added ffmpeg-0.5_p22846.ebuild

Signed-off-by: Zhang Le <r0bertz@gentoo.org>

5 files changed, 3596 insertions, 0 deletions

diff --git a/media-video/ffmpeg/ChangeLog b/media-video/ffmpeg/ChangeLog
new file mode 100644
index 0000000..83619bf
--- /dev/null
+++ b/media-video/ffmpeg/ChangeLog
@@ -0,0 +1,1520 @@
+# ChangeLog for media-video/ffmpeg
+# Copyright 1999-2010 Gentoo Foundation; Distributed under the GPL v2
+# $Header: /var/cvsroot/gentoo-x86/media-video/ffmpeg/ChangeLog,v 1.351 2010/05/11 10:18:21 aballier Exp $
+
+  11 May 2010; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild,
+  metadata.xml:
+  add vaapi useflag, by Olivier Huber <oli.huber@gmail.com>, bug #308857
+
+  11 May 2010; Alexis Ballier <aballier@gentoo.org>
+  ffmpeg-0.5_p20373.ebuild, ffmpeg-0.5_p22846.ebuild:
+  force disabling vaapi to avoid automagics on unmasked versions, bug
+  #308857
+
+  11 May 2010; Alexis Ballier <aballier@gentoo.org>
+  -ffmpeg-0.5_p20601-r1.ebuild, -ffmpeg-0.5_p21602.ebuild,
+  -files/ffmpeg-0.5_p20601-r1-libtheora.patch:
+  remove old
+
+  23 Apr 2010; Samuli Suominen <ssuominen@gentoo.org>
+  ffmpeg-0.5_p20373.ebuild, ffmpeg-0.5_p20601-r1.ebuild,
+  ffmpeg-0.5_p21602.ebuild, ffmpeg-0.5_p22846.ebuild, metadata.xml:
+  Rename USE opencore-amr to amr.
+
+  23 Apr 2010; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild:
+  rename opencore-amr useflag to amr, we should get rid of the old non free
+  one
+
+*ffmpeg-0.5_p22846 (12 Apr 2010)
+
+  12 Apr 2010; Alexis Ballier <aballier@gentoo.org>
+  +ffmpeg-0.5_p22846.ebuild:
+  push a new snapshot
+
+  13 Mar 2010; Luca Barbato <lu_zero@gentoo.org> ffmpeg-0.5_p20373.ebuild,
+  ffmpeg-0.5_p20601-r1.ebuild, ffmpeg-0.5_p21602.ebuild,
+  ffmpeg-9999-r1.ebuild:
+  Add dep on pkg-config, fixes bug #306943
+
+  10 Mar 2010; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild:
+  simplify libgsm code as upstream now uses a more standard way for
+  #including it
+
+  04 Feb 2010; Alexis Ballier <aballier@gentoo.org>
+  ffmpeg-0.5_p21602.ebuild, ffmpeg-9999-r1.ebuild:
+  Remove ipv6 useflag, it cannot be disabled separately anymore and is
+  automatically enabled when available with USE=network, reported by marienz
+
+*ffmpeg-0.5_p21602 (02 Feb 2010)
+
+  02 Feb 2010; Alexis Ballier <aballier@gentoo.org>
+  +ffmpeg-0.5_p21602.ebuild:
+  bump a new snapshot
+
+  31 Jan 2010; Raúl Porcel <armin76@gentoo.org> ffmpeg-0.5_p20373.ebuild:
+  ia64/sparc stable
+
+  26 Jan 2010; Dror Levin <spatz@gentoo.org> ffmpeg-0.5_p20601-r1.ebuild,
+  ffmpeg-9999-r1.ebuild:
+  Handle -march=*-sse3 properly, bug 283968; Install the manpages, bug
+  288467.
+
+  26 Jan 2010; Dror Levin <spatz@gentoo.org> ffmpeg-0.5_p20601-r1.ebuild,
+  ffmpeg-9999-r1.ebuild:
+  Fix faac handling if USE=-encode, bug 300364.
+
+  24 Jan 2010; Tobias Klausmann <klausman@gentoo.org>
+  ffmpeg-0.5_p20373.ebuild:
+  Stable on alpha, bug #297847
+
+  19 Jan 2010; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild:
+  Adjust x264 deps
+
+  09 Jan 2010; Christian Faulhammer <fauli@gentoo.org>
+  ffmpeg-0.5_p20373.ebuild:
+  stable x86, bug 297847
+
+  07 Jan 2010; Brent Baude <ranger@gentoo.org> ffmpeg-0.5_p20373.ebuild:
+  Marking ffmpeg-0.5_p20373 ppc for bug 297847
+
+  06 Jan 2010; Brent Baude <ranger@gentoo.org> ffmpeg-0.5_p20373.ebuild:
+  Marking ffmpeg-0.5_p20373 ppc64 for bug 297847
+
+  04 Jan 2010; Markus Meier <maekke@gentoo.org> ffmpeg-0.5_p20373.ebuild:
+  arm stable, bug #297847
+
+  24 Dec 2009; Jeroen Roovers <jer@gentoo.org> ffmpeg-0.5_p20373.ebuild:
+  Stable for HPPA (bug #297847).
+
+  22 Dec 2009; Samuli Suominen <ssuominen@gentoo.org>
+  ffmpeg-0.5_p20373.ebuild:
+  amd64 stable wrt #297847
+
+*ffmpeg-0.5_p20601-r1 (26 Nov 2009)
+
+  26 Nov 2009; Steve Dibb <beandog@gentoo.org> -ffmpeg-0.5_p20601.ebuild,
+  +ffmpeg-0.5_p20601-r1.ebuild, +files/ffmpeg-0.5_p20601-r1-libtheora.patch:
+  Fix x264 compiling; libtheora patch for 1.0; bug 294469
+
+  25 Nov 2009; Steve Dibb <beandog@gentoo.org> ffmpeg-0.5_p20601.ebuild,
+  ffmpeg-9999-r1.ebuild:
+  Fix libtheora dep
+
+  24 Nov 2009; Steve Dibb <beandog@gentoo.org> ffmpeg-9999-r1.ebuild:
+  Update vdpau deps
+
+*ffmpeg-0.5_p20601 (24 Nov 2009)
+
+  24 Nov 2009; Steve Dibb <beandog@gentoo.org> +ffmpeg-0.5_p20601.ebuild:
+  Snapshot bump
+
+  18 Nov 2009; Raúl Porcel <armin76@gentoo.org> ffmpeg-0.5_p20373.ebuild:
+  Re-add ~sparc wrt #283953
+
+  17 Nov 2009; Alexis Ballier <aballier@gentoo.org>
+  -ffmpeg-0.5_p19928.ebuild:
+  remove old
+
+  06 Nov 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild:
+  Force disabling doc with use=-doc
+
+*ffmpeg-0.5_p20373 (26 Oct 2009)
+
+  26 Oct 2009; Alexis Ballier <aballier@gentoo.org>
+  +ffmpeg-0.5_p20373.ebuild:
+  Bump a new snapshot
+
+  22 Oct 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild:
+  bump deps on x264
+
+  22 Oct 2009; Samuli Suominen <ssuominen@gentoo.org> ffmpeg-0.5-r1.ebuild,
+  ffmpeg-0.5_p19928.ebuild:
+  Fix media-libs/x264 depend wrt #290080, thanks to Diego E. 'Flameeyes'
+  Pettenò for reporting.
+
+  28 Sep 2009; Alexis Ballier <aballier@gentoo.org>
+  -ffmpeg-0.5_p19787.ebuild:
+  remove old
+
+  28 Sep 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild,
+  metadata.xml:
+  Add a pic useflag
+
+  25 Sep 2009; Mounir Lamouri <volkmar@gentoo.org> ffmpeg-0.5_p19928.ebuild:
+  Keywording for ppc. Bug 283953
+
+  24 Sep 2009; Raúl Porcel <armin76@gentoo.org> ffmpeg-0.5_p19787.ebuild,
+  ffmpeg-0.5_p19928.ebuild:
+  Re-add ~ia64 wrt #283953
+
+  23 Sep 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild:
+  bump x264 requirements
+
+  20 Sep 2009; Markus Meier <maekke@gentoo.org> ffmpeg-0.5_p19928.ebuild:
+  add ~arm
+
+  20 Sep 2009; Markus Meier <maekke@gentoo.org> ffmpeg-0.5_p19787.ebuild:
+  add ~arm, bug #283953
+
+*ffmpeg-0.5_p19928 (20 Sep 2009)
+
+  20 Sep 2009; Alexis Ballier <aballier@gentoo.org>
+  +ffmpeg-0.5_p19928.ebuild:
+  bump a new snapshot
+
+  15 Sep 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-0.5-r1.ebuild:
+  Fix dependencies on x264 for 0.5, it doesn't detect the latest version,
+  bug #284937
+
+  07 Sep 2009; Alexis Ballier <aballier@gentoo.org>
+  ffmpeg-0.5_p19787.ebuild, ffmpeg-9999-r1.ebuild:
+  disable tests without USE=encode, they fail if it is not enabled
+
+  07 Sep 2009; Alexis Ballier <aballier@gentoo.org>
+  ffmpeg-0.5_p19787.ebuild:
+  drop keywords for arches that dont have opencore-amr keyworded
+
+*ffmpeg-0.5_p19787 (06 Sep 2009)
+
+  06 Sep 2009; Alexis Ballier <aballier@gentoo.org>
+  +ffmpeg-0.5_p19787.ebuild:
+  Add today's ffmpeg svn snapshot.
+
+  06 Sep 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild:
+  Use ffmpeg.org svn repo and add PV conditionals so that we can make
+  snapshots by simply copying -9999
+
+  06 Sep 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild:
+  Correctly detect cpu by upstream checks when march=native is used, bug
+  #273421
+
+  04 Aug 2009; Samuli Suominen <ssuominen@gentoo.org> ffmpeg-9999-r1.ebuild,
+  metadata.xml:
+  Rename opencore-amrnb and opencore-amrwb to opencore-amr to keep it in
+  sync with MPlayer.
+
+  27 Jul 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild:
+  Set svn revision
+
+  25 Jul 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild:
+  set target os when crosscompiling for a different os, support for linux,
+  freebsd and mingw32 so far
+
+  22 Jul 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild:
+  Set cross-prefix when cross compiling so that CHOST-prefixed tools are
+  used
+
+  09 Jul 2009; Samuli Suominen <ssuominen@gentoo.org> ffmpeg-0.5-r1.ebuild,
+  ffmpeg-9999-r1.ebuild:
+  USE vdpau with VIDEO_CARDS=nvidia wrt #262433, thanks to Bob Deblier for
+  reporting.
+
+  08 Jul 2009; Samuli Suominen <ssuominen@gentoo.org>
+  ffmpeg-0.4.9_p20090201.ebuild, ffmpeg-0.5-r1.ebuild,
+  ffmpeg-9999-r1.ebuild:
+  Rename EXTRA_ECONF to EXTRA_FFMPEG_CONF because this package isn't using
+  econf and fails with --disable-dependency-tracking in it.
+
+  07 Jul 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild:
+  Drop libamr support, its been removed upstream
+
+  04 Jul 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild:
+  Reflect upstream changes for the way of disabling libavdevice in/outdevs
+
+  03 Jul 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild:
+  force a non free license only when needed, ie, when faac or amr is
+  requested
+
+  03 Jul 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild,
+  metadata.xml:
+  Add opencore-amr support, switch to GPL-3
+
+  30 Jun 2009; Raúl Porcel <armin76@gentoo.org> ffmpeg-0.5-r1.ebuild:
+  arm/ia64/sparc stable
+
+  20 Jun 2009; Jeroen Roovers <jer@gentoo.org> ffmpeg-0.5-r1.ebuild:
+  Stable for HPPA (bug #274666).
+
+  10 Jun 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild:
+  remove useless warning, it must not be the user's duty to figure out abi
+  breaks
+
+  10 Jun 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999-r1.ebuild:
+  make tests work when upgrading, they should pass so make them also fatal
+
+  10 Jun 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-0.5-r1.ebuild,
+  ffmpeg-9999-r1.ebuild:
+  eapi2ify and depend on libtheora[encode] when needed, reported by remi
+
+  21 May 2009; Brent Baude <ranger@gentoo.org> ffmpeg-0.5-r1.ebuild:
+  Marking ffmpeg-0.5-r1 ppc64 stable for bug 266986
+
+  21 May 2009; Brent Baude <ranger@gentoo.org> ffmpeg-0.5-r1.ebuild:
+  Marking ffmpeg-0.5-r1 ppc stable for bug 266986
+
+  16 May 2009; Mounir Lamouri <volkmar@gentoo.org> metadata.xml:
+  Use global gsm USE flag instead of local one.
+
+  14 May 2009; Markus Meier <maekke@gentoo.org> ffmpeg-0.5-r1.ebuild:
+  amd64 stable, bug #266986
+
+  12 May 2009; Christian Faulhammer <fauli@gentoo.org> ffmpeg-0.5-r1.ebuild:
+  stable x86, bug 266986
+
+  03 May 2009; Peter Alfredsen <loki_val@gentoo.org> ffmpeg-0.5-r1.ebuild:
+  Update deps for USE="lame" w.r.t. bug 265830.
+
+  03 May 2009; Tobias Klausmann <klausman@gentoo.org> ffmpeg-0.5-r1.ebuild:
+  Stable on alpha, bug #266986
+
+*ffmpeg-9999-r1 (02 May 2009)
+
+  02 May 2009; Steve Dibb <beandog@gentoo.org> metadata.xml,
+  -ffmpeg-9999.ebuild, +ffmpeg-9999-r1.ebuild:
+  Add cpudetection use flag; require nonfree option for faac support
+
+  04 Apr 2009; Alexis Ballier <aballier@gentoo.org>
+  -files/ffmpeg-0.4.9_p20081014-sparc-gcc43.patch,
+  -files/ffmpeg-shared-gcc4.1.patch, -ffmpeg-0.4.9_p20081014.ebuild,
+  -ffmpeg-0.4.9_p20081219.ebuild, -ffmpeg-0.4.9_p20090121.ebuild,
+  -ffmpeg-0.4.9_p20090201-r1.ebuild:
+  remove old
+
+  04 Apr 2009; Raúl Porcel <armin76@gentoo.org>
+  ffmpeg-0.4.9_p20090201.ebuild, ffmpeg-0.4.9_p20090201-r1.ebuild,
+  ffmpeg-0.5-r1.ebuild, ffmpeg-9999.ebuild:
+  Enable arm optimizations again, now that i've found out that using -march
+  works fine
+
+  04 Apr 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999.ebuild:
+  add jack useflag
+
+  31 Mar 2009; Raúl Porcel <armin76@gentoo.org>
+  ffmpeg-0.4.9_p20090201.ebuild, ffmpeg-0.4.9_p20090201-r1.ebuild,
+  ffmpeg-0.5-r1.ebuild, ffmpeg-9999.ebuild:
+  arm stable, fix some arm-related stuff
+
+  11 Mar 2009; Steve Dibb <beandog@gentoo.org> metadata.xml,
+  ffmpeg-0.5-r1.ebuild, ffmpeg-9999.ebuild:
+  Fix faac local use flag desc; don't disable by default in IUSE
+
+*ffmpeg-0.5-r1 (10 Mar 2009)
+
+  10 Mar 2009; Steve Dibb <beandog@gentoo.org> metadata.xml,
+  -ffmpeg-0.5.ebuild, +ffmpeg-0.5-r1.ebuild, ffmpeg-9999.ebuild:
+  Drop aac use flag, split to faac / faad use flags - ffmpeg has native AAC
+  encoding / decoding which should be used by default; drop unused vhook use
+  flag on live ebuild
+
+*ffmpeg-0.5 (10 Mar 2009)
+
+  10 Mar 2009; Alexis Ballier <aballier@gentoo.org> +ffmpeg-0.5.ebuild:
+  Bump to the 0.5 release.
+
+  05 Mar 2009; Diego E. Pettenò <flameeyes@gentoo.org> metadata.xml,
+  ffmpeg-9999.ebuild:
+  Fix build with recent changes upstream (vhook removed in favour of
+  libavfilter, swscale cannot be disabled anylonger); merge the EAPI=1
+  changes from the snapshotted version; add 3dnowext USE flag.
+
+  28 Feb 2009; Markus Meier <maekke@gentoo.org> metadata.xml:
+  custom-cflags is a global USE-flag
+
+*ffmpeg-0.4.9_p20090201-r1 (21 Feb 2009)
+
+  21 Feb 2009; Luca Barbato <lu_zero@gentoo.org>
+  +ffmpeg-0.4.9_p20090201-r1.ebuild:
+  Set some useflag defaults and other minor changes
+
+  11 Feb 2009; Tobias Scherbaum <dertobi123@gentoo.org>
+  ffmpeg-0.4.9_p20090201.ebuild:
+  ppc stable, bug #257217
+
+  10 Feb 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999.ebuild:
+  add openjpeg support
+
+  09 Feb 2009; Raúl Porcel <armin76@gentoo.org>
+  ffmpeg-0.4.9_p20090201.ebuild:
+  ia64/sparc stable wrt #257217
+
+  05 Feb 2009; Jeroen Roovers <jer@gentoo.org>
+  ffmpeg-0.4.9_p20090201.ebuild:
+  Stable for HPPA (bug #257217).
+
+  04 Feb 2009; Markus Meier <maekke@gentoo.org>
+  ffmpeg-0.4.9_p20090201.ebuild:
+  x86 stable, bug #257217
+
+  02 Feb 2009; Tobias Heinlein <keytoaster@gentoo.org>
+  ffmpeg-0.4.9_p20090201.ebuild:
+  amd64 stable wrt security bug 257217
+
+  02 Feb 2009; Brent Baude <ranger@gentoo.org>
+  ffmpeg-0.4.9_p20090201.ebuild:
+  Marking ffmpeg-0.4.9_p20090201 ppc64 for bug 257217
+
+  02 Feb 2009; Tobias Klausmann <klausman@gentoo.org>
+  ffmpeg-0.4.9_p20090201.ebuild:
+  Stable on alpha, bug #257217
+
+*ffmpeg-0.4.9_p20090201 (01 Feb 2009)
+
+  01 Feb 2009; Alexis Ballier <aballier@gentoo.org>
+  +ffmpeg-0.4.9_p20090201.ebuild:
+  new snapshot, bug #257217
+
+  31 Jan 2009; Tobias Klausmann <klausman@gentoo.org>
+  ffmpeg-0.4.9_p20081219.ebuild:
+  Stable on alpha, bug #255008
+
+  27 Jan 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999.ebuild:
+  no need for local variables when we use them only once
+
+  27 Jan 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999.ebuild:
+  Add alsa/oss/3dnow useflags and disable servertest as its mainly broken
+
+  25 Jan 2009; Markus Meier <maekke@gentoo.org>
+  ffmpeg-0.4.9_p20081219.ebuild:
+  amd64/x86 stable, bug #255008
+
+  25 Jan 2009; Brent Baude <ranger@gentoo.org>
+  ffmpeg-0.4.9_p20081219.ebuild:
+  stable ppc64, bug 255008
+
+*ffmpeg-0.4.9_p20090121 (21 Jan 2009)
+
+  21 Jan 2009; Alexis Ballier <aballier@gentoo.org>
+  +ffmpeg-0.4.9_p20090121.ebuild:
+  push a new snapshot
+
+  20 Jan 2009; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999.ebuild:
+  set empty keywords for the live ebuild
+
+  19 Jan 2009; Alexis Ballier <aballier@gentoo.org>
+  ffmpeg-0.4.9_p20081014.ebuild, ffmpeg-0.4.9_p20081219.ebuild,
+  ffmpeg-9999.ebuild:
+  require gnu make 3.81, by Bernd Kaiser <meldron@gmx.net>, bug #255376
+
+  17 Jan 2009; Alexis Ballier <aballier@gentoo.org>
+  ffmpeg-0.4.9_p20081219.ebuild, ffmpeg-9999.ebuild:
+  require at least speex 1.2, should fix #254922
+
+  15 Jan 2009; Alexis Ballier <aballier@gentoo.org>
+  ffmpeg-0.4.9_p20081014.ebuild, ffmpeg-0.4.9_p20081219.ebuild,
+  ffmpeg-9999.ebuild:
+  make sure version.h gets created, bug #252269
+
+  23 Dec 2008; Alexis Ballier <aballier@gentoo.org> ffmpeg-9999.ebuild:
+  Sync live ebuild with latest snapshot; merely add speex useflag and remove
+  unused make targets
+
+  22 Dec 2008; Raúl Porcel <armin76@gentoo.org>
+  ffmpeg-0.4.9_p20081014.ebuild, ffmpeg-0.4.9_p20081219.ebuild:
+  arm/ia64 stable wrt #245285
+
+  21 Dec 2008; nixnut <nixnut@gentoo.org> ffmpeg-0.4.9_p20081014.ebuild:
+  Stable on ppc wrt bug 245313
+
+  20 Dec 2008; Ferris McCormick <fmccor@gentoo.org>
+  ffmpeg-0.4.9_p20081014.ebuild:
+  Sparc stable, Security Bug #245313.
+
+  20 Dec 2008; Thomas Anderson <gentoofan23@gentoo.org>
+  ffmpeg-0.4.9_p20081014.ebuild:
+  stable amd64, security bug #245313
+
+*ffmpeg-0.4.9_p20081219 (19 Dec 2008)
+
+  19 Dec 2008; Alexis Ballier <aballier@gentoo.org>
+  +ffmpeg-0.4.9_p20081219.ebuild:
+  push a new snapshot
+
+  16 Dec 2008; Jeroen Roovers <jer@gentoo.org>
+  ffmpeg-0.4.9_p20081014.ebuild:
+  Stable for HPPA (bug #245291).
+
+  16 Dec 2008; Brent Baude <ranger@gentoo.org>
+  ffmpeg-0.4.9_p20081014.ebuild:
+  Marking ffmpeg-0.4.9_p20081014 ppc64 for bug 245285
+
+  15 Dec 2008; <ssuominen@gentoo.org> ffmpeg-0.4.9_p20081014.ebuild:
+  Change faad2 dep. to >= 2.6.1 wrt #251010.
