1.64         {d0},    [r3,:64], r2
++        vqmovun.s16     d2, q1
++        vld1.64         {d19},   [r1,:64], r2
++        vqmovun.s16     d4, q2
++        vld1.64         {d6-d7}, [r0,:128]!
++        vaddw.u8        q3, q3, d19
++        vst1.64         {d2},    [r3,:64], r2
++        vqmovun.s16     d6, q3
++        vst1.64         {d4},    [r3,:64], r2
++        vst1.64         {d6},    [r3,:64], r2
++        bx              lr
++        .endfunc
++
+ function ff_float_to_int16_neon, export=1
+         subs            r2,  r2,  #8
+         vld1.64         {d0-d1},  [r1,:128]!
+-- 
+1.6.3.3
+

debian/patches/neon/0002-ARM-NEON-optimized-put_signed_pixels_clamped.patch

+From 5bf2745fa3bdc996f2201c06eeff1d242d81cc2a Mon Sep 17 00:00:00 2001
+From: conrad <conrad@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Sat, 4 Apr 2009 21:02:48 +0000
+Subject: [PATCH 02/27] ARM: NEON optimized put_signed_pixels_clamped
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@18333 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+ libavcodec/arm/dsputil_neon.c   |    2 ++
+ libavcodec/arm/dsputil_neon_s.S |   37 +++++++++++++++++++++++++++++++++++++
+ 2 files changed, 39 insertions(+), 0 deletions(-)
+
+diff --git a/libavcodec/arm/dsputil_neon.c b/libavcodec/arm/dsputil_neon.c
+index 2e56308..37425a3 100644
+--- a/libavcodec/arm/dsputil_neon.c
++++ b/libavcodec/arm/dsputil_neon.c
+@@ -42,6 +42,7 @@ void ff_put_pixels8_xy2_no_rnd_neon(uint8_t *, const uint8_t *, int, int);
+ void ff_avg_pixels16_neon(uint8_t *, const uint8_t *, int, int);
+ 
+ void ff_add_pixels_clamped_neon(const DCTELEM *, uint8_t *, int);
++void ff_put_signed_pixels_clamped_neon(const DCTELEM *, uint8_t *, int);
+ 
+ void ff_put_h264_qpel16_mc00_neon(uint8_t *, uint8_t *, int);
+ void ff_put_h264_qpel16_mc10_neon(uint8_t *, uint8_t *, int);
+@@ -179,6 +180,7 @@ void ff_dsputil_init_neon(DSPContext *c, AVCodecContext *avctx)
+     c->avg_pixels_tab[0][0] = ff_avg_pixels16_neon;
+ 
+     c->add_pixels_clamped = ff_add_pixels_clamped_neon;
++    c->put_signed_pixels_clamped = ff_put_signed_pixels_clamped_neon;
+ 
+     c->put_h264_chroma_pixels_tab[0] = ff_put_h264_chroma_mc8_neon;
+     c->put_h264_chroma_pixels_tab[1] = ff_put_h264_chroma_mc4_neon;
+diff --git a/libavcodec/arm/dsputil_neon_s.S b/libavcodec/arm/dsputil_neon_s.S
+index c305210..5b95717 100644
+--- a/libavcodec/arm/dsputil_neon_s.S
++++ b/libavcodec/arm/dsputil_neon_s.S
+@@ -273,6 +273,43 @@ function ff_put_h264_qpel8_mc00_neon, export=1
+         pixfunc2 put_ pixels8_y2,   _no_rnd, vhadd.u8
+         pixfunc2 put_ pixels8_xy2,  _no_rnd, vshrn.u16, 1
+ 
++function ff_put_signed_pixels_clamped_neon, export=1
++        vmov.u8         d31, #128
++        vld1.64         {d16-d17}, [r0,:128]!
++        vqmovn.s16      d0, q8
++        vld1.64         {d18-d19}, [r0,:128]!
++        vqmovn.s16      d1, q9
++        vld1.64         {d16-d17}, [r0,:128]!
++        vqmovn.s16      d2, q8
++        vld1.64         {d18-d19}, [r0,:128]!
++        vadd.u8         d0, d0, d31
++        vld1.64         {d20-d21}, [r0,:128]!
++        vadd.u8         d1, d1, d31
++        vld1.64         {d22-d23}, [r0,:128]!
++        vadd.u8         d2, d2, d31
++        vst1.64         {d0},      [r1,:64], r2
++        vqmovn.s16      d3, q9
++        vst1.64         {d1},      [r1,:64], r2
++        vqmovn.s16      d4, q10
++        vst1.64         {d2},      [r1,:64], r2
++        vqmovn.s16      d5, q11
++        vld1.64         {d24-d25}, [r0,:128]!
++        vadd.u8         d3, d3, d31
++        vld1.64         {d26-d27}, [r0,:128]!
++        vadd.u8         d4, d4, d31
++        vadd.u8         d5, d5, d31
++        vst1.64         {d3},      [r1,:64], r2
++        vqmovn.s16      d6, q12
++        vst1.64         {d4},      [r1,:64], r2
++        vqmovn.s16      d7, q13
++        vst1.64         {d5},      [r1,:64], r2
++        vadd.u8         d6, d6, d31
++        vadd.u8         d7, d7, d31
++        vst1.64         {d6},      [r1,:64], r2
++        vst1.64         {d7},      [r1,:64], r2
++        bx              lr
++        .endfunc
++
+ function ff_add_pixels_clamped_neon, export=1
+         mov             r3, r1
+         vld1.64         {d16},   [r1,:64], r2
+-- 
+1.6.3.3
+

debian/patches/neon/0003-Add-guaranteed-alignment-for-loading-dest-pixels-in-.patch

+From ce7bbcc96ff94b2fdfbcff35f517f03512bc147b Mon Sep 17 00:00:00 2001
+From: conrad <conrad@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Thu, 16 Apr 2009 08:39:13 +0000
+Subject: [PATCH 03/27] Add guaranteed alignment for loading dest pixels in avg_pixels16_neon
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@18535 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+ libavcodec/arm/dsputil_neon_s.S |    8 ++++----
+ 1 files changed, 4 insertions(+), 4 deletions(-)
+
+diff --git a/libavcodec/arm/dsputil_neon_s.S b/libavcodec/arm/dsputil_neon_s.S
+index 5b95717..f16293d 100644
+--- a/libavcodec/arm/dsputil_neon_s.S
++++ b/libavcodec/arm/dsputil_neon_s.S
+@@ -38,13 +38,13 @@
+         pld             [r1, r2]
+         pld             [r1, r2, lsl #1]
+ .if \avg
+-        vld1.64         {d16,d17}, [ip], r2
++        vld1.64         {d16,d17}, [ip,:128], r2
+         vrhadd.u8       q0,  q0,  q8
+-        vld1.64         {d18,d19}, [ip], r2
++        vld1.64         {d18,d19}, [ip,:128], r2
+         vrhadd.u8       q1,  q1,  q9
+-        vld1.64         {d20,d21}, [ip], r2
++        vld1.64         {d20,d21}, [ip,:128], r2
+         vrhadd.u8       q2,  q2,  q10
+-        vld1.64         {d22,d23}, [ip], r2
++        vld1.64         {d22,d23}, [ip,:128], r2
+         vrhadd.u8       q3,  q3,  q11
+ .endif
+         subs            r3,  r3,  #4
+-- 
+1.6.3.3
+

debian/patches/neon/0005-ARM-asm-for-AV_RN.patch

+From 8e2336d4b05585ed3a533b388751bff4f3cfef05 Mon Sep 17 00:00:00 2001
+From: mru <mru@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Sat, 18 Apr 2009 00:00:28 +0000
+Subject: [PATCH 05/27] ARM asm for AV_RN*()
+
+ARMv6 and later support unaligned loads and stores for single
+word/halfword but not double/multiple.  GCC is ignorant of this and
+will always use bytewise accesses for unaligned data.  Casting to an
+int32_t pointer is dangerous since a load/store double or multiple
+instruction might be used (this happens with some code in FFmpeg).
+Implementing the AV_[RW]* macros with inline asm using only supported
+instructions gives fast and safe unaligned accesses.  ARM RVCT does
+the right thing with generic code.
+
+This gives an overall speedup of up to 10%.
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@18601 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+ libavutil/arm/intreadwrite.h |   78 ++++++++++++++++++++++++++++++++++++++++++
+ libavutil/intreadwrite.h     |    3 ++
+ 2 files changed, 81 insertions(+), 0 deletions(-)
+ create mode 100644 libavutil/arm/intreadwrite.h
+
+diff --git a/libavutil/arm/intreadwrite.h b/libavutil/arm/intreadwrite.h
+new file mode 100644
+index 0000000..de2e553
+--- /dev/null
++++ b/libavutil/arm/intreadwrite.h
+@@ -0,0 +1,78 @@
++/*
++ * This file is part of FFmpeg.
++ *
++ * FFmpeg is free software; you can redistribute it and/or
++ * modify it under the terms of the GNU Lesser General Public
++ * License as published by the Free Software Foundation; either
++ * version 2.1 of the License, or (at your option) any later version.
++ *
++ * FFmpeg is distributed in the hope that it will be useful,
++ * but WITHOUT ANY WARRANTY; without even the implied warranty of
++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
++ * Lesser General Public License for more details.
++ *
++ * You should have received a copy of the GNU Lesser General Public
++ * License along with FFmpeg; if not, write to the Free Software
++ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
++ */
++
++#ifndef AVUTIL_ARM_INTREADWRITE_H
++#define AVUTIL_ARM_INTREADWRITE_H
++
++#include <stdint.h>
++#include "config.h"
++
++#if HAVE_FAST_UNALIGNED && HAVE_INLINE_ASM
++
++#define AV_RN16 AV_RN16
++static inline uint16_t AV_RN16(const void *p)
++{
++    uint16_t v;
++    __asm__ ("ldrh %0, %1" : "=r"(v) : "m"(*(const uint16_t *)p));
++    return v;
++}
++
++#define AV_WN16 AV_WN16
++static inline void AV_WN16(void *p, uint16_t v)
++{
++    __asm__ ("strh %1, %0" : "=m"(*(uint16_t *)p) : "r"(v));
++}
++
++#define AV_RN32 AV_RN32
++static inline uint32_t AV_RN32(const void *p)
++{
++    uint32_t v;
++    __asm__ ("ldr  %0, %1" : "=r"(v) : "m"(*(const uint32_t *)p));
++    return v;
++}
++
++#define AV_WN32 AV_WN32
++static inline void AV_WN32(void *p, uint32_t v)
++{
++    __asm__ ("str  %1, %0" : "=m"(*(uint32_t *)p) : "r"(v));
++}
++
++#define AV_RN64 AV_RN64
++static inline uint64_t AV_RN64(const void *p)
++{
++    union { uint64_t v; uint32_t hl[2]; } v;
++    __asm__ ("ldr   %0, %2  \n\t"
++             "ldr   %1, %3  \n\t"
++             : "=r"(v.hl[0]), "=r"(v.hl[1])
++             : "m"(*(const uint32_t*)p), "m"(*((const uint32_t*)p+1)));
++    return v.v;
++}
++
++#define AV_WN64 AV_WN64
++static inline void AV_WN64(void *p, uint64_t v)
++{
++    union { uint64_t v; uint32_t hl[2]; } vv = { v };
++    __asm__ ("str  %2, %0  \n\t"
++             "str  %3, %1  \n\t"
++             : "=m"(*(uint32_t*)p), "=m"(*((uint32_t*)p+1))
++             : "r"(vv.hl[0]), "r"(vv.hl[1]));
++}
++
++#endif /* HAVE_INLINE_ASM */
++
++#endif /* AVUTIL_ARM_INTREADWRITE_H */
+diff --git a/libavutil/intreadwrite.h b/libavutil/intreadwrite.h
+index b1c5c2a..42fb890 100644
+--- a/libavutil/intreadwrite.h
++++ b/libavutil/intreadwrite.h
+@@ -29,6 +29,9 @@
+  * defined, even if these are implemented as inline functions.
+  */
+ 
++#if   ARCH_ARM
++#   include "arm/intreadwrite.h"
++#endif
+ 
+ /*
+  * Define AV_[RW]N helper macros to simplify definitions not provided
+-- 
+1.6.3.3
+

debian/patches/neon/0006-ARM-NEON-put_pixels_clamped.patch

+From 9744978dd7b9ca9dda0846fa82faf6fa264a52a7 Mon Sep 17 00:00:00 2001
+From: conrad <conrad@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Wed, 29 Apr 2009 11:31:43 +0000
+Subject: [PATCH 06/27] ARM: NEON put_pixels_clamped
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@18712 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+ libavcodec/arm/dsputil_neon.c   |    2 ++
+ libavcodec/arm/dsputil_neon_s.S |   24 ++++++++++++++++++++++++
+ 2 files changed, 26 insertions(+), 0 deletions(-)
+
+diff --git a/libavcodec/arm/dsputil_neon.c b/libavcodec/arm/dsputil_neon.c
+index 37425a3..9b95130 100644
+--- a/libavcodec/arm/dsputil_neon.c
++++ b/libavcodec/arm/dsputil_neon.c
+@@ -42,6 +42,7 @@ void ff_put_pixels8_xy2_no_rnd_neon(uint8_t *, const uint8_t *, int, int);
+ void ff_avg_pixels16_neon(uint8_t *, const uint8_t *, int, int);
+ 
+ void ff_add_pixels_clamped_neon(const DCTELEM *, uint8_t *, int);
++void ff_put_pixels_clamped_neon(const DCTELEM *, uint8_t *, int);
+ void ff_put_signed_pixels_clamped_neon(const DCTELEM *, uint8_t *, int);
+ 
+ void ff_put_h264_qpel16_mc00_neon(uint8_t *, uint8_t *, int);
+@@ -180,6 +181,7 @@ void ff_dsputil_init_neon(DSPContext *c, AVCodecContext *avctx)
+     c->avg_pixels_tab[0][0] = ff_avg_pixels16_neon;
+ 
+     c->add_pixels_clamped = ff_add_pixels_clamped_neon;
++    c->put_pixels_clamped = ff_put_pixels_clamped_neon;
+     c->put_signed_pixels_clamped = ff_put_signed_pixels_clamped_neon;
+ 
+     c->put_h264_chroma_pixels_tab[0] = ff_put_h264_chroma_mc8_neon;
+diff --git a/libavcodec/arm/dsputil_neon_s.S b/libavcodec/arm/dsputil_neon_s.S
+index f16293d..a55e05f 100644
+--- a/libavcodec/arm/dsputil_neon_s.S
++++ b/libavcodec/arm/dsputil_neon_s.S
+@@ -273,6 +273,30 @@ function ff_put_h264_qpel8_mc00_neon, export=1
+         pixfunc2 put_ pixels8_y2,   _no_rnd, vhadd.u8
+         pixfunc2 put_ pixels8_xy2,  _no_rnd, vshrn.u16, 1
+ 
++function ff_put_pixels_clamped_neon, export=1
++        vld1.64         {d16-d19}, [r0,:128]!
++        vqmovun.s16     d0, q8
++        vld1.64         {d20-d23}, [r0,:128]!
++        vqmovun.s16     d1, q9
++        vld1.64         {d24-d27}, [r0,:128]!
++        vqmovun.s16     d2, q10
++        vld1.64         {d28-d31}, [r0,:128]!
++        vqmovun.s16     d3, q11
++        vst1.64         {d0},      [r1,:64], r2
++        vqmovun.s16     d4, q12
++        vst1.64         {d1},      [r1,:64], r2
++        vqmovun.s16     d5, q13
++        vst1.64         {d2},      [r1,:64], r2
++        vqmovun.s16     d6, q14
++        vst1.64         {d3},      [r1,:64], r2
++        vqmovun.s16     d7, q15
++        vst1.64         {d4},      [r1,:64], r2
++        vst1.64         {d5},      [r1,:64], r2
++        vst1.64         {d6},      [r1,:64], r2
++        vst1.64         {d7},      [r1,:64], r2
++        bx              lr
++        .endfunc
++
+ function ff_put_signed_pixels_clamped_neon, export=1
+         vmov.u8         d31, #128
+         vld1.64         {d16-d17}, [r0,:128]!
+-- 
+1.6.3.3
+

debian/patches/neon/0007-ARM-Use-fewer-register-in-NEON-put_pixels-_y2-and-_x.patch