+
+  15 Dec 2008; <ssuominen@gentoo.org> ffmpeg-0.4.9_p20081014.ebuild:
+  x86 stable wrt #245285
+
+  13 Dec 2008; <ssuominen@gentoo.org>
+  +files/ffmpeg-0.4.9_p20081014-sparc-gcc43.patch,
+  ffmpeg-0.4.9_p20081014.ebuild:
+  Backport upstream patch for sparc to allow compilation with GCC 4.3 wrt
+  #247653.
+
+  12 Dec 2008; Tobias Klausmann <klausman@gentoo.org>
+  ffmpeg-0.4.9_p20081014.ebuild:
+  Stable on alpha, bug #245285
+
+*ffmpeg-9999 (07 Dec 2008)
+
+  07 Dec 2008; Luca Barbato <lu_zero@gentoo.org> +ffmpeg-9999.ebuild:
+  live ebuild
+
+  10 Nov 2008; Raúl Porcel <armin76@gentoo.org>
+  ffmpeg-0.4.9_p20081014.ebuild:
+  Add ~ia64 wrt #241908
+
+  30 Oct 2008; Raúl Porcel <armin76@gentoo.org>
+  ffmpeg-0.4.9_p20070616-r3.ebuild:
+  alpha/arm/ia64 stable #231831
+
+  27 Oct 2008; Alexis Ballier <aballier@gentoo.org>
+  ffmpeg-0.4.9_p20081014.ebuild:
+  Fixup crosscompilation, bug #237662
+
+  27 Oct 2008; Mike Frysinger <vapier@gentoo.org>
+  ffmpeg-0.4.9_p20081014.ebuild:
+  Add arm/ppc love #241908.
+
+  18 Oct 2008; Peter Alfredsen <loki_val@gentoo.org>
+  ffmpeg-0.4.9_p20080326.ebuild:
+  Fixup 20080326 dependency on x264 to be <0.0.20081006.
+
+  18 Oct 2008; Alexis Ballier <aballier@gentoo.org>
+  ffmpeg-0.4.9_p20081014.ebuild:
+  Drop arm/ia64/ppc ~arch keywords per bug #241908 to allow unmasking
+
+  14 Oct 2008; Jeroen Roovers <jer@gentoo.org>
+  +files/ffmpeg-0.4.9_p20080326-hppa.patch, ffmpeg-0.4.9_p20080326.ebuild:
+  Build PIC on HPPA (bug #241124).
+
+  14 Oct 2008; Alexis Ballier <aballier@gentoo.org>
+  ffmpeg-0.4.9_p20081014.ebuild:
+  Give the configure the --cpu option based on CFLAGS; if it is unknown it
+  will not hurt because the configure script will warn and ignore it. This
+  helps getting better support for given cpu, like using CMOV on i686 and
+  later. Bug #172723.
+
+  14 Oct 2008; Alexis Ballier <aballier@gentoo.org>
+  ffmpeg-0.4.9_p20081014.ebuild:
+  Set the ffmpeg version to the exported revision number, bug #233667,
+  borrowed from mplayer ebuild
+
+  14 Oct 2008; Alexis Ballier <aballier@gentoo.org>
+  ffmpeg-0.4.9_p20081014.ebuild:
+  Broaden libdc1394 dep as it works with v2 too, by Fabio Correa
+  <facorread@gmail.com>, bug #237687
+
+  14 Oct 2008; Alexis Ballier <aballier@gentoo.org>
+  ffmpeg-0.4.9_p20070616.ebuild, ffmpeg-0.4.9_p20070616-r1.ebuild,
+  ffmpeg-0.4.9_p20070616-r2.ebuild, ffmpeg-0.4.9_p20070616-r3.ebuild,
+  ffmpeg-0.4.9_p20070616-r20.ebuild, ffmpeg-0.4.9_p20080206.ebuild,
+  ffmpeg-0.4.9_p20080326.ebuild, ffmpeg-0.4.9_p20081014.ebuild:
+  Bump imlib2 dep for bug #196525
+
+*ffmpeg-0.4.9_p20081014 (14 Oct 2008)
+
+  14 Oct 2008; Alexis Ballier <aballier@gentoo.org> metadata.xml,
+  +ffmpeg-0.4.9_p20081014.ebuild:
+  Add a new snapshot, masked because it breaks a lot of consumers.
+
+  11 Oct 2008; Tobias Scherbaum <dertobi123@gentoo.org>
+  ffmpeg-0.4.9_p20070616-r3.ebuild:
+  ppc stable, bug #231831
+
+  06 Oct 2008; Markus Meier <maekke@gentoo.org>
+  ffmpeg-0.4.9_p20070616-r3.ebuild:
+  amd64/x86 stable, bug #231831
+
+  06 Oct 2008; Friedrich Oslage <bluebird@gentoo.org>
+  ffmpeg-0.4.9_p20070616-r3.ebuild:
+  Stable on sparc, security bug #231831
+
+  06 Oct 2008; Jeroen Roovers <jer@gentoo.org>
+  ffmpeg-0.4.9_p20070616-r3.ebuild:
+  Stable for HPPA (bug #231831). Fixed some quoting issues.
+
+  04 Oct 2008; Markus Rothe <corsair@gentoo.org>
+  ffmpeg-0.4.9_p20070616-r3.ebuild:
+  Stable on ppc64; bug #231831
+
+*ffmpeg-0.4.9_p20070616-r20 (17 Jul 2008)
+*ffmpeg-0.4.9_p20070616-r3 (17 Jul 2008)
+
+  17 Jul 2008; Alexis Ballier <aballier@gentoo.org>
+  +files/CVE-2008-3162.patch, +ffmpeg-0.4.9_p20070616-r3.ebuild,
+  +ffmpeg-0.4.9_p20070616-r20.ebuild:
+  Add patches for security bug #231831. -r3 is -r0 with the patch, stable
+  candidate without swscaler. -r20 is -r2 with the patch, with swscaler.
+
+  07 Jul 2008; Peter Alfredsen <loki_val@gentoo.org>
+  ffmpeg-0.4.9_p20080326.ebuild:
+  Fix build failure when disabling MMX, which would fail to disable MMX2
+  code on processors supporting it, resulting in #ifdef spaghetti build
+  failure. Bug 229981.
+
+  20 Apr 2008; Mike Frysinger <vapier@gentoo.org>
+  +files/ffmpeg-arm-pld.patch, ffmpeg-0.4.9_p20070616.ebuild,
+  ffmpeg-0.4.9_p20070616-r1.ebuild, ffmpeg-0.4.9_p20070616-r2.ebuild,
+  ffmpeg-0.4.9_p20080206.ebuild:
+  Fix from upstream svn for building for older arm cpus.
+
+  07 Apr 2008; Ben de Groot <yngwin@gentoo.org>
+  ffmpeg-0.4.9_p20070616.ebuild, ffmpeg-0.4.9_p20070616-r1.ebuild,
+  ffmpeg-0.4.9_p20070616-r2.ebuild, ffmpeg-0.4.9_p20080206.ebuild,
+  ffmpeg-0.4.9_p20080326.ebuild:
+  Pkgmove x264-svn to x264
+
+  27 Mar 2008; Joerg Bornkessel <hd_brummy@gentoo.org> Manifest:
+  manifest fixed; #bug 214967
+
+*ffmpeg-0.4.9_p20080326 (26 Mar 2008)
+
+  26 Mar 2008; Luca Barbato <lu_zero@gentoo.org>
+  -files/ffmpeg-0.4.9_p20051216-asneeded-configure.patch,
+  -ffmpeg-0.4.9_p20050226-r3.ebuild, -ffmpeg-0.4.9_p20061016.ebuild,
+  -ffmpeg-0.4.9_p20070330.ebuild, -ffmpeg-0.4.9_p20070525.ebuild,
+  +ffmpeg-0.4.9_p20080326.ebuild:
+  Furter cleanup and new version
+
+  28 Feb 2008; Alexis Ballier <aballier@gentoo.org>
+  ffmpeg-0.4.9_p20080206.ebuild:
+  fix build with use ieee1394, dc1394->libdc1394. Bug #209610
+
+  14 Feb 2008; Samuli Suominen <drac@gentoo.org>
+  -ffmpeg-0.4.9_p20051216.ebuild, -ffmpeg-0.4.9_p20060302.ebuild,
+  -ffmpeg-0.4.9_p20060530.ebuild, -ffmpeg-0.4.9_p20060816.ebuild,
+  ffmpeg-0.4.9_p20061016.ebuild, -ffmpeg-0.4.9_p20070129.ebuild,
+  -ffmpeg-0.4.9_p20070325.ebuild, ffmpeg-0.4.9_p20070330.ebuild:
+  Remove USE dts from some old ebuilds, and remove some entirely because
+  libdts is replaced by libdca.
+
+*ffmpeg-0.4.9_p20080206 (06 Feb 2008)
+
+  06 Feb 2008; Luca Barbato <lu_zero@gentoo.org>
+  +ffmpeg-0.4.9_p20080206.ebuild:
+  New snapshot
+
+  31 Oct 2007; Matthias Schwarzott <zzam@gentoo.org>
+  ffmpeg-0.4.9_p20051216.ebuild, ffmpeg-0.4.9_p20060302.ebuild,
+  ffmpeg-0.4.9_p20060530.ebuild, ffmpeg-0.4.9_p20060816.ebuild,
+  ffmpeg-0.4.9_p20061016.ebuild, ffmpeg-0.4.9_p20070129.ebuild,
+  ffmpeg-0.4.9_p20070325.ebuild, ffmpeg-0.4.9_p20070330.ebuild,
+  ffmpeg-0.4.9_p20070525.ebuild, ffmpeg-0.4.9_p20070616.ebuild,
+  ffmpeg-0.4.9_p20070616-r1.ebuild, ffmpeg-0.4.9_p20070616-r2.ebuild:
+  Improved text about possible ABI changes, noticed by mark_alec.
+
+*ffmpeg-0.4.9_p20070616-r2 (26 Oct 2007)
+
+  26 Oct 2007; Steve Dibb <beandog@gentoo.org>
+  +ffmpeg-0.4.9_p20070616-r2.ebuild:
+  Add ipv6 use flag
+
+  21 Oct 2007; Steve Dibb <beandog@gentoo.org>
+  ffmpeg-0.4.9_p20070616.ebuild:
+  amd64 stable, bug 193563
+
+  20 Oct 2007; Ferris McCormick <fmccor@gentoo.org>
+  ffmpeg-0.4.9_p20070616.ebuild:
+  sparc stable --- Bug #193563 --- builds and works.
+
+  08 Oct 2007; Markus Rothe <corsair@gentoo.org>
+  ffmpeg-0.4.9_p20070616.ebuild:
+  Stable on ppc64; bug #193563
+
+  24 Sep 2007; Raúl Porcel <armin76@gentoo.org>
+  ffmpeg-0.4.9_p20070616.ebuild:
+  alpha/ia64 stable wrt #193563
+
+  24 Sep 2007; Lars Weiler <pylon@gentoo.org> ffmpeg-0.4.9_p20070616.ebuild:
+  stable ppc, bug #193563
+
+  24 Sep 2007; Jeroen Roovers <jer@gentoo.org>
+  ffmpeg-0.4.9_p20070616.ebuild:
+  Stable for HPPA (bug #193563).
+
+  24 Sep 2007; Christian Faulhammer <opfer@gentoo.org>
+  ffmpeg-0.4.9_p20070616.ebuild:
+  stable x86, bug #193563
+
+  07 Jul 2007; Samuli Suominen <drac@gentoo.org>
+  ffmpeg-0.4.9_p20070616-r1.ebuild:
+  Fix building with USE altivec for bug 183687. Thanks to Laurent G. for
+  reporting, Joe Jezak for solution, Kimura Masaru for testing.
+
+*ffmpeg-0.4.9_p20070616-r1 (26 Jun 2007)
+
+  26 Jun 2007; Luca Barbato <lu_zero@gentoo.org>
+  +ffmpeg-0.4.9_p20070616-r1.ebuild:
+  Enable swscaler
+
+  16 Jun 2007; Samuli Suominen <drac@gentoo.org>
+  ffmpeg-0.4.9_p20070616.ebuild:
+  enable xvid to enable libxvid.
+
+*ffmpeg-0.4.9_p20070616 (16 Jun 2007)
+
+  16 Jun 2007; Luca Barbato <lu_zero@gentoo.org>
+  +ffmpeg-0.4.9_p20070616.ebuild:
+  New snapshot, SVN revision 9330
+
+  01 Jun 2007; Samuli Suominen <drac@gentoo.org>
+  ffmpeg-0.4.9_p20070525.ebuild:
+  Append -DBROKEN_RELOCATIONS which is a hack to workaround bug 179872.
+
+  01 Jun 2007; Samuli Suominen <drac@gentoo.org>
+  ffmpeg-0.4.9_p20070525.ebuild:
+  Remove USE dts for bug 180129.
+
+*ffmpeg-0.4.9_p20070525 (26 May 2007)
+
+  26 May 2007; Luca Barbato <lu_zero@gentoo.org>
+  +ffmpeg-0.4.9_p20070525.ebuild:
+  New snapshot, using libamr
+
+  16 May 2007; Jeroen Roovers <jer@gentoo.org>
+  ffmpeg-0.4.9_p20070330.ebuild:
+  Stable for HPPA (bug #174909).
+
+  05 May 2007; Fabian Groffen <grobian@gentoo.org>
+  ffmpeg-0.4.9_p20051216.ebuild, ffmpeg-0.4.9_p20060302.ebuild,
+  ffmpeg-0.4.9_p20060530.ebuild:
+  Dropped ppc-macos keyword, see you in prefix
+
+  22 Apr 2007; Bryan Østergaard <kloeri@gentoo.org>
+  ffmpeg-0.4.9_p20070330.ebuild:
+  Stable on Alpha, bug 174909.
+
+  17 Apr 2007; Tobias Scherbaum <dertobi123@gentoo.org>
+  ffmpeg-0.4.9_p20070330.ebuild:
+  ppc stable, bug #168907
+
+  16 Apr 2007; Raúl Porcel <armin76@gentoo.org>
+  ffmpeg-0.4.9_p20070330.ebuild:
+  ia64 stable
+
+  12 Apr 2007; Raúl Porcel <armin76@gentoo.org>
+  ffmpeg-0.4.9_p20070330.ebuild:
+  x86 stable
+
+  12 Apr 2007; Peter Weller <welp@gentoo.org> ffmpeg-0.4.9_p20070330.ebuild:
+  Stable on amd64 wrt bug 168907
+
+  05 Apr 2007; Alexis Ballier <aballier@gentoo.org>
+  +files/ffmpeg-0.4.9_p20070330-asmpic.patch, ffmpeg-0.4.9_p20070330.ebuild:
+  Disable non pic safe asm on x86 and amd64, bug #172845, bug #172877 and dupes
+
+  04 Apr 2007; Markus Rothe <corsair@gentoo.org>
+  ffmpeg-0.4.9_p20070330.ebuild:
+  Stable on ppc64
+
+  04 Apr 2007; Gustavo Zacarias <gustavoz@gentoo.org>
+  ffmpeg-0.4.9_p20070330.ebuild:
+  Stable on sparc wrt #170860
+
+*ffmpeg-0.4.9_p20070330 (30 Mar 2007)
+
+  30 Mar 2007; Luca Barbato <lu_zero@gentoo.org>
+  +ffmpeg-0.4.9_p20070330.ebuild:
+  New snapshot
+
+  27 Mar 2007; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20070325.ebuild:
+  s/a52/liba52
+
+  25 Mar 2007; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20070325.ebuild:
+  Minor fixes and update
+
+*ffmpeg-0.4.9_p20070325 (25 Mar 2007)
+
+  25 Mar 2007; Luca Barbato <lu_zero@gentoo.org>
+  +ffmpeg-0.4.9_p20070325.ebuild:
+  New snapshot
+
+  22 Mar 2007; Gustavo Zacarias <gustavoz@gentoo.org>
+  ffmpeg-0.4.9_p20070129.ebuild:
+  Stable on sparc
+
+  21 Mar 2007; Christian Faulhammer <opfer@gentoo.org>
+  ffmpeg-0.4.9_p20070129.ebuild:
+  stable x86, security bug 170208
+
+  16 Feb 2007; Simon Stelling <blubb@gentoo.org>
+  ffmpeg-0.4.9_p20061016.ebuild, ffmpeg-0.4.9_p20070129.ebuild:
+  depend on >=portage-2.1.2 on amd64 to make sure the mmx USE flag is unmasked
+
+  29 Jan 2007; Steve Dibb <beandog@gentoo.org>
+  ffmpeg-0.4.9_p20070129.ebuild:
+  Rekeyword ~
+
+  29 Jan 2007; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20070129.ebuild:
+  Put it back in ~, not the right time, closing bug #164445, thanks to
+  Sebastian <sebastian_ml@gmx.net> for notifying.
+
+  29 Jan 2007; Luca Barbato <lu_zero@gentoo.org>
+  files/ffmpeg-libdir-2007.patch:
+  Make the patch working everywhere, thanks to Markus Trippelsdorf
+  <markus@trippelsdorf.de> for pointing the issue
+
+*ffmpeg-0.4.9_p20070129 (29 Jan 2007)
+
+  29 Jan 2007; Luca Barbato <lu_zero@gentoo.org>
+  +files/ffmpeg-libdir-2007.patch, +ffmpeg-0.4.9_p20070129.ebuild:
+  New snapshot
+
+  29 Jan 2007; Bryan Østergaard <kloeri@gentoo.org>
+  ffmpeg-0.4.9_p20061016.ebuild:
+  Stable on Alpha + IA64.
+
+  06 Jan 2007; Michael Cummings <mcummings@gentoo.org>
+  ffmpeg-0.4.9_p20061016.ebuild:
+  amd64 stable
+
+  04 Jan 2007; Christian Faulhammer <opfer@gentoo.org>
+  ffmpeg-0.4.9_p20061016.ebuild:
+  stable x86, bug #157814
+
+  30 Dec 2006; Bryan Østergaard <kloeri@gentoo.org>
+  ffmpeg-0.4.9_p20061016.ebuild:
+  Add ~alpha and ~ia64 keywords.
+
+  23 Dec 2006; Tobias Scherbaum <dertobi123@gentoo.org>
+  ffmpeg-0.4.9_p20061016.ebuild:
+  Stable on ppc wrt bug #157814.
+
+  21 Dec 2006; Gustavo Zacarias <gustavoz@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild, ffmpeg-0.4.9_p20061016.ebuild:
+  Stable on sparc wrt #157814
+
+  21 Dec 2006; Markus Rothe <corsair@gentoo.org>
+  ffmpeg-0.4.9_p20061016.ebuild:
+  Stable on ppc64; bug #157814
+
+  10 Dec 2006; Matthias Schwarzott <zzam@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r3.ebuild, ffmpeg-0.4.9_p20051216.ebuild,
+  ffmpeg-0.4.9_p20060302.ebuild, ffmpeg-0.4.9_p20060517.ebuild,
+  ffmpeg-0.4.9_p20060530.ebuild, ffmpeg-0.4.9_p20060816.ebuild,
+  ffmpeg-0.4.9_p20061016.ebuild:
+  Corrected installation of source-Changelog-file, thanks to Josef Reidinger
+  <queen.killer@seznam.cz> for reporting, Bug #157628.
+
+  07 Dec 2006; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20061016.ebuild:
+  Fix reg depletion in x86, see bug #154922
+
+  01 Dec 2006; Jeroen Roovers <jer@gentoo.org>
+  ffmpeg-0.4.9_p20061016.ebuild:
+  Stable for HPPA (or xine-lib will not even configure).
+
+*ffmpeg-0.4.9_p20061016 (17 Oct 2006)
+
+  17 Oct 2006; Luca Barbato <lu_zero@gentoo.org>
+  +ffmpeg-0.4.9_p20061016.ebuild:
+  New version
+
+  03 Oct 2006; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild, ffmpeg-0.4.9_p20060816.ebuild:
+  Fix deps, see bug #134555
+
+  11 Sep 2006; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20060816.ebuild:
+  emake -j1 calls and add emake depend just in case
+
+  08 Sep 2006; Rene Nussbaumer <killerfox@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild:
+  Stable on hppa. See bug #133520.
+
+  08 Sep 2006; Thomas Cort <tcort@gentoo.org> ffmpeg-0.4.9_p20060530.ebuild:
+  Stable on alpha wrt security Bug #133520.
+
+  08 Sep 2006; Tobias Scherbaum <dertobi123@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild:
+  ppc stable, bug #133520
+
+  07 Sep 2006; Joshua Jackson <tsunam@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild:
+  Stable x86; for security bug #133520
+
+  07 Sep 2006; Thomas Cort <tcort@gentoo.org> ffmpeg-0.4.9_p20060530.ebuild:
+  Stable on amd64.
+
+  07 Sep 2006; Gustavo Zacarias <gustavoz@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild:
+  Stable on sparc wrt #133520
+
+  07 Sep 2006; Markus Rothe <corsair@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild:
+  Stable on ppc64; bug #133520
+
+  04 Sep 2006; Diego Pettenò <flameeyes@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild, ffmpeg-0.4.9_p20060816.ebuild:
+  Export true as LDCONFIG value during install phases, so that it does not
+  trigger sandbox on Gentoo/FreeBSD.
+
+  30 Aug 2006; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20060816.ebuild:
+  Fix multilib-strict and support EXTRA_ECONF
+
+  17 Aug 2006; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20060816.ebuild:
+  fix sdl dep
+
+  16 Aug 2006; Luca Barbato <lu_zero@gentoo.org> ChangeLog:
+  New version
+
+  14 Aug 2006; Luca Barbato <lu_zero@gentoo.org>
+  +files/ffmpeg-0.4.9_p20060530-snow-mmx.patch,
+  ffmpeg-0.4.9_p20060530.ebuild:
+  Make snow compile on x86 with pic and mmx enabled at the same time, many
+  thanks to Martin von Gagern <Martin.vGagern@gmx.net> for the patch and the
+  overall help
+
+  12 Aug 2006; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild:
+  Marked ~ppc64
+
+  12 Aug 2006; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild:
+  Pin x264 version to avoid build issue with snapshots too new
+
+  01 Aug 2006; Joshua Jackson <tsunam@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild:
+  Adding ~x86 as it works and all that jazz
+
+  31 May 2006; Gustavo Zacarias <gustavoz@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild:
+  Gave it a ~sparc
+
+  31 May 2006; Diego Pettenò <flameeyes@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild:
+  Re-add ~x86-fbsd and ~amd64 keywords.
+
+  31 May 2006; Diego Pettenò <flameeyes@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild:
+  Remove keywording comments, we always add and drop them every other release.
+
+  31 May 2006; Diego Pettenò <flameeyes@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild:
+  Remove old hack for $LDFLAGS passing (fixed upstream), remove the
+  append-flags (fixed upstream).
+
+  12 Jun 2006; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild:
+  more files to the sed PIC
+
+  31 May 2006; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20060530.ebuild:
+  Missing patch
+
+*ffmpeg-0.4.9_p20060530 (30 May 2006)
+
+  30 May 2006; Luca Barbato <lu_zero@gentoo.org>
+  +ffmpeg-0.4.9_p20060530.ebuild:
+  New snapshot
+
+  28 May 2006; Diego Pettenò <flameeyes@gentoo.org>
+  -files/ffmpeg-0.4.9_p20050226-gcc4.patch,
+  -files/ffmpeg-0.4.9_p20050906-osx.patch, -files/ffmpeg-a52.patch,
+  -files/ffmpeg-configure.patch, -files/ffmpeg-missing_links.patch,
+  -files/ffmpeg-osx.patch, -files/gentoo-ffmpeg001.patch,
+  -ffmpeg-0.4.9_p20050226-r5.ebuild, -ffmpeg-0.4.9_p20050906.ebuild,
+  -ffmpeg-0.4.9_p20051120.ebuild:
+  Remove old versions.
+
+  25 May 2006; Diego Pettenò <flameeyes@gentoo.org>
+  ffmpeg-0.4.9_p20060302.ebuild:
+  Add ~x86-fbsd keyword.
+
+*ffmpeg-0.4.9_p20060517 (17 May 2006)
+
+  17 May 2006; Luca Barbato <lu_zero@gentoo.org>
+  +ffmpeg-0.4.9_p20060517.ebuild:
+  new snapshot, to be tested
+
+  08 May 2006; Diego Pettenò <flameeyes@gentoo.org>
+  ffmpeg-0.4.9_p20060302.ebuild:
+  Use -rpath-link to make sure that the libraries are linked correctly, should
+  fixes bug #132171.
+
+  03 May 2006; Diego Pettenò <flameeyes@gentoo.org> metadata.xml:
+  Update metadata to list media-video@gentoo.org as maintainer.
+
+  20 Apr 2006; Daniel Gryniewicz <dang@gentoo.org>
+  +files/ffmpeg-0.4.9_p20060302-amr-64bit.patch,
+  ffmpeg-0.4.9_p20060302.ebuild:
+  Fix amr on 64-bit arches per bug# 130530
+
+  14 Apr 2006; Diego Pettenò <flameeyes@gentoo.org>
+  +files/ffmpeg-0.4.9_p20060302-fbsd-flags.patch,
+  ffmpeg-0.4.9_p20060302.ebuild:
+  Add patch to respect CFLAGS on FreeBSD, and don't use preplib that's no more
+  needed anyway.
+
+  31 Mar 2006; Simon Stelling <blubb@gentoo.org> +files/ffmpeg-libdir.patch,
+  ffmpeg-0.4.9_p20060302.ebuild:
+  fix wrong libdir paths in .pc files
+
+  21 Mar 2006; Diego Pettenò <flameeyes@gentoo.org>
+  +files/ffmpeg-0.4.9_p20051216-asneeded-configure.patch,
+  ffmpeg-0.4.9_p20051216.ebuild, ffmpeg-0.4.9_p20060302.ebuild:
+  Use --cc option instead of overwriting CC variable to respect the compiler,
+  pass --extra-ldflags to respect LDFLAGS (only 20060302 respect them
+  entirely). Add patch to allow using --as-needed flag.
+
+  08 Mar 2006; Diego Pettenò <flameeyes@gentoo.org>
+  ffmpeg-0.4.9_p20051216.ebuild, ffmpeg-0.4.9_p20060302.ebuild:
+  Disable strip, leaving to portage decide if and when to strip.
+
+  06 Mar 2006; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20060302.ebuild:
+  merged v4l and v4l2
+
+  05 Mar 2006; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20060302.ebuild:
+  Improved ebuild with amr and v4l2 support, thanks to Philip Kovacs
+  <kovacsp3@comcast.net> for the improved ebuild, other minor fixes
+
+  04 Mar 2006; Luca Barbato <lu_zero@gentoo.org>
+  files/ffmpeg-shared-gcc4.1.patch:
+  fix
+
+  04 Mar 2006; Luca Barbato <lu_zero@gentoo.org>
+  +files/ffmpeg-shared-gcc4.1.patch, ffmpeg-0.4.9_p20060302.ebuild:
+  ffmpeg-shared-gcc4 patch is back
+
+*ffmpeg-0.4.9_p20060302 (03 Mar 2006)
+
+  03 Mar 2006; Luca Barbato <lu_zero@gentoo.org>
+  +ffmpeg-0.4.9_p20060302.ebuild:
+  New snapshot
+
+  15 Feb 2006; Diego Pettenò <flameeyes@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r3.ebuild:
+  Add missing sdl useflag in IUSE.
+
+  30 Jan 2006; Luca Barbato <lu_zero@gentoo.org>
+  +files/ffmpeg-shared-gcc4.patch, ffmpeg-0.4.9_p20051216.ebuild:
+  x86 workaround from Kevin F. Quinn <kevquinn@gentoo.org>, I'll share the
+  blame if breaks something
+
+  09 Jan 2006; Bryan Østergaard <kloeri@gentoo.org
+  ffmpeg-0.4.9_p20051216.ebuild:
+  Stable on alpha, bug 116181.
+
+  06 Jan 2006; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20051216.ebuild:
+  Warning added
+
+  06 Jan 2006; Luca Barbato <lu_zero@gentoo.org>
+  files/ffmpeg-soname-symlink.patch:
+  minor fix for osx
+
+  03 Jan 2006; Bryan Østergaard <kloeri@gentoo.org
+  ffmpeg-0.4.9_p20051216.ebuild:
+  ~alpha keyword, bug 116181.
+
+  02 Jan 2006; Fabian Groffen <grobian@gentoo.org>
+  ffmpeg-0.4.9_p20051216.ebuild:
+  Marked ~ppc-macos (bug #116181)
+
+  02 Jan 2006; Michael Hanselmann <hansmi@gentoo.org>
+  ffmpeg-0.4.9_p20051216.ebuild:
+  Stable on ppc.
+
+  01 Jan 2006; Simon Stelling <blubb@gentoo.org>
+  ffmpeg-0.4.9_p20051216.ebuild:
+  stable on amd64 wrt bug 116181
+
+  31 Dec 2005; Markus Rothe <corsair@gentoo.org>
+  ffmpeg-0.4.9_p20051216.ebuild:
+  Stable on ppc64
+
+  30 Dec 2005; Mark Loeser <halcy0n@gentoo.org>
+  ffmpeg-0.4.9_p20051216.ebuild:
+  Stable on x86; bug #116181
+
+  30 Dec 2005; Gustavo Zacarias <gustavoz@gentoo.org>
+  ffmpeg-0.4.9_p20051216.ebuild:
+  Stable on sparc wrt security #116181
+
+  28 Dec 2005; Guy Martin <gmsoft@gentoo.org> ffmpeg-0.4.9_p20051216.ebuild:
+  Fix compilation of xine-lib on hppa.
+
+  26 Dec 2005; Luca Barbato <lu_zero@gentoo.org> -ffmpeg-0.4.8.ebuild,
+  -ffmpeg-0.4.9_pre1-r1.ebuild, ffmpeg-0.4.9_p20050226-r3.ebuild:
+  Cleanup
+
+  21 Dec 2005; Luis Medinas <metalgod@gentoo.org>
+  ffmpeg-0.4.9_p20051216.ebuild:
+  Add ~amd64 keyword for bug #116181.
+
+  21 Dec 2005; Markus Rothe <corsair@gentoo.org>
+  ffmpeg-0.4.9_p20051216.ebuild:
+  Added ~ppc64; bug #116181
+
+  21 Dec 2005; Mark Loeser <halcy0n@gentoo.org>
+  ffmpeg-0.4.9_p20051216.ebuild:
+  Adding ~x86; bug #116181
+
+  20 Dec 2005; Rene Nussbaumer <killerfox@gentoo.org>
+  ffmpeg-0.4.9_p20051216.ebuild:
+  Unstable on hppa. See bug #116181.
+
+  20 Dec 2005; Gustavo Zacarias <gustavoz@gentoo.org>
+  ffmpeg-0.4.9_p20051216.ebuild:
+  Keyworded ~sparc wrt #116181
+
+  17 Dec 2005; Luca Barbato <lu_zero@gentoo.org>
+  files/ffmpeg-soname-symlink.patch:
+  yet another fix part 2
+
+  17 Dec 2005; Luca Barbato <lu_zero@gentoo.org>
+  files/ffmpeg-soname-symlink.patch:
+  yet another fix
+
+  16 Dec 2005; Luca Barbato <lu_zero@gentoo.org>
+  files/ffmpeg-soname-symlink.patch:
+  Fix in the patch
+
+*ffmpeg-0.4.9_p20051216 (16 Dec 2005)
+
+  16 Dec 2005; Luca Barbato <lu_zero@gentoo.org>
+  +files/ffmpeg-soname-symlink.patch, +files/ffmpeg-unknown-options.patch,
+  +ffmpeg-0.4.9_p20051216.ebuild:
+  New snapshot
+
+  04 Dec 2005; Lina Pezzella <j4rg0n@gentoo.org>
+  ffmpeg-0.4.9_p20051120.ebuild:
+  Testing ppc-macos, Bug #113107
+
+  29 Nov 2005; Jason Wever <weeve@gentoo.org> ffmpeg-0.4.9_p20051120.ebuild:
+  Added ~sparc keyword wrt bug #113107.
+
+  21 Nov 2005; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20051120.ebuild:
+  Marked ~ppc
+
+  21 Nov 2005; Markus Rothe <corsair@gentoo.org>
+  ffmpeg-0.4.9_p20051120.ebuild:
+  Added ~ppc64 keyword; bug #113107
+
+  21 Nov 2005; Herbie Hopkins <herbs@gentoo.org>
+  ffmpeg-0.4.9_p20051120.ebuild:
+  Marked ~amd64 wrt bug #113107.
+
+  21 Nov 2005; Chris White <chriswhite@gentoo.org>
+  ffmpeg-0.4.9_p20051120.ebuild:
+  Added ~x86 keyword as per bug #113107.
+
+*ffmpeg-0.4.9_p20051120 (20 Nov 2005)
+
+  20 Nov 2005; Luca Barbato <lu_zero@gentoo.org>
+  +files/ffmpeg-configure.patch, +ffmpeg-0.4.9_p20051120.ebuild:
+  New experimental snapshot
+
+  18 Sep 2005; Bryan Østergaard <kloeri@gentoo.org>
+  ffmpeg-0.4.9_p20050906.ebuild:
+  Add ~alpha keyword.
+
+  17 Sep 2005; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20050906.ebuild:
+  ~alpha temporary dropped
+
+  17 Sep 2005; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20050906.ebuild:
+  Dep fix for xvid, thanks to Staffan Palmroos <spalmroos@gmail.com>
+
+  15 Sep 2005; Aron Griffis <agriffis@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r5.ebuild:
+  Mark 0.4.9_p20050226-r5 stable on alpha
+
+  15 Sep 2005; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r5.ebuild:
+  ieee1394 issues, fixed the stable ebuild.
+
+  12 Sep 2005; Luca Barbato <lu_zero@gentoo.org>
+  +files/ffmpeg-0.4.9_p20050906-osx.patch, ffmpeg-0.4.9_p20050906.ebuild:
+  Other fixes and dylib support
+
+  12 Sep 2005; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20050906.ebuild:
+  Fix ieee1394 issues
+
+  10 Sep 2005; Diego Pettenò <flameeyes@gentoo.org>
+  ffmpeg-0.4.9_p20050906.ebuild:
+  Comments cleanup for missing keywords. Don't install INSTALL file.
+
+  10 Sep 2005; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20050906.ebuild:
+  Fix Ogg support
+
+*ffmpeg-0.4.9_p20050906 (08 Sep 2005)
+
+  08 Sep 2005; Luca Barbato <lu_zero@gentoo.org>
+  +ffmpeg-0.4.9_p20050906.ebuild:
+  New snapshot, src_test added.
+
+  26 Aug 2005; Seemant Kulleen <seemant@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r5.ebuild:
+  moved libdc1394 from media-plugins to media-libs
+
+  24 Aug 2005; Aron Griffis <agriffis@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r5.ebuild:
+  stable on ia64
+
+  12 Jun 2005; Bryan Østergaard <kloeri@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r5.ebuild:
+  Add ~alpha keyword.
+
+  10 Jun 2005; Diego Pettenò <flameeyes@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r5.ebuild:
+  Use memalign hack on FreeBSD.
+
+  06 Jun 2005; Markus Rothe <corsair@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r5.ebuild:
+  Stable on ppc64
+
+  17 May 2005; Gustavo Zacarias <gustavoz@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r5.ebuild:
+  Stable on sparc
+
+  15 May 2005; Diego Pettenò <flameeyes@gentoo.org>
+  ffmpeg-0.4.9_pre1-r1.ebuild, ffmpeg-0.4.9_p20050226-r3.ebuild,
+  ffmpeg-0.4.9_p20050226-r5.ebuild:
+  Removed unused gcc inheriting. Inherited toolchain-funcs where needed.
+
+  09 May 2005; Aron Griffis <agriffis@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r3.ebuild, ffmpeg-0.4.9_p20050226-r5.ebuild:
+  mark 0.4.9_p20050226-r3 stable on ia64.  add ~ia64 to 0.4.9_p20050226-r5
+
+  06 May 2005; Diego Pettenò <flameeyes@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r5.ebuild:
+  Marked amd64.
+
+  06 May 2005; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r5.ebuild:
+  Marked ppc and x86
+
+  05 May 2005; Donnie Berkholz <spyderous@gentoo.org>;
+  +files/0.4.8-gcc3.4-magicF2W.patch:
+  Re-add missing patch.
+
+*ffmpeg-0.4.9_pre1-r1 (04 May 2005)
+
+  04 May 2005; Lina Pezzella <j4rg0n@gentoo.org> +files/ffmpeg-osx.patch,
+  +ffmpeg-0.4.9_pre1-r1.ebuild:
+  Re-added the pre1 ebuild since there are extensive patches made against it
+  that do not yet work with upstream's CVS. We're working on the latter, but
+  I'm not going to hold up ffmpeg any longer.
+
+  02 May 2005; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r5.ebuild:
+  Forced -fomit-frame-pointer
+
+*ffmpeg-0.4.9_p20050226-r5 (26 Apr 2005)
+
+  26 Apr 2005; Jeremy Huddleston <eradicator@gentoo.org>
+  -files/ffmpeg-0.4.7-2.6.patch, -files/0.4.8-gcc3.4-magicF2W.patch,
+  -files/alpha-idct.patch, +files/ffmpeg-a52.patch,
+  -files/ffmpeg-configure-extralibs.patch,
+  +files/ffmpeg-missing_links.patch, -ffmpeg-0.4.7.ebuild,
+  -ffmpeg-0.4.8.20040222.ebuild, -ffmpeg-0.4.8.20040322.ebuild,
+  -ffmpeg-0.4.8.20040322-r1.ebuild, -ffmpeg-0.4.9_pre1.ebuild,
+  -ffmpeg-0.4.9_p20050226-r1.ebuild, -ffmpeg-0.4.9_p20050226.ebuild,
+  -ffmpeg-0.4.9_p20050226-r2.ebuild, ffmpeg-0.4.9_p20050226-r3.ebuild,
+  -ffmpeg-0.4.9_p20050226-r4.ebuild, +ffmpeg-0.4.9_p20050226-r5.ebuild:
+  Revbump to add missing links to needed libs in libavformat. Handle a52 and
+  faad properly. Removing old versions.
+
+  25 Apr 2005; Martin Schlemmer <azarah@gentoo.org>
+  files/ffmpeg-0.4.9_p20050226-gcc4.patch, ffmpeg-0.4.9_p20050226-r4.ebuild:
+  Fixup mafteah's patch the other way around, else it breaks avifile among
+  things.
+
+  25 Apr 2005; Martin Schlemmer <azarah@gentoo.org>
+  +files/ffmpeg-0.4.9_p20050226-gcc4.patch,
+  ffmpeg-0.4.9_p20050226-r4.ebuild:
+  Fix building with gcc4. Patch from Genady Okrain (mafteah) 's overlay, with
+  some added hunks that was needed this side.
+
+  24 Apr 2005; Diego Pettenò <flameeyes@gentoo.org>
+  +files/ffmpeg-configure-extralibs.patch, ffmpeg-0.4.9_p20050226-r4.ebuild:
+  Added patch to link to libdc1394 when ieee1394 is enabled. Fixes #90150.
+
+  24 Apr 2005; Michael Hanselmann <hansmi@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r3.ebuild:
+  Stable on hppa.
+
+  24 Apr 2005; Bryan Østergaard <kloeri@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r3.ebuild:
+  Stable on alpha.
+
+  24 Apr 2005; Markus Rothe <corsair@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r3.ebuild:
+  Stable on ppc64
+
+  22 Apr 2005; Diego Pettenò <flameeyes@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r4.ebuild:
+  Fixed typo on pic/mmx if condition.
+
+*ffmpeg-0.4.9_p20050226-r4 (21 Apr 2005)
+
+  21 Apr 2005; Diego Pettenò <flameeyes@gentoo.org>
+  +ffmpeg-0.4.9_p20050226-r4.ebuild:
+  Added new revision which applies cleanups on bug #89172, and patch to enable
+  mmx on amd64 as for bug #88965.
+
+  19 Apr 2005; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r3.ebuild:
+  Marked ppc and x86
+
+  19 Apr 2005; Diego Pettenò <flameeyes@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r3.ebuild:
+  Mark ffmpeg-0.4.9_p20050226-r3 stable on amd64.
+
+  18 Apr 2005; Diego Pettenò <flameeyes@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r3.ebuild:
+  Changed xvid dependency to 1.0. Fixes #88879.
+
+*ffmpeg-0.4.9_p20050226-r3 (03 Apr 2005)
+
+  03 Apr 2005; Jan Brinkmann <luckyduck@gentoo.org>
+  +ffmpeg-0.4.9_p20050226-r3.ebuild:
+  revision bump to indicate the latest changes, fixes #87681
+
+  30 Mar 2005; Chris White <chriswhite@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r1.ebuild, ffmpeg-0.4.9_p20050226-r2.ebuild,
+  ffmpeg-0.4.9_p20050226.ebuild:
+  Fixed bug #87069.
+
+  21 Mar 2005; Jeremy Huddleston <eradicator@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r2.ebuild:
+  Use the right toolchain compiler.
+
+  20 Mar 2005; Chris White <chriswhite@gentoo.org>
+  files/gentoo-ffmpeg001.patch, ffmpeg-0.4.9_p20050226-r2.ebuild:
+  Fixed a52 linking (again).  Closes bugs #85929 and #85952.
+
+*ffmpeg-0.4.9_p20050226-r2 (19 Mar 2005)
+
+  19 Mar 2005; Chris White <chriswhite@gentoo.org>
+  +files/ffmpeg-libdir-pic.patch, files/gentoo-ffmpeg001.patch,
+  +ffmpeg-0.4.9_p20050226-r2.ebuild:
+  Fix bug #84241 (multilib and pic).
+
+  16 Mar 2005; Chris White <chriswhite@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r1.ebuild:
+  Added missing emake.
+
+  11 Mar 2005; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20050226-r1.ebuild:
+  Put back xvid support
+
+*ffmpeg-0.4.9_p20050226-r1 (10 Mar 2005)
+
+  10 Mar 2005; Chris White <chriswhite@gentoo.org>
+  +ffmpeg-0.4.9_p20050226-r1.ebuild:
+  Fixed bug #67947. -fPIC logic makes sense now. Thanks to Kevin Quin for
+  supplying the logic.
+
+  07 Mar 2005; Chris White <chriswhite@gentoo.org>
+  ffmpeg-0.4.9_p20050226.ebuild:
+  Added threading support.
+
+  03 Mar 2005; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.9_p20050226.ebuild:
+  Removed xvid useflag
+
+  01 Mar 2005; Chris White <chriswhite@gentoo.org>
+  ffmpeg-0.4.9_p20050226.ebuild:
+  Fix bad spelling.
+
+*ffmpeg-0.4.9_p20050226 (01 Mar 2005)
+
+  01 Mar 2005; Chris White <chriswhite@gentoo.org>
+  +files/gentoo-ffmpeg001.patch, +ffmpeg-0.4.9_p20050226.ebuild:
+  Bumped to 20050226 snapshot release.
+
+  24 Feb 2005; Jan Brinkmann <luckyduck@gentoo.org> ffmpeg-0.4.8.ebuild:
+  fixed #82873, emerge failed due to missing patch.
+
+  06 Feb 2005; Jan Brinkmann <luckyduck@gentoo.org> ffmpeg-0.4.7.ebuild,
+  ffmpeg-0.4.8.20040222.ebuild, ffmpeg-0.4.8.20040322-r1.ebuild,
+  ffmpeg-0.4.8.20040322.ebuild, ffmpeg-0.4.8.ebuild, ffmpeg-0.4.9_pre1.ebuild:
+  added dummy src_test() to fix #77212
+
+  16 Dec 2004; Markus Rothe <corsair@gentoo.org> ffmpeg-0.4.9_pre1.ebuild:
+  Stable on ppc64
+
+  14 Dec 2004; Jeremy Huddleston <eradicator@gentoo.org>
+  ffmpeg-0.4.7.ebuild, ffmpeg-0.4.8.20040222.ebuild,
+  ffmpeg-0.4.8.20040322-r1.ebuild, ffmpeg-0.4.8.20040322.ebuild,
+  ffmpeg-0.4.8.ebuild, ffmpeg-0.4.9_pre1.ebuild:
+  Using 'aac' instead of faad and faac.