+From dcd9d80a563a1877e1e5f9182d1bf0ac08ccc7d4 Mon Sep 17 00:00:00 2001
+From: conrad <conrad@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Wed, 29 Apr 2009 11:38:09 +0000
+Subject: [PATCH 07/27] ARM: Use fewer register in NEON put_pixels _y2 and _xy2
+ Approved by Mans on IRC
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@18713 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+ libavcodec/arm/dsputil_neon_s.S |   64 ++++++++++++++++-----------------------
+ 1 files changed, 26 insertions(+), 38 deletions(-)
+
+diff --git a/libavcodec/arm/dsputil_neon_s.S b/libavcodec/arm/dsputil_neon_s.S
+index a55e05f..303b11c 100644
+--- a/libavcodec/arm/dsputil_neon_s.S
++++ b/libavcodec/arm/dsputil_neon_s.S
+@@ -73,35 +73,29 @@
+         .endm
+ 
+         .macro pixels16_y2 vhadd=vrhadd.u8
+-        push            {lr}
+-        add             ip,  r1,  r2
+-        lsl             lr,  r2,  #1
+-        vld1.64         {d0, d1},  [r1], lr
+-        vld1.64         {d2, d3},  [ip], lr
++        vld1.64         {d0, d1},  [r1], r2
++        vld1.64         {d2, d3},  [r1], r2
+ 1:      subs            r3,  r3,  #2
+         \vhadd          q2,  q0,  q1
+-        vld1.64         {d0, d1},  [r1],      lr
++        vld1.64         {d0, d1},  [r1], r2
+         \vhadd          q3,  q0,  q1
+-        vld1.64         {d2, d3},  [ip],      lr
++        vld1.64         {d2, d3},  [r1], r2
+         pld             [r1]
+-        pld             [ip]
++        pld             [r1, r2]
+         vst1.64         {d4, d5},  [r0,:128], r2
+         vst1.64         {d6, d7},  [r0,:128], r2
+         bne             1b
+-        pop             {pc}
++        bx              lr
+         .endm
+ 
+         .macro pixels16_xy2 vshrn=vrshrn.u16 no_rnd=0
+-        push            {lr}
+-        lsl             lr,  r2,  #1
+-        add             ip,  r1,  r2
+-        vld1.64         {d0-d2},   [r1], lr
+-        vld1.64         {d4-d6},   [ip], lr
++        vld1.64         {d0-d2},   [r1], r2
++        vld1.64         {d4-d6},   [r1], r2
+ .if \no_rnd
+         vmov.i16        q13, #1
+ .endif
+         pld             [r1]
+-        pld             [ip]
++        pld             [r1, r2]
+         vext.8          q1,  q0,  q1,  #1
+         vext.8          q3,  q2,  q3,  #1
+         vaddl.u8        q8,  d0,  d2
+@@ -109,7 +103,7 @@
+         vaddl.u8        q9,  d4,  d6
+         vaddl.u8        q11, d5,  d7
+ 1:      subs            r3,  r3,  #2
+-        vld1.64         {d0-d2},   [r1], lr
++        vld1.64         {d0-d2},   [r1], r2
+         vadd.u16        q12, q8,  q9
+         pld             [r1]
+ .if \no_rnd
+@@ -123,11 +117,11 @@
+ .endif
+         \vshrn          d29, q1,  #2
+         vaddl.u8        q8,  d0,  d30
+-        vld1.64         {d2-d4},   [ip], lr
++        vld1.64         {d2-d4},   [r1], r2
+         vaddl.u8        q10, d1,  d31
+         vst1.64         {d28,d29}, [r0,:128], r2
+         vadd.u16        q12, q8,  q9
+-        pld             [ip]
++        pld             [r1, r2]
+ .if \no_rnd
+         vadd.u16        q12, q12, q13
+ .endif
+@@ -142,7 +136,7 @@
+         vaddl.u8        q11, d3,  d5
+         vst1.64         {d30,d31}, [r0,:128], r2
+         bgt             1b
+-        pop             {pc}
++        bx              lr
+         .endm
+ 
+         .macro pixels8
+@@ -180,41 +174,35 @@
+         .endm
+ 
+         .macro pixels8_y2 vhadd=vrhadd.u8
+-        push            {lr}
+-        add             ip,  r1,  r2
+-        lsl             lr,  r2,  #1
+-        vld1.64         {d0},      [r1], lr
+-        vld1.64         {d1},      [ip], lr
++        vld1.64         {d0},      [r1], r2
++        vld1.64         {d1},      [r1], r2
+ 1:      subs            r3,  r3,  #2
+         \vhadd          d4,  d0,  d1
+-        vld1.64         {d0},      [r1],     lr
++        vld1.64         {d0},      [r1], r2
+         \vhadd          d5,  d0,  d1
+-        vld1.64         {d1},      [ip],     lr
++        vld1.64         {d1},      [r1], r2
+         pld             [r1]
+-        pld             [ip]
++        pld             [r1, r2]
+         vst1.64         {d4},      [r0,:64], r2
+         vst1.64         {d5},      [r0,:64], r2
+         bne             1b
+-        pop             {pc}
++        bx              lr
+         .endm
+ 
+         .macro pixels8_xy2 vshrn=vrshrn.u16 no_rnd=0
+-        push            {lr}
+-        lsl             lr,  r2,  #1
+-        add             ip,  r1,  r2
+-        vld1.64         {d0, d1},  [r1], lr
+-        vld1.64         {d2, d3},  [ip], lr
++        vld1.64         {d0, d1},  [r1], r2
++        vld1.64         {d2, d3},  [r1], r2
+ .if \no_rnd
+         vmov.i16        q11, #1
+ .endif
+         pld             [r1]
+-        pld             [ip]
++        pld             [r1, r2]
+         vext.8          d4,  d0,  d1,  #1
+         vext.8          d6,  d2,  d3,  #1
+         vaddl.u8        q8,  d0,  d4
+         vaddl.u8        q9,  d2,  d6
+ 1:      subs            r3,  r3,  #2
+-        vld1.64         {d0, d1},  [r1], lr
++        vld1.64         {d0, d1},  [r1], r2
+         pld             [r1]
+         vadd.u16        q10, q8,  q9
+         vext.8          d4,  d0,  d1,  #1
+@@ -223,9 +211,9 @@
+ .endif
+         vaddl.u8        q8,  d0,  d4
+         \vshrn          d5,  q10, #2
+-        vld1.64         {d2, d3},  [ip], lr
++        vld1.64         {d2, d3},  [r1], r2
+         vadd.u16        q10, q8,  q9
+-        pld             [ip]
++        pld             [r1, r2]
+ .if \no_rnd
+         vadd.u16        q10, q10, q11
+ .endif
+@@ -235,7 +223,7 @@
+         vaddl.u8        q9,  d2,  d6
+         vst1.64         {d7},      [r0,:64], r2
+         bgt             1b
+-        pop             {pc}
++        bx              lr
+         .endm
+ 
+         .macro pixfunc pfx name suf rnd_op args:vararg
+-- 
+1.6.3.3
+