+
+  02 Nov 2004; Markus Rothe <corsair@gentoo.org> ffmpeg-0.4.9_pre1.ebuild:
+  Marked ~ppc64
+
+  25 Aug 2004; Sven Wegener <swegener@gentoo.org> ffmpeg-0.4.9_pre1.ebuild:
+  Changed SRC_URI to use mirror:// syntax.
+
+*ffmpeg-0.4.9_pre1 (21 Jul 2004)
+
+  21 Jul 2004; Jeremy Huddleston <eradicator@gentoo.org>
+  +ffmpeg-0.4.9_pre1.ebuild:
+  Version bump closes bug #56613.
+
+  25 Jun 2004; Daniel Goller <morfic@gentoo.org> ffmpeg-0.4.8.20040322-r1.ebuild:
+    ffmpeg-0.4.8.20040322-r1.ebuild adding patch for gcc 3.4.x 
+    closes 49383 Credit goes to Ed Catmur for the patch
+
+  24 Jun 2004; Martin Holzer <mholzer@gentoo.org>
+  ffmpeg-0.4.8.20040322-r1.ebuild, ffmpeg-0.4.8.20040322.ebuild:
+  removing symblink from ffplay to ffmpeg. closes 51014
+
+  09 Jun 2004; Luca Barbato <lu_zero@gentoo.org>
+  ffmpeg-0.4.8.20040322-r1.ebuild:
+  Same fix from amd64
+
+  09 Jun 2004; Thomas Raschbacher <lordvan@gentoo.org>
+  ffmpeg-0.4.8.20040322-r1.ebuild:
+  only enable faac when we are not on ia64 or alpha in src_compile
+
+*ffmpeg-0.4.8.20040322-r1 (09 Jun 2004)
+
+  09 Jun 2004; Thomas Raschbacher <lordvan@gentoo.org>
+  +ffmpeg-0.4.8.20040322-r1.ebuild:
+  -r1 adds faac support, added local USE flag
+  fixes bug #48090, thanks to Stefan Briesenick <sbriesen@gmx.de>
+
+  07 Jun 2004; Travis Tilley <lv@gentoo.org> ffmpeg-0.4.8.20040322.ebuild:
+  stable on amd64
+
+  05 Jun 2004; <tester@gentoo.org> ffmpeg-0.4.8.20040322.ebuild:
+  No longer available from SRC_URI, changed the keywords to -*.. oops
+  we have it on our mirrors... sorry..
+
+  30 Mar 2004; Sven Blumenstein <bazik@gentoo.org>
+  ffmpeg-0.4.8.20040322.ebuild:
+  Stable on sparc.
+
+  26 Mar 2004; Joel Martin <kanaka@gentoo.org> ffmpeg-0.4.8.20040222.ebuild,
+  ffmpeg-0.4.8.20040322.ebuild:
+  Filter out -momit-leaf-frame-pointer flag from libpostproc compilation
+
+*ffmpeg-0.4.8.20040322 (22 Mar 2004)
+
+  22 Mar 2004; Luca Barbato <lu_zero@gentoo.org> ffmpeg-0.4.8.20040322.ebuild:
+  New snapshot, works with ppc
+
+  15 Mar 2004; Luca Barbato <lu_zero@gentoo.org> ffmpeg-0.4.8.20040222.ebuild:
+  Marked -ppc pending a fix for misbuild.
+
+*ffmpeg-0.4.8.20040222 (11 Mar 2004)
+
+  11 Mar 2004; <kanaka@gentoo.org> ffmpeg-0.4.8.20040222.ebuild:
+  Install libpostproc library. Update to newer snapshot. Bug 27051
+
+  17 Feb 2004; Aron Griffis <agriffis@gentoo.org> ffmpeg-0.4.8.ebuild:
+  stable on alpha and ia64
+
+  16 Jan 2004; Bartosch Pixa <darkspecter@gentoo.org> ffmpeg-0.4.8.ebuild:
+  set ppc in keywords
+
+  29 Nov 2003; Brad House <brad_mssw@gentoo.org> ffmpeg-0.4.8.ebuild:
+  amd64 needs -fPIC to compile properly
+
+  22 Nov 2003; Luca Barbato <lu_zero@gentoo.org> ffmpeg-0.4.8.ebuild:
+  altivec related fix, thanks to Olivier <ocastan@noos.fr> for the report
+  and the patch
+
+*ffmpeg-0.4.8 (18 Oct 2003)
+
+  18 Oct 2003; <iggy@gentoo.org> ffmpeg-0.4.8.ebuild:
+  version bump, fix a bug where it was trying to include X11/Xlib.h for no
+  apparent reason
+
+*ffmpeg-0.4.7 (09 Sep 2003)
+
+  09 Sep 2003; Martin Holzer <mholzer@gentoo.org> ffmpeg-0.4.7.ebuild:
+  Version bumped.
+
+  14 Jul 2003; Alastair Tse <liquidx@gentoo.org>
+  ffmpeg-0.4.7_pre20030624.ebuild:
+  depend on a non-masked version of imlib2
+
+  13 Jul 2003; Nick Hadaway <raker@gentoo.org> ffmpeg-0.4.7_pre20030624.ebuild:
+  Adding sdl, imlib, and truetype use flags.
+
+  06 Jul 2003; Nick Hadaway <raker@gentoo.org> ffmpeg-0.4.7_pre20030624.ebuild:
+  Marked stble for x86
+
+  02 Jul 2003; Nick Hadaway <raker@gentoo.org> ffmpeg-0.4.7_pre20030624.ebuild,
+  files/alpha-idct.patch:
+  Addresses bug #23563.  This problem was fixed upstream in cvs.
+  Added a small patch.  
+
+*ffmpeg-0.4.7_pre20030624 (26 Jun 2003)
+
+  26 Jun 2003; Nick Hadaway <raker@gentoo.org> ffmpeg-0.4.7_pre20030624.ebuild:
+  As of June 9th, 0.4.6 is considered obsolete by the upstream author.
+  This build pulled from a cvs snapshot now has support for the dvd
+  and static use variables.  Also added faad as a local use variable.
+
+*ffmpeg-0.4.6-r1 (4 Jan 2003)
+
+  10 Mar 2003; Will Woods <wwoods@gentoo.org> ffmpeg-0.4.6-r1.ebuild:
+  Added ~alpha to keywords and -fPIC to flags, fixes bug #16281
+
+  19 Feb 2003; Martin Holzer <mholzer@gentoo.org> ffmpeg-0.4.6-r1.ebuild :
+  Marked as stable.
+ 
+  26 Jan 2003; Seemant Kulleen <seemant@gentoo.org> ffmpeg-0.4.6-r1.ebuild :
+
+  Added -fPIC to the filter-flags to sort out compiling issues such as
+  those reported in bug #14500 by atomicdog@akier.net (Conrad Akier)
+ 
+  4 Jan 2003; Joshua Brindle <method@gentoo.org> ffmpeg-0.4.6-r1.ebuild :
+  merge patches to ebuild and package from Rene Wagner <reenoo@gmx.de>
+  to install libavcodec as a .so. 
+
+*ffmpeg-0.4.6 (3 Jan 2003)
+  
+  3 Jan 2003; Joshua Brindle <method@gentoo.org> ffmpeg-0.4.6.ebuild :
+  version bump, this is a release non-snapshot version, recommend upgrade
+
+*ffmpeg-0.4.5.20021212 (13 Dec 2002)
+
+  5 Jan 2002; Joshua Brindle <method@gentoo.org ffmpeg-0.4.5.20021212.ebuild :
+  removed this ebuild from portage, it is old and was a cvs snapshot, 0.4.6 
+  contains everything this did and is a release version
+
+  14 Dec 2002; Joshua Brindle <method@gentoo.org> ffmpeg-0.4.5.20021212.ebuild :
+  Changing url to reflect where author moved his.
+  
+  13 Dec 2002; Bryon Roche <kain@gentoo.org> ffmpeg-0.4.5.20021212.ebuild :
+  ffmpeg isn't REALLY x86 specific.  Giving ppc and sparc a nice big ~.
+  
+  12 Dec 2002; Joshua Brindle <method@gentoo.org> ffmpeg-0.4.5.20021212.ebuild :
+  cvs snapshot so that streaming will work, also added mp3lame and oggvorbis
+  support with encode and oggvorbis use vars.
+
+  06 Dec 2002; Rodney Rees <manson@gentoo.org> : changed sparc ~sparc keywords
+ 
+*ffmpeg-0.4.5 (1 Feb 2002)
+
+  25 Sep 2002; Daniel Ahlberg <aliz@gentoo.org> ffmpeg-0.4.5.ebuild :
+  Filter out "-fforce-addr".
+
+  05 Sep 2002; Seemant Kulleen <seemant@gentoo.org> ffmpeg-0.4.5.ebuild :
+
+  Added nasm to DEPEND. Thanks to: jfelice@cronosys.com (Jay 'Eraserhead'
+  Felice) in bug #7481.  Also tagged it x86 only due to this.
+
+  1 Feb 2002; G.Bevin <gbevin@gentoo.org> ChangeLog :
+  
+  Added initial ChangeLog which should be updated whenever the package is
+  updated in any way. This changelog is targetted to users. This means that the
+  comments should well explained and written in clean English. The details about
+  writing correct changelogs are explained in the skel.ChangeLog file which you
+  can find in the root directory of the portage repository.
diff --git a/media-video/ffmpeg/Manifest b/media-video/ffmpeg/Manifest
new file mode 100644
index 0000000..75afea2
--- /dev/null
+++ b/media-video/ffmpeg/Manifest
@@ -0,0 +1,5 @@
+AUX ffmpeg-loongson.patch 55657 RMD160 ec3487a22679200b1aae60700d6453387a8fc71f SHA1 c17afd72e1a211a9ec33acdfbaf6ae08f6050437 SHA256 c115f49a271f7f69923527af70311b76d05c12ce6659bd4cdaaae6ba0da73157
+DIST ffmpeg-0.5_p22846.tar.bz2 3675057 RMD160 3ebd997321d59fa0baf87329518e925ccbed3cf8 SHA1 d207570ed009b63c94485ccc64439107cf700f39 SHA256 a52d82476afa7c8ed62c1a52760e49907947c1ef1ecfc3e9022fa713e9783f2f
+EBUILD ffmpeg-0.5_p22846.ebuild 7658 RMD160 788bfaddfb9c1d883393b6d46bb05b6d54ec8e47 SHA1 e21405b2a6002cab5249d5516ff87bd17feec615 SHA256 a9caa8701e3383c454378580d1a9bf0612484c75335fadc5bd18039ddf31d323
+MISC ChangeLog 52729 RMD160 855d3b5b6c5c08956fbec01454c41e276c89fd08 SHA1 a75e9b54685d91b944bda6174839f75efd53117b SHA256 6212eddd7cc186cca8bf1f7fcc489f59435fc9f4ee373947d89f7d02be9b11db
+MISC metadata.xml 1830 RMD160 88d52f0a9539230e1addcab5fd704400d34b84d3 SHA1 9e2bdcf8205744623d964f1e426dadacbf24b0e1 SHA256 eb541891f7dab5e5ed4b46975a578d4ff19b88d9d80e44668f2195774fb157c8
diff --git a/media-video/ffmpeg/ffmpeg-0.5_p22846.ebuild b/media-video/ffmpeg/ffmpeg-0.5_p22846.ebuild
new file mode 100644
index 0000000..9697f47
--- /dev/null
+++ b/media-video/ffmpeg/ffmpeg-0.5_p22846.ebuild
@@ -0,0 +1,249 @@
+# Copyright 1999-2010 Gentoo Foundation
+# Distributed under the terms of the GNU General Public License v2
+# $Header: /var/cvsroot/gentoo-x86/media-video/ffmpeg/ffmpeg-0.5_p22846.ebuild,v 1.3 2010/05/11 09:50:42 aballier Exp $
+
+EAPI=2
+SCM=""
+if [ "${PV#9999}" != "${PV}" ] ; then
+	SCM=subversion
+	ESVN_REPO_URI="svn://svn.ffmpeg.org/ffmpeg/trunk"
+fi
+
+inherit eutils flag-o-matic multilib toolchain-funcs ${SCM}
+
+DESCRIPTION="Complete solution to record, convert and stream audio and video. Includes libavcodec."
+HOMEPAGE="http://ffmpeg.org/"
+if [ "${PV#9999}" != "${PV}" ] ; then
+	SRC_URI=""
+elif [ "${PV%_p*}" != "${PV}" ] ; then # Snapshot
+	SRC_URI="mirror://gentoo/${P}.tar.bz2"
+else # Release
+	SRC_URI="http://ffmpeg.org/releases/${P}.tar.bz2"
+fi
+FFMPEG_REVISION="${PV#*_p}"
+
+LICENSE="GPL-3"
+SLOT="0"
+KEYWORDS="~alpha ~amd64 ~arm ~hppa ~ia64 ~ppc ~ppc64 ~sparc ~x86 ~x86-fbsd"
+IUSE="+3dnow +3dnowext alsa altivec cpudetection custom-cflags debug dirac
+	  doc ieee1394 +encode faac faad gsm jack +mmx +mmxext vorbis test
+	  theora threads x264 xvid network zlib sdl X mp3 amr
+	  oss pic schroedinger +hardcoded-tables bindist v4l v4l2
+	  speex +ssse3 jpeg2k vdpau"
+
+VIDEO_CARDS="nvidia"
+
+for x in ${VIDEO_CARDS}; do
+	IUSE="${IUSE} video_cards_${x}"
+done
+
+RDEPEND="sdl? ( >=media-libs/libsdl-1.2.10 )
+	alsa? ( media-libs/alsa-lib )
+	encode? (
+		faac? ( media-libs/faac )
+		mp3? ( media-sound/lame )
+		vorbis? ( media-libs/libvorbis media-libs/libogg )
+		theora? ( >=media-libs/libtheora-1.1.1[encode] media-libs/libogg )
+		x264? ( >=media-libs/x264-0.0.20100118 )
+		xvid? ( >=media-libs/xvid-1.1.0 ) )
+	faad? ( >=media-libs/faad2-2.6.1 )
+	zlib? ( sys-libs/zlib )
+	ieee1394? ( media-libs/libdc1394
+				sys-libs/libraw1394 )
+	dirac? ( media-video/dirac )
+	gsm? ( >=media-sound/gsm-1.0.12-r1 )
+	jpeg2k? ( >=media-libs/openjpeg-1.3-r2 )
+	amr? ( media-libs/opencore-amr )
+	schroedinger? ( media-libs/schroedinger )
+	speex? ( >=media-libs/speex-1.2_beta3 )
+	jack? ( media-sound/jack-audio-connection-kit )
+	X? ( x11-libs/libX11 x11-libs/libXext )
+	video_cards_nvidia? (
+		vdpau? ( x11-libs/libvdpau )
+	)"
+
+DEPEND="${RDEPEND}
+	>=sys-devel/make-3.81
+	dirac? ( dev-util/pkgconfig )
+	schroedinger? ( dev-util/pkgconfig )
+	mmx? ( dev-lang/yasm )
+	doc? ( app-text/texi2html )
+	test? ( net-misc/wget )
+	v4l? ( sys-kernel/linux-headers )
+	v4l2? ( sys-kernel/linux-headers )"
+
+src_prepare() {
+	if [[ ${PV} = *9999* ]]; then
+		# Set SVN version manually
+		subversion_wc_info
+		sed -i s/UNKNOWN/SVN-r${ESVN_WC_REVISION}/ "${S}/version.sh"
+	elif [ "${PV%_p*}" != "${PV}" ] ; then # Snapshot
+		sed -i s/UNKNOWN/SVN-r${FFMPEG_REVISION}/ "${S}/version.sh"
+	fi
+	cd ${S}
+	epatch ${FILESDIR}/${PN}-loongson.patch
+}
+
+src_configure() {
+	local myconf="${EXTRA_FFMPEG_CONF}"
+
+	# enabled by default
+	use debug || myconf="${myconf} --disable-debug"
+	use zlib || myconf="${myconf} --disable-zlib"
+	use sdl || myconf="${myconf} --disable-ffplay"
+	use network || myconf="${myconf} --disable-network"
+
+	use custom-cflags && myconf="${myconf} --disable-optimizations"
+	use cpudetection && myconf="${myconf} --enable-runtime-cpudetect"
+
+	# enabled by default
+	if use encode
+	then
+		use mp3 && myconf="${myconf} --enable-libmp3lame"
+		use vorbis && myconf="${myconf} --enable-libvorbis"
+		use theora && myconf="${myconf} --enable-libtheora"
+		use x264 && myconf="${myconf} --enable-libx264"
+		use xvid && myconf="${myconf} --enable-libxvid"
+		if use bindist
+		then
+			use faac && ewarn "faac is nonfree and cannot be distributed;
+			disabling faac support."
+		else
+			use faac && myconf="${myconf} --enable-libfaac --enable-nonfree"
+		fi
+	else
+		myconf="${myconf} --disable-encoders"
+	fi
+
+	# libavdevice options
+	use ieee1394 && myconf="${myconf} --enable-libdc1394"
+	# Indevs
+	for i in v4l v4l2 alsa oss jack ; do
+		use $i || myconf="${myconf} --disable-indev=$i"
+	done
+	# Outdevs
+	for i in alsa oss ; do
+		use $i || myconf="${myconf} --disable-outdev=$i"
+	done
+	use X && myconf="${myconf} --enable-x11grab"
+
+	# Threads; we only support pthread for now but ffmpeg supports more
+	use threads && myconf="${myconf} --enable-pthreads"
+
+	# Decoders
+	use amr && myconf="${myconf} --enable-libopencore-amrwb
+		--enable-libopencore-amrnb"
+	for i in gsm faad dirac schroedinger speex; do
+		use $i && myconf="${myconf} --enable-lib$i"
+	done
+	use jpeg2k && myconf="${myconf} --enable-libopenjpeg"
+
+	#for i in h264_vdpau mpeg1_vdpau mpeg_vdpau vc1_vdpau wmv3_vdpau; do
+	#	use video_cards_nvidia || myconf="${myconf} --disable-decoder=$i"
+	#	use vdpau || myconf="${myconf} --disable-decoder=$i"
+	#done
+	use video_cards_nvidia || myconf="${myconf} --disable-vdpau"
+	use vdpau || myconf="${myconf} --disable-vdpau"
+	myconf="${myconf} --disable-vaapi"
+
+	# CPU features
+	for i in mmx ssse3 altivec ; do
+		use $i ||  myconf="${myconf} --disable-$i"
+	done
+	use mmxext || myconf="${myconf} --disable-mmx2"
+	use 3dnow || myconf="${myconf} --disable-amd3dnow"
+	use 3dnowext || myconf="${myconf} --disable-amd3dnowext"
+	# disable mmx accelerated code if PIC is required
+	# as the provided asm decidedly is not PIC.
+	if gcc-specs-pie ; then
+		myconf="${myconf} --disable-mmx --disable-mmx2"
+	fi
+
+	# Option to force building pic
+	use pic && myconf="${myconf} --enable-pic"
+
+	# Try to get cpu type based on CFLAGS.
+	# Bug #172723
+	# We need to do this so that features of that CPU will be better used
+	# If they contain an unknown CPU it will not hurt since ffmpeg's configure
+	# will just ignore it.
+	for i in $(get-flag march) $(get-flag mcpu) $(get-flag mtune) ; do
+		[ "${i}" = "native" ] && i="host" # bug #273421
+		[[ ${i} = *-sse3 ]] && i="${i%-sse3}" # bug 283968
+		myconf="${myconf} --cpu=$i"
+		break
+	done
+
+	# Mandatory configuration
+	myconf="${myconf} --enable-gpl --enable-version3 --enable-postproc \
+			--enable-avfilter --enable-avfilter-lavf \
+			--disable-stripping"
+
+	# cross compile support
+	if tc-is-cross-compiler ; then
+		myconf="${myconf} --enable-cross-compile --arch=$(tc-arch-kernel) --cross-prefix=${CHOST}-"
+		case ${CHOST} in
+			*freebsd*)
+				myconf="${myconf} --target-os=freebsd"
+				;;
+			mingw32*)
+				myconf="${myconf} --target-os=mingw32"
+				;;
+			*linux*)
+				myconf="${myconf} --target-os=linux"
+				;;
+		esac
+	fi
+
+	# Misc stuff
+	use hardcoded-tables && myconf="${myconf} --enable-hardcoded-tables"
+	use doc || myconf="${myconf} --disable-doc"
+
+	# Specific workarounds for too-few-registers arch...
+	if [[ $(tc-arch) == "x86" ]]; then
+		filter-flags -fforce-addr -momit-leaf-frame-pointer
+		append-flags -fomit-frame-pointer
+		is-flag -O? || append-flags -O2
+		if (use debug); then
+			# no need to warn about debug if not using debug flag
+			ewarn ""
+			ewarn "Debug information will be almost useless as the frame pointer is omitted."
+			ewarn "This makes debugging harder, so crashes that has no fixed behavior are"
+			ewarn "difficult to fix. Please have that in mind."