debian/patches/neon/0008-ARM-NEON-VP3-Loop-Filter.patch

+From 91dafadca2744957e6cc3ae13688c9033ec3c330 Mon Sep 17 00:00:00 2001
+From: conrad <conrad@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Sat, 23 May 2009 18:36:20 +0000
+Subject: [PATCH 08/27] ARM: NEON VP3 Loop Filter
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@18916 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+ libavcodec/Makefile           |    4 ++++
+ libavcodec/arm/dsputil_neon.c |    8 ++++++++
+ 2 files changed, 12 insertions(+), 0 deletions(-)
+
+diff --git a/libavcodec/Makefile b/libavcodec/Makefile
+index cf742d1..a213309 100644
+--- a/libavcodec/Makefile
++++ b/libavcodec/Makefile
+@@ -479,11 +479,15 @@ OBJS-$(HAVE_ARMVFP)                    += arm/dsputil_vfp.o             \
+ OBJS-$(HAVE_IWMMXT)                    += arm/dsputil_iwmmxt.o          \
+                                           arm/mpegvideo_iwmmxt.o        \
+ 
++NEON-OBJS-$(CONFIG_THEORA_DECODER)     += arm/vp3dsp_neon.o
++NEON-OBJS-$(CONFIG_VP3_DECODER)        += arm/vp3dsp_neon.o
++
+ OBJS-$(HAVE_NEON)                      += arm/dsputil_neon.o            \
+                                           arm/dsputil_neon_s.o          \
+                                           arm/h264dsp_neon.o            \
+                                           arm/h264idct_neon.o           \
+                                           arm/simple_idct_neon.o        \
++					  $(NEON-OBJS-yes)
+ 
+ OBJS-$(ARCH_BFIN)                      += bfin/dsputil_bfin.o           \
+                                           bfin/fdct_bfin.o              \
+diff --git a/libavcodec/arm/dsputil_neon.c b/libavcodec/arm/dsputil_neon.c
+index 9b95130..20425c1 100644
+--- a/libavcodec/arm/dsputil_neon.c
++++ b/libavcodec/arm/dsputil_neon.c
+@@ -150,6 +150,9 @@ void ff_h264_idct_add8_neon(uint8_t **dest, const int *block_offset,
+                             DCTELEM *block, int stride,
+                             const uint8_t nnzc[6*8]);
+ 
++void ff_vp3_v_loop_filter_neon(uint8_t *, int, int *);
++void ff_vp3_h_loop_filter_neon(uint8_t *, int, int *);
++
+ void ff_vector_fmul_neon(float *dst, const float *src, int len);
+ void ff_vector_fmul_window_neon(float *dst, const float *src0,
+                                 const float *src1, const float *win,
+@@ -255,6 +258,11 @@ void ff_dsputil_init_neon(DSPContext *c, AVCodecContext *avctx)
+     c->h264_idct_add16intra = ff_h264_idct_add16intra_neon;
+     c->h264_idct_add8       = ff_h264_idct_add8_neon;
+ 
++    if (CONFIG_VP3_DECODER || CONFIG_THEORA_DECODER) {
++        c->vp3_v_loop_filter = ff_vp3_v_loop_filter_neon;
++        c->vp3_h_loop_filter = ff_vp3_h_loop_filter_neon;
++    }
++
+     c->vector_fmul = ff_vector_fmul_neon;
+     c->vector_fmul_window = ff_vector_fmul_window_neon;
+ 
+-- 
+1.6.3.3
+

debian/patches/neon/0009-ARM-actually-add-VP3-loop-filter.patch

+From 2ca1dc1b533f6916ac593d435e1af8a299944d2c Mon Sep 17 00:00:00 2001
+From: conrad <conrad@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Sat, 23 May 2009 18:47:26 +0000
+Subject: [PATCH 09/27] ARM: actually add VP3 loop filter
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@18917 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+ libavcodec/arm/vp3dsp_neon.S |   94 ++++++++++++++++++++++++++++++++++++++++++
+ 1 files changed, 94 insertions(+), 0 deletions(-)
+ create mode 100644 libavcodec/arm/vp3dsp_neon.S
+
+diff --git a/libavcodec/arm/vp3dsp_neon.S b/libavcodec/arm/vp3dsp_neon.S
+new file mode 100644
+index 0000000..cce80dd
+--- /dev/null
++++ b/libavcodec/arm/vp3dsp_neon.S
+@@ -0,0 +1,94 @@
++/*
++ * Copyright (c) 2009 David Conrad
++ *
++ * This file is part of FFmpeg.
++ *
++ * FFmpeg is free software; you can redistribute it and/or
++ * modify it under the terms of the GNU Lesser General Public
++ * License as published by the Free Software Foundation; either
++ * version 2.1 of the License, or (at your option) any later version.
++ *
++ * FFmpeg is distributed in the hope that it will be useful,
++ * but WITHOUT ANY WARRANTY; without even the implied warranty of
++ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
++ * Lesser General Public License for more details.
++ *
++ * You should have received a copy of the GNU Lesser General Public
++ * License along with FFmpeg; if not, write to the Free Software
++ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
++ */
++
++#include "asm.S"
++
++.macro vp3_loop_filter
++    vsubl.u8        q3,  d18, d17
++    vsubl.u8        q2,  d16, d19
++    vadd.i16        q1,  q3,  q3
++    vadd.i16        q2,  q2,  q3
++    vadd.i16        q0,  q1,  q2
++    vrshr.s16       q0,  q0,  #3
++    vmovl.u8        q9,  d18
++    vdup.u16        q15, r2
++
++    vabs.s16        q1,  q0
++    vshr.s16        q0,  q0,  #15
++    vqsub.u16       q2,  q15, q1
++    vqsub.u16       q3,  q2,  q1
++    vsub.i16        q1,  q2,  q3
++    veor            q1,  q1,  q0
++    vsub.i16        q0,  q1,  q0
++
++    vaddw.u8        q2,  q0,  d17
++    vsub.i16        q3,  q9,  q0
++    vqmovun.s16     d0,  q2
++    vqmovun.s16     d1,  q3
++.endm
++
++function ff_vp3_v_loop_filter_neon, export=1
++    sub             ip,  r0,  r1
++    sub             r0,  r0,  r1,  lsl #1
++    vld1.64         {d16}, [r0,:64], r1
++    vld1.64         {d17}, [r0,:64], r1
++    vld1.64         {d18}, [r0,:64], r1
++    vld1.64         {d19}, [r0,:64], r1
++    ldrb            r2,    [r2, #129*4]
++
++    vp3_loop_filter
++
++    vst1.64         {d0},  [ip,:64], r1
++    vst1.64         {d1},  [ip,:64], r1
++    bx              lr
++.endfunc
++
++function ff_vp3_h_loop_filter_neon, export=1
++    sub             ip,  r0,  #1
++    sub             r0,  r0,  #2
++    vld1.32         {d16[]},  [r0], r1
++    vld1.32         {d17[]},  [r0], r1
++    vld1.32         {d18[]},  [r0], r1
++    vld1.32         {d19[]},  [r0], r1
++    vld1.32         {d16[1]}, [r0], r1
++    vld1.32         {d17[1]}, [r0], r1
++    vld1.32         {d18[1]}, [r0], r1
++    vld1.32         {d19[1]}, [r0], r1
++    ldrb            r2,  [r2, #129*4]
++
++    vtrn.8          d16, d17
++    vtrn.8          d18, d19
++    vtrn.16         d16, d18
++    vtrn.16         d17, d19
++
++    vp3_loop_filter
++
++    vtrn.8          d0,  d1
++
++    vst1.16         {d0[0]}, [ip], r1
++    vst1.16         {d1[0]}, [ip], r1
++    vst1.16         {d0[1]}, [ip], r1
++    vst1.16         {d1[1]}, [ip], r1
++    vst1.16         {d0[2]}, [ip], r1
++    vst1.16         {d1[2]}, [ip], r1
++    vst1.16         {d0[3]}, [ip], r1
++    vst1.16         {d1[3]}, [ip], r1
++    bx              lr
++.endfunc
+-- 
+1.6.3.3
+

debian/patches/neon/0010-ARM-add-some-PLD-in-NEON-IDCT.patch

+From 8d8bd2247af043b8b61af8bbc4c131524182531a Mon Sep 17 00:00:00 2001
+From: mru <mru@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Thu, 28 May 2009 17:19:28 +0000
+Subject: [PATCH 10/27] ARM: add some PLD in NEON IDCT
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@18972 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+ libavcodec/arm/simple_idct_neon.S |   17 +++++++++++++++--
+ 1 files changed, 15 insertions(+), 2 deletions(-)
+
+diff --git a/libavcodec/arm/simple_idct_neon.S b/libavcodec/arm/simple_idct_neon.S
+index e7099a2..0882481 100644
+--- a/libavcodec/arm/simple_idct_neon.S
++++ b/libavcodec/arm/simple_idct_neon.S
+@@ -68,6 +68,19 @@
+         .text
+         .align 6
+ 
++function idct_row4_pld_neon
++        pld             [r0]
++        add             r3,  r0,  r1,  lsl #2
++        pld             [r0, r1]
++        pld             [r0, r1, lsl #1]
++        pld             [r3, -r1]
++        pld             [r3]
++        pld             [r3, r1]
++        add             r3,  r3,  r1,  lsl #1
++        pld             [r3]
++        pld             [r3, r1]
++        .endfunc
++
+ function idct_row4_neon
+         vmov.i32        q15, #(1<<(ROW_SHIFT-1))
+         vld1.64         {d2-d5},  [r2,:128]!
+@@ -252,7 +265,7 @@ idct_coeff_neon:
+ function ff_simple_idct_put_neon, export=1
+         idct_start      r2
+ 
+-        bl              idct_row4_neon
++        bl              idct_row4_pld_neon
+         bl              idct_row4_neon
+         add             r2,  r2,  #-128
+         bl              idct_col4_neon
+@@ -307,7 +320,7 @@ function idct_col4_add8_neon
+ function ff_simple_idct_add_neon, export=1
+         idct_start      r2
+ 
+-        bl              idct_row4_neon
++        bl              idct_row4_pld_neon
+         bl              idct_row4_neon
+         add             r2,  r2,  #-128
+         bl              idct_col4_neon
+-- 
+1.6.3.3
+

debian/patches/neon/0011-ARM-slightly-faster-NEON-H264-horizontal-loop-filter.patch

+From 56f78a98e66f1bd2bc29b00a048421e2f1760785 Mon Sep 17 00:00:00 2001
+From: mru <mru@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Wed, 17 Jun 2009 22:33:04 +0000
+Subject: [PATCH 11/27] ARM: slightly faster NEON H264 horizontal loop filter
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@19216 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+ libavcodec/arm/h264dsp_neon.S |   49 +++++++++++++++++++++--------------------
+ 1 files changed, 25 insertions(+), 24 deletions(-)
+
+diff --git a/libavcodec/arm/h264dsp_neon.S b/libavcodec/arm/h264dsp_neon.S
+index 44a1373..03e21f1 100644
+--- a/libavcodec/arm/h264dsp_neon.S
++++ b/libavcodec/arm/h264dsp_neon.S
+@@ -37,6 +37,13 @@
+         vtrn.8          \r6, \r7
+         .endm
+ 
++        .macro transpose_4x4 r0 r1 r2 r3
++        vtrn.16         \r0, \r2
++        vtrn.16         \r1, \r3
++        vtrn.8          \r0, \r1
++        vtrn.8          \r2, \r3
++        .endm
++
+         .macro swap4 r0 r1 r2 r3 r4 r5 r6 r7
+         vswp            \r0, \r4
+         vswp            \r1, \r5
+@@ -469,35 +476,29 @@ function ff_h264_h_loop_filter_luma_neon, export=1
+         transpose_8x8   q3, q10, q9, q8, q0, q1, q2, q13
+ 
+         align_push_regs
+-        sub             sp,  sp,  #16
+-        vst1.64         {d4, d5},  [sp,:128]
+-        sub             sp,  sp,  #16
+-        vst1.64         {d20,d21}, [sp,:128]
+ 
+         h264_loop_filter_luma
+ 
+-        vld1.64         {d20,d21}, [sp,:128]!
+-        vld1.64         {d4, d5},  [sp,:128]!
+-
+-        transpose_8x8   q3, q10, q4, q8, q0, q5, q2, q13
++        transpose_4x4   q4, q8, q0, q5
+ 
+         sub             r0,  r0,  r1, lsl #4
+-        vst1.64         {d6},  [r0], r1
+-        vst1.64         {d20}, [r0], r1
+-        vst1.64         {d8},  [r0], r1
+-        vst1.64         {d16}, [r0], r1
+-        vst1.64         {d0},  [r0], r1
+-        vst1.64         {d10}, [r0], r1
+-        vst1.64         {d4},  [r0], r1
+-        vst1.64         {d26}, [r0], r1
+-        vst1.64         {d7},  [r0], r1
+-        vst1.64         {d21}, [r0], r1
+-        vst1.64         {d9},  [r0], r1
+-        vst1.64         {d17}, [r0], r1
+-        vst1.64         {d1},  [r0], r1
+-        vst1.64         {d11}, [r0], r1
+-        vst1.64         {d5},  [r0], r1
+-        vst1.64         {d27}, [r0], r1
++        add             r0,  r0,  #2
++        vst1.32         {d8[0]},  [r0], r1
++        vst1.32         {d16[0]}, [r0], r1
++        vst1.32         {d0[0]},  [r0], r1
++        vst1.32         {d10[0]}, [r0], r1
++        vst1.32         {d8[1]},  [r0], r1
++        vst1.32         {d16[1]}, [r0], r1
++        vst1.32         {d0[1]},  [r0], r1
++        vst1.32         {d10[1]}, [r0], r1
++        vst1.32         {d9[0]},  [r0], r1
++        vst1.32         {d17[0]}, [r0], r1
++        vst1.32         {d1[0]},  [r0], r1
++        vst1.32         {d11[0]}, [r0], r1
++        vst1.32         {d9[1]},  [r0], r1
++        vst1.32         {d17[1]}, [r0], r1
++        vst1.32         {d1[1]},  [r0], r1
++        vst1.32         {d11[1]}, [r0], r1
+ 
+         align_pop_regs
+         bx              lr
+-- 
+1.6.3.3
+

debian/patches/neon/0012-ARM-enable-fast_unaligned-when-cpu-armv-67-is-specif.patch

+From 8f934d1aa05a5db078dc732b37f2d46faa92e866 Mon Sep 17 00:00:00 2001
+From: mru <mru@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Tue, 30 Jun 2009 12:46:09 +0000
+Subject: [PATCH 12/27] ARM: enable fast_unaligned when --cpu=armv[67] is specified
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@19308 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+ configure |    4 ++++
+ 1 files changed, 4 insertions(+), 0 deletions(-)
+
+diff --git a/configure b/configure
+index fd11501..7459215 100755
+--- a/configure
++++ b/configure
+@@ -1792,6 +1792,10 @@ if test $cpu != "generic"; then
+             add_cflags -mcpu=$cpu
+             enable fast_unaligned
+         ;;
++        armv[67]*)
++            add_cflags -march=$cpu
++            enable fast_unaligned
++        ;;
+         armv*)
+             add_cflags -march=$cpu
+         ;;
+-- 
+1.6.3.3
+