+			ewarn ""
+		fi
+	fi
+
+	cd "${S}"
+	./configure \
+		--prefix=/usr \
+		--libdir=/usr/$(get_libdir) \
+		--shlibdir=/usr/$(get_libdir) \
+		--mandir=/usr/share/man \
+		--enable-static --enable-shared \
+		--cc="$(tc-getCC)" \
+		${myconf} || die "configure failed"
+}
+
+src_compile() {
+	emake version.h || die #252269
+	emake || die "make failed"
+}
+
+src_install() {
+	emake DESTDIR="${D}" install install-man || die "Install Failed"
+
+	dodoc Changelog README INSTALL
+	dodoc doc/*
+}
+
+src_test() {
+	if use encode ; then
+		for t in codectest lavftest seektest ; do
+			LD_LIBRARY_PATH="${S}/libpostproc:${S}/libswscale:${S}/libavcodec:${S}/libavdevice:${S}/libavfilter:${S}/libavformat:${S}/libavutil" \
+				emake ${t} || die "Some tests in ${t} failed"
+		done
+	else
+		ewarn "Tests fail without USE=encode, skipping"
+	fi
+}
diff --git a/media-video/ffmpeg/files/ffmpeg-loongson.patch b/media-video/ffmpeg/files/ffmpeg-loongson.patch
new file mode 100644
index 0000000..1bc4146
--- /dev/null
+++ b/media-video/ffmpeg/files/ffmpeg-loongson.patch
@@ -0,0 +1,1794 @@
+diff --git a/configure b/configure
+index 25e8cef..1d6c652 100755
+--- a/configure
++++ b/configure
+@@ -230,6 +230,7 @@ Advanced options (experts only):
+   --disable-armvfp         disable ARM VFP optimizations
+   --disable-iwmmxt         disable iwmmxt optimizations
+   --disable-mmi            disable MMI optimizations
++  --disable-loongson2mmi   disable LOONGSON2 Multi-Media Instructions usage"
+   --disable-neon           disable neon optimizations
+   --disable-vis            disable VIS optimizations
+   --disable-yasm           disable use of yasm assembler
+@@ -995,6 +996,7 @@ ARCH_EXT_LIST='
+     armvfp
+     iwmmxt
+     mmi
++    loongson2mmi
+     mmx
+     mmx2
+     neon
+@@ -2862,6 +2864,7 @@ if enabled arm; then
+ fi
+ if enabled mips; then
+     echo "MMI enabled               ${mmi-no}"
++    echo "LOONGSON2MMI enabled               ${loongson2mmi-no}"
+ fi
+ if enabled ppc; then
+     echo "AltiVec enabled           ${altivec-no}"
+diff --git a/libavcodec/avcodec.h b/libavcodec/avcodec.h
+index add4b10..8244e51 100644
+--- a/libavcodec/avcodec.h
++++ b/libavcodec/avcodec.h
+@@ -1586,6 +1586,8 @@ typedef struct AVCodecContext {
+ #define FF_IDCT_SIMPLENEON    22
+ #define FF_IDCT_SIMPLEALPHA   23
+ #define FF_IDCT_BINK          24
++#define FF_IDCT_LIBMPEG2LOONGSON2	25
++#define FF_IDCT_XVIDLOONGSON2 26
+ 
+     /**
+      * slice count
+diff --git a/libavcodec/dsputil.c b/libavcodec/dsputil.c
+index bbfdb6a..dfc3452 100644
+--- a/libavcodec/dsputil.c
++++ b/libavcodec/dsputil.c
+@@ -4525,6 +4525,7 @@ av_cold void dsputil_init(DSPContext* c, AVCodecContext *avctx)
+     if (HAVE_MMI)        dsputil_init_mmi   (c, avctx);
+     if (ARCH_SH4)        dsputil_init_sh4   (c, avctx);
+     if (ARCH_BFIN)       dsputil_init_bfin  (c, avctx);
++    if (HAVE_LOONGSON2MMI)	dsputil_init_loongson2(c, avctx);
+ 
+     for(i=0; i<64; i++){
+         if(!c->put_2tap_qpel_pixels_tab[0][i])
+diff --git a/libavcodec/dsputil.h b/libavcodec/dsputil.h
+index d1816e6..1a72ae9 100644
+--- a/libavcodec/dsputil.h
++++ b/libavcodec/dsputil.h
+@@ -636,6 +636,7 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx);
+ void dsputil_init_ppc(DSPContext* c, AVCodecContext *avctx);
+ void dsputil_init_sh4(DSPContext* c, AVCodecContext *avctx);
+ void dsputil_init_vis(DSPContext* c, AVCodecContext *avctx);
++void dsputil_init_loongson2(DSPContext* c, AVCodecContext *avctx);
+ 
+ void ff_dsputil_init_dwt(DSPContext *c);
+ void ff_cavsdsp_init(DSPContext* c, AVCodecContext *avctx);
+diff --git a/libavcodec/loongson2/dsputil_loongson2.c b/libavcodec/loongson2/dsputil_loongson2.c
+new file mode 100644
+index 0000000..01bd3ac
+--- /dev/null
++++ b/libavcodec/loongson2/dsputil_loongson2.c
+@@ -0,0 +1,221 @@
++/*
++ *  Copyright(C) 2006-2010 comcat <jiankemeng@gmail.com>
++ *
++ *  Optimized for Loongson2 CPUs by comcat <jiankemeng@gmail.com>
++ * 
++ */
++
++#include "dsputil_loongson2.h"
++#include "../simple_idct.h"
++#include "../mpegvideo.h"
++
++//extern void ff_idct_xvid_loongson2(short *block);
++
++extern void ff_loongson2_idct(DCTELEM *block);
++extern void ff_idct_xvid_loongson2(short *block);
++
++static void add_pixels_clamped_loongson2(const DCTELEM *block, uint8_t *restrict pixels, int line_size)
++{
++    const DCTELEM *p;
++    uint8_t *pix;
++    int i,j;
++    p = block;
++    pix = pixels;
++    i = 4;
++    j = line_size << 1;
++	__asm __volatile("xor  $f14, $f14, $f14\n\t");
++	do {
++		__asm __volatile(
++//				".set mips3					\n\t"
++				"ldc1   $f0, 0(%2)			\n\t"
++				"ldc1   $f2, 8(%2)			\n\t"
++				"ldc1   $f4, 16(%2)			\n\t"
++				"ldc1   $f6, 24(%2)			\n\t"
++				"ldc1   $f8, %0				\n\t"
++				"ldc1   $f12, %1			\n\t"
++				"mov.d  $f10, $f8			\n\t"
++
++				"punpcklbh  $f8, $f8, $f14 	\n\t"
++				"punpckhbh  $f10, $f10, $f14\n\t"
++
++				"paddsh $f0, $f0, $f8		\n\t"
++				"paddsh $f2, $f2, $f10		\n\t"
++
++				"mov.d  $f10, $f12			\n\t"
++
++				"punpcklbh  $f12, $f12, $f14\n\t"
++				"punpckhbh  $f10, $f10, $f14\n\t"
++
++				"paddsh $f4, $f4, $f12		\n\t"
++				"paddsh $f6, $f6, $f10		\n\t"
++
++				"packushb   $f0, $f0, $f2	\n\t"
++				"packushb   $f4, $f4, $f6	\n\t"
++
++				"sdc1   $f0, %0				\n\t"
++				"sdc1   $f4, %1				\n\t"
++//				".set mips2					\n\t"
++				:"+m"(*pix), "+m"(*(pix+line_size))
++				:"r"(p)
++				:"$f0","$f2","$f4","$f6","$f8","$f10","$f12","$f14","memory");
++		pix += j;
++		p += 16;
++	} while (--i);
++
++}
++
++static void put_pixels_clamped_loongson2(const DCTELEM *block, uint8_t *restrict pixels, int line_size)
++{
++    const DCTELEM *p;
++    uint8_t *pix;
++    int tmp = line_size * 3;
++    p = block;
++    pix = pixels;
++	__asm __volatile
++	   (
++//		   ".set mips3 \n\t"
++		   //"dadd	$12, $0, $0\n\t"
++		   //"dadd	$13, $0, $0\n\t"
++		   //"dadd	$14, $0, $0\n\t"
++
++		   "ldc1    $f0, 0(%3)\n\t"
++		   "ldc1    $f2, 8(%3)\n\t"
++		   "ldc1    $f4, 16(%3)\n\t"
++		   "ldc1    $f6, 24(%3)\n\t"
++		   "ldc1    $f8, 32(%3)\n\t"
++		   "ldc1    $f10, 40(%3)\n\t"
++		   "ldc1    $f16, 48(%3)\n\t"
++		   "ldc1    $f18, 56(%3)\n\t"
++
++		   "packushb    $f0, $f0, $f2\n\t"
++		   "packushb    $f4, $f4, $f6\n\t"
++		   "packushb    $f8, $f8, $f10\n\t"
++		   "packushb    $f16, $f16, $f18\n\t"
++
++		   "add    $12, %0, %1\n\t"
++		   "add    $13, $12, %1\n\t"
++		   "add    $14, %0, %2\n\t"
++
++		   "sdc1    $f0, 0(%0)\n\t"
++		   "sdc1    $f4, 0($12)\n\t"
++		   "sdc1    $f8, 0($13)\n\t"
++		   "sdc1    $f16, 0($14)\n\t"
++//		   ".set mips2\n\t"
++		   :
++		   :"r" (pix), "r" (line_size), "r" (tmp), "r"(p)
++		   :"$12","$13","$14","$f0","$f2","$f4","$f6","$f8","$f10","$16","$18"
++		);
++
++		pix += line_size*4;
++		p += 32;
++
++	__asm __volatile
++		(
++//			".set mips3 \n\t"
++
++		    "dadd	$12, $0, $0\n\t"
++		    "dadd	$13, $0, $0\n\t"
++		    "dadd	$14, $0, $0\n\t"
++			"lw      $12, %3\n\t"
++
++			"ldc1    $f0, 0($12)\n\t"
++			"ldc1    $f2, 8($12)\n\t"
++			"ldc1    $f4, 16($12)\n\t"
++			"ldc1    $f6, 24($12)\n\t"
++			"ldc1    $f8, 32($12)\n\t"
++			"ldc1    $f10, 40($12)\n\t"
++			"ldc1    $f16, 48($12)\n\t"
++			"ldc1    $f18, 56($12)\n\t"
++
++			"packushb        $f0, $f0, $f2\n\t"
++			"packushb        $f4, $f4, $f6\n\t"
++			"packushb        $f8, $f8, $f10\n\t"
++			"packushb        $f16, $f16, $f18\n\t"
++
++			"add    $12, %1, %0\n\t"
++			"add    $13, $12, %1\n\t"
++			"add    $15, %2, %0\n\t"
++
++			"sdc1    $f0, 0(%0)\n\t"
++			"sdc1    $f4, 0($12)\n\t"
++
++			"sdc1    $f8, 0($13)\n\t"
++			"sdc1    $f16, 0($15)\n\t"
++//			".set mips2\n\t"
++			:
++			:"r" (pix), "r" (line_size), "r" (tmp), "m"(p)
++			:"$12","$13","$15","$f0","$f2","$f4","$f6","$f8","$f10","$16","$18","memory"
++		);
++
++}
++
++/*
++void put_signed_pixels_clamped_loongson2(const DCTELEM *block, uint8_t *pixels, int line_size)
++{
++
++}
++
++
++void ff_loongson2_idct_put(uint8_t *dest, int line_size, DCTELEM *block)
++{
++	ff_loongson2_idct(block);
++	put_pixels_clamped_loongson2(block, dest, line_size);
++}
++
++void ff_loongson2_idct_add(uint8_t *dest, int line_size, DCTELEM *block)
++{
++	ff_loongson2_idct(block);
++	add_pixels_clamped_loongson2(block, dest, line_size);	
++}*/
++
++static void ff_idct_xvid_loongson2_put(uint8_t *dest, int line_size, DCTELEM *block)
++{
++	ff_idct_xvid_loongson2(block);	
++	put_pixels_clamped_loongson2(block, dest, line_size);
++}
++
++static void ff_idct_xvid_loongson2_add(uint8_t *dest, int line_size, DCTELEM *block)
++{
++	ff_idct_xvid_loongson2(block);	
++	add_pixels_clamped_loongson2(block, dest, line_size);	
++}
++
++void dsputil_init_loongson2(DSPContext *c, AVCodecContext *avctx)
++{
++
++	const int idct_algo = avctx->idct_algo;
++
++/*
++#ifdef CONFIG_ENCODERS
++	const int dct_algo = avctx->dct_algo;
++	if(dct_algo == FF_DCT_AUTO || dct_algo == FF_DCT_LOONGSON2)
++		c->fdct = ff_fdct_loongson2;
++#endif
++*/
++
++#if 0
++	if(avctx->lowres==0)
++	{
++		if(idct_algo == FF_IDCT_LIBMPEG2LOONGSON2)
++		{
++			c->idct_add = ff_loongson2_idct_add;
++			c->idct_put = ff_loongson2_idct_put;
++			c->idct = ff_loongson2_idct;
++		}
++		else if(idct_algo == FF_IDCT_XVIDLOONGSON2)
++		{
++#endif
++			c->idct_add = ff_idct_xvid_loongson2_add;
++			c->idct_put = ff_idct_xvid_loongson2_put;
++			c->idct = ff_idct_xvid_loongson2;
++		//}
++	//}
++
++	c->put_pixels_clamped = put_pixels_clamped_loongson2;
++	c->add_pixels_clamped = add_pixels_clamped_loongson2;
++
++#ifdef	CONFIG_ENCODERS
++	dsputil_init_pix_loongson2(c, avctx);
++#endif
++
++}
+diff --git a/libavcodec/loongson2/dsputil_loongson2.d b/libavcodec/loongson2/dsputil_loongson2.d
+new file mode 100644
+index 0000000..808f0a3
+--- /dev/null
++++ b/libavcodec/loongson2/dsputil_loongson2.d
+@@ -0,0 +1,18 @@
++libavcodec/loongson2/dsputil_loongson2.o: \
++ libavcodec/loongson2/dsputil_loongson2.c \
++ libavcodec/loongson2/dsputil_loongson2.h libavcodec/dsputil.h \
++ libavutil/intreadwrite.h config.h libavutil/bswap.h \
++ libavutil/attributes.h libavutil/common.h libavutil/intmath.h \
++ libavutil/mem.h libavutil/internal.h libavutil/timer.h libavutil/libm.h \
++ libavutil/mips/intreadwrite.h libavcodec/avcodec.h libavutil/avutil.h \
++ libavutil/error.h libavutil/avutil.h libavutil/mathematics.h \
++ libavutil/rational.h libavutil/intfloat_readwrite.h libavutil/log.h \
++ libavutil/pixfmt.h libavutil/avconfig.h \
++ libavcodec/loongson2/../simple_idct.h libavcodec/loongson2/../dsputil.h \
++ libavcodec/loongson2/../mpegvideo.h libavcodec/loongson2/../get_bits.h \
++ libavutil/bswap.h libavutil/common.h libavutil/log.h \
++ libavcodec/loongson2/../mathops.h libavcodec/loongson2/../mips/mathops.h \
++ libavcodec/loongson2/../put_bits.h libavcodec/loongson2/../ratecontrol.h \
++ libavcodec/loongson2/../eval.h libavcodec/loongson2/../parser.h \
++ libavcodec/loongson2/../avcodec.h libavcodec/loongson2/../mpeg12data.h \
++ libavutil/rational.h libavcodec/loongson2/../rl.h
+diff --git a/libavcodec/loongson2/dsputil_loongson2.h b/libavcodec/loongson2/dsputil_loongson2.h
+new file mode 100644
+index 0000000..87c7bd9
+--- /dev/null
++++ b/libavcodec/loongson2/dsputil_loongson2.h
+@@ -0,0 +1,3 @@
++#include "libavcodec/dsputil.h"
++
++void dsputil_init_pix_loongson2(DSPContext* c, AVCodecContext *avctx);
+diff --git a/libavcodec/loongson2/dsputil_loongson2.o b/libavcodec/loongson2/dsputil_loongson2.o
+new file mode 100644
+index 0000000..fca0b55
+Binary files /dev/null and b/libavcodec/loongson2/dsputil_loongson2.o differ
+diff --git a/libavcodec/loongson2/idct_loongson2.c b/libavcodec/loongson2/idct_loongson2.c
+new file mode 100644
+index 0000000..539cab5
+--- /dev/null
++++ b/libavcodec/loongson2/idct_loongson2.c
+@@ -0,0 +1,336 @@
++/*
++ * Copyright (C) 2000-2003 Michel Lespinasse <walken@zoy.org>
++ * Copyright (C) 1999-2000 Aaron Holtzman <aholtzma@ess.engr.uvic.ca>
++ *
++ * Copyright (c) 2007-2010 comcat <jiankemeng@gmail.com>.