debian/patches/neon/0014-ARM-NEON-optimised-vorbis_inverse_coupling.patch

+From d32e115cb8c43fe4531567b1c668dba6dc76274d Mon Sep 17 00:00:00 2001
+From: mru <mru@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Fri, 14 Aug 2009 01:02:06 +0000
+Subject: [PATCH 14/27] ARM: NEON optimised vorbis_inverse_coupling
+
+12% faster Vorbis decoding on Cortex-A8.
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@19637 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+ libavcodec/arm/dsputil_neon.c   |    5 +++
+ libavcodec/arm/dsputil_neon_s.S |   64 +++++++++++++++++++++++++++++++++++++++
+ 2 files changed, 69 insertions(+), 0 deletions(-)
+
+diff --git a/libavcodec/arm/dsputil_neon.c b/libavcodec/arm/dsputil_neon.c
+index 20425c1..eb9aba1 100644
+--- a/libavcodec/arm/dsputil_neon.c
++++ b/libavcodec/arm/dsputil_neon.c
+@@ -161,6 +161,8 @@ void ff_vector_fmul_window_neon(float *dst, const float *src0,
+ void ff_float_to_int16_neon(int16_t *, const float *, long);
+ void ff_float_to_int16_interleave_neon(int16_t *, const float **, long, int);
+ 
++void ff_vorbis_inverse_coupling_neon(float *mag, float *ang, int blocksize);
++
+ void ff_dsputil_init_neon(DSPContext *c, AVCodecContext *avctx)
+ {
+     c->put_pixels_tab[0][0] = ff_put_pixels16_neon;
+@@ -270,4 +272,7 @@ void ff_dsputil_init_neon(DSPContext *c, AVCodecContext *avctx)
+         c->float_to_int16 = ff_float_to_int16_neon;
+         c->float_to_int16_interleave = ff_float_to_int16_interleave_neon;
+     }
++
++    if (CONFIG_VORBIS_DECODER)
++        c->vorbis_inverse_coupling = ff_vorbis_inverse_coupling_neon;
+ }
+diff --git a/libavcodec/arm/dsputil_neon_s.S b/libavcodec/arm/dsputil_neon_s.S
+index 303b11c..2bc07fa 100644
+--- a/libavcodec/arm/dsputil_neon_s.S
++++ b/libavcodec/arm/dsputil_neon_s.S
+@@ -19,6 +19,7 @@
+  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+  */
+ 
++#include "config.h"
+ #include "asm.S"
+ 
+         preserve8
+@@ -793,3 +794,66 @@ function ff_vector_fmul_window_neon, export=1
+         vst1.64         {d22,d23},[ip,:128], r5
+         pop             {r4,r5,pc}
+         .endfunc
++
++#if CONFIG_VORBIS_DECODER
++function ff_vorbis_inverse_coupling_neon, export=1
++        vmov.i32        q10, #1<<31
++        subs            r2,  r2,  #4
++        tst             r2,  #4
++        mov             r3,  r0
++        mov             r12, r1
++        beq             3f
++
++        vld1.32         {d24-d25},[r1,:128]!
++        vld1.32         {d22-d23},[r0,:128]!
++        vcle.s32        q8,  q12, #0
++        vand            q9,  q11, q10
++        veor            q12, q12, q9
++        vand            q2,  q12, q8
++        vbic            q3,  q12, q8
++        vadd.f32        q12, q11, q2
++        vsub.f32        q11, q11, q3
++1:      vld1.32         {d2-d3},  [r1,:128]!
++        vld1.32         {d0-d1},  [r0,:128]!
++        vcle.s32        q8,  q1,  #0
++        vand            q9,  q0,  q10
++        veor            q1,  q1,  q9
++        vst1.32         {d24-d25},[r3, :128]!
++        vst1.32         {d22-d23},[r12,:128]!
++        vand            q2,  q1,  q8
++        vbic            q3,  q1,  q8
++        vadd.f32        q1,  q0,  q2
++        vsub.f32        q0,  q0,  q3
++        subs            r2,  r2,  #8
++        ble             2f
++        vld1.32         {d24-d25},[r1,:128]!
++        vld1.32         {d22-d23},[r0,:128]!
++        vcle.s32        q8,  q12, #0
++        vand            q9,  q11, q10
++        veor            q12, q12, q9
++        vst1.32         {d2-d3},  [r3, :128]!
++        vst1.32         {d0-d1},  [r12,:128]!
++        vand            q2,  q12, q8
++        vbic            q3,  q12, q8
++        vadd.f32        q12, q11, q2
++        vsub.f32        q11, q11, q3
++        b               1b
++
++2:      vst1.32         {d2-d3},  [r3, :128]!
++        vst1.32         {d0-d1},  [r12,:128]!
++        bxlt            lr
++
++3:      vld1.32         {d2-d3},  [r1,:128]
++        vld1.32         {d0-d1},  [r0,:128]
++        vcle.s32        q8,  q1,  #0
++        vand            q9,  q0,  q10
++        veor            q1,  q1,  q9
++        vand            q2,  q1,  q8
++        vbic            q3,  q1,  q8
++        vadd.f32        q1,  q0,  q2
++        vsub.f32        q0,  q0,  q3
++        vst1.32         {d2-d3},  [r0,:128]!
++        vst1.32         {d0-d1},  [r1,:128]!
++        bx              lr
++        .endfunc
++#endif
+-- 
+1.6.3.3
+

debian/patches/neon/0015-ARM-handle-VFP-register-arguments-in-ff_vector_fmul_.patch

+From e965241fea3348a8205bd38f91efbfcd13e8cd31 Mon Sep 17 00:00:00 2001
+From: mru <mru@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Mon, 20 Jul 2009 22:30:27 +0000
+Subject: [PATCH 15/27] ARM: handle VFP register arguments in ff_vector_fmul_window_neon()
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@19475 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+ libavcodec/arm/dsputil_neon_s.S |    6 ++++--
+ 1 files changed, 4 insertions(+), 2 deletions(-)
+
+diff --git a/libavcodec/arm/dsputil_neon_s.S b/libavcodec/arm/dsputil_neon_s.S
+index 2bc07fa..71d09c6 100644
+--- a/libavcodec/arm/dsputil_neon_s.S
++++ b/libavcodec/arm/dsputil_neon_s.S
+@@ -746,9 +746,11 @@ function ff_vector_fmul_neon, export=1
+         .endfunc
+ 
+ function ff_vector_fmul_window_neon, export=1
+-        vld1.32         {d16[],d17[]}, [sp,:32]
++VFP     vdup.32         q8,  d0[0]
++NOVFP   vld1.32         {d16[],d17[]}, [sp,:32]
+         push            {r4,r5,lr}
+-        ldr             lr,  [sp, #16]
++VFP     ldr             lr,  [sp, #12]
++NOVFP   ldr             lr,  [sp, #16]
+         sub             r2,  r2,  #8
+         sub             r5,  lr,  #2
+         add             r2,  r2,  r5, lsl #2
+-- 
+1.6.3.3
+

debian/patches/neon/0017-ARM-faster-NEON-IMDCT.patch

+From d49bcbe5ba20f6130b2dcad078b0ee27b2660a8f Mon Sep 17 00:00:00 2001
+From: mru <mru@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Fri, 11 Sep 2009 02:01:18 +0000
+Subject: [PATCH 17/27] ARM: faster NEON IMDCT
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@19817 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+ libavcodec/arm/mdct_neon.S |   46 ++++++++++++++++---------------------------
+ 1 files changed, 17 insertions(+), 29 deletions(-)
+
+diff --git a/libavcodec/arm/mdct_neon.S b/libavcodec/arm/mdct_neon.S
+index 6d1dcfd..d84eccd 100644
+--- a/libavcodec/arm/mdct_neon.S
++++ b/libavcodec/arm/mdct_neon.S
+@@ -38,30 +38,28 @@ function ff_imdct_half_neon, export=1
+         mov             r12,  #-16
+         sub             r7,  r7,  #16
+ 
+-        vld1.32         {d16-d17},[r7,:128],r12 @ d16=x,n1 d17=x,n0
+-        vld1.32         {d0-d1},  [r2,:128]!    @ d0 =m0,x d1 =m1,x
++        vld2.32         {d16-d17},[r7,:128],r12 @ d16=x,n1 d17=x,n0
++        vld2.32         {d0-d1},  [r2,:128]!    @ d0 =m0,x d1 =m1,x
++        vrev64.32       d17, d17
+         vld1.32         {d2},     [r4,:64]!     @ d2=c0,c1
++        vmul.f32        d6,  d17, d2
+         vld1.32         {d3},     [r5,:64]!     @ d3=s0,s1
+-        vuzp.32         d17, d16
+-        vuzp.32         d0,  d1
+-        vmul.f32        d6,  d16, d2
+         vmul.f32        d7,  d0,  d2
+ 1:
+         subs            lr,  lr,  #2
+         ldr             r6,  [r3], #4
+         vmul.f32        d4,  d0,  d3
+-        vmul.f32        d5,  d16, d3
++        vmul.f32        d5,  d17, d3
+         vsub.f32        d4,  d6,  d4
+         vadd.f32        d5,  d5,  d7
+         uxtah           r8,  r1,  r6,  ror #16
+         uxtah           r6,  r1,  r6
+         beq             1f
+-        vld1.32         {d16-d17},[r7,:128],r12
+-        vld1.32         {d0-d1},  [r2,:128]!
+-        vuzp.32         d17, d16
++        vld2.32         {d16-d17},[r7,:128],r12
++        vld2.32         {d0-d1},  [r2,:128]!
++        vrev64.32       d17, d17
+         vld1.32         {d2},     [r4,:64]!
+-        vuzp.32         d0,  d1
+-        vmul.f32        d6,  d16, d2
++        vmul.f32        d6,  d17, d2
+         vld1.32         {d3},     [r5,:64]!
+         vmul.f32        d7,  d0,  d2
+         vst2.32         {d4[0],d5[0]}, [r6,:64]
+@@ -95,11 +93,9 @@ function ff_imdct_half_neon, export=1
+         mov             r8,  r6
+         mov             r0,  r3
+ 
+-        vld1.32         {d0-d1},  [r3,:128], r7 @ d0 =i1,r1 d1 =i0,r0
+-        vld1.32         {d20-d21},[r6,:128]!    @ d20=i2,r2 d21=i3,r3
++        vld2.32         {d0-d1},  [r3,:128], r7 @ d0 =i1,r1 d1 =i0,r0
++        vld2.32         {d20-d21},[r6,:128]!    @ d20=i2,r2 d21=i3,r3
+         vld1.32         {d18},    [r2,:64], r12 @ d18=s1,s0
+-        vuzp.32         d20, d21
+-        vuzp.32         d0,  d1
+ 1:
+         subs            lr,  lr,  #2
+         vmul.f32        d7,  d0,  d18
+@@ -118,25 +114,17 @@ function ff_imdct_half_neon, export=1
+         vsub.f32        d4,  d4,  d24
+         vsub.f32        d5,  d5,  d25
+         beq             1f
+-        vld1.32         {d0-d1},  [r3,:128], r7
+-        vld1.32         {d20-d21},[r6,:128]!
++        vld2.32         {d0-d1},  [r3,:128], r7
++        vld2.32         {d20-d21},[r6,:128]!
+         vld1.32         {d18},    [r2,:64], r12
+-        vuzp.32         d20, d21
+-        vuzp.32         d0,  d1
+         vrev64.32       q3,  q3
+-        vtrn.32         d4,  d6
+-        vtrn.32         d5,  d7
+-        vswp            d5,  d6
+-        vst1.32         {d4-d5},  [r0,:128], r7
+-        vst1.32         {d6-d7},  [r8,:128]!
++        vst2.32         {d4,d6},  [r0,:128], r7
++        vst2.32         {d5,d7},  [r8,:128]!
+         b               1b
+ 1:
+         vrev64.32       q3,  q3
+-        vtrn.32         d4,  d6
+-        vtrn.32         d5,  d7
+-        vswp            d5,  d6
+-        vst1.32         {d4-d5},  [r0,:128]
+-        vst1.32         {d6-d7},  [r8,:128]
++        vst2.32         {d4,d6},  [r0,:128]
++        vst2.32         {d5,d7},  [r8,:128]
+ 
+         pop             {r4-r8,pc}
+ .endfunc
+-- 
+1.6.3.3
+

debian/patches/neon/0018-Prepare-for-optimised-forward-MDCT-implementations.patch

+From 696f65e4e5d92a8777d922269558cc9f03ca1145 Mon Sep 17 00:00:00 2001
+From: mru <mru@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Fri, 11 Sep 2009 02:24:19 +0000
+Subject: [PATCH 18/27] Prepare for optimised forward MDCT implementations
+
+This adds a function pointer for forward MDCT to FFTContext and
+initialises it with the existing C function.  ff_calc_mdct() is
+changed to an inline function calling the selected version as
+done for other fft/mdct functions.
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@19818 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+ libavcodec/dsputil.h |    9 ++++++++-
+ libavcodec/fft.c     |    1 +
+ libavcodec/mdct.c    |    2 +-
+ 3 files changed, 10 insertions(+), 2 deletions(-)
+
+diff --git a/libavcodec/dsputil.h b/libavcodec/dsputil.h
+index c4abd14..e299ade 100644
+--- a/libavcodec/dsputil.h
++++ b/libavcodec/dsputil.h
+@@ -680,6 +680,7 @@ typedef struct FFTContext {
+     void (*fft_calc)(struct FFTContext *s, FFTComplex *z);
+     void (*imdct_calc)(struct MDCTContext *s, FFTSample *output, const FFTSample *input);
+     void (*imdct_half)(struct MDCTContext *s, FFTSample *output, const FFTSample *input);
++    void (*mdct_calc)(struct MDCTContext *s, FFTSample *output, const FFTSample *input);
+ } FFTContext;
+ 
+ extern FFTSample* ff_cos_tabs[13];
+@@ -737,6 +738,12 @@ static inline void ff_imdct_half(MDCTContext *s, FFTSample *output, const FFTSam
+     s->fft.imdct_half(s, output, input);
+ }
+ 
++static inline void ff_mdct_calc(MDCTContext *s, FFTSample *output,
++                                const FFTSample *input)
++{
++    s->fft.mdct_calc(s, output, input);
++}
++
+ /**
+  * Generate a Kaiser-Bessel Derived Window.
+  * @param   window  pointer to half window
+@@ -762,6 +769,7 @@ extern float *ff_sine_windows[6];
+ int ff_mdct_init(MDCTContext *s, int nbits, int inverse);
+ void ff_imdct_calc_c(MDCTContext *s, FFTSample *output, const FFTSample *input);
+ void ff_imdct_half_c(MDCTContext *s, FFTSample *output, const FFTSample *input);
++void ff_mdct_calc_c(MDCTContext *s, FFTSample *output, const FFTSample *input);
+ void ff_imdct_calc_3dn(MDCTContext *s, FFTSample *output, const FFTSample *input);
+ void ff_imdct_half_3dn(MDCTContext *s, FFTSample *output, const FFTSample *input);
+ void ff_imdct_calc_3dn2(MDCTContext *s, FFTSample *output, const FFTSample *input);
+@@ -770,7 +778,6 @@ void ff_imdct_calc_sse(MDCTContext *s, FFTSample *output, const FFTSample *input
+ void ff_imdct_half_sse(MDCTContext *s, FFTSample *output, const FFTSample *input);
+ void ff_imdct_calc_neon(MDCTContext *s, FFTSample *output, const FFTSample *input);
+ void ff_imdct_half_neon(MDCTContext *s, FFTSample *output, const FFTSample *input);
+-void ff_mdct_calc(MDCTContext *s, FFTSample *out, const FFTSample *input);
+ void ff_mdct_end(MDCTContext *s);
+ 
+ /* Real Discrete Fourier Transform */
+diff --git a/libavcodec/fft.c b/libavcodec/fft.c
+index b05ddda..655acd3 100644
+--- a/libavcodec/fft.c
++++ b/libavcodec/fft.c
+@@ -86,6 +86,7 @@ av_cold int ff_fft_init(FFTContext *s, int nbits, int inverse)
+     s->fft_calc    = ff_fft_calc_c;
+     s->imdct_calc  = ff_imdct_calc_c;
+     s->imdct_half  = ff_imdct_half_c;
++    s->mdct_calc   = ff_mdct_calc_c;
+     s->exptab1     = NULL;
+ 
+ #if HAVE_MMX && HAVE_YASM
+diff --git a/libavcodec/mdct.c b/libavcodec/mdct.c
+index cb3388f..b0816b0 100644
+--- a/libavcodec/mdct.c
++++ b/libavcodec/mdct.c
+@@ -179,7 +179,7 @@ void ff_imdct_calc_c(MDCTContext *s, FFTSample *output, const FFTSample *input)
+  * @param input N samples
+  * @param out N/2 samples
+  */
+-void ff_mdct_calc(MDCTContext *s, FFTSample *out, const FFTSample *input)
++void ff_mdct_calc_c(MDCTContext *s, FFTSample *out, const FFTSample *input)
+ {
+     int i, j, n, n8, n4, n2, n3;
+     FFTSample re, im;
+-- 
+1.6.3.3
+