++ *
++ * Optimized for Loongson2 CPUs by comcat <jiankemeng@gmail.com>
++ * 
++ * Based on i386
++ */
++
++#include "libavutil/common.h"
++#include "dsputil_loongson2.h"
++
++
++#define ATTR_ALIGN(align) __attribute__ ((__aligned__ (align)))
++
++
++#define ROW_SHIFT 11
++#define COL_SHIFT 6
++
++#define round(bias) ((int)(((bias)+0.5) * (1<<ROW_SHIFT)))
++#define rounder(bias) {round (bias), round (bias)}
++
++
++
++#define loongson2_table(c1,c2,c3,c4,c5,c6,c7) { c4,c2,-c4,-c2, \
++					       c4,c6,c4,c6, \
++					       c1,c3,-c1,-c5,\
++					       c5,c7,c3,-c7, \
++					       c4,-c6,c4,-c6, \
++					       -c4,c2,c4,-c2, \
++					       c5,-c1,c3,-c1, \
++					       c7,c3,c7,-c5 }
++
++
++static inline void loongson2_row_head(int16_t * const row, const int offset,
++					const int16_t * const table)
++{
++	__asm__ volatile(
++//		".set\tmips3\n"
++		".set noreorder\n"
++		"ldc1	$f6,%0\n"
++		"ldc1	$f14,%1\n"
++		"ldc1	$f2,%2\n"
++		"ldc1	$f8,%3\n"
++		"dli $12,%4\n"
++		"dmtc1	$12,$f16\n"
++		"mov.d	$f4,$f6\n"
++		"mov.d	$f10,$f14\n"
++		"pmaddhw	$f2,$f2,$f4\n"
++		"pshufh	$f6,$f6,$f16\n"
++		".set reorder\n"
++//		".set\tmips0\n"
++		:	
++		:"m"(*(row+offset)),"m"(*(row+offset+4)),"m"(*table),"m"(*(table+4)),"i"(0x4e)
++		:"$f2","$f4","$f6","$f8","$f10","$f14","$f16","$12"
++	);
++}
++
++
++static inline void loongson2_row(const int16_t * const table,
++				const int32_t * const rounder)
++{
++	__asm__ volatile (
++//	".set\tmips3\n"
++	".set\tnoreorder\n"
++	"ldc1	$f0,%0\n"
++	"pmaddhw $f8,$f8,$f6\n"
++	"ldc1	$f16,%1\n"
++	"dli	$13,%8\n"
++	"ldc1	$f20,%2\n"
++	"pmaddhw $f0,$f0,$f14\n"
++	"ldc1	$f22,%3\n"
++	"pmaddhw $f4,$f4,$f16\n"
++	"paddw	$f2,$f2,$f22\n"
++	"ldc1	$f22,%4\n"
++	"dmtc1	$13,$f16\n"
++	"paddw	$f2,$f2,$f8\n"
++	"pmaddhw	$f14,$f14,$f22\n"
++	"mov.d	$f8,$f2\n"
++	"pshufh	$f10,$f10,$f16\n"
++	"ldc1	$f22,%3\n"
++	"pmaddhw	$f20,$f20,$f10\n"
++	"ldc1	$f16,%5\n"
++	"paddw	$f4,$f4,$f22\n"
++	"paddw	$f0,$f0,$f20\n"
++	"dli	$12,%6\n"
++	"pmaddhw	$f6,$f6,$f16\n"
++	"psubw	$f2,$f2,$f0\n"
++	"ldc1	$f16,%7\n"
++	"paddw	$f0,$f0,$f8\n"
++	"paddw	$f4,$f4,$f6\n"
++	"pmaddhw	$f10,$f10,$f16\n"
++	"mov.d	$f8,$f4\n"
++	"dmtc1	$12,$f16\n"
++	"paddw	$f14,$f14,$f10\n"
++	"psraw	$f2,$f2,$f16\n"
++	"psraw	$f0,$f0,$f16\n"
++	"paddw	$f4,$f4,$f14\n"
++	"psubw	$f8,$f8,$f14\n"
++	".set\treorder\n"
++//	".set\tmips0\n"
++	:
++	:"m"(*(table+8)),"m"(*(table+16)),"m"(*(table+12)),"m"(*rounder),"m"(*(table+24)),"m"(*(table+20)),"i"(ROW_SHIFT),"m"(*(table+16)),"i"(0x4e)
++	:"$f0","$f2","$f4","$f6","$f8","$f10","$f14","$f16","$f20","$f22","$12","$13","memory"
++	);
++}
++
++static inline void loongson2_row_tail(int16_t * const row, const int store)
++{
++	__asm__ volatile (
++//	".set\tmips3\n"
++	".set\tnoreorder\n"
++	"dli	$12,%2\n"
++	"dmtc1	$12,$f16\n"
++	"psraw	$f4,$f4,$f16\n"
++	"psraw	$f8,$f8,$f16\n"
++	"packsswh	$f0,$f0,$f4\n"
++	"packsswh	$f8,$f8,$f2\n"
++	"sdc1	$f0,%0\n"
++	"dli $13,%3\n"
++	"dmtc1	$13,$f22\n"
++	"pshufh	$f8,$f8,$f22\n"
++	"sdc1	$f8,%1\n"
++	".set\treorder\n"
++//	".set\tmips0\n"
++	:"=m"(*(row+store)),"=m"(*(row+store+4))
++	:"i"(ROW_SHIFT),"i"(0xb1)
++	:"$f0","$f2","$f4","$f6","$f8","$f16","$f22","$12","$13","memory"
++	);
++}
++
++static inline void loongson2_row_mid(int16_t * const row, const int store,
++					const int offset,
++					const int16_t * const table)
++{
++	__asm__ volatile (
++//	".set\tmips3\n"
++	".set\tnoreorder\n"
++	"ldc1	$f6,%2\n"
++	"dli $12,%3\n"
++	"dmtc1	$12,$f16\n"
++	"psraw	$f4,$f4,$f16\n"
++	"ldc1	$f14,%4\n"
++	"psraw	$f8,$f8,$f16\n"
++	"packsswh	$f0,$f0,$f4\n"
++	"mov.d	$f10,$f14\n"
++	"packsswh	$f8,$f8,$f2\n"
++	"mov.d	$f4,$f6\n"
++	"sdc1	$f0,%0\n"
++	"dli $13,%5\n"
++	"dmtc1	$13,$f22\n"
++	"pshufh	$f8,$f8,$f22\n"
++	"ldc1	$f2,%6\n"
++	"sdc1	$f8,%1\n"
++	"pmaddhw	$f2,$f2,$f4\n"
++	"ldc1	$f8,%7\n"
++	"dli $12,%8\n"
++	"dmtc1	$12,$f16\n"
++	"pshufh	$f6,$f6,$f16\n"
++	".set\treorder\n"
++//	".set\tmips0\n"
++	:"=m"(*(row+store)),"=m"(*(row+store+4))
++	: "m"(*(row+offset)),"i"(ROW_SHIFT),"m"(*(row+offset+4)),"i"(0xb1),"m"(*table),"m"(*(table+4)),"i"(0x4e)
++	:"$f0","$f2","$f4","$f6","$f8","$f10","$14","$f16","$f22","$12","$13","memory"
++	);
++}
++
++static inline void idct_col(int16_t * const col, const int offset)
++{
++#define T1 13036
++#define T2 27146
++#define T3 43790
++#define C4 23170
++	static const short _T1[] ATTR_ALIGN(8) = {T1,T1,T1,T1};
++	static const short _T2[] ATTR_ALIGN(8) = {T2,T2,T2,T2};
++	static const short _T3[] ATTR_ALIGN(8) = {T3,T3,T3,T3};
++	static const short _C4[] ATTR_ALIGN(8) = {C4,C4,C4,C4};
++
++	__asm__ volatile (
++//	".set\tmips3\n"
++	".set\tnoreorder\n"
++	"ldc1	$f4,%8\n"
++	"ldc1	$f0,%9\n"
++	"mov.d	$f6,$f4\n"
++	"ldc1	$f8,%10\n"
++	"pmulhh	$f4,$f4,$f0\n"
++	"ldc1	$f14,%11\n"
++	"pmulhh	$f6,$f6,$f8\n"
++	"ldc1	$f10,%12\n"
++	"mov.d	$f20,$f14\n"
++	"ldc1	$f2,%13\n"
++	"psubsh	$f4,$f4,$f8\n"
++	"ldc1	$f8,%14\n"
++	"pmulhh	$f14,$f14,$f2\n"
++	"paddsh	$f0,$f0,$f6\n"
++	"pmulhh	$f20,$f20,$f10\n"
++	"mov.d	$f6,$f8\n"
++	"paddsh	$f14,$f14,$f2\n"
++	"ldc1	$f16,%15\n"
++	"pmulhh	$f8,$f8,$f16\n"
++	"paddsh	$f20,$f20,$f10\n"
++	"psubsh	$f14,$f14,$f10\n"
++	"paddsh	$f20,$f20,$f2\n"
++	"ldc1	$f2,%16\n"
++	"mov.d	$f10,$f4\n"
++	"pmulhh	$f6,$f6,$f2\n"
++	"psubsh	$f4,$f4,$f14\n"
++	"psubsh	$f8,$f8,$f2\n"
++	"paddsh	$f14,$f14,$f10\n"
++	"sdc1	$f4,%0\n"
++	"mov.d	$f10,$f0\n"
++	"ldc1	$f22,%15\n"
++	"paddsh	$f6,$f6,$f22\n"
++	"paddsh	$f10,$f10,$f20\n"
++	"psubsh	$f0,$f0,$f20\n"
++	"mov.d	$f20,$f0\n"
++	"ldc1	$f2,%17\n"
++	"paddsh	$f0,$f0,$f14\n"
++	"ldc1	$f4,%18\n"
++	"psubsh	$f20,$f20,$f14\n"
++	"sdc1	$f10,%1\n"
++	"pmulhh	$f0,$f0,$f4\n"
++	"mov.d	$f10,$f8\n"
++	"pmulhh	$f20,$f20,$f4\n"
++	"ldc1	$f14,%19\n"
++	"mov.d	$f4,$f2\n"
++	"psubsh	$f2,$f2,$f14\n"
++	"paddsh	$f4,$f4,$f14\n"
++	"paddsh	$f8,$f8,$f2\n"
++	"mov.d	$f14,$f4\n"
++	"psubsh	$f2,$f2,$f10\n"
++	"paddsh	$f14,$f14,$f6\n"
++	"paddsh	$f0,$f0,$f0\n"
++	"psubsh	$f4,$f4,$f6\n"
++	"paddsh	$f20,$f20,$f20\n"
++	"mov.d	$f6,$f2\n"
++	"mov.d	$f10,$f8\n"
++	"paddsh	$f2,$f2,$f20\n"
++	"dli $12,%20\n"
++	"dmtc1	$12,$f16\n"
++	"psrah	$f2,$f2,$f16\n"
++	"paddsh	$f8,$f8,$f0\n"
++	"psrah	$f8,$f8,$f16\n"
++	"psubsh	$f10,$f10,$f0\n"
++	"ldc1	$f0,%12\n"
++	"psubsh	$f6,$f6,$f20\n"
++	"psrah	$f10,$f10,$f16\n"
++	"mov.d	$f20,$f14\n"
++	"sdc1	$f8,%2\n"
++	"psrah	$f6,$f6,$f16\n"
++	"sdc1	$f2,%3\n"
++	"paddsh	$f14,$f14,$f0\n"
++	"ldc1	$f8,%13\n"
++	"psubsh	$f20,$f20,$f0\n"
++	"psrah	$f14,$f14,$f16\n"
++	"mov.d	$f2,$f4\n"
++	"sdc1	$f6,%1\n"
++	"psubsh	$f2,$f2,$f8\n"
++	"psrah	$f20,$f20,$f16\n"
++	"paddsh	$f8,$f8,$f4\n"
++	"sdc1	$f14,%4\n"
++	"psrah	$f2,$f2,$f16\n"
++	"sdc1	$f10,%5\n"
++	"psrah	$f8,$f8,$f16\n"
++	"sdc1	$f20,%6\n"
++	"sdc1	$f2,%7\n"
++	"sdc1	$f8,%0\n"
++	".set\treorder\n"
++//	".set\tmips0\n"
++	:"=m"(*(col+offset+3*8)),"=m"(*(col+offset+5*8)),"=m"(*(col+offset+1*8)),"=m"(*(col+offset+2*8)),"=m"(*(col+offset+0*8)),"=m"(*(col+offset+6*8)),"=m"(*(col+offset+7*8)),"=m"(*(col+offset+4*8))
++	:"m"(*_T1),"m"(*(col+offset+1*8)),"m"(*(col+offset+7*8)),"m"(*_T3),"m"(*(col+offset+5*8)),"m"(*(col+offset+3*8)),"m"(*_T2),"m"(*(col+offset+2*8)),"m"(*(col+offset+6*8)),"m"(*(col+offset+0*8)),"m"(*_C4),"m"(*(col+offset+4*8)),"i"(COL_SHIFT)
++	:"$f0","$f2","$f4","$f6","$f8","$f10","$14","$f16","$20","$f22","$12","memory" 
++	);
++}
++
++static const int32_t rounder0[] ATTR_ALIGN(8) =
++    rounder ((1 << (COL_SHIFT - 1)) - 0.5);
++static const int32_t rounder4[] ATTR_ALIGN(8) = rounder (0);
++static const int32_t rounder1[] ATTR_ALIGN(8) =
++    rounder (1.25683487303);	/* C1*(C1/C4+C1+C7)/2 */
++static const int32_t rounder7[] ATTR_ALIGN(8) =
++    rounder (-0.25);		/* C1*(C7/C4+C7-C1)/2 */
++static const int32_t rounder2[] ATTR_ALIGN(8) =
++    rounder (0.60355339059);	/* C2 * (C6+C2)/2 */
++static const int32_t rounder6[] ATTR_ALIGN(8) =
++    rounder (-0.25);		/* C2 * (C6-C2)/2 */
++static const int32_t rounder3[] ATTR_ALIGN(8) =
++    rounder (0.087788325588);	/* C3*(-C3/C4+C3+C5)/2 */
++static const int32_t rounder5[] ATTR_ALIGN(8) =
++    rounder (-0.441341716183);	/* C3*(-C5/C4+C5-C3)/2 */
++
++
++#undef COL_SHIFT
++#undef ROW_SHIFT
++
++
++#define declare_idct(idct,table,idct_row_head,idct_row,idct_row_tail,idct_row_mid)	\
++inline void idct (int16_t * const block)				\
++{									\
++    static const int16_t table04[] ATTR_ALIGN(16) =			\
++	table (22725, 21407, 19266, 16384, 12873,  8867, 4520);		\
++    static const int16_t table17[] ATTR_ALIGN(16) =			\
++	table (31521, 29692, 26722, 22725, 17855, 12299, 6270);		\
++    static const int16_t table26[] ATTR_ALIGN(16) =			\
++	table (29692, 27969, 25172, 21407, 16819, 11585, 5906);		\
++    static const int16_t table35[] ATTR_ALIGN(16) =			\
++	table (26722, 25172, 22654, 19266, 15137, 10426, 5315);		\
++									\
++    idct_row_head (block, 0*8, table04);				\
++    idct_row (table04, rounder0);					\
++    idct_row_mid (block, 0*8, 4*8, table04);				\
++    idct_row (table04, rounder4);					\
++    idct_row_mid (block, 4*8, 1*8, table17);				\
++    idct_row (table17, rounder1);					\
++    idct_row_mid (block, 1*8, 7*8, table17);				\
++    idct_row (table17, rounder7);					\
++    idct_row_mid (block, 7*8, 2*8, table26);				\
++    idct_row (table26, rounder2);					\
++    idct_row_mid (block, 2*8, 6*8, table26);				\
++    idct_row (table26, rounder6);					\
++    idct_row_mid (block, 6*8, 3*8, table35);				\
++    idct_row (table35, rounder3);					\
++    idct_row_mid (block, 3*8, 5*8, table35);				\
++    idct_row (table35, rounder5);					\
++    idct_row_tail (block, 5*8);						\
++									\
++    idct_col (block, 0);						\
++    idct_col (block, 4);						\
++}
++
++void ff_loongson2_idct(DCTELEM *block);
++
++declare_idct (ff_loongson2_idct, loongson2_table,
++	      loongson2_row_head, loongson2_row, loongson2_row_tail, loongson2_row_mid)
+diff --git a/libavcodec/loongson2/idct_loongson2_xvid.c b/libavcodec/loongson2/idct_loongson2_xvid.c
+new file mode 100644
+index 0000000..4a1ee1e
+--- /dev/null
++++ b/libavcodec/loongson2/idct_loongson2_xvid.c
+@@ -0,0 +1,301 @@
++/*
++ *  XVID MPEG-4 VIDEO CODEC
++ *
++ *  Copyright(C) 2006-2010 comcat <jiankemeng@gmail.com>
++ *
++ *  Optimized for Loongson2 CPUs by comcat <jiankemeng@gmail.com>
++ * 
++ *  Based on i386
++ *
++ */
++
++
++#include <inttypes.h>
++#include "../avcodec.h"
++
++void ff_idct_xvid_loongson2(short *block);
++
++//=============================================================================
++// Macros and other preprocessor constants
++//=============================================================================
++
++#define BITS_INV_ACC    5                              // 4 or 5 for IEEE
++#define SHIFT_INV_ROW   (16 - BITS_INV_ACC) //11
++#define SHIFT_INV_COL   (1 + BITS_INV_ACC) //6
++#define RND_INV_ROW     (1024 * (6 - BITS_INV_ACC))
++#define RND_INV_COL     (16 * (BITS_INV_ACC - 3))
++#define RND_INV_CORR    (RND_INV_COL - 1)
++
++#define BITS_FRW_ACC    3                              // 2 or 3 for accuracy
++#define SHIFT_FRW_COL   BITS_FRW_ACC
++#define SHIFT_FRW_ROW   (BITS_FRW_ACC + 17)
++#define RND_FRW_ROW     (262144*(BITS_FRW_ACC - 1))
++
++
++//-----------------------------------------------------------------------------
++// Various memory constants (trigonometric values or rounding values)
++//-----------------------------------------------------------------------------
++
++static const int16_t tg_1_16[4*4] attribute_used __attribute__ ((aligned(8))) = {
++  13036,13036,13036,13036,        // tg * (2<<16) + 0.5
++  27146,27146,27146,27146,        // tg * (2<<16) + 0.5
++  -21746,-21746,-21746,-21746,    // tg * (2<<16) + 0.5
++  23170,23170,23170,23170};       // cos * (2<<15) + 0.5
++
++static const int32_t rounder_0[2*8] attribute_used __attribute__ ((aligned(8))) = {
++  65536,65536,
++  3597,3597,
++  2260,2260,
++  1203,1203,
++  0,0,
++  120,120,
++  512,512,
++  512,512};
++
++
++// Table for rows 0,4 - constants are multiplied by cos_4_16
++static const int16_t tab_i_04_mmx[32*4] attribute_used __attribute__ ((aligned(8))) = {
++  16384,16384,16384,-16384,       // movq-> w06 w04 w02 w00
++  21407,8867,8867,-21407,         // w07 w05 w03 w01
++  16384,-16384,16384,16384,       // w14 w12 w10 w08
++  -8867,21407,-21407,-8867,       // w15 w13 w11 w09
++  22725,12873,19266,-22725,       // w22 w20 w18 w16
++  19266,4520,-4520,-12873,        // w23 w21 w19 w17
++  12873,4520,4520,19266,          // w30 w28 w26 w24
++  -22725,19266,-12873,-22725,     // w31 w29 w27 w25
++// Table for rows 1,7 - constants are multiplied by cos_1_16
++  22725,22725,22725,-22725,       // movq-> w06 w04 w02 w00
++  29692,12299,12299,-29692,       // w07 w05 w03 w01
++  22725,-22725,22725,22725,       // w14 w12 w10 w08
++  -12299,29692,-29692,-12299,     // w15 w13 w11 w09
++  31521,17855,26722,-31521,       // w22 w20 w18 w16
++  26722,6270,-6270,-17855,        // w23 w21 w19 w17
++  17855,6270,6270,26722,          // w30 w28 w26 w24
++  -31521,26722,-17855,-31521,     // w31 w29 w27 w25
++// Table for rows 2,6 - constants are multiplied by cos_2_16
++  21407,21407,21407,-21407,       // movq-> w06 w04 w02 w00
++  27969,11585,11585,-27969,       // w07 w05 w03 w01
++  21407,-21407,21407,21407,       // w14 w12 w10 w08
++  -11585,27969,-27969,-11585,     // w15 w13 w11 w09
++  29692,16819,25172,-29692,       // w22 w20 w18 w16
++  25172,5906,-5906,-16819,        // w23 w21 w19 w17
++  16819,5906,5906,25172,          // w30 w28 w26 w24
++  -29692,25172,-16819,-29692,     // w31 w29 w27 w25
++// Table for rows 3,5 - constants are multiplied by cos_3_16
++  19266,19266,19266,-19266,       // movq-> w06 w04 w02 w00
++  25172,10426,10426,-25172,       // w07 w05 w03 w01
++  19266,-19266,19266,19266,       // w14 w12 w10 w08
++  -10426,25172,-25172,-10426,     // w15 w13 w11 w09
++  26722,15137,22654,-26722,       // w22 w20 w18 w16
++  22654,5315,-5315,-15137,        // w23 w21 w19 w17
++  15137,5315,5315,22654,          // w30 w28 w26 w24
++  -26722,22654,-15137,-26722,     // w31 w29 w27 w25
++};
++
++
++// %3 for rows 0,4 - constants are multiplied by cos_4_16
++static const int16_t tab_i_04_xmm[32*4] attribute_used __attribute__ ((aligned(8))) = {
++  16384,21407,16384,8867,      // movq-> w05 w04 w01 w00
++  16384,8867,-16384,-21407,    // w07 w06 w03 w02
++  16384,-8867,16384,-21407,    // w13 w12 w09 w08
++  -16384,21407,16384,-8867,    // w15 w14 w11 w10
++  22725,19266,19266,-4520,     // w21 w20 w17 w16
++  12873,4520,-22725,-12873,    // w23 w22 w19 w18
++  12873,-22725,4520,-12873,    // w29 w28 w25 w24
++  4520,19266,19266,-22725,     // w31 w30 w27 w26
++// %3 for rows 1,7 - constants are multiplied by cos_1_16
++  22725,29692,22725,12299,     // movq-> w05 w04 w01 w00
++  22725,12299,-22725,-29692,   // w07 w06 w03 w02
++  22725,-12299,22725,-29692,   // w13 w12 w09 w08
++  -22725,29692,22725,-12299,   // w15 w14 w11 w10
++  31521,26722,26722,-6270,     // w21 w20 w17 w16
++  17855,6270,-31521,-17855,    // w23 w22 w19 w18
++  17855,-31521,6270,-17855,    // w29 w28 w25 w24
++  6270,26722,26722,-31521,     // w31 w30 w27 w26
++// %3 for rows 2,6 - constants are multiplied by cos_2_16
++  21407,27969,21407,11585,     // movq-> w05 w04 w01 w00
++  21407,11585,-21407,-27969,   // w07 w06 w03 w02
++  21407,-11585,21407,-27969,   // w13 w12 w09 w08
++  -21407,27969,21407,-11585,   // w15 w14 w11 w10
++  29692,25172,25172,-5906,     // w21 w20 w17 w16
++  16819,5906,-29692,-16819,    // w23 w22 w19 w18
++  16819,-29692,5906,-16819,    // w29 w28 w25 w24
++  5906,25172,25172,-29692,     // w31 w30 w27 w26
++// %3 for rows 3,5 - constants are multiplied by cos_3_16
++  19266,25172,19266,10426,     // movq-> w05 w04 w01 w00
++  19266,10426,-19266,-25172,   // w07 w06 w03 w02
++  19266,-10426,19266,-25172,   // w13 w12 w09 w08
++  -19266,25172,19266,-10426,   // w15 w14 w11 w10
++  26722,22654,22654,-5315,     // w21 w20 w17 w16
++  15137,5315,-26722,-15137,    // w23 w22 w19 w18
++  15137,-26722,5315,-15137,    // w29 w28 w25 w24
++  5315,22654,22654,-26722,     // w31 w30 w27 w26
++};
++
++
++
++#define DCT_8_INV_ROW_LOONGSON2(A1,A2,A3,A4)\
++  "ldc1		$f0, " #A1 "           \n\t"/* 0     ; x3 x2 x1 x0*/\
++  "ldc1		$f2, 8+" #A1 "         \n\t"/* 1     ; x7 x6 x5 x4*/\
++  "mov.d	$f4, $f0               \n\t"/* 2     ; x3 x2 x1 x0*/\
++  "ldc1		$f6, " #A3 "           \n\t"/* 3     ; w05 w04 w01 w00*/\
++  "li		$12, 0x88				\n\t"\
++  "dmtc1	$12, $f16				\n\t"\
++  "pshufh	$f0, $f0, $f16         \n\t"/* x2 x0 x2 x0*/\
++  "ldc1		$f8, 8+" #A3 "         \n\t"/* 4     ; w07 w06 w03 w02*/\
++  "mov.d	$f10, $f2              \n\t"/* 5     ; x7 x6 x5 x4*/\
++  "pmaddhw	$f6, $f6, $f0          \n\t"/* x2*w05+x0*w04 x2*w01+x0*w00*/\
++  "ldc1		$f12, 32+" #A3 "       \n\t"/* 6     ; w21 w20 w17 w16*/\
++  "pshufh	$f2, $f2, $f16        \n\t"/* x6 x4 x6 x4*/\
++  "pmaddhw	$f8, $f8, $f2          \n\t"/* x6*w07+x4*w06 x6*w03+x4*w02*/\
++  "li		$12, 0xdd			   \n\t"\
++  "dmtc1	$12, $f16			   \n\t"\
++  "ldc1		$f14, 40+" #A3 "       \n\t"/* 7    ; w23 w22 w19 w18*/\
++  "pshufh	$f4, $f4, $f16         \n\t"/* x3 x1 x3 x1*/\
++  "pmaddhw	$f12, $f12, $f4        \n\t"/* x3*w21+x1*w20 x3*w17+x1*w16*/\
++  "ldc1		$f18, " #A4 "          \n\t" \
++  "ldc1		$f20, 16+" #A3 "	   \n\t" \
++  "ldc1		$f22, 24+" #A3 "	   \n\t" \
++  "ldc1		$f24, 48+" #A3 "	   \n\t" \
++  "ldc1		$f26, 56+" #A3 "	   \n\t" \
++  "pshufh	$f10, $f10, $f16       \n\t"/* x7 x5 x7 x5*/\
++  "pmaddhw	$f14, $f14, $f10       \n\t"/* x7*w23+x5*w22 x7*w19+x5*w18*/\
++  "paddw	$f6, $f6, $f18         \n\t"/* +%4*/\
++  "pmaddhw	$f0, $f0, $f20         \n\t"/* x2*w13+x0*w12 x2*w09+x0*w08*/\
++  "paddw	$f6, $f6, $f8	       \n\t"/* 4     ; a1=sum(even1) a0=sum(even0)*/\
++  "pmaddhw	$f2, $f2, $f22         \n\t"/* x6*w15+x4*w14 x6*w11+x4*w10*/\
++  "mov.d	$f8, $f6               \n\t"/* 4     ; a1 a0*/\
++  "li		$12, 11				   \n\t"\
++  "dmtc1	$12, $f16			   \n\t"\
++  "pmaddhw	$f4, $f4, $f24		   \n\t"/* x3*w29+x1*w28 x3*w25+x1*w24*/\
++  "paddw	$f12, $f12, $f14       \n\t"/* 7     ; b1=sum(odd1) b0=sum(odd0)*/\
++  "pmaddhw	$f10, $f10, $f26	   \n\t"/* x7*w31+x5*w30 x7*w27+x5*w26*/\
++  "paddw	$f6, $f6, $f12         \n\t"/* a1+b1 a0+b0*/\
++  "paddw	$f0, $f0, $f18         \n\t"/* +%4*/\
++  "psraw	$f6, $f6, $f16		   \n\t"/* y1=a1+b1 y0=a0+b0*/\
++  "paddw	$f0, $f0, $f2          \n\t"/* 1     ; a3=sum(even3) a2=sum(even2)*/\
++  "psubw	$f8, $f8, $f12         \n\t"/* 6     ; a1-b1 a0-b0*/\
++  "mov.d	$f14, $f0              \n\t"/* 7     ; a3 a2*/\
++  "paddw	$f4, $f4, $f10         \n\t"/* 5     ; b3=sum(odd3) b2=sum(odd2)*/\
++  "paddw	$f0, $f0, $f4          \n\t"/* a3+b3 a2+b2*/\
++  "psraw	$f8, $f8, $f16	       \n\t"/* y6=a1-b1 y7=a0-b0*/\
++  "psubw 	$f14, $f14, $f4        \n\t"/* 2     ; a3-b3 a2-b2*/\
++  "psraw	$f0, $f0, $f16     	   \n\t"/* y3=a3+b3 y2=a2+b2*/\
++  "psraw	$f14, $f14, $f16	   \n\t"/* y4=a3-b3 y5=a2-b2*/\
++  "li		$12, 0xb1			   \n\t"\
++  "dmtc1	$12, $f20			   \n\t"\
++  "packsswh $f6, $f6, $f0          \n\t"/* 0     ; y3 y2 y1 y0*/\
++  "packsswh $f14, $f14, $f8        \n\t"/* 4     ; y6 y7 y4 y5*/\
++  "sdc1		$f6, " #A2 "           \n\t"/* 3     ; save y3 y2 y1 y0*/\
++  "pshufh	$f14, $f14, $f20      \n\t"/* y7 y6 y5 y4*/\
++  "sdc1		$f14, 8 +" #A2 "	   \n\t"/* 7     ; save y7 y6 y5 y4*/\
++
++
++#define DCT_8_INV_COL(A1,A2)\
++  "ldc1		$f0, 2*8(%3)			\n\t"/* */\
++  "ldc1		$f6, 16*3+" #A1 "		\n\t"/* x3 */\
++  "mov.d	$f2, $f0				\n\t"/* tg_3_16*/\
++  "ldc1		$f10, 16*5+" #A1 "		\n\t"/* x5 */\
++  "pmulhh 	$f0, $f0, $f6 			\n\t"/* x3*(tg_3_16-1)*/\
++  "ldc1 	$f8, (%3)				\n\t"\
++  "pmulhh	$f2, $f2, $f10			\n\t"/* x5*(tg_3_16-1)*/\
++  "ldc1 	$f14, 16*7+" #A1 "		\n\t"/* x7 */\
++  "mov.d	$f4, $f8				\n\t"/* tg_1_16*/\
++  "ldc1		$f12, 16*1+" #A1 "		\n\t"/* x1 */\
++  "pmulhh 	$f8, $f8, $f14			\n\t"/* x7*tg_1_16*/\
++  "paddsh	$f0, $f0, $f6			\n\t"/* x3*tg_3_16*/\
++  "pmulhh	$f4, $f4, $f12			\n\t"/* x1*tg_1_16*/\
++  "paddsh	$f2, $f2, $f6			\n\t"/* x3+x5*(tg_3_16-1)*/\
++  "psubsh	$f0, $f0, $f10			\n\t"/* x3*tg_3_16-x5 = tm35*/\
++  "ldc1		$f6, 3*8(%3)			\n\t"\
++  "paddsh	$f2, $f2, $f10			\n\t"/* x3+x5*tg_3_16 = tp35*/\
++  "paddsh	$f8, $f8, $f12			\n\t"/* x1+tg_1_16*x7 = tp17*/\
++  "psubsh	$f4, $f4, $f14			\n\t"/* x1*tg_1_16-x7 = tm17*/\
++  "mov.d	$f10, $f8				\n\t"/* tp17*/\
++  "mov.d	$f12, $f4				\n\t"/* tm17*/\
++  "paddsh	$f10, $f10, $f2			\n\t"/* tp17+tp35 = b0*/\
++  "psubsh	$f12, $f12, $f0			\n\t"/* tm17-tm35 = b3*/\
++  "psubsh	$f8, $f8, $f2			\n\t"/* tp17-tp35 = t1*/\
++  "paddsh	$f4, $f4, $f0			\n\t"/* tm17+tm35 = t2*/\
++  "ldc1		$f14, 1*8(%3)			\n\t"\
++  "mov.d	$f2, $f8				\n\t"/* t1*/\
++  "sdc1		$f10, 3*16+" #A2 "		\n\t"/* save b0*/\
++  "paddsh	$f2, $f2, $f4			\n\t"/* t1+t2*/\
++  "sdc1		$f12, 5*16+" #A2 "		\n\t"/* save b3*/\
++  "psubsh	$f8, $f8, $f4			\n\t"/* t1-t2*/\
++  "ldc1		$f10, 2*16+" #A1 "		\n\t"\
++  "mov.d	$f0, $f14				\n\t"/* tg_2_16*/\
++  "ldc1		$f12, 6*16+" #A1 "		\n\t"\
++  "pmulhh	$f0, $f0, $f10			\n\t"/* x2*tg_2_16*/\
++  "pmulhh	$f14, $f14, $f12		\n\t"/* x6*tg_2_16*/\
++  "pmulhh	$f2, $f2, $f6			\n\t"/* ocos_4_16*(t1+t2) = b1/2*/\
++  "ldc1		$f4, 0*16+" #A1 "		\n\t"\
++  "pmulhh 	$f8, $f8, $f6			\n\t"/* ocos_4_16*(t1-t2) = b2/2*/\
++  "psubsh	$f0, $f0, $f12			\n\t"/* t2*tg_2_16-x6 = tm26*/\
++  "mov.d 	$f6, $f4				\n\t"/* x0*/\
++  "ldc1		$f12, 4*16+" #A1 "		\n\t"\
++  "paddsh 	$f14, $f14, $f10		\n\t"/* x2+x6*tg_2_16 = tp26*/\
++  "paddsh 	$f4, $f4, $f12			\n\t"/* x0+x4 = tp04*/\
++  "psubsh 	$f6, $f6, $f12			\n\t"/* x0-x4 = tm04*/\
++  "mov.d	$f10, $f4				\n\t"/* tp04*/\
++  "mov.d 	$f12, $f6				\n\t"/* tm04*/\
++  "psubsh 	$f4, $f4, $f14			\n\t"/* tp04-tp26 = a3*/\
++  "paddsh 	$f6, $f6, $f0			\n\t"/* tm04+tm26 = a1*/\
++  "paddsh 	$f2, $f2, $f2			\n\t"/* b1*/\
++  "paddsh 	$f8, $f8, $f8			\n\t"/* b2*/\
++  "paddsh	$f10, $f10, $f14		\n\t"/* tp04+tp26 = a0*/\
++  "psubsh	$f12, $f12, $f0			\n\t"/* tm04-tm26 = a2*/\
++  "li		$12, 6					\n\t"\
++  "dmtc1	$12, $f18				\n\t"\
++  "mov.d	$f14, $f6				\n\t"/* a1*/\
++  "mov.d	$f0, $f12				\n\t"/* a2*/\
++  "paddsh 	$f6, $f6, $f2			\n\t"/* a1+b1*/\
++  "paddsh 	$f12, $f12, $f8			\n\t"/* a2+b2*/\
++  "psrah 	$f6, $f6, $f18			\n\t"/* dst1*/\
++  "psubsh 	$f14, $f14, $f2			\n\t"/* a1-b1*/\
++  "psrah 	$f12, $f12, $f18		\n\t"/* dst2*/\
++  "psubsh 	$f0, $f0, $f8			\n\t"/* a2-b2*/\
++  "ldc1		$f2, 3*16+" #A2 "		\n\t"/* load b0*/\
++  "psrah 	$f14, $f14, $f18		\n\t"/* dst6*/\
++  "mov.d	$f8, $f10				\n\t"/* a0*/\
++  "psrah 	$f0, $f0, $f18			\n\t"/* dst5*/\
++  "sdc1		$f6, 1*16+" #A2 "		\n\t"\
++  "paddsh 	$f10, $f10, $f2			\n\t"/* a0+b0*/\
++  "sdc1		$f12, 2*16+" #A2 "		\n\t"\
++  "psubsh 	$f8, $f8, $f2			\n\t"/* a0-b0*/\
++  "ldc1		$f6, 5*16+" #A2 "		\n\t"/* load b3*/\
++  "psrah	$f10, $f10, $f18		\n\t"/* dst0*/\
++  "mov.d	$f12, $f4				\n\t"/* a3*/\
++  "psrah 	$f8, $f8, $f18			\n\t"/* dst7*/\
++  "sdc1		$f0, 5*16+" #A2 "		\n\t"\
++  "paddsh 	$f4, $f4, $f6			\n\t"/* a3+b3*/\
++  "sdc1		$f14, 6*16+" #A2 "		\n\t"\
++  "psubsh 	$f12, $f12, $f6			\n\t"/* a3-b3*/\
++  "sdc1		$f10, 0*16+" #A2 "		\n\t"\
++  "psrah 	$f4, $f4, $f18 			\n\t"/* dst3*/\
++  "sdc1		$f8, 7*16+" #A2 "		\n\t"\
++  "psrah 	$f12, $f12, $f18		\n\t"/* dst4*/\
++  "sdc1		$f4, 3*16+" #A2 "		\n\t"\
++  "sdc1		$f12, 4*16+" #A2 "		\n\t"
++
++
++
++void ff_idct_xvid_loongson2(short *block){
++	__asm__ volatile(
++	//# Process each row
++    DCT_8_INV_ROW_LOONGSON2(0*16(%0), 0*16(%0), 64*0(%2), 8*0(%1))
++    DCT_8_INV_ROW_LOONGSON2(1*16(%0), 1*16(%0), 64*1(%2), 8*1(%1))
++    DCT_8_INV_ROW_LOONGSON2(2*16(%0), 2*16(%0), 64*2(%2), 8*2(%1))
++    DCT_8_INV_ROW_LOONGSON2(3*16(%0), 3*16(%0), 64*3(%2), 8*3(%1))
++    DCT_8_INV_ROW_LOONGSON2(4*16(%0), 4*16(%0), 64*0(%2), 8*4(%1))
++    DCT_8_INV_ROW_LOONGSON2(5*16(%0), 5*16(%0), 64*3(%2), 8*5(%1))
++    DCT_8_INV_ROW_LOONGSON2(6*16(%0), 6*16(%0), 64*2(%2), 8*6(%1))
++    DCT_8_INV_ROW_LOONGSON2(7*16(%0), 7*16(%0), 64*1(%2), 8*7(%1))
++
++	//# Process the columns (4 at a time)
++    DCT_8_INV_COL(0(%0), 0(%0))
++    DCT_8_INV_COL(8(%0), 8(%0))
++    :
++    : "r"(block), "r"(rounder_0), "r"(tab_i_04_xmm), "r"(tg_1_16)
++    :"$f0","$f2","$f4","$f6","$f8","$f10","$f12","$f14","$f18","$f16","$20","$22","$24","$26");
++}
++
+diff --git a/libavcodec/loongson2/motion_est_loongson2.c b/libavcodec/loongson2/motion_est_loongson2.c
+new file mode 100644
+index 0000000..bb67290
+--- /dev/null
++++ b/libavcodec/loongson2/motion_est_loongson2.c
+@@ -0,0 +1,365 @@
++/*
++ * Loongson2E MMI optimized motion estimation
++ * Copyright (c) 2007 comcat <jiankemeng@gmail.com>.
++ *
++ * based on Michael Niedermayer <michaelni@gmx.at>
++ *
++ */
++
++#include "dsputil_loongson2.h"
++#include "../avcodec.h"
++
++static const __attribute__ ((aligned(8))) uint64_t round_tab[3]={
++	0x0000000000000000ULL,
++	0x0001000100010001ULL,
++	0x0002000200020002ULL,
++};
++
++static attribute_used __attribute__ ((aligned(8))) uint64_t bone= 0x0101010101010101LL;
++
++static inline void sad8_1_loongson2(uint8_t *blk1, uint8_t *blk2, int stride, int h)
++{
++    long len= -(stride*h);
++    __asm__ volatile(
++        
++//		".set mips3						\n\t"
++		".align 4						\n\t"
++
++		"move	$8, %0					\n\t"
++		"move	$21, %1					\n\t"
++		"move	$22, %2					\n\t"
++		"move	$23, %3					\n\t"
++
++        "1:                             \n\t"
++
++		"add	$9, $8, $21				\n\t"
++		"add	$10, $8, $22				\n\t"
++        
++		"uld	$11, ($9)				\n\t"
++		"dmtc1	$11, $f0				\n\t"
++
++		"uld	$12, ($9)				\n\t"
++		"dmtc1	$12, $f4				\n\t"
++        
++		"pasubub $f10, $f0, $f4			\n\t"
++		"biadd	 $f0, $f10				\n\t"
++
++		"add	$8, $8, $23				\n\t"
++
++		"add	$9, $8, $21				\n\t"
++		"add	$10, $8, $22				\n\t"
++        
++		"uld	$11, ($9)				\n\t"
++		"dmtc1	$11, $f2				\n\t"
++		
++		"uld	$12, ($10)				\n\t"
++		"dmtc1	$12, $f6				\n\t"
++
++		"pasubub $f16, $f2, $f6			\n\t"
++		"biadd	 $f6, $f16				\n\t"
++        
++		"paddh	 $f0, $f0, $f6			\n\t"
++        
++		"paddh	 $f12, $f12, $f0		\n\t"
++		
++		"bltz	$8, 1b					\n\t"		
++		"add	$8, $8, $23				\n\t" 
++
++        : "+r" (len)
++        : "r" (blk1 - len), "r" (blk2 - len), "r" ((long)stride)
++		: "$8", "$9", "$10", "$21", "$22", "$23", "$f0", "$f2", "$f4", "$f6", "$f10", "$f16"
++    );
++}
++
++static inline void sad8_2_loongson2(uint8_t *blk1a, uint8_t *blk1b, uint8_t *blk2, int stride, int h)
++{
++    long len= -(stride*h);
++    __asm__ volatile(
++        
++//		".set mips3						\n\t"
++		".align 4						\n\t"
++
++		"move	$8, %0					\n\t"
++
++        "1:                             \n\t"
++		"add	$9, $8, %1				\n\t"
++		"add	$10, $8, %2				\n\t"
++		"add	$11, $8, %3				\n\t"
++
++		"uld	$12, ($9)				\n\t"
++		"dmtc1	$12, $f0				\n\t"
++		"uld	$13, ($10)				\n\t"
++		"dmtc1	$13, $f4				\n\t"
++
++		"pavgb	$f0, $f0, $f4			\n\t"
++		
++		"uld	$12, ($11)				\n\t"
++		"dmtc1	$12, $f4				\n\t"
++
++		"pasubub $f10, $f0, $f4			\n\t"
++		"biadd	 $f0, $f10				\n\t"
++        
++		"add	$8, $8, %4				\n\t"
++
++		"add	$9, $8, %1				\n\t"
++		"add	$10, $8, %2				\n\t"
++		"add	$11, $8, %3				\n\t"
++
++		"uld	$12, ($9)				\n\t"
++		"dmtc1	$12, $f2				\n\t"
++		"uld	$13, ($10)				\n\t"
++		"dmtc1	$13, $f6				\n\t"
++        
++		"pavgb	$f6, $f6, $f2			\n\t"
++		
++		"uld	$12, ($11)				\n\t"
++		"dmtc1	$12, $f2				\n\t"
++        
++		"pasubub $f16, $f6, $f2			\n\t"
++		"biadd	 $f6, $f16				\n\t"
++        
++		"paddh	 $f0, $f0, $f6			\n\t"
++		"paddh	 $f12, $f12, $f0		\n\t"
++
++		"bltz	$8, 1b					\n\t"
++		"add	$8, $8, %4				\n\t"
++        : "+r" (len)
++        : "r" (blk1a - len), "r" (blk1b -len), "r" (blk2 - len), "r" ((long)stride)
++		: "$8", "$9", "$10", "$11", "$12", "$13", "$f0", "$f2", "$f4", "$f6", "$f10", "$f16"
++    );
++}
++
++static inline void sad8_4_loongson2(uint8_t *blk1, uint8_t *blk2, int stride, int h)
++{ 
++    long len= -(stride*h);
++    __asm__ volatile(
++        
++        
++//		".set mips3						\n\t"
++		".align 4						\n\t"
++
++		"ldc1	$f10, "MANGLE(bone)"	\n\t"
++
++		"move	$8, %0					\n\t"
++
++        "1:                             \n\t"
++		"add	$9, $8, %1				\n\t"
++		"add	$10, $8, %2				\n\t"
++		"add	$11, $8, %3				\n\t"
++        
++		"uld	$12, ($9)				\n\t"
++		"dmtc1	$12, $f0				\n\t"
++		
++		"uld	$13, ($10)				\n\t"
++		"dmtc1	$13, $f4				\n\t"
++		
++		"uld	$12, 1($9)				\n\t"
++		"dmtc1	$12, $f2				\n\t"
++		
++		"uld	$13, 1($10)				\n\t"
++		"dmtc1	$13, $f6				\n\t"
++
++		"pavgb	$f0, $f0, $f4			\n\t"
++		"pavgb	$f6, $f6, $f2			\n\t"
++        
++		"psubusb $f6, $f6, $f10			\n\t"
++		"pavgb	 $f0, $f0, $f6			\n\t"
++
++		"uld	$13, 1($11)				\n\t"
++		"dmtc1	$13, $f4				\n\t"
++        
++		"pasubub $f16, $f0, $f4			\n\t"
++		"biadd	 $f0, $f16				\n\t"
++
++		"add	 $8, $8, %4				\n\t"
++
++		"add	$9, $8, %1				\n\t"
++		"add	$10, $8, %2				\n\t"
++		"add	$11, $8, %3				\n\t"
++
++		"uld	$12, ($9)				\n\t"
++		"dmtc1	$12, $f2				\n\t"
++		"uld	$13, ($10)				\n\t"
++		"dmtc1	$12, $f6				\n\t"
++		"uld	$12, 1($9)				\n\t"
++		"dmtc1	$12, $f4				\n\t"
++		"uld	$13, 1($10)				\n\t"
++		"dmtc1	$12, $f8				\n\t"
++
++		"pavgb	$f2, $f2, $f6			\n\t"
++		"pavgb	$f4, $f4, $f8			\n\t"
++        
++		"psubusb $f4, $f4, $f10			\n\t"
++		"pavgb	 $f4, $f4, $f2			\n\t"
++
++		"uld	$13, ($11)				\n\t"
++		"dmtc1	$13, $f2				\n\t"
++
++		"pasubub $f18, $f4, $f2			\n\t"
++		"biadd	 $f4, $f18				\n\t"
++        
++		"paddh	 $f0, $f0, $f4			\n\t"
++		"paddh	 $f12, $f12, $f0		\n\t"
++        
++		"bltz	 $8, 1b					\n\t"
++		"add	 $8, $8, %4				\n\t"
++        : "+r" (len)
++        : "r" (blk1 - len), "r" (blk1 - len + stride), "r" (blk2 - len), "r" ((long)stride)
++		: "$8", "$9", "$10", "$11", "$12", "$13", "$f0", "$f2", "$f4", "$f6", "$f8", "$f10", "$f16", "$f18"
++    );
++}
++
++static inline int sum_loongson2(void)
++{
++    int ret;
++    __asm__ volatile(
++//		".set mips3						\n\t"
++        
++		"dmfc1	%0, $f12				\n\t"
++        : "=r" (ret)
++    );
++    return ret;
++}
++
++
++static int sad8_loongson2(void *v, uint8_t *blk2, uint8_t *blk1, int stride, int h)
++{
++    assert(h==8);
++    __asm__ volatile(
++//			".set mips3						\n\t"
++			"xor	$f14, $f14, $f14 \n\t"
++			"xor	$f12, $f12, $f12 \n\t"
++			:
++	);
++
++    sad8_1_loongson2(blk1, blk2, stride, 8);
++
++    return sum_loongson2();
++}
++
++static int sad8_x2_loongson2(void *v, uint8_t *blk2, uint8_t *blk1, int stride, int h)
++{
++    assert(h==8);
++    __asm__ volatile(
++//			".set mips3						\n\t"
++			"xor	$f14, $f14, $f14	\n\t"
++			"xor	$f12, $f12, $f12	\n\t"
++            
++			"ldc1	$f10, %0		 \n\t"
++            :: "m"(round_tab[1]) 
++    );
++
++    sad8_2_loongson2(blk1, blk1+1, blk2, stride, 8);
++
++    return sum_loongson2();
++}
++
++static int sad8_y2_loongson2(void *v, uint8_t *blk2, uint8_t *blk1, int stride, int h)
++{
++    assert(h==8);
++    __asm__ volatile(
++//				 ".set mips3				\n\t"
++				 "xor	$f14, $f14, $f14	\n\t"
++				 "xor	$f12, $f12, $f12	\n\t"
++                 
++				 "ldc1	$f10, %0		 \n\t"
++                 :: "m"(round_tab[1]) 
++                 );
++
++    sad8_2_loongson2(blk1, blk1+stride, blk2, stride, 8);
++
++    return sum_loongson2();
++}
++
++static int sad8_xy2_loongson2(void *v, uint8_t *blk2, uint8_t *blk1, int stride, int h)
++{
++    assert(h==8);
++    __asm__ volatile(
++//				 ".set mips3				\n\t"
++				 "xor	$f14, $f14, $f14	\n\t"
++				 "xor	$f12, $f12, $f12	\n\t"
++				 "ldc1	$f10, %0		 \n\t"
++                 :: "m"(round_tab[2]) 
++                 );
++
++    sad8_4_loongson2(blk1, blk2, stride, 8);
++
++    return sum_loongson2();
++}
++
++static int sad16_loongson2(void *v, uint8_t *blk2, uint8_t *blk1, int stride, int h)
++{
++    __asm__ volatile(
++//				 ".set mips3			\n\t"
++				 "xor	$f14, $f14, $f14	\n\t"
++				 "xor	$f12, $f12, $f12	\n\t":);
++
++    sad8_1_loongson2(blk1  , blk2  , stride, h);
++    sad8_1_loongson2(blk1+8, blk2+8, stride, h);
++
++    return sum_loongson2();
++}
++
++static int sad16_x2_loongson2(void *v, uint8_t *blk2, uint8_t *blk1, int stride, int h)
++{
++    __asm__ volatile(
++//				 ".set mips3				\n\t"
++				 "xor	$f14, $f14, $f14	\n\t"
++				 "xor	$f12, $f12, $f12	\n\t"
++				 "ldc1	$f10, %0		 \n\t"
++                 :: "m"(round_tab[1]) 
++                 );
++
++    sad8_2_loongson2(blk1  , blk1+1, blk2  , stride, h);
++    sad8_2_loongson2(blk1+8, blk1+9, blk2+8, stride, h);
++
++    return sum_loongson2();
++}
++
++static int sad16_y2_loongson2(void *v, uint8_t *blk2, uint8_t *blk1, int stride, int h)
++{
++    __asm__ volatile(
++//				 ".set mips3				\n\t"
++				 "xor	$f14, $f14, $f14	\n\t"
++				 "xor	$f12, $f12, $f12	\n\t"
++				 "ldc1	$f10, %0		 \n\t"
++                 :: "m"(round_tab[1]) 
++                 );
++
++    sad8_2_loongson2(blk1  , blk1+stride,  blk2  , stride, h);
++    sad8_2_loongson2(blk1+8, blk1+stride+8,blk2+8, stride, h);
++
++    return sum_loongson2();
++}
++
++static int sad16_xy2_loongson2(void *v, uint8_t *blk2, uint8_t *blk1, int stride, int h)
++{
++    __asm__ volatile(
++//				 ".set mips3				\n\t"
++				 "xor	$f14, $f14, $f14	\n\t"
++				 "xor	$f12, $f12, $f12	\n\t"
++				 "ldc1	$f10, %0		 \n\t"
++                 :: "m"(round_tab[2]) 
++                 );
++
++    sad8_4_loongson2(blk1  , blk2  , stride, h);
++    sad8_4_loongson2(blk1+8, blk2+8, stride, h);
++
++    return sum_loongson2();
++}
++
++
++void dsputil_init_pix_loongson2(DSPContext* c, AVCodecContext *avctx)
++{
++        c->pix_abs[0][0] = sad16_loongson2;
++        c->pix_abs[0][1] = sad16_x2_loongson2;
++        c->pix_abs[0][2] = sad16_y2_loongson2;
++        c->pix_abs[0][3] = sad16_xy2_loongson2;
++        c->pix_abs[1][0] = sad8_loongson2;
++        c->pix_abs[1][1] = sad8_x2_loongson2;
++        c->pix_abs[1][2] = sad8_y2_loongson2;
++        c->pix_abs[1][3] = sad8_xy2_loongson2;
++
++        c->sad[0]= sad16_loongson2;
++        c->sad[1]= sad8_loongson2;
++}
+diff --git a/libavcodec/loongson2/mpegvideo_loongson2.c b/libavcodec/loongson2/mpegvideo_loongson2.c
+new file mode 100644
+index 0000000..18d070a
+--- /dev/null
++++ b/libavcodec/loongson2/mpegvideo_loongson2.c
+@@ -0,0 +1,385 @@
++/*
++ * The simplest mpeg encoder (well, it was the simplest!)