debian/patches/neon/0020-ARM-remove-unnecessary-.fpu-neon-directives.patch

+From cf79e9ce120293a3d8fd6887298283c27ee866d4 Mon Sep 17 00:00:00 2001
+From: mru <mru@9553f0bf-9b14-0410-a0b8-cfaf0461ba5b>
+Date: Fri, 2 Oct 2009 19:35:12 +0000
+Subject: [PATCH 20/20] ARM: remove unnecessary .fpu neon directives
+
+git-svn-id: file:///var/local/repositories/ffmpeg/trunk@20151 9553f0bf-9b14-0410-a0b8-cfaf0461ba5b
+---
+ libavcodec/arm/dsputil_neon_s.S   |    1 -
+ libavcodec/arm/h264dsp_neon.S     |    2 --
+ libavcodec/arm/h264idct_neon.S    |    2 --
+ libavcodec/arm/mdct_neon.S        |    1 -
+ libavcodec/arm/simple_idct_neon.S |    2 --
+ 5 files changed, 0 insertions(+), 8 deletions(-)
+
+diff --git a/libavcodec/arm/dsputil_neon_s.S b/libavcodec/arm/dsputil_neon_s.S
+index 71d09c6..f039349 100644
+--- a/libavcodec/arm/dsputil_neon_s.S
++++ b/libavcodec/arm/dsputil_neon_s.S
+@@ -23,7 +23,6 @@
+ #include "asm.S"
+ 
+         preserve8
+-        .fpu neon
+         .text
+ 
+         .macro pixels16 avg=0
+diff --git a/libavcodec/arm/h264dsp_neon.S b/libavcodec/arm/h264dsp_neon.S
+index 03e21f1..78f312d 100644
+--- a/libavcodec/arm/h264dsp_neon.S
++++ b/libavcodec/arm/h264dsp_neon.S
+@@ -20,8 +20,6 @@
+ 
+ #include "asm.S"
+ 
+-        .fpu neon
+-
+         .macro transpose_8x8 r0 r1 r2 r3 r4 r5 r6 r7
+         vtrn.32         \r0, \r4
+         vtrn.32         \r1, \r5
+diff --git a/libavcodec/arm/h264idct_neon.S b/libavcodec/arm/h264idct_neon.S
+index 6527390..d30e703 100644
+--- a/libavcodec/arm/h264idct_neon.S
++++ b/libavcodec/arm/h264idct_neon.S
+@@ -21,8 +21,6 @@
+ #include "asm.S"
+ 
+         preserve8
+-        .fpu neon
+-
+         .text
+ 
+ function ff_h264_idct_add_neon, export=1
+diff --git a/libavcodec/arm/mdct_neon.S b/libavcodec/arm/mdct_neon.S
+index 5cd4647..f089519 100644
+--- a/libavcodec/arm/mdct_neon.S
++++ b/libavcodec/arm/mdct_neon.S
+@@ -21,7 +21,6 @@
+ 
+ #include "asm.S"
+ 
+-        .fpu neon
+         .text
+ 
+ function ff_imdct_half_neon, export=1
+diff --git a/libavcodec/arm/simple_idct_neon.S b/libavcodec/arm/simple_idct_neon.S
+index 0882481..4c329d8 100644
+--- a/libavcodec/arm/simple_idct_neon.S
++++ b/libavcodec/arm/simple_idct_neon.S
+@@ -45,8 +45,6 @@
+ #define w7 d1[2]
+ #define w4c d1[3]
+ 
+-        .fpu neon
+-
+         .macro idct_col4_top
+         vmull.s16       q7,  d6,  w2    /* q9   = W2 * col[2] */
+         vmull.s16       q8,  d6,  w6    /* q10  = W6 * col[2] */
+-- 
+1.6.3.3
+

debian/patches/series

+# fix-qtrle-encoding-when-previous-frame-linesize.patch
+# fix-dv-seeking.patch
+# fix-constant-quant-encoding.patch
+# increase-VOFW-on-x86.patch
+# use-normal-check_func-test-for-math-functions.patch
+900_doxyfile
+# 901-fix-misc-typos.patch
+# fpic-ftbfs-fix.patch
+# fpic-libpostproc-fix.patch
+# fpic-libswscale-fix.patch
+
+# NEON backport
+# neon/0001-ARM-NEON-optimised-add_pixels_clamped.patch
+# neon/0002-ARM-NEON-optimized-put_signed_pixels_clamped.patch
+# neon/0003-Add-guaranteed-alignment-for-loading-dest-pixels-in-.patch
+# neon/0004-Reorganise-intreadwrite.h.patch
+# neon/0005-ARM-asm-for-AV_RN.patch
+# neon/0006-ARM-NEON-put_pixels_clamped.patch
+# neon/0007-ARM-Use-fewer-register-in-NEON-put_pixels-_y2-and-_x.patch
+# neon/0008-ARM-NEON-VP3-Loop-Filter.patch
+# neon/0009-ARM-actually-add-VP3-loop-filter.patch
+# neon/0010-ARM-add-some-PLD-in-NEON-IDCT.patch
+# neon/0011-ARM-slightly-faster-NEON-H264-horizontal-loop-filter.patch
+# neon/0012-ARM-enable-fast_unaligned-when-cpu-armv-67-is-specif.patch
+# neon/0013-ARM-NEON-VP3-IDCT.patch
+# neon/0014-ARM-NEON-optimised-vorbis_inverse_coupling.patch
+# neon/0015-ARM-handle-VFP-register-arguments-in-ff_vector_fmul_.patch
+# neon/0016-ARM-NEON-optimised-FFT-and-MDCT.patch
+# neon/0017-ARM-faster-NEON-IMDCT.patch
+# neon/0018-Prepare-for-optimised-forward-MDCT-implementations.patch
+# neon/0019-ARM-NEON-optimised-MDCT.patch
+# neon/0020-ARM-remove-unnecessary-.fpu-neon-directives.patch
+# neon/0021-ARM-check-for-VFP-register-arguments.patch
+

debian/source.lintian-overrides

+# The dependencies for packages within ffmpeg are different than the
+# dependencies for packages that depend on the ffmpeg libraries.
+ffmpeg-debian source: debian-rules-calls-debhelper-in-odd-order dh_makeshlibs (line 174)
+ffmpeg-debian source: debian-rules-calls-debhelper-in-odd-order dh_makeshlibs (line 178)

debian/watch

+version=3
+opts="uversionmangle=s/.*-snapshot//i" \
+http://www.ffmpeg.org/releases/ffmpeg-(.*)\.tar\.bz2