++ * Copyright (c) 2007-2010 comcat <jiankemeng@gmail.com>.
++ *
++ * Optimized for Loongson2 CPUs by comcat <jiankemeng@gmail.com>
++ * 
++ * Based on i386
++ */
++
++#include "dsputil_loongson2.h"
++#include "../mpegvideo.h"
++#include "../avcodec.h"
++
++extern uint8_t zigzag_direct_noperm[64];
++extern uint16_t inv_zigzag_direct16[64];
++
++static const unsigned long long int mm_wabs __attribute__ ((aligned(8))) = 0xffffffffffffffffULL;
++static const unsigned long long int mm_wone __attribute__ ((aligned(8))) = 0x0001000100010001ULL;
++
++
++static void dct_unquantize_h263_intra_loongson2(MpegEncContext *s,
++                                  DCTELEM *block, int n, int qscale)
++{
++    long level, qmul, qadd, nCoeffs;
++
++    qmul = qscale << 1;
++
++    assert(s->block_last_index[n]>=0 || s->h263_aic); 
++    if (!s->h263_aic) {
++        if (n < 4)
++            level = block[0] * s->y_dc_scale;
++        else
++            level = block[0] * s->c_dc_scale;
++        qadd = (qscale - 1) | 1;
++    }else{
++        qadd = 0;
++        level= block[0];
++    }
++    if(s->ac_pred)
++        nCoeffs=63;
++    else
++        nCoeffs= s->inter_scantable.raster_end[ s->block_last_index[n] ];
++
++
++	__asm__ volatile(
++//			".set mips3						\n\t"
++			
++			"xor		$f12, $f12, $f12	\n\t"
++			"lwc1		$f12, %1			\n\t"
++			
++			"xor		$f10, $f10, $f10	\n\t"
++			
++			"packsswh	$f12, $f12, $f12	\n\t"
++			
++			"lwc1		$f10, %2			\n\t"
++			
++			"packsswh	$f10, $f10, $f10	\n\t"
++			
++			"packsswh	$f12, $f12, $f12	\n\t"
++			
++			"xor 		$f14, $f14, $f14	\n\t"
++			
++			"packsswh	$f10, $f10, $f10	\n\t"
++			
++			"xor		$f8, $f8, $f8		\n\t"
++			
++			"psubh		$f14, $f14, $f10	\n\t"
++
++			
++			"1:                             \n\t"
++			"add		$12, %0, %3			\n\t"
++			
++			"ldc1		$f0, ($12)			\n\t"
++			
++			"ldc1		$f2, 8($12)			\n\t"
++
++			"mov.d		$f4, $f0			\n\t"
++			"mov.d		$f6, $f2			\n\t"
++			
++			"pmullh		$f0, $f0, $f12		\n\t"
++			"pmullh		$f2, $f2, $f12		\n\t"
++
++			"pcmpgth	$f4, $f4, $f8		\n\t"
++			"pcmpgth	$f6, $f6, $f8		\n\t"
++			
++			"xor		$f0, $f0, $f4		\n\t"
++			"xor		$f2, $f2, $f6		\n\t"
++
++			
++			"paddh		$f0, $f0, $f14      \n\t"
++			
++			"paddh		$f2, $f2, $f14		\n\t"
++
++			
++			"xor		$f4, $f4, $f0		\n\t"
++			
++			"xor		$f6, $f6, $f2		\n\t"
++
++			
++			"pcmpeqh	$f0, $f0, $f14		\n\t"
++			
++			"pcmpeqh	$f2, $f2, $f14		\n\t"
++
++			
++			"pandn		$f0, $f0, $f4		\n\t"	
++			
++			"pandn		$f2, $f2, $f6		\n\t"
++
++			
++			"sdc1		$f0, ($12)			\n\t"
++			
++			"sdc1		$f2, 8($12)			\n\t"
++
++			
++			"addiu		%3, %3, 16			\n\t"
++			
++			"blez		%3, 1b				\n\t"
++			"nop							\n\t"
++			::"r" (block+nCoeffs), "m"(qmul), "m" (qadd), "r" (2*(-nCoeffs))
++			: "memory"
++        );
++        block[0]= level;
++}
++
++
++static void dct_unquantize_h263_inter_loongson2(MpegEncContext *s,
++                                  DCTELEM *block, int n, int qscale)
++{
++    long qmul, qadd, nCoeffs;
++
++    qmul = qscale << 1;
++    qadd = (qscale - 1) | 1;
++
++    assert(s->block_last_index[n]>=0 || s->h263_aic);
++
++    nCoeffs= s->inter_scantable.raster_end[ s->block_last_index[n] ];
++
++   __asm__ volatile(
++//				".set mips3						\n\t"
++                
++				"xor		$f12, $f12, $f12	\n\t"
++				"lwc1		$f12, %1			\n\t"
++                
++				"xor		$f10, $f10, $f10	\n\t"
++                
++				"packsswh	$f12, $f12, $f12	\n\t"
++
++				"lwc1		$f10, %2			\n\t"
++                
++				"packsswh	$f10, $f10, $f10	\n\t"
++                
++				"xor		$f14, $f14, $f14	\n\t"
++                
++				"packsswh	$f12, $f12, $f12	\n\t"
++                
++				"packsswh	$f10, $f10, $f10	\n\t"
++                
++				"xor		$f8, $f8, $f8		\n\t"
++
++				"psubh		$f14, $f14, $f10	\n\t"
++                
++
++                "1:                             \n\t"
++				"add		$12, %0, %3			\n\t"
++                
++				"ldc1		$f0, ($12)			\n\t"
++                
++				"ldc1		$f2, 8($12)			\n\t"
++
++				"mov.d		$f4, $f0			\n\t"
++				"mov.d		$f6, $f2			\n\t"
++                
++				"pmullh		$f0, $f0, $f12		\n\t"
++                
++				"pmullh		$f2, $f2, $f12		\n\t"
++
++				"pcmpgth	$f4, $f4, $f8		\n\t"
++                
++				"pcmpgth	$f6, $f6, $f8		\n\t"
++                
++				"xor		$f0, $f0, $f4		\n\t"
++                
++				"xor		$f2, $f2, $f6		\n\t"
++                
++				"paddh		$f0, $f0, $f14		\n\t"
++                
++				"paddh		$f2, $f2, $f14		\n\t"
++                
++				"xor		$f4, $f4, $f0		\n\t"
++                
++				"xor		$f6, $f6, $f2		\n\t"
++                
++				"pcmpeqh	$f0, $f0, $f14		\n\t"
++                
++				"pcmpeqh	$f2, $f2, $f14		\n\t"
++                
++				"pandn		$f0, $f0, $f4		\n\t"
++                
++				"pandn		$f2, $f2, $f6		\n\t"
++                
++				"sdc1		$f0, ($12)			\n\t"
++                
++				"sdc1		$f2, 8($12)			\n\t"
++
++                
++				"addiu		%3, %3, 16			\n\t"
++                
++				"blez		%3, 1b				\n\t"
++				"nop							\n\t"
++                ::"r" (block+nCoeffs), "m"(qmul), "m" (qadd), "r" (2*(-nCoeffs))
++                : "memory"
++        );
++}
++
++
++/* draw the edges of width 'w' of an image of size width, height
++   this mmx version can only handle w==8 || w==16 */
++
++static void draw_edges_loongson2(uint8_t *buf, int wrap, int width, int height, int w)
++{
++    uint8_t *ptr, *last_line;
++    int i;
++
++    last_line = buf + (height - 1) * wrap;
++    
++    ptr = buf;
++    if(w==8)
++    {
++        __asm__ volatile(
++//				".set mips3						\n\t"
++
++				"move		$9, %0				\n\t"
++
++                "1:                             \n\t"
++                
++				"xor		$f0, $f0, $f0		\n\t"
++				"lwc1		$f0, ($9)			\n\t"
++                
++				"punpcklbh	$f0, $f0, $f0		\n\t"
++
++				"add		$12, $9, %2			\n\t"
++                
++				"punpcklhw	$f0, $f0, $f0		\n\t"
++
++				"punpcklwd	$f0, $f0, $f0		\n\t"
++
++				"ldc1		$f2, -8($12)		\n\t"
++
++				"sdc1		$f0, -8($9)			\n\t"
++                
++				"punpckhbh	$f2, $f2, $f2		\n\t"
++                
++				"add		$9, $9, %1			\n\t"
++
++				"punpckhhw	$f2, $f2, $f2		\n\t"
++                
++				"sub		$13, $9, %3			\n\t"
++
++				"punpckhwd	$f2, $f2, $f2		\n\t"
++
++				"bltz		$13, 1b				\n\t"
++                
++				"sdc1		$f2, ($12)			\n\t"
++
++                : "+r" (ptr)
++                : "r" ((long)wrap), "r" ((long)width), "r" (ptr + wrap*height)
++				: "$9", "$13", "$12", "$f2", "$f0"
++        );
++    }
++    else
++    {
++        __asm__ volatile(
++//				".set mips3						\n\t"
++
++				"move		$8, %0				\n\t"
++
++                "1:                             \n\t"
++                
++				"xor		$f0, $f0, $f0		\n\t"
++				"lwc1		$f0, ($8)			\n\t"
++                
++				"punpcklbh	$f0, $f0, $f0		\n\t"
++				"punpcklhw	$f0, $f0, $f0		\n\t"
++				"punpcklwd	$f0, $f0, $f0		\n\t"
++
++				"sdc1		$f0, -8($8)			\n\t"
++				"sdc1		$f0, -16($8)		\n\t"
++                
++				"add		$15, $8, %2			\n\t"
++				"ldc1		$f2, -8($15)		\n\t"
++                
++				"punpckhbh	$f2, $f2, $f2		\n\t"
++				"punpckhhw	$f2, $f2, $f2		\n\t"
++				"punpckhwd	$f2, $f2, $f2		\n\t"
++
++				"sdc1		$f2, ($15)			\n\t"
++				"sdc1		$f2, 8($15)			\n\t"
++
++				"add		$8, $8, %1			\n\t"
++                
++				"sub		$16, $8, %3			\n\t"
++				"bltz		$16, 1b				\n\t"
++				"nop							\n\t"
++                : "+r" (ptr)
++                : "r" ((long)wrap), "r" ((long)width), "r" (ptr + wrap*height)
++				: "$8", "$15", "$16", "$f0", "$f2"
++        );
++    }
++
++    for(i=0;i<w;i+=4) {
++        
++        ptr= buf - (i + 1) * wrap - w;
++        __asm__ volatile(
++//				".set mips3						\n\t"
++				"move		$8, %0				\n\t"
++
++                "1:                             \n\t"
++                
++				"add		$9, $8, %1			\n\t"
++				"ldc1		$f0, ($9)			\n\t"
++
++				"add		$10, $8, %2			\n\t"
++				"add		$11, $10, %2		\n\t"
++				"add		$12, $8, %3			\n\t"
++                
++				"sdc1		$f0, ($8)			\n\t"
++				"sdc1		$f0, ($10)			\n\t"
++				"sdc1		$f0, ($11)			\n\t"
++				"sdc1		$f0, ($12)			\n\t"
++
++				"addiu		$8, $8, 8			\n\t"
++                
++				"sub		$13, $8, %4			\n\t"
++                
++				"bltz		$13, 1b				\n\t"
++				"nop							\n\t"
++
++                : "+r" (ptr)
++                : "r" (((long)buf - (long)ptr - w)), "r" ((long)-wrap), "r" ((long)-wrap*3), "r" (*(ptr+width+2*w))
++				: "$8", "$9", "$10", "$11", "$12", "$13", "$f0"
++		);
++
++        ptr= last_line + (i + 1) * wrap - w;
++
++        __asm__ volatile(
++//				".set mips3						\n\t"
++
++				"move		$9, %0				\n\t"
++
++                "1:                             \n\t"
++                
++				"add		$10, $9, %1			\n\t"
++				"ldc1		$f0, ($10)				\n\t"
++
++				"add		$11, $9, %2			\n\t"
++				"add		$12, $11, %2		\n\t"
++				"add		$13, $9, %3			\n\t"
++				
++				"sdc1		$f0, ($9)			\n\t"
++				"sdc1		$f0, ($11)			\n\t"
++				"sdc1		$f0, ($12)			\n\t"
++				"sdc1		$f0, ($13)			\n\t"
++                
++				"addiu		$9, $9, 8			\n\t"
++                
++				"sub		$14, $9, %4			\n\t"
++                
++				"bltz		$14, 1b				\n\t"
++				"nop							\n\t"
++                : "+r" (ptr)
++                : "r" ((long)last_line - (long)ptr - w), "r" ((long)wrap), "r" ((long)wrap*3), "r" (ptr+width+2*w)
++				: "$9", "$10", "$11", "$12", "$13", "$14", "$f0"
++
++		);
++    }
++}
++
++void MPV_common_init_loongson2(MpegEncContext *s)
++{
++	s->dct_unquantize_h263_intra = dct_unquantize_h263_intra_loongson2;
++	s->dct_unquantize_h263_inter = dct_unquantize_h263_inter_loongson2;
++
++//	draw_edges = draw_edges_loongson2;
++
++}
+diff --git a/libavcodec/mips/Makefile b/libavcodec/mips/Makefile
+index 3f4da68..73e4d56 100644
+--- a/libavcodec/mips/Makefile
++++ b/libavcodec/mips/Makefile
+@@ -1,3 +1,9 @@
+ OBJS-$(HAVE_MMI)                       += ps2/dsputil_mmi.o             \
+                                           ps2/idct_mmi.o                \
+                                           ps2/mpegvideo_mmi.o           \
++
++OBJS-$(HAVE_LOONGSON2MMI)              += loongson2/idct_loongson2.o   \
++                                          loongson2/dsputil_loongson2.o \
++                                          loongson2/idct_loongson2_xvid.o \
++                                          loongson2/mpegvideo_loongson2.o \
++                                          loongson2/motion_est_loongson2.o
+diff --git a/libavcodec/mpegvideo.c b/libavcodec/mpegvideo.c
+index b47ff9a..af92552 100644
+--- a/libavcodec/mpegvideo.c
++++ b/libavcodec/mpegvideo.c
+@@ -176,6 +176,9 @@ av_cold int ff_dct_common_init(MpegEncContext *s)
+ #elif ARCH_BFIN
+     MPV_common_init_bfin(s);
+ #endif
++#ifdef HAVE_LOONGSON2MMI
++    MPV_common_init_loongson2(s);
++#endif
+ 
+     /* load & permutate scantables
+        note: only wmv uses different ones
+diff --git a/libavcodec/mpegvideo.h b/libavcodec/mpegvideo.h
+index 5302be9..8d09906 100644
+--- a/libavcodec/mpegvideo.h
++++ b/libavcodec/mpegvideo.h
+@@ -689,6 +689,7 @@ int MPV_encode_picture(AVCodecContext *avctx, unsigned char *buf, int buf_size,
+ void MPV_common_init_mmx(MpegEncContext *s);
+ void MPV_common_init_axp(MpegEncContext *s);
+ void MPV_common_init_mlib(MpegEncContext *s);
++void MPV_common_init_loongson2(MpegEncContext *s);
+ void MPV_common_init_mmi(MpegEncContext *s);
+ void MPV_common_init_arm(MpegEncContext *s);
+ void MPV_common_init_altivec(MpegEncContext *s);
+diff --git a/libavcodec/options.c b/libavcodec/options.c
+index 7ca1062..c05b3f4 100644
+--- a/libavcodec/options.c
++++ b/libavcodec/options.c
+@@ -205,6 +205,8 @@ static const AVOption options[]={
+ {"simple", NULL, 0, FF_OPT_TYPE_CONST, FF_IDCT_SIMPLE, INT_MIN, INT_MAX, V|E|D, "idct"},
+ {"simplemmx", NULL, 0, FF_OPT_TYPE_CONST, FF_IDCT_SIMPLEMMX, INT_MIN, INT_MAX, V|E|D, "idct"},
+ {"libmpeg2mmx", NULL, 0, FF_OPT_TYPE_CONST, FF_IDCT_LIBMPEG2MMX, INT_MIN, INT_MAX, V|E|D, "idct"},
++{"libmpeg2loongson2", NULL, 0, FF_OPT_TYPE_CONST, FF_IDCT_LIBMPEG2LOONGSON2, INT_MIN, INT_MAX, V|E|D, "idct"},
++{"xvidloongson2", NULL, 0, FF_OPT_TYPE_CONST, FF_IDCT_XVIDLOONGSON2, INT_MIN, INT_MAX, V|E|D, "idct"},
+ {"ps2", NULL, 0, FF_OPT_TYPE_CONST, FF_IDCT_PS2, INT_MIN, INT_MAX, V|E|D, "idct"},
+ {"mlib", NULL, 0, FF_OPT_TYPE_CONST, FF_IDCT_MLIB, INT_MIN, INT_MAX, V|E|D, "idct"},
+ {"arm", NULL, 0, FF_OPT_TYPE_CONST, FF_IDCT_ARM, INT_MIN, INT_MAX, V|E|D, "idct"},
diff --git a/media-video/ffmpeg/metadata.xml b/media-video/ffmpeg/metadata.xml
new file mode 100644
index 0000000..afa3011
--- /dev/null
+++ b/media-video/ffmpeg/metadata.xml
@@ -0,0 +1,28 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE pkgmetadata SYSTEM "http://www.gentoo.org/dtd/metadata.dtd">
+<pkgmetadata>
+  <herd>video</herd>
+  <maintainer>
+    <email>media-video@gentoo.org</email>
+  </maintainer>
+  <use>
+    <flag name="3dnowext">
+      Enable manually-optimised routines using the AMD 3DNow!Ex SIMD
+      instruction set, present in modern AMD CPUs.  (Check for
+      3dnowext in /proc/cpuinfo to know whether your CPU supports it).
+    </flag>
+    <flag name="amr">Enables Adaptive Multi-Rate Audio support</flag>
+    <flag name="cpudetection">Enables runtime CPU detection (useful for bindist, compatibility on other CPUs)</flag>
+    <flag name="dirac">Enable Dirac video support (an advanced royalty-free video compression format) via the reference library: dirac.</flag>
+	<flag name="faac">Use external faac library for AAC encoding</flag>
+	<flag name="faad">Use external faad library for AAC decoding (instead of internal libavcodec support)</flag>
+    <flag name="hardcoded-tables">Use pre-calculated tables rather than calculating them on the fly.</flag>
+    <flag name="mmxext">Enables mmx2 support</flag>
+    <flag name="network">Enables network streaming support</flag>
+    <flag name="pic">Force shared libraries to be built as PIC (this is slower)</flag>
+    <flag name="schroedinger">Enable Dirac video support (an advanced royalty-free video compression format) via libschroedinger (high-speed implementation in C of the Dirac codec).</flag>
+    <flag name="ssse3">faster floating point optimization for SSSE3 capable chips (Intel Core 2 and later chips)</flag>
+	<flag name="vaapi">Enables VAAPI (Video Acceleration API) for hardware decoding</flag>
+	<flag name="vdpau">Enables VDPAU decoders (requires nVidia video cards to offload CPU processing to video card</flag>
+  </use>
+</pkgmetadata